]> git.ipfire.org Git - people/arne_f/kernel.git/commitdiff
Merge branch 'linus' into locking/core, to pick up fixes
authorIngo Molnar <mingo@kernel.org>
Fri, 25 Aug 2017 09:04:51 +0000 (11:04 +0200)
committerIngo Molnar <mingo@kernel.org>
Fri, 25 Aug 2017 09:04:51 +0000 (11:04 +0200)
Signed-off-by: Ingo Molnar <mingo@kernel.org>
412 files changed:
Documentation/fb/efifb.txt
Documentation/networking/switchdev.txt
Documentation/printk-formats.txt
Documentation/sysctl/net.txt
MAINTAINERS
Makefile
arch/arc/Kconfig
arch/arc/Makefile
arch/arc/boot/dts/axc001.dtsi
arch/arc/boot/dts/axc003.dtsi
arch/arc/boot/dts/axc003_idu.dtsi
arch/arc/boot/dts/axs10x_mb.dtsi
arch/arc/configs/haps_hs_defconfig
arch/arc/configs/haps_hs_smp_defconfig
arch/arc/configs/nps_defconfig
arch/arc/configs/nsim_700_defconfig
arch/arc/configs/nsim_hs_defconfig
arch/arc/configs/nsim_hs_smp_defconfig
arch/arc/configs/nsimosci_defconfig
arch/arc/configs/nsimosci_hs_defconfig
arch/arc/configs/nsimosci_hs_smp_defconfig
arch/arc/configs/tb10x_defconfig
arch/arc/include/asm/cache.h
arch/arc/include/asm/mmu.h
arch/arc/kernel/intc-arcv2.c
arch/arc/kernel/intc-compact.c
arch/arc/mm/cache.c
arch/arc/mm/dma.c
arch/arc/mm/tlb.c
arch/arc/plat-sim/Kconfig [deleted file]
arch/arc/plat-sim/platform.c
arch/arm/boot/dts/exynos5422-odroidxu3-common.dtsi
arch/arm/boot/dts/imx25.dtsi
arch/arm/boot/dts/imx6qdl-nitrogen6_som2.dtsi
arch/arm/boot/dts/imx7d-sdb.dts
arch/arm/boot/dts/sama5d2.dtsi
arch/arm/mach-at91/Kconfig
arch/arm/mach-at91/pm.c
arch/arm64/boot/dts/allwinner/sun50i-a64-bananapi-m64.dts
arch/arm64/boot/dts/allwinner/sun50i-a64-pine64.dts
arch/arm64/boot/dts/allwinner/sun50i-a64-sopine-baseboard.dts
arch/arm64/boot/dts/allwinner/sun50i-h5.dtsi
arch/arm64/boot/dts/renesas/salvator-common.dtsi
arch/arm64/include/asm/arch_timer.h
arch/arm64/include/asm/elf.h
arch/arm64/kernel/fpsimd.c
arch/arm64/kernel/head.S
arch/arm64/kernel/kaslr.c
arch/arm64/mm/fault.c
arch/mips/Kconfig
arch/mips/Makefile
arch/mips/boot/compressed/.gitignore [new file with mode: 0644]
arch/mips/cavium-octeon/octeon-usb.c
arch/mips/dec/int-handler.S
arch/mips/include/asm/cache.h
arch/mips/include/asm/cpu-features.h
arch/mips/include/asm/octeon/cvmx-l2c-defs.h
arch/mips/include/asm/octeon/cvmx-l2d-defs.h [new file with mode: 0644]
arch/mips/include/asm/octeon/cvmx.h
arch/mips/kernel/smp.c
arch/mips/mm/uasm-mips.c
arch/mips/pci/pci.c
arch/mips/vdso/gettimeofday.c
arch/powerpc/Kconfig
arch/powerpc/configs/powernv_defconfig
arch/powerpc/configs/ppc64_defconfig
arch/powerpc/configs/pseries_defconfig
arch/powerpc/kernel/entry_64.S
arch/powerpc/kernel/process.c
arch/powerpc/kernel/smp.c
arch/powerpc/kernel/watchdog.c
arch/powerpc/platforms/powernv/idle.c
arch/sparc/include/asm/page_32.h
arch/sparc/kernel/pci_sun4v.c
arch/sparc/kernel/pcic.c
arch/sparc/lib/multi3.S
arch/x86/Kconfig
arch/x86/crypto/sha1_avx2_x86_64_asm.S
arch/x86/crypto/sha1_ssse3_glue.c
arch/x86/entry/entry_64.S
arch/x86/events/core.c
arch/x86/events/intel/bts.c
arch/x86/events/intel/p4.c
arch/x86/events/intel/rapl.c
arch/x86/events/intel/uncore.c
arch/x86/events/intel/uncore_nhmex.c
arch/x86/events/intel/uncore_snb.c
arch/x86/events/intel/uncore_snbep.c
arch/x86/include/asm/cpufeatures.h
arch/x86/include/asm/elf.h
arch/x86/include/asm/hypervisor.h
arch/x86/kernel/cpu/aperfmperf.c
arch/x86/kernel/cpu/mcheck/therm_throt.c
arch/x86/kernel/cpu/microcode/core.c
arch/x86/kernel/cpu/mtrr/main.c
arch/x86/kernel/head64.c
arch/x86/kernel/ksysfs.c
arch/x86/kernel/smpboot.c
arch/x86/kvm/svm.c
arch/x86/mm/init.c
arch/x86/mm/mmap.c
arch/x86/platform/uv/tlb_uv.c
arch/x86/xen/enlighten_hvm.c
block/bio-integrity.c
block/blk-mq-pci.c
block/blk-mq.c
drivers/acpi/acpica/nsxfeval.c
drivers/acpi/ec.c
drivers/acpi/internal.h
drivers/acpi/property.c
drivers/acpi/scan.c
drivers/acpi/spcr.c
drivers/base/firmware_class.c
drivers/block/xen-blkfront.c
drivers/clocksource/Kconfig
drivers/clocksource/arm_arch_timer.c
drivers/clocksource/em_sti.c
drivers/clocksource/timer-of.c
drivers/cpufreq/intel_pstate.c
drivers/cpuidle/cpuidle-powernv.c
drivers/crypto/ixp4xx_crypto.c
drivers/gpio/gpio-mvebu.c
drivers/gpio/gpiolib-sysfs.c
drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
drivers/gpu/drm/i915/i915_debugfs.c
drivers/gpu/drm/i915/i915_gem_context.c
drivers/gpu/drm/i915/i915_gem_render_state.c
drivers/gpu/drm/i915/intel_ddi.c
drivers/gpu/drm/i915/intel_display.c
drivers/gpu/drm/i915/intel_lrc.h
drivers/iio/accel/bmc150-accel-core.c
drivers/iio/accel/st_accel_core.c
drivers/iio/adc/aspeed_adc.c
drivers/iio/adc/axp288_adc.c
drivers/iio/adc/sun4i-gpadc-iio.c
drivers/iio/adc/vf610_adc.c
drivers/iio/common/st_sensors/st_sensors_core.c
drivers/iio/light/tsl2563.c
drivers/iio/pressure/st_pressure_core.c
drivers/infiniband/core/device.c
drivers/infiniband/core/uverbs_cmd.c
drivers/infiniband/core/uverbs_main.c
drivers/infiniband/core/verbs.c
drivers/infiniband/hw/cxgb4/mem.c
drivers/infiniband/hw/hns/hns_roce_ah.c
drivers/infiniband/hw/i40iw/i40iw_ctrl.c
drivers/infiniband/hw/i40iw/i40iw_d.h
drivers/infiniband/hw/i40iw/i40iw_puda.c
drivers/infiniband/hw/i40iw/i40iw_status.h
drivers/infiniband/hw/i40iw/i40iw_uk.c
drivers/infiniband/hw/mlx5/main.c
drivers/infiniband/hw/mlx5/qp.c
drivers/infiniband/hw/vmw_pvrdma/pvrdma_cq.c
drivers/input/mouse/elan_i2c_core.c
drivers/input/mouse/trackpoint.c
drivers/iommu/arm-smmu.c
drivers/irqchip/irq-atmel-aic-common.c
drivers/irqchip/irq-atmel-aic-common.h
drivers/irqchip/irq-atmel-aic.c
drivers/irqchip/irq-atmel-aic5.c
drivers/irqchip/irq-brcmstb-l2.c
drivers/irqchip/irq-gic-v3-its-platform-msi.c
drivers/irqchip/irq-gic-v3-its.c
drivers/irqchip/irq-gic-v3.c
drivers/irqchip/irq-gic.c
drivers/isdn/mISDN/fsm.c
drivers/isdn/mISDN/fsm.h
drivers/isdn/mISDN/layer1.c
drivers/isdn/mISDN/layer2.c
drivers/isdn/mISDN/tei.c
drivers/md/md.c
drivers/md/raid5-cache.c
drivers/memory/atmel-ebi.c
drivers/mfd/atmel-smc.c
drivers/mfd/da9062-core.c
drivers/misc/mei/pci-me.c
drivers/misc/mei/pci-txe.c
drivers/mmc/core/block.c
drivers/mmc/core/mmc.c
drivers/mmc/host/omap_hsmmc.c
drivers/mtd/mtd_blkdevs.c
drivers/net/bonding/bond_main.c
drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
drivers/net/ethernet/chelsio/cxgb4/sge.c
drivers/net/ethernet/chelsio/cxgb4vf/adapter.h
drivers/net/ethernet/chelsio/cxgb4vf/cxgb4vf_main.c
drivers/net/ethernet/chelsio/cxgb4vf/sge.c
drivers/net/ethernet/mellanox/mlx4/main.c
drivers/net/ethernet/netronome/nfp/flower/cmsg.c
drivers/net/ethernet/netronome/nfp/nfp_net_common.c
drivers/net/ethernet/qlogic/netxen/netxen_nic_hw.c
drivers/net/ethernet/sfc/mcdi_port.c
drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
drivers/net/tun.c
drivers/net/wireless/broadcom/brcm80211/brcmfmac/feature.c
drivers/net/wireless/intel/iwlwifi/cfg/9000.c
drivers/net/wireless/intel/iwlwifi/fw/file.h
drivers/net/wireless/intel/iwlwifi/iwl-config.h
drivers/net/wireless/intel/iwlwifi/iwl-drv.c
drivers/net/wireless/intel/iwlwifi/iwl-nvm-parse.c
drivers/net/wireless/intel/iwlwifi/mvm/fw.c
drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c
drivers/net/wireless/intel/iwlwifi/mvm/rs.c
drivers/net/wireless/intel/iwlwifi/mvm/rxmq.c
drivers/net/wireless/intel/iwlwifi/mvm/sta.c
drivers/net/wireless/intel/iwlwifi/mvm/tx.c
drivers/net/wireless/intel/iwlwifi/pcie/drv.c
drivers/nvme/host/core.c
drivers/nvme/host/fabrics.c
drivers/nvme/host/pci.c
drivers/nvme/target/admin-cmd.c
drivers/nvme/target/fc.c
drivers/of/device.c
drivers/parisc/dino.c
drivers/pci/pci.c
drivers/pci/probe.c
drivers/pci/quirks.c
drivers/rtc/rtc-ds1307.c
drivers/scsi/Kconfig
drivers/scsi/aacraid/aachba.c
drivers/scsi/aacraid/aacraid.h
drivers/scsi/csiostor/csio_hw.c
drivers/scsi/csiostor/csio_init.c
drivers/scsi/cxgbi/cxgb4i/cxgb4i.c
drivers/scsi/ipr.c
drivers/scsi/ipr.h
drivers/scsi/lpfc/lpfc_attr.c
drivers/scsi/lpfc/lpfc_debugfs.c
drivers/scsi/lpfc/lpfc_nvmet.c
drivers/scsi/lpfc/lpfc_nvmet.h
drivers/scsi/megaraid/megaraid_sas_base.c
drivers/scsi/qla2xxx/qla_tmpl.c
drivers/scsi/qla2xxx/tcm_qla2xxx.c
drivers/scsi/scsi.c
drivers/scsi/sd.c
drivers/scsi/sd_zbc.c
drivers/scsi/ses.c
drivers/scsi/st.c
drivers/soc/imx/gpcv2.c
drivers/soc/ti/knav_qmss_queue.c
drivers/soc/ti/ti_sci_pm_domains.c
drivers/staging/comedi/comedi_fops.c
drivers/staging/iio/resolver/ad2s1210.c
drivers/target/iscsi/cxgbit/cxgbit_cm.c
drivers/target/iscsi/cxgbit/cxgbit_target.c
drivers/target/iscsi/iscsi_target.c
drivers/target/iscsi/iscsi_target_login.c
drivers/target/target_core_tpg.c
drivers/target/target_core_transport.c
drivers/target/target_core_user.c
drivers/thunderbolt/eeprom.c
drivers/tty/pty.c
drivers/tty/serial/8250/8250_core.c
drivers/tty/serial/amba-pl011.c
drivers/tty/tty_io.c
drivers/usb/core/hcd.c
drivers/usb/core/hub.c
drivers/usb/core/quirks.c
drivers/usb/dwc3/gadget.c
drivers/usb/gadget/udc/renesas_usb3.c
drivers/usb/host/pci-quirks.c
drivers/usb/musb/musb_host.c
drivers/usb/phy/phy-msm-usb.c
drivers/usb/renesas_usbhs/mod_gadget.c
drivers/usb/renesas_usbhs/rcar3.c
drivers/usb/serial/cp210x.c
drivers/usb/serial/option.c
drivers/usb/serial/pl2303.c
drivers/usb/serial/pl2303.h
drivers/usb/storage/unusual_uas.h
drivers/usb/storage/usb.c
drivers/video/fbdev/efifb.c
drivers/video/fbdev/imxfb.c
drivers/video/fbdev/omap2/omapfb/dss/core.c
drivers/xen/Makefile
drivers/xen/biomerge.c
drivers/xen/events/events_base.c
drivers/xen/xenbus/xenbus_xs.c
fs/binfmt_elf.c
fs/btrfs/disk-io.c
fs/btrfs/inode.c
fs/btrfs/raid56.c
fs/btrfs/volumes.c
fs/btrfs/volumes.h
fs/devpts/inode.c
fs/ext4/mballoc.c
fs/ext4/xattr.c
fs/fuse/file.c
fs/fuse/fuse_i.h
fs/iomap.c
fs/nfs/Kconfig
fs/nfs/flexfilelayout/flexfilelayoutdev.c
fs/nfs/nfs4proc.c
fs/quota/dquot.c
fs/xfs/libxfs/xfs_ialloc.c
fs/xfs/xfs_log.c
fs/xfs/xfs_mount.c
include/asm-generic/vmlinux.lds.h
include/linux/acpi.h
include/linux/devpts_fs.h
include/linux/iio/common/st_sensors.h
include/linux/memblock.h
include/linux/memcontrol.h
include/linux/net.h
include/linux/nmi.h
include/linux/nvme-fc-driver.h
include/linux/oom.h
include/linux/pci.h
include/linux/perf_event.h
include/linux/pid.h
include/linux/platform_data/st_sensors_pdata.h
include/linux/ptr_ring.h
include/linux/sched.h
include/linux/skb_array.h
include/linux/wait.h
include/net/addrconf.h
include/net/bonding.h
include/net/busy_poll.h
include/net/ip.h
include/net/mac80211.h
include/net/sch_generic.h
include/net/sock.h
include/net/udp.h
include/rdma/ib_verbs.h
include/scsi/scsi_cmnd.h
include/target/iscsi/iscsi_target_core.h
kernel/audit_watch.c
kernel/events/core.c
kernel/irq/chip.c
kernel/irq/ipi.c
kernel/kmod.c
kernel/pid.c
kernel/signal.c
kernel/trace/bpf_trace.c
kernel/trace/ftrace.c
kernel/trace/ring_buffer.c
kernel/trace/ring_buffer_benchmark.c
kernel/trace/trace.c
kernel/trace/trace_events_filter.c
kernel/trace/tracing_map.c
kernel/watchdog.c
kernel/watchdog_hld.c
lib/Kconfig.debug
mm/cma_debug.c
mm/huge_memory.c
mm/memblock.c
mm/memcontrol.c
mm/memory.c
mm/mempolicy.c
mm/migrate.c
mm/nobootmem.c
mm/page-writeback.c
mm/page_alloc.c
mm/slub.c
mm/vmalloc.c
net/core/datagram.c
net/core/filter.c
net/dccp/proto.c
net/dsa/tag_ksz.c
net/ipv4/fib_semantics.c
net/ipv4/igmp.c
net/ipv4/route.c
net/ipv4/tcp_input.c
net/ipv4/tcp_ipv4.c
net/ipv4/tcp_ulp.c
net/ipv4/udp.c
net/ipv6/ip6_fib.c
net/ipv6/route.c
net/ipv6/tcp_ipv6.c
net/ipv6/udp.c
net/irda/af_irda.c
net/key/af_key.c
net/mac80211/agg-rx.c
net/openvswitch/actions.c
net/openvswitch/datapath.c
net/openvswitch/datapath.h
net/rxrpc/call_accept.c
net/sched/act_ipt.c
net/sched/cls_api.c
net/sched/sch_api.c
net/sched/sch_atm.c
net/sched/sch_cbq.c
net/sched/sch_hfsc.c
net/sched/sch_htb.c
net/sched/sch_sfq.c
net/sctp/ipv6.c
net/tipc/bearer.c
net/tipc/msg.c
net/tipc/netlink_compat.c
net/unix/af_unix.c
scripts/Kbuild.include
scripts/Makefile.asm-generic
scripts/Makefile.build
scripts/Makefile.dtbinst
scripts/basic/Makefile
scripts/basic/fixdep.c
sound/core/seq/Kconfig
sound/core/seq/seq_clientmgr.c
sound/core/seq/seq_queue.c
sound/core/seq/seq_queue.h
sound/pci/emu10k1/emufx.c
sound/pci/hda/patch_realtek.c
sound/usb/mixer.c
sound/usb/mixer.h
sound/usb/mixer_quirks.c
sound/usb/quirks.c
tools/lib/bpf/libbpf.c
tools/testing/selftests/futex/Makefile
tools/testing/selftests/kmod/kmod.sh [changed mode: 0644->0755]
tools/testing/selftests/sysctl/sysctl.sh [changed mode: 0644->0755]
tools/testing/selftests/timers/freq-step.c

index a59916c29b3312cd4946a1d9a8da2331819e7845..1a85c1bdaf38a9ae7fb8b6555afc30abae661a20 100644 (file)
@@ -27,5 +27,11 @@ You have to add the following kernel parameters in your elilo.conf:
        Macbook Pro 17", iMac 20" :
                video=efifb:i20
 
+Accepted options:
+
+nowc   Don't map the framebuffer write combined. This can be used
+       to workaround side-effects and slowdowns on other CPU cores
+       when large amounts of console data are written.
+
 --
 Edgar Hucek <gimli@dark-green.com>
index 3e7b946dea2778441ce0599643535188699e16ca..5e40e1f68873b0f2b594c22a6e69ddb218ce56ec 100644 (file)
@@ -228,7 +228,7 @@ Learning on the device port should be enabled, as well as learning_sync:
        bridge link set dev DEV learning on self
        bridge link set dev DEV learning_sync on self
 
-Learning_sync attribute enables syncing of the learned/forgotton FDB entry to
+Learning_sync attribute enables syncing of the learned/forgotten FDB entry to
 the bridge's FDB.  It's possible, but not optimal, to enable learning on the
 device port and on the bridge port, and disable learning_sync.
 
@@ -245,7 +245,7 @@ the responsibility of the port driver/device to age out these entries.  If the
 port device supports ageing, when the FDB entry expires, it will notify the
 driver which in turn will notify the bridge with SWITCHDEV_FDB_DEL.  If the
 device does not support ageing, the driver can simulate ageing using a
-garbage collection timer to monitor FBD entries.  Expired entries will be
+garbage collection timer to monitor FDB entries.  Expired entries will be
 notified to the bridge using SWITCHDEV_FDB_DEL.  See rocker driver for
 example of driver running ageing timer.
 
index 65ea5915178b4b9842474b7908d06660444aae34..074670b98bac784c27eecc18f672b2f1ad24f489 100644 (file)
@@ -58,20 +58,23 @@ Symbols/Function Pointers
        %ps     versatile_init
        %pB     prev_fn_of_versatile_init+0x88/0x88
 
-For printing symbols and function pointers. The ``S`` and ``s`` specifiers
-result in the symbol name with (``S``) or without (``s``) offsets. Where
-this is used on a kernel without KALLSYMS - the symbol address is
-printed instead.
+The ``F`` and ``f`` specifiers are for printing function pointers,
+for example, f->func, &gettimeofday. They have the same result as
+``S`` and ``s`` specifiers. But they do an extra conversion on
+ia64, ppc64 and parisc64 architectures where the function pointers
+are actually function descriptors.
+
+The ``S`` and ``s`` specifiers can be used for printing symbols
+from direct addresses, for example, __builtin_return_address(0),
+(void *)regs->ip. They result in the symbol name with (``S``) or
+without (``s``) offsets. If KALLSYMS are disabled then the symbol
+address is printed instead.
 
 The ``B`` specifier results in the symbol name with offsets and should be
 used when printing stack backtraces. The specifier takes into
 consideration the effect of compiler optimisations which may occur
 when tail-call``s are used and marked with the noreturn GCC attribute.
 
-On ia64, ppc64 and parisc64 architectures function pointers are
-actually function descriptors which must first be resolved. The ``F`` and
-``f`` specifiers perform this resolution and then provide the same
-functionality as the ``S`` and ``s`` specifiers.
 
 Kernel Pointers
 ===============
index 14db18c970b1b048c0ecc50d75e10efb674715ca..28596e03220b5528891c19316b7cfe2b53fcf331 100644 (file)
@@ -35,9 +35,34 @@ Table : Subdirectories in /proc/sys/net
 bpf_jit_enable
 --------------
 
-This enables Berkeley Packet Filter Just in Time compiler.
-Currently supported on x86_64 architecture, bpf_jit provides a framework
-to speed packet filtering, the one used by tcpdump/libpcap for example.
+This enables the BPF Just in Time (JIT) compiler. BPF is a flexible
+and efficient infrastructure allowing to execute bytecode at various
+hook points. It is used in a number of Linux kernel subsystems such
+as networking (e.g. XDP, tc), tracing (e.g. kprobes, uprobes, tracepoints)
+and security (e.g. seccomp). LLVM has a BPF back end that can compile
+restricted C into a sequence of BPF instructions. After program load
+through bpf(2) and passing a verifier in the kernel, a JIT will then
+translate these BPF proglets into native CPU instructions. There are
+two flavors of JITs, the newer eBPF JIT currently supported on:
+  - x86_64
+  - arm64
+  - ppc64
+  - sparc64
+  - mips64
+  - s390x
+
+And the older cBPF JIT supported on the following archs:
+  - arm
+  - mips
+  - ppc
+  - sparc
+
+eBPF JITs are a superset of cBPF JITs, meaning the kernel will
+migrate cBPF instructions into eBPF instructions and then JIT
+compile them transparently. Older cBPF JITs can only translate
+tcpdump filters, seccomp rules, etc, but not mentioned eBPF
+programs loaded through bpf(2).
+
 Values :
        0 - disable the JIT (default value)
        1 - enable the JIT
@@ -46,9 +71,9 @@ Values :
 bpf_jit_harden
 --------------
 
-This enables hardening for the Berkeley Packet Filter Just in Time compiler.
-Supported are eBPF JIT backends. Enabling hardening trades off performance,
-but can mitigate JIT spraying.
+This enables hardening for the BPF JIT compiler. Supported are eBPF
+JIT backends. Enabling hardening trades off performance, but can
+mitigate JIT spraying.
 Values :
        0 - disable JIT hardening (default value)
        1 - enable JIT hardening for unprivileged users only
@@ -57,11 +82,11 @@ Values :
 bpf_jit_kallsyms
 ----------------
 
-When Berkeley Packet Filter Just in Time compiler is enabled, then compiled
-images are unknown addresses to the kernel, meaning they neither show up in
-traces nor in /proc/kallsyms. This enables export of these addresses, which
-can be used for debugging/tracing. If bpf_jit_harden is enabled, this feature
-is disabled.
+When BPF JIT compiler is enabled, then compiled images are unknown
+addresses to the kernel, meaning they neither show up in traces nor
+in /proc/kallsyms. This enables export of these addresses, which can
+be used for debugging/tracing. If bpf_jit_harden is enabled, this
+feature is disabled.
 Values :
        0 - disable JIT kallsyms export (default value)
        1 - enable JIT kallsyms export for privileged users only
index 6f7721d1634c2eb7247538f2cb4d85fa1be1a458..1c3feffb1c1cfd2b46685907300a9f026fb67e6a 100644 (file)
@@ -7110,7 +7110,6 @@ M:        Marc Zyngier <marc.zyngier@arm.com>
 L:     linux-kernel@vger.kernel.org
 S:     Maintained
 T:     git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq/core
-T:     git git://git.infradead.org/users/jcooper/linux.git irqchip/core
 F:     Documentation/devicetree/bindings/interrupt-controller/
 F:     drivers/irqchip/
 
index 6eba23bcb5ad031d6365d938984dd3140a596018..dda88e744d5f3132ab926c43f512fe49f5a0c3b4 100644 (file)
--- a/Makefile
+++ b/Makefile
@@ -1,7 +1,7 @@
 VERSION = 4
 PATCHLEVEL = 13
 SUBLEVEL = 0
-EXTRAVERSION = -rc4
+EXTRAVERSION = -rc6
 NAME = Fearless Coyote
 
 # *DOCUMENTATION*
@@ -396,7 +396,7 @@ LINUXINCLUDE    := \
 KBUILD_CPPFLAGS := -D__KERNEL__
 
 KBUILD_CFLAGS   := -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \
-                  -fno-strict-aliasing -fno-common \
+                  -fno-strict-aliasing -fno-common -fshort-wchar \
                   -Werror-implicit-function-declaration \
                   -Wno-format-security \
                   -std=gnu89 $(call cc-option,-fno-PIE)
@@ -442,7 +442,7 @@ export RCS_TAR_IGNORE := --exclude SCCS --exclude BitKeeper --exclude .svn \
 # ===========================================================================
 # Rules shared between *config targets and build targets
 
-# Basic helpers built in scripts/
+# Basic helpers built in scripts/basic/
 PHONY += scripts_basic
 scripts_basic:
        $(Q)$(MAKE) $(build)=scripts/basic
@@ -505,7 +505,7 @@ ifeq ($(KBUILD_EXTMOD),)
                 endif
         endif
 endif
-# install and module_install need also be processed one by one
+# install and modules_install need also be processed one by one
 ifneq ($(filter install,$(MAKECMDGOALS)),)
         ifneq ($(filter modules_install,$(MAKECMDGOALS)),)
                mixed-targets := 1
@@ -964,7 +964,7 @@ export KBUILD_VMLINUX_MAIN := $(core-y) $(libs-y2) $(drivers-y) $(net-y) $(virt-
 export KBUILD_VMLINUX_LIBS := $(libs-y1)
 export KBUILD_LDS          := arch/$(SRCARCH)/kernel/vmlinux.lds
 export LDFLAGS_vmlinux
-# used by scripts/pacmage/Makefile
+# used by scripts/package/Makefile
 export KBUILD_ALLDIRS := $(sort $(filter-out arch/%,$(vmlinux-alldirs)) arch Documentation include samples scripts tools)
 
 vmlinux-deps := $(KBUILD_LDS) $(KBUILD_VMLINUX_INIT) $(KBUILD_VMLINUX_MAIN) $(KBUILD_VMLINUX_LIBS)
@@ -992,8 +992,8 @@ include/generated/autoksyms.h: FORCE
 ARCH_POSTLINK := $(wildcard $(srctree)/arch/$(SRCARCH)/Makefile.postlink)
 
 # Final link of vmlinux with optional arch pass after final link
-    cmd_link-vmlinux =                                                 \
-       $(CONFIG_SHELL) $< $(LD) $(LDFLAGS) $(LDFLAGS_vmlinux) ;       \
+cmd_link-vmlinux =                                                 \
+       $(CONFIG_SHELL) $< $(LD) $(LDFLAGS) $(LDFLAGS_vmlinux) ;    \
        $(if $(ARCH_POSTLINK), $(MAKE) -f $(ARCH_POSTLINK) $@, true)
 
 vmlinux: scripts/link-vmlinux.sh vmlinux_prereq $(vmlinux-deps) FORCE
@@ -1184,6 +1184,7 @@ PHONY += kselftest
 kselftest:
        $(Q)$(MAKE) -C tools/testing/selftests run_tests
 
+PHONY += kselftest-clean
 kselftest-clean:
        $(Q)$(MAKE) -C tools/testing/selftests clean
 
index a5459698f0ee3762917c82c8cbf133e34643f021..7db85ab00c522315c3108930365fc57c59fd311c 100644 (file)
@@ -96,7 +96,6 @@ menu "ARC Architecture Configuration"
 
 menu "ARC Platform/SoC/Board"
 
-source "arch/arc/plat-sim/Kconfig"
 source "arch/arc/plat-tb10x/Kconfig"
 source "arch/arc/plat-axs10x/Kconfig"
 #New platform adds here
index 44ef35d339564623b2beb4a380a81da6a7790087..3a61cfcc38c0dd78f58247512a4d31231a2daa4b 100644 (file)
@@ -107,7 +107,7 @@ core-y              += arch/arc/
 # w/o this dtb won't embed into kernel binary
 core-y         += arch/arc/boot/dts/
 
-core-$(CONFIG_ARC_PLAT_SIM)    += arch/arc/plat-sim/
+core-y                         += arch/arc/plat-sim/
 core-$(CONFIG_ARC_PLAT_TB10X)  += arch/arc/plat-tb10x/
 core-$(CONFIG_ARC_PLAT_AXS10X) += arch/arc/plat-axs10x/
 core-$(CONFIG_ARC_PLAT_EZNPS)  += arch/arc/plat-eznps/
index 53ce226f77a59857615f8fc53fca3ba1c4f09749..a380ffa1a4589b11de92a5a0c21bc831e6dc751b 100644 (file)
 
 / {
        compatible = "snps,arc";
-       #address-cells = <1>;
-       #size-cells = <1>;
+       #address-cells = <2>;
+       #size-cells = <2>;
 
        cpu_card {
                compatible = "simple-bus";
                #address-cells = <1>;
                #size-cells = <1>;
 
-               ranges = <0x00000000 0xf0000000 0x10000000>;
+               ranges = <0x00000000 0x0 0xf0000000 0x10000000>;
 
                core_clk: core_clk {
                        #clock-cells = <0>;
        mb_intc: dw-apb-ictl@0xe0012000 {
                #interrupt-cells = <1>;
                compatible = "snps,dw-apb-ictl";
-               reg = < 0xe0012000 0x200 >;
+               reg = < 0x0 0xe0012000 0x0 0x200 >;
                interrupt-controller;
                interrupt-parent = <&core_intc>;
                interrupts = < 7 >;
        };
 
        memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
-               ranges = <0x00000000 0x80000000 0x20000000>;
                device_type = "memory";
-               reg = <0x80000000 0x1b000000>;  /* (512 - 32) MiB */
+               /* CONFIG_KERNEL_RAM_BASE_ADDRESS needs to match low mem start */
+               reg = <0x0 0x80000000 0x0 0x1b000000>;  /* (512 - 32) MiB */
        };
 
        reserved-memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
+               #address-cells = <2>;
+               #size-cells = <2>;
                ranges;
                /*
                 * We just move frame buffer area to the very end of
                 */
                frame_buffer: frame_buffer@9e000000 {
                        compatible = "shared-dma-pool";
-                       reg = <0x9e000000 0x2000000>;
+                       reg = <0x0 0x9e000000 0x0 0x2000000>;
                        no-map;
                };
        };
index 14df46f141bf3450b8df0660bc0f248beef4669b..cc9239ef8d08c9f6efbe6dbee0323f283c4f5e37 100644 (file)
 
 / {
        compatible = "snps,arc";
-       #address-cells = <1>;
-       #size-cells = <1>;
+       #address-cells = <2>;
+       #size-cells = <2>;
 
        cpu_card {
                compatible = "simple-bus";
                #address-cells = <1>;
                #size-cells = <1>;
 
-               ranges = <0x00000000 0xf0000000 0x10000000>;
+               ranges = <0x00000000 0x0 0xf0000000 0x10000000>;
 
                core_clk: core_clk {
                        #clock-cells = <0>;
        mb_intc: dw-apb-ictl@0xe0012000 {
                #interrupt-cells = <1>;
                compatible = "snps,dw-apb-ictl";
-               reg = < 0xe0012000 0x200 >;
+               reg = < 0x0 0xe0012000 0x0 0x200 >;
                interrupt-controller;
                interrupt-parent = <&core_intc>;
                interrupts = < 24 >;
        };
 
        memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
-               ranges = <0x00000000 0x80000000 0x40000000>;
                device_type = "memory";
-               reg = <0x80000000 0x20000000>;  /* 512MiB */
+               /* CONFIG_KERNEL_RAM_BASE_ADDRESS needs to match low mem start */
+               reg = <0x0 0x80000000 0x0 0x20000000    /* 512 MiB low mem */
+                      0x1 0xc0000000 0x0 0x40000000>;  /* 1 GiB highmem */
        };
 
        reserved-memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
+               #address-cells = <2>;
+               #size-cells = <2>;
                ranges;
                /*
                 * Move frame buffer out of IOC aperture (0x8z-0xAz).
                 */
                frame_buffer: frame_buffer@be000000 {
                        compatible = "shared-dma-pool";
-                       reg = <0xbe000000 0x2000000>;
+                       reg = <0x0 0xbe000000 0x0 0x2000000>;
                        no-map;
                };
        };
index 695f9fa1996bcbb2689e1bf98dc8fa661cc1c216..4ebb2170abecc7e00cd581a7ca6dcef6e94231e0 100644 (file)
 
 / {
        compatible = "snps,arc";
-       #address-cells = <1>;
-       #size-cells = <1>;
+       #address-cells = <2>;
+       #size-cells = <2>;
 
        cpu_card {
                compatible = "simple-bus";
                #address-cells = <1>;
                #size-cells = <1>;
 
-               ranges = <0x00000000 0xf0000000 0x10000000>;
+               ranges = <0x00000000 0x0 0xf0000000 0x10000000>;
 
                core_clk: core_clk {
                        #clock-cells = <0>;
        mb_intc: dw-apb-ictl@0xe0012000 {
                #interrupt-cells = <1>;
                compatible = "snps,dw-apb-ictl";
-               reg = < 0xe0012000 0x200 >;
+               reg = < 0x0 0xe0012000 0x0 0x200 >;
                interrupt-controller;
                interrupt-parent = <&idu_intc>;
                interrupts = <0>;
        };
 
        memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
-               ranges = <0x00000000 0x80000000 0x40000000>;
                device_type = "memory";
-               reg = <0x80000000 0x20000000>;  /* 512MiB */
+               /* CONFIG_KERNEL_RAM_BASE_ADDRESS needs to match low mem start */
+               reg = <0x0 0x80000000 0x0 0x20000000    /* 512 MiB low mem */
+                      0x1 0xc0000000 0x0 0x40000000>;  /* 1 GiB highmem */
        };
 
        reserved-memory {
-               #address-cells = <1>;
-               #size-cells = <1>;
+               #address-cells = <2>;
+               #size-cells = <2>;
                ranges;
                /*
                 * Move frame buffer out of IOC aperture (0x8z-0xAz).
                 */
                frame_buffer: frame_buffer@be000000 {
                        compatible = "shared-dma-pool";
-                       reg = <0xbe000000 0x2000000>;
+                       reg = <0x0 0xbe000000 0x0 0x2000000>;
                        no-map;
                };
        };
index 41cfb29b62c142ca48193359e8fedf20151cc9bf..0ff7e07edcd4d28bcf6266cedd590034b8aebd13 100644 (file)
@@ -13,7 +13,7 @@
                compatible = "simple-bus";
                #address-cells = <1>;
                #size-cells = <1>;
-               ranges = <0x00000000 0xe0000000 0x10000000>;
+               ranges = <0x00000000 0x0 0xe0000000 0x10000000>;
                interrupt-parent = <&mb_intc>;
 
                i2sclk: i2sclk@100a0 {
index 57b3e599322f83d5babd0014c7d9ccf3a03578eb..db04ea4dd2d97237b843fd4b90706f2e0c8ef6d9 100644 (file)
@@ -21,7 +21,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_ARC_BUILTIN_DTB_NAME="haps_hs"
 CONFIG_PREEMPT=y
index f85985adebb245981d8a4130e8c0b91a45abe7dd..821a2e562f3f12422844556630d7136ad63097a8 100644 (file)
@@ -23,7 +23,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_SMP=y
 CONFIG_ARC_BUILTIN_DTB_NAME="haps_hs_idu"
index ede625c76216616b097365a3274e09ccea97d3d6..7c9c706ae7f66eb29d4cf48ca0b95d5dd44630f5 100644 (file)
@@ -39,7 +39,6 @@ CONFIG_IP_PNP=y
 # CONFIG_INET_XFRM_MODE_TRANSPORT is not set
 # CONFIG_INET_XFRM_MODE_TUNNEL is not set
 # CONFIG_INET_XFRM_MODE_BEET is not set
-# CONFIG_INET_LRO is not set
 # CONFIG_INET_DIAG is not set
 # CONFIG_IPV6 is not set
 # CONFIG_WIRELESS is not set
index b0066a749d4c49d8a23e3bd33b4a52789789c913..6dff83a238b8591ba08f889e7d7013734ff587db 100644 (file)
@@ -23,7 +23,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ARC_BUILTIN_DTB_NAME="nsim_700"
 CONFIG_PREEMPT=y
 # CONFIG_COMPACTION is not set
index ebe9ebb92933302af79f98dae000e8567b86ec6b..31ee51b987e7c5b97c2f044794f9d1c5b4eea73e 100644 (file)
@@ -26,7 +26,6 @@ CONFIG_MODULE_FORCE_UNLOAD=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_ARC_BUILTIN_DTB_NAME="nsim_hs"
 CONFIG_PREEMPT=y
index 4bde43278be6757c5c739bfd58a9f030a80c7568..8d3b1f67cae421c423e55c2622b5a78ea3d67794 100644 (file)
@@ -24,7 +24,6 @@ CONFIG_MODULE_FORCE_UNLOAD=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_SMP=y
 CONFIG_ARC_BUILTIN_DTB_NAME="nsim_hs_idu"
index f6fb3d26557eb7c63b2f77263acfed80f0b45387..6168ce2ac2efdd869705a8a75e78f7f9b9c55af4 100644 (file)
@@ -23,7 +23,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ARC_BUILTIN_DTB_NAME="nsimosci"
 # CONFIG_COMPACTION is not set
 CONFIG_NET=y
index b9f0fe00044b6c44d62a81a91064583cd9badee3..a70bdeb2b3fd03bdefd10388e69cd8e5b120ac0f 100644 (file)
@@ -23,7 +23,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_ARC_BUILTIN_DTB_NAME="nsimosci_hs"
 # CONFIG_COMPACTION is not set
index 155add7761ed63080bd81e830c819483fb55ee61..ef96406c446e823d4706e5c84b6ac0626d4b5812 100644 (file)
@@ -18,7 +18,6 @@ CONFIG_MODULES=y
 # CONFIG_BLK_DEV_BSG is not set
 # CONFIG_IOSCHED_DEADLINE is not set
 # CONFIG_IOSCHED_CFQ is not set
-CONFIG_ARC_PLAT_SIM=y
 CONFIG_ISA_ARCV2=y
 CONFIG_SMP=y
 # CONFIG_ARC_TIMERS_64BIT is not set
index 4c5118384eb5f1c393208cde601fddc1be773494..f3018254939508e373662a2ae98f9d08be9425a2 100644 (file)
@@ -38,7 +38,6 @@ CONFIG_IP_MULTICAST=y
 # CONFIG_INET_XFRM_MODE_TRANSPORT is not set
 # CONFIG_INET_XFRM_MODE_TUNNEL is not set
 # CONFIG_INET_XFRM_MODE_BEET is not set
-# CONFIG_INET_LRO is not set
 # CONFIG_INET_DIAG is not set
 # CONFIG_IPV6 is not set
 # CONFIG_WIRELESS is not set
index 19ebddffb279db05ca8e53e2c4753665376cc9ad..02fd1cece6ef339209c993a6b56b87312e39d648 100644 (file)
@@ -96,7 +96,9 @@ extern unsigned long perip_base, perip_end;
 #define ARC_REG_SLC_FLUSH      0x904
 #define ARC_REG_SLC_INVALIDATE 0x905
 #define ARC_REG_SLC_RGN_START  0x914
+#define ARC_REG_SLC_RGN_START1 0x915
 #define ARC_REG_SLC_RGN_END    0x916
+#define ARC_REG_SLC_RGN_END1   0x917
 
 /* Bit val in SLC_CONTROL */
 #define SLC_CTRL_DIS           0x001
index db7319e9b506e9ece826fde788aba60adada59cb..efb79fafff1d12214dc41923a75b237a969214f5 100644 (file)
@@ -94,6 +94,8 @@ static inline int is_pae40_enabled(void)
        return IS_ENABLED(CONFIG_ARC_HAS_PAE40);
 }
 
+extern int pae40_exist_but_not_enab(void);
+
 #endif /* !__ASSEMBLY__ */
 
 #endif
index f928795fd07a813e9f6f3e6d86f325b45561b3dc..cf90714a676d6baabbdc1beb2119fe530e1c84f3 100644 (file)
@@ -75,10 +75,13 @@ void arc_init_IRQ(void)
         * Set a default priority for all available interrupts to prevent
         * switching of register banks if Fast IRQ and multiple register banks
         * are supported by CPU.
+        * Also disable all IRQ lines so faulty external hardware won't
+        * trigger interrupt that kernel is not ready to handle.
         */
        for (i = NR_EXCEPTIONS; i < irq_bcr.irqs + NR_EXCEPTIONS; i++) {
                write_aux_reg(AUX_IRQ_SELECT, i);
                write_aux_reg(AUX_IRQ_PRIORITY, ARCV2_IRQ_DEF_PRIO);
+               write_aux_reg(AUX_IRQ_ENABLE, 0);
        }
 
        /* setup status32, don't enable intr yet as kernel doesn't want */
index 7e608c6b0a018697fb22c52c7fb4ac46a45fe2f0..cef388025adf43aed091869f102e01e1bee72092 100644 (file)
@@ -27,7 +27,7 @@
  */
 void arc_init_IRQ(void)
 {
-       int level_mask = 0;
+       int level_mask = 0, i;
 
        /* Is timer high priority Interrupt (Level2 in ARCompact jargon) */
        level_mask |= IS_ENABLED(CONFIG_ARC_COMPACT_IRQ_LEVELS) << TIMER0_IRQ;
@@ -40,6 +40,18 @@ void arc_init_IRQ(void)
 
        if (level_mask)
                pr_info("Level-2 interrupts bitset %x\n", level_mask);
+
+       /*
+        * Disable all IRQ lines so faulty external hardware won't
+        * trigger interrupt that kernel is not ready to handle.
+        */
+       for (i = TIMER0_IRQ; i < NR_CPU_IRQS; i++) {
+               unsigned int ienb;
+
+               ienb = read_aux_reg(AUX_IENABLE);
+               ienb &= ~(1 << i);
+               write_aux_reg(AUX_IENABLE, ienb);
+       }
 }
 
 /*
index a867575a758b9845c09a84fb4515f17f08f1272b..7db283b46ebde8daccf779525e82526e47ad2722 100644 (file)
@@ -665,6 +665,7 @@ noinline void slc_op(phys_addr_t paddr, unsigned long sz, const int op)
        static DEFINE_SPINLOCK(lock);
        unsigned long flags;
        unsigned int ctrl;
+       phys_addr_t end;
 
        spin_lock_irqsave(&lock, flags);
 
@@ -694,8 +695,19 @@ noinline void slc_op(phys_addr_t paddr, unsigned long sz, const int op)
         * END needs to be setup before START (latter triggers the operation)
         * END can't be same as START, so add (l2_line_sz - 1) to sz
         */
-       write_aux_reg(ARC_REG_SLC_RGN_END, (paddr + sz + l2_line_sz - 1));
-       write_aux_reg(ARC_REG_SLC_RGN_START, paddr);
+       end = paddr + sz + l2_line_sz - 1;
+       if (is_pae40_enabled())
+               write_aux_reg(ARC_REG_SLC_RGN_END1, upper_32_bits(end));
+
+       write_aux_reg(ARC_REG_SLC_RGN_END, lower_32_bits(end));
+
+       if (is_pae40_enabled())
+               write_aux_reg(ARC_REG_SLC_RGN_START1, upper_32_bits(paddr));
+
+       write_aux_reg(ARC_REG_SLC_RGN_START, lower_32_bits(paddr));
+
+       /* Make sure "busy" bit reports correct stataus, see STAR 9001165532 */
+       read_aux_reg(ARC_REG_SLC_CTRL);
 
        while (read_aux_reg(ARC_REG_SLC_CTRL) & SLC_CTRL_BUSY);
 
@@ -1111,6 +1123,13 @@ noinline void __init arc_ioc_setup(void)
        __dc_enable();
 }
 
+/*
+ * Cache related boot time checks/setups only needed on master CPU:
+ *  - Geometry checks (kernel build and hardware agree: e.g. L1_CACHE_BYTES)
+ *    Assume SMP only, so all cores will have same cache config. A check on
+ *    one core suffices for all
+ *  - IOC setup / dma callbacks only need to be done once
+ */
 void __init arc_cache_init_master(void)
 {
        unsigned int __maybe_unused cpu = smp_processor_id();
@@ -1190,12 +1209,27 @@ void __ref arc_cache_init(void)
 
        printk(arc_cache_mumbojumbo(0, str, sizeof(str)));
 
-       /*
-        * Only master CPU needs to execute rest of function:
-        *  - Assume SMP so all cores will have same cache config so
-        *    any geomtry checks will be same for all
-        *  - IOC setup / dma callbacks only need to be setup once
-        */
        if (!cpu)
                arc_cache_init_master();
+
+       /*
+        * In PAE regime, TLB and cache maintenance ops take wider addresses
+        * And even if PAE is not enabled in kernel, the upper 32-bits still need
+        * to be zeroed to keep the ops sane.
+        * As an optimization for more common !PAE enabled case, zero them out
+        * once at init, rather than checking/setting to 0 for every runtime op
+        */
+       if (is_isa_arcv2() && pae40_exist_but_not_enab()) {
+
+               if (IS_ENABLED(CONFIG_ARC_HAS_ICACHE))
+                       write_aux_reg(ARC_REG_IC_PTAG_HI, 0);
+
+               if (IS_ENABLED(CONFIG_ARC_HAS_DCACHE))
+                       write_aux_reg(ARC_REG_DC_PTAG_HI, 0);
+
+               if (l2_line_sz) {
+                       write_aux_reg(ARC_REG_SLC_RGN_END1, 0);
+                       write_aux_reg(ARC_REG_SLC_RGN_START1, 0);
+               }
+       }
 }
index 71d3efff99d35c56c6f30bea8a6ab1cf1702d753..e9d93604ad0ff7b76a5458235b3ab4d0ef18aefb 100644 (file)
@@ -153,6 +153,19 @@ static void _dma_cache_sync(phys_addr_t paddr, size_t size,
        }
 }
 
+/*
+ * arc_dma_map_page - map a portion of a page for streaming DMA
+ *
+ * Ensure that any data held in the cache is appropriately discarded
+ * or written back.
+ *
+ * The device owns this memory once this call has completed.  The CPU
+ * can regain ownership by calling dma_unmap_page().
+ *
+ * Note: while it takes struct page as arg, caller can "abuse" it to pass
+ * a region larger than PAGE_SIZE, provided it is physically contiguous
+ * and this still works correctly
+ */
 static dma_addr_t arc_dma_map_page(struct device *dev, struct page *page,
                unsigned long offset, size_t size, enum dma_data_direction dir,
                unsigned long attrs)
@@ -165,6 +178,24 @@ static dma_addr_t arc_dma_map_page(struct device *dev, struct page *page,
        return plat_phys_to_dma(dev, paddr);
 }
 
+/*
+ * arc_dma_unmap_page - unmap a buffer previously mapped through dma_map_page()
+ *
+ * After this call, reads by the CPU to the buffer are guaranteed to see
+ * whatever the device wrote there.
+ *
+ * Note: historically this routine was not implemented for ARC
+ */
+static void arc_dma_unmap_page(struct device *dev, dma_addr_t handle,
+                              size_t size, enum dma_data_direction dir,
+                              unsigned long attrs)
+{
+       phys_addr_t paddr = plat_dma_to_phys(dev, handle);
+
+       if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
+               _dma_cache_sync(paddr, size, dir);
+}
+
 static int arc_dma_map_sg(struct device *dev, struct scatterlist *sg,
           int nents, enum dma_data_direction dir, unsigned long attrs)
 {
@@ -178,6 +209,18 @@ static int arc_dma_map_sg(struct device *dev, struct scatterlist *sg,
        return nents;
 }
 
+static void arc_dma_unmap_sg(struct device *dev, struct scatterlist *sg,
+                            int nents, enum dma_data_direction dir,
+                            unsigned long attrs)
+{
+       struct scatterlist *s;
+       int i;
+
+       for_each_sg(sg, s, nents, i)
+               arc_dma_unmap_page(dev, sg_dma_address(s), sg_dma_len(s), dir,
+                                  attrs);
+}
+
 static void arc_dma_sync_single_for_cpu(struct device *dev,
                dma_addr_t dma_handle, size_t size, enum dma_data_direction dir)
 {
@@ -223,7 +266,9 @@ const struct dma_map_ops arc_dma_ops = {
        .free                   = arc_dma_free,
        .mmap                   = arc_dma_mmap,
        .map_page               = arc_dma_map_page,
+       .unmap_page             = arc_dma_unmap_page,
        .map_sg                 = arc_dma_map_sg,
+       .unmap_sg               = arc_dma_unmap_sg,
        .sync_single_for_device = arc_dma_sync_single_for_device,
        .sync_single_for_cpu    = arc_dma_sync_single_for_cpu,
        .sync_sg_for_cpu        = arc_dma_sync_sg_for_cpu,
index d0126fdfe2d8546d62883cbef419e4b76589cf18..b181f3ee38aab54565d8de8c1e5136375c6358cc 100644 (file)
 /* A copy of the ASID from the PID reg is kept in asid_cache */
 DEFINE_PER_CPU(unsigned int, asid_cache) = MM_CTXT_FIRST_CYCLE;
 
+static int __read_mostly pae_exists;
+
 /*
  * Utility Routine to erase a J-TLB entry
  * Caller needs to setup Index Reg (manually or via getIndex)
@@ -784,7 +786,7 @@ void read_decode_mmu_bcr(void)
                mmu->u_dtlb = mmu4->u_dtlb * 4;
                mmu->u_itlb = mmu4->u_itlb * 4;
                mmu->sasid = mmu4->sasid;
-               mmu->pae = mmu4->pae;
+               pae_exists = mmu->pae = mmu4->pae;
        }
 }
 
@@ -809,6 +811,11 @@ char *arc_mmu_mumbojumbo(int cpu_id, char *buf, int len)
        return buf;
 }
 
+int pae40_exist_but_not_enab(void)
+{
+       return pae_exists && !is_pae40_enabled();
+}
+
 void arc_mmu_init(void)
 {
        char str[256];
@@ -859,6 +866,9 @@ void arc_mmu_init(void)
        /* swapper_pg_dir is the pgd for the kernel, used by vmalloc */
        write_aux_reg(ARC_REG_SCRATCH_DATA0, swapper_pg_dir);
 #endif
+
+       if (pae40_exist_but_not_enab())
+               write_aux_reg(ARC_REG_TLBPD1HI, 0);
 }
 
 /*
diff --git a/arch/arc/plat-sim/Kconfig b/arch/arc/plat-sim/Kconfig
deleted file mode 100644 (file)
index ac6af96..0000000
+++ /dev/null
@@ -1,13 +0,0 @@
-#
-# Copyright (C) 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
-#
-# This program is free software; you can redistribute it and/or modify
-# it under the terms of the GNU General Public License version 2 as
-# published by the Free Software Foundation.
-#
-
-menuconfig ARC_PLAT_SIM
-       bool "ARC nSIM based simulation virtual platforms"
-       help
-         Support for nSIM based ARC simulation platforms
-         This includes the standalone nSIM (uart only) vs. System C OSCI VP
index aea87389e44bd517b196ec87de0edb4c0216b4b1..5cda56b1a2ead7bc54ca1ac63296fcd1f54fc43f 100644 (file)
  */
 
 static const char *simulation_compat[] __initconst = {
+#ifdef CONFIG_ISA_ARCOMPACT
        "snps,nsim",
-       "snps,nsim_hs",
        "snps,nsimosci",
+#else
+       "snps,nsim_hs",
        "snps,nsimosci_hs",
        "snps,zebu_hs",
+#endif
        NULL,
 };
 
index f92f957412073279a3be9713794bb2336dbf4752..a183b56283f8ff9e9d41d616edfc00a82d200267 100644 (file)
 
 &hdmicec {
        status = "okay";
+       needs-hpd;
 };
 
 &hsi2c_4 {
index dfcc8e00cf1c53a753dc753c2c5598b220b480a1..0ade3619f3c3f1332b8d89c8d06ed4a67f29025b 100644 (file)
                                #address-cells = <1>;
                                #size-cells = <1>;
                                status = "disabled";
+                               ranges;
 
                                adc: adc@50030800 {
                                        compatible = "fsl,imx25-gcq";
index aeaa5a6e4fcf462bc8a381ab20ba85c3579dea5c..a24e4f1911abe77917726d8ff8342f7dcb88dab2 100644 (file)
        pinctrl_pcie: pciegrp {
                fsl,pins = <
                        /* PCIe reset */
-                       MX6QDL_PAD_EIM_BCLK__GPIO6_IO31 0x030b0
+                       MX6QDL_PAD_EIM_DA0__GPIO3_IO00  0x030b0
                        MX6QDL_PAD_EIM_DA4__GPIO3_IO04  0x030b0
                >;
        };
 &pcie {
        pinctrl-names = "default";
        pinctrl-0 = <&pinctrl_pcie>;
-       reset-gpio = <&gpio6 31 GPIO_ACTIVE_LOW>;
+       reset-gpio = <&gpio3 0 GPIO_ACTIVE_LOW>;
        status = "okay";
 };
 
index 54c45402286b10dc06240aef9ec7d015a3fd5a58..0a24d1bf3c393463148919e685a879df39af58ef 100644 (file)
                        >;
                };
 
+               pinctrl_spi4: spi4grp {
+                       fsl,pins = <
+                               MX7D_PAD_GPIO1_IO09__GPIO1_IO9  0x59
+                               MX7D_PAD_GPIO1_IO12__GPIO1_IO12 0x59
+                               MX7D_PAD_GPIO1_IO13__GPIO1_IO13 0x59
+                       >;
+               };
+
                pinctrl_tsc2046_pendown: tsc2046_pendown {
                        fsl,pins = <
                                MX7D_PAD_EPDC_BDR1__GPIO2_IO29          0x59
                fsl,pins = <
                        MX7D_PAD_LPSR_GPIO1_IO01__PWM1_OUT              0x110b0
                >;
-
-               pinctrl_spi4: spi4grp {
-                       fsl,pins = <
-                               MX7D_PAD_GPIO1_IO09__GPIO1_IO9  0x59
-                               MX7D_PAD_GPIO1_IO12__GPIO1_IO12 0x59
-                               MX7D_PAD_GPIO1_IO13__GPIO1_IO13 0x59
-                       >;
-               };
        };
 };
index cc06da3943668415e9b6379dbe38d2fde36518ec..60e69aeacbdbf4dff78923f227f96f8bc49c1fbc 100644 (file)
                        #size-cells = <1>;
                        atmel,smc = <&hsmc>;
                        reg = <0x10000000 0x10000000
-                              0x40000000 0x30000000>;
+                              0x60000000 0x30000000>;
                        ranges = <0x0 0x0 0x10000000 0x10000000
                                  0x1 0x0 0x60000000 0x10000000
                                  0x2 0x0 0x70000000 0x10000000
                        };
 
                        hsmc: hsmc@f8014000 {
-                               compatible = "atmel,sama5d3-smc", "syscon", "simple-mfd";
+                               compatible = "atmel,sama5d2-smc", "syscon", "simple-mfd";
                                reg = <0xf8014000 0x1000>;
-                               interrupts = <5 IRQ_TYPE_LEVEL_HIGH 6>;
+                               interrupts = <17 IRQ_TYPE_LEVEL_HIGH 6>;
                                clocks = <&hsmc_clk>;
                                #address-cells = <1>;
                                #size-cells = <1>;
                                ranges;
 
-                               pmecc: ecc-engine@ffffc070 {
+                               pmecc: ecc-engine@f8014070 {
                                        compatible = "atmel,sama5d2-pmecc";
-                                       reg = <0xffffc070 0x490>,
-                                             <0xffffc500 0x100>;
+                                       reg = <0xf8014070 0x490>,
+                                             <0xf8014500 0x100>;
                                };
                        };
 
index d735e5fc47727b0536c943b6fdb15223d7394360..195da38cb9a220f22c2f890c5bcb8fc1f2c7c621 100644 (file)
@@ -1,7 +1,7 @@
 menuconfig ARCH_AT91
        bool "Atmel SoCs"
        depends on ARCH_MULTI_V4T || ARCH_MULTI_V5 || ARCH_MULTI_V7 || ARM_SINGLE_ARMV7M
-       select ARM_CPU_SUSPEND if PM
+       select ARM_CPU_SUSPEND if PM && ARCH_MULTI_V7
        select COMMON_CLK_AT91
        select GPIOLIB
        select PINCTRL
index 667fddac38561eff3f25788ccbf9e05ac8214d84..5036f996e694a533d66f82743b901385ae048007 100644 (file)
@@ -608,6 +608,9 @@ static void __init at91_pm_init(void (*pm_idle)(void))
 
 void __init at91rm9200_pm_init(void)
 {
+       if (!IS_ENABLED(CONFIG_SOC_AT91RM9200))
+               return;
+
        at91_dt_ramc();
 
        /*
@@ -620,18 +623,27 @@ void __init at91rm9200_pm_init(void)
 
 void __init at91sam9_pm_init(void)
 {
+       if (!IS_ENABLED(CONFIG_SOC_AT91SAM9))
+               return;
+
        at91_dt_ramc();
        at91_pm_init(at91sam9_idle);
 }
 
 void __init sama5_pm_init(void)
 {
+       if (!IS_ENABLED(CONFIG_SOC_SAMA5))
+               return;
+
        at91_dt_ramc();
        at91_pm_init(NULL);
 }
 
 void __init sama5d2_pm_init(void)
 {
+       if (!IS_ENABLED(CONFIG_SOC_SAMA5D2))
+               return;
+
        at91_pm_backup_init();
        sama5_pm_init();
 }
index 0d1f026d831aac7b7ecda69ac5ec0f57e693ec2b..ba2fde2909f949531bf4e63e25fe22856b0cec3b 100644 (file)
@@ -51,6 +51,7 @@
        compatible = "sinovoip,bananapi-m64", "allwinner,sun50i-a64";
 
        aliases {
+               ethernet0 = &emac;
                serial0 = &uart0;
                serial1 = &uart1;
        };
index 08cda24ea194cbffd08fc8de50dc7a4d525e5019..827168bc22ed2c6f9b831e4b3b92d75980becf1d 100644 (file)
@@ -51,6 +51,7 @@
        compatible = "pine64,pine64", "allwinner,sun50i-a64";
 
        aliases {
+               ethernet0 = &emac;
                serial0 = &uart0;
                serial1 = &uart1;
                serial2 = &uart2;
index 17eb1cc5bf6b4061a8ea8cf40e54b4eb52c1ebe3..216e3a5dafaef892fc764a076a0e0e5bdfaf396e 100644 (file)
@@ -53,6 +53,7 @@
                     "allwinner,sun50i-a64";
 
        aliases {
+               ethernet0 = &emac;
                serial0 = &uart0;
        };
 
index 732e2e06f503c83e42c0c4925b06295e6e47dc64..d9a720bff05d39ae93941140a3a8298f0bfa6570 100644 (file)
 };
 
 &pio {
+       interrupts = <GIC_SPI 11 IRQ_TYPE_LEVEL_HIGH>,
+                    <GIC_SPI 17 IRQ_TYPE_LEVEL_HIGH>,
+                    <GIC_SPI 23 IRQ_TYPE_LEVEL_HIGH>;
        compatible = "allwinner,sun50i-h5-pinctrl";
 };
index a451996f590a5173a3e9dbe56831a623e3c0639a..f903957da504a9e02fa910243b60388dc85ec959 100644 (file)
@@ -45,7 +45,7 @@
                stdout-path = "serial0:115200n8";
        };
 
-       audio_clkout: audio_clkout {
+       audio_clkout: audio-clkout {
                /*
                 * This is same as <&rcar_sound 0>
                 * but needed to avoid cs2000/rcar_sound probe dead-lock
index 74d08e44a651b9e58de7d8a6075f6101083f5262..a652ce0a5cb2c33178ac2e447fad1159efb0c6a9 100644 (file)
@@ -65,13 +65,13 @@ DECLARE_PER_CPU(const struct arch_timer_erratum_workaround *,
        u64 _val;                                                       \
        if (needs_unstable_timer_counter_workaround()) {                \
                const struct arch_timer_erratum_workaround *wa;         \
-               preempt_disable();                                      \
+               preempt_disable_notrace();                              \
                wa = __this_cpu_read(timer_unstable_counter_workaround); \
                if (wa && wa->read_##reg)                               \
                        _val = wa->read_##reg();                        \
                else                                                    \
                        _val = read_sysreg(reg);                        \
-               preempt_enable();                                       \
+               preempt_enable_notrace();                               \
        } else {                                                        \
                _val = read_sysreg(reg);                                \
        }                                                               \
index acae781f7359ece14f713e74e168deb6e47dd68b..3288c2b3673149b728ea52ee87bc65ba723c9811 100644 (file)
 
 /*
  * This is the base location for PIE (ET_DYN with INTERP) loads. On
- * 64-bit, this is raised to 4GB to leave the entire 32-bit address
+ * 64-bit, this is above 4GB to leave the entire 32-bit address
  * space open for things that want to use the area for 32-bit pointers.
  */
-#define ELF_ET_DYN_BASE                0x100000000UL
+#define ELF_ET_DYN_BASE                (2 * TASK_SIZE_64 / 3)
 
 #ifndef __ASSEMBLY__
 
index 06da8ea16bbe5e150bb7330ab0049c7438b25f8d..c7b4995868e14d78216c4b06ce87476dfe8a4789 100644 (file)
@@ -161,9 +161,11 @@ void fpsimd_flush_thread(void)
 {
        if (!system_supports_fpsimd())
                return;
+       preempt_disable();
        memset(&current->thread.fpsimd_state, 0, sizeof(struct fpsimd_state));
        fpsimd_flush_task_state(current);
        set_thread_flag(TIF_FOREIGN_FPSTATE);
+       preempt_enable();
 }
 
 /*
index 973df7de7bf8d3e4fcde6dc2ba1590fa051964a2..adb0910b88f5f26c1562a7fef6f00f29ee9ec391 100644 (file)
@@ -354,7 +354,6 @@ __primary_switched:
        tst     x23, ~(MIN_KIMG_ALIGN - 1)      // already running randomized?
        b.ne    0f
        mov     x0, x21                         // pass FDT address in x0
-       mov     x1, x23                         // pass modulo offset in x1
        bl      kaslr_early_init                // parse FDT for KASLR options
        cbz     x0, 0f                          // KASLR disabled? just proceed
        orr     x23, x23, x0                    // record KASLR offset
index a9710efb8c017a735f9aa6518463d11b8d6264c9..47080c49cc7e77a3df120f061f7b10f90a8c2ec5 100644 (file)
@@ -75,7 +75,7 @@ extern void *__init __fixmap_remap_fdt(phys_addr_t dt_phys, int *size,
  * containing function pointers) to be reinitialized, and zero-initialized
  * .bss variables will be reset to 0.
  */
-u64 __init kaslr_early_init(u64 dt_phys, u64 modulo_offset)
+u64 __init kaslr_early_init(u64 dt_phys)
 {
        void *fdt;
        u64 seed, offset, mask, module_range;
@@ -131,15 +131,17 @@ u64 __init kaslr_early_init(u64 dt_phys, u64 modulo_offset)
        /*
         * The kernel Image should not extend across a 1GB/32MB/512MB alignment
         * boundary (for 4KB/16KB/64KB granule kernels, respectively). If this
-        * happens, increase the KASLR offset by the size of the kernel image
-        * rounded up by SWAPPER_BLOCK_SIZE.
+        * happens, round down the KASLR offset by (1 << SWAPPER_TABLE_SHIFT).
+        *
+        * NOTE: The references to _text and _end below will already take the
+        *       modulo offset (the physical displacement modulo 2 MB) into
+        *       account, given that the physical placement is controlled by
+        *       the loader, and will not change as a result of the virtual
+        *       mapping we choose.
         */
-       if ((((u64)_text + offset + modulo_offset) >> SWAPPER_TABLE_SHIFT) !=
-           (((u64)_end + offset + modulo_offset) >> SWAPPER_TABLE_SHIFT)) {
-               u64 kimg_sz = _end - _text;
-               offset = (offset + round_up(kimg_sz, SWAPPER_BLOCK_SIZE))
-                               & mask;
-       }
+       if ((((u64)_text + offset) >> SWAPPER_TABLE_SHIFT) !=
+           (((u64)_end + offset) >> SWAPPER_TABLE_SHIFT))
+               offset = round_down(offset, 1 << SWAPPER_TABLE_SHIFT);
 
        if (IS_ENABLED(CONFIG_KASAN))
                /*
index 2509e4fe699225675f74876032e22b24b338a3b1..1f22a41565a38ac08083b29fcad216ca4ea1a2c9 100644 (file)
@@ -435,8 +435,11 @@ retry:
                 * the mmap_sem because it would already be released
                 * in __lock_page_or_retry in mm/filemap.c.
                 */
-               if (fatal_signal_pending(current))
+               if (fatal_signal_pending(current)) {
+                       if (!user_mode(regs))
+                               goto no_context;
                        return 0;
+               }
 
                /*
                 * Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk of
index 8dd20358464f8efabcad2185d2b6f56c4fe805fa..48d91d5be4e9b6cf56c1cd0849830e2708967aa2 100644 (file)
@@ -2260,7 +2260,7 @@ config CPU_R4K_CACHE_TLB
 
 config MIPS_MT_SMP
        bool "MIPS MT SMP support (1 TC on each available VPE)"
-       depends on SYS_SUPPORTS_MULTITHREADING && !CPU_MIPSR6
+       depends on SYS_SUPPORTS_MULTITHREADING && !CPU_MIPSR6 && !CPU_MICROMIPS
        select CPU_MIPSR2_IRQ_VI
        select CPU_MIPSR2_IRQ_EI
        select SYNC_R4K
index 04343625b9292807d886951eaed6866426dfc42d..bc2708c9ada40cf4206cd228d84a6ebf634161b3 100644 (file)
@@ -243,8 +243,21 @@ include arch/mips/Kbuild.platforms
 ifdef CONFIG_PHYSICAL_START
 load-y                                 = $(CONFIG_PHYSICAL_START)
 endif
-entry-y                                = 0x$(shell $(NM) vmlinux 2>/dev/null \
+
+entry-noisa-y                          = 0x$(shell $(NM) vmlinux 2>/dev/null \
                                        | grep "\bkernel_entry\b" | cut -f1 -d \ )
+ifdef CONFIG_CPU_MICROMIPS
+  #
+  # Set the ISA bit, since the kernel_entry symbol in the ELF will have it
+  # clear which would lead to images containing addresses which bootloaders may
+  # jump to as MIPS32 code.
+  #
+  entry-y = $(patsubst %0,%1,$(patsubst %2,%3,$(patsubst %4,%5, \
+              $(patsubst %6,%7,$(patsubst %8,%9,$(patsubst %a,%b, \
+              $(patsubst %c,%d,$(patsubst %e,%f,$(entry-noisa-y)))))))))
+else
+  entry-y = $(entry-noisa-y)
+endif
 
 cflags-y                       += -I$(srctree)/arch/mips/include/asm/mach-generic
 drivers-$(CONFIG_PCI)          += arch/mips/pci/
diff --git a/arch/mips/boot/compressed/.gitignore b/arch/mips/boot/compressed/.gitignore
new file mode 100644 (file)
index 0000000..ebae133
--- /dev/null
@@ -0,0 +1,2 @@
+ashldi3.c
+bswapsi.c
index 542be1cd0f32c6cbe14cb0c3936ffe9cd11bd594..bfdfaf32d2c49742066329e800a6b535f363c3e1 100644 (file)
@@ -13,9 +13,9 @@
 #include <linux/mutex.h>
 #include <linux/delay.h>
 #include <linux/of_platform.h>
+#include <linux/io.h>
 
 #include <asm/octeon/octeon.h>
-#include <asm/octeon/cvmx-gpio-defs.h>
 
 /* USB Control Register */
 union cvm_usbdrd_uctl_ctl {
index 1910223a9c02ba1474929905b235e43a07501aaa..cea2bb1621e68b211d5dc99fdad78f208b1ad8a3 100644 (file)
                 * Find irq with highest priority
                 */
                # open coded PTR_LA t1, cpu_mask_nr_tbl
-#if (_MIPS_SZPTR == 32)
+#if defined(CONFIG_32BIT) || defined(KBUILD_64BIT_SYM32)
                # open coded la t1, cpu_mask_nr_tbl
                lui     t1, %hi(cpu_mask_nr_tbl)
                addiu   t1, %lo(cpu_mask_nr_tbl)
-
-#endif
-#if (_MIPS_SZPTR == 64)
-               # open coded dla t1, cpu_mask_nr_tbl
-               .set    push
-               .set    noat
-               lui     t1, %highest(cpu_mask_nr_tbl)
-               lui     AT, %hi(cpu_mask_nr_tbl)
-               daddiu  t1, t1, %higher(cpu_mask_nr_tbl)
-               daddiu  AT, AT, %lo(cpu_mask_nr_tbl)
-               dsll    t1, 32
-               daddu   t1, t1, AT
-               .set    pop
+#else
+#error GCC `-msym32' option required for 64-bit DECstation builds
 #endif
 1:             lw      t2,(t1)
                nop
                 * Find irq with highest priority
                 */
                # open coded PTR_LA t1,asic_mask_nr_tbl
-#if (_MIPS_SZPTR == 32)
+#if defined(CONFIG_32BIT) || defined(KBUILD_64BIT_SYM32)
                # open coded la t1, asic_mask_nr_tbl
                lui     t1, %hi(asic_mask_nr_tbl)
                addiu   t1, %lo(asic_mask_nr_tbl)
-
-#endif
-#if (_MIPS_SZPTR == 64)
-               # open coded dla t1, asic_mask_nr_tbl
-               .set    push
-               .set    noat
-               lui     t1, %highest(asic_mask_nr_tbl)
-               lui     AT, %hi(asic_mask_nr_tbl)
-               daddiu  t1, t1, %higher(asic_mask_nr_tbl)
-               daddiu  AT, AT, %lo(asic_mask_nr_tbl)
-               dsll    t1, 32
-               daddu   t1, t1, AT
-               .set    pop
+#else
+#error GCC `-msym32' option required for 64-bit DECstation builds
 #endif
 2:             lw      t2,(t1)
                nop
index fc67947ed6586c2e416b2f6cc588a4445b3761ab..8b14c2706aa52ca83e07ff4dcb37c71850d133bc 100644 (file)
@@ -9,6 +9,8 @@
 #ifndef _ASM_CACHE_H
 #define _ASM_CACHE_H
 
+#include <kmalloc.h>
+
 #define L1_CACHE_SHIFT         CONFIG_MIPS_L1_CACHE_SHIFT
 #define L1_CACHE_BYTES         (1 << L1_CACHE_SHIFT)
 
index 8baa9033b181d2ca2e3f6e469315bdb8bf5345b5..721b698bfe3cf7e0274bb9f0bb4bf58bfecef29e 100644 (file)
 #ifndef cpu_scache_line_size
 #define cpu_scache_line_size() cpu_data[0].scache.linesz
 #endif
+#ifndef cpu_tcache_line_size
+#define cpu_tcache_line_size() cpu_data[0].tcache.linesz
+#endif
 
 #ifndef cpu_hwrena_impl_bits
 #define cpu_hwrena_impl_bits           0
index d045973ddb336abbe9b54b21a65b4fb1a6c4f5fe..3ea84acf1814cb2c874ac17dafc73c64db50b5dc 100644 (file)
 #define CVMX_L2C_DBG (CVMX_ADD_IO_SEG(0x0001180080000030ull))
 #define CVMX_L2C_CFG (CVMX_ADD_IO_SEG(0x0001180080000000ull))
 #define CVMX_L2C_CTL (CVMX_ADD_IO_SEG(0x0001180080800000ull))
+#define CVMX_L2C_ERR_TDTX(block_id)                                           \
+       (CVMX_ADD_IO_SEG(0x0001180080A007E0ull) + ((block_id) & 3) * 0x40000ull)
+#define CVMX_L2C_ERR_TTGX(block_id)                                           \
+       (CVMX_ADD_IO_SEG(0x0001180080A007E8ull) + ((block_id) & 3) * 0x40000ull)
 #define CVMX_L2C_LCKBASE (CVMX_ADD_IO_SEG(0x0001180080000058ull))
 #define CVMX_L2C_LCKOFF (CVMX_ADD_IO_SEG(0x0001180080000060ull))
 #define CVMX_L2C_PFCTL (CVMX_ADD_IO_SEG(0x0001180080000090ull))
                ((offset) & 1) * 8)
 #define CVMX_L2C_WPAR_PPX(offset) (CVMX_ADD_IO_SEG(0x0001180080840000ull)    + \
                ((offset) & 31) * 8)
-#define CVMX_L2D_FUS3 (CVMX_ADD_IO_SEG(0x00011800800007B8ull))
 
 
+union cvmx_l2c_err_tdtx {
+       uint64_t u64;
+       struct cvmx_l2c_err_tdtx_s {
+               __BITFIELD_FIELD(uint64_t dbe:1,
+               __BITFIELD_FIELD(uint64_t sbe:1,
+               __BITFIELD_FIELD(uint64_t vdbe:1,
+               __BITFIELD_FIELD(uint64_t vsbe:1,
+               __BITFIELD_FIELD(uint64_t syn:10,
+               __BITFIELD_FIELD(uint64_t reserved_22_49:28,
+               __BITFIELD_FIELD(uint64_t wayidx:18,
+               __BITFIELD_FIELD(uint64_t reserved_2_3:2,
+               __BITFIELD_FIELD(uint64_t type:2,
+               ;)))))))))
+       } s;
+};
+
+union cvmx_l2c_err_ttgx {
+       uint64_t u64;
+       struct cvmx_l2c_err_ttgx_s {
+               __BITFIELD_FIELD(uint64_t dbe:1,
+               __BITFIELD_FIELD(uint64_t sbe:1,
+               __BITFIELD_FIELD(uint64_t noway:1,
+               __BITFIELD_FIELD(uint64_t reserved_56_60:5,
+               __BITFIELD_FIELD(uint64_t syn:6,
+               __BITFIELD_FIELD(uint64_t reserved_22_49:28,
+               __BITFIELD_FIELD(uint64_t wayidx:15,
+               __BITFIELD_FIELD(uint64_t reserved_2_6:5,
+               __BITFIELD_FIELD(uint64_t type:2,
+               ;)))))))))
+       } s;
+};
+
 union cvmx_l2c_cfg {
        uint64_t u64;
        struct cvmx_l2c_cfg_s {
diff --git a/arch/mips/include/asm/octeon/cvmx-l2d-defs.h b/arch/mips/include/asm/octeon/cvmx-l2d-defs.h
new file mode 100644 (file)
index 0000000..a951ad5
--- /dev/null
@@ -0,0 +1,60 @@
+/***********************license start***************
+ * Author: Cavium Networks
+ *
+ * Contact: support@caviumnetworks.com
+ * This file is part of the OCTEON SDK
+ *
+ * Copyright (c) 2003-2017 Cavium, Inc.
+ *
+ * This file is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, Version 2, as
+ * published by the Free Software Foundation.
+ *
+ * This file is distributed in the hope that it will be useful, but
+ * AS-IS and WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, TITLE, or
+ * NONINFRINGEMENT.  See the GNU General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this file; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ * or visit http://www.gnu.org/licenses/.
+ *
+ * This file may also be available under a different license from Cavium.
+ * Contact Cavium Networks for more information
+ ***********************license end**************************************/
+
+#ifndef __CVMX_L2D_DEFS_H__
+#define __CVMX_L2D_DEFS_H__
+
+#define CVMX_L2D_ERR   (CVMX_ADD_IO_SEG(0x0001180080000010ull))
+#define CVMX_L2D_FUS3  (CVMX_ADD_IO_SEG(0x00011800800007B8ull))
+
+
+union cvmx_l2d_err {
+       uint64_t u64;
+       struct cvmx_l2d_err_s {
+               __BITFIELD_FIELD(uint64_t reserved_6_63:58,
+               __BITFIELD_FIELD(uint64_t bmhclsel:1,
+               __BITFIELD_FIELD(uint64_t ded_err:1,
+               __BITFIELD_FIELD(uint64_t sec_err:1,
+               __BITFIELD_FIELD(uint64_t ded_intena:1,
+               __BITFIELD_FIELD(uint64_t sec_intena:1,
+               __BITFIELD_FIELD(uint64_t ecc_ena:1,
+               ;)))))))
+       } s;
+};
+
+union cvmx_l2d_fus3 {
+       uint64_t u64;
+       struct cvmx_l2d_fus3_s {
+               __BITFIELD_FIELD(uint64_t reserved_40_63:24,
+               __BITFIELD_FIELD(uint64_t ema_ctl:3,
+               __BITFIELD_FIELD(uint64_t reserved_34_36:3,
+               __BITFIELD_FIELD(uint64_t q3fus:34,
+               ;))))
+       } s;
+};
+
+#endif
index 9742202f2a326c45c504de6d77b51eb523421c0c..e638735cc3ac56c84fc861f293e6f2217b7ba3cb 100644 (file)
@@ -62,6 +62,7 @@ enum cvmx_mips_space {
 #include <asm/octeon/cvmx-iob-defs.h>
 #include <asm/octeon/cvmx-ipd-defs.h>
 #include <asm/octeon/cvmx-l2c-defs.h>
+#include <asm/octeon/cvmx-l2d-defs.h>
 #include <asm/octeon/cvmx-l2t-defs.h>
 #include <asm/octeon/cvmx-led-defs.h>
 #include <asm/octeon/cvmx-mio-defs.h>
index 770d4d1516cbb1ff34cca30cdccbe6c1b0aeebdc..6bace7695788fbc3b7663aeb353a08a3c45503af 100644 (file)
@@ -376,9 +376,6 @@ asmlinkage void start_secondary(void)
        cpumask_set_cpu(cpu, &cpu_coherent_mask);
        notify_cpu_starting(cpu);
 
-       complete(&cpu_running);
-       synchronise_count_slave(cpu);
-
        set_cpu_online(cpu, true);
 
        set_cpu_sibling_map(cpu);
@@ -386,6 +383,9 @@ asmlinkage void start_secondary(void)
 
        calculate_cpu_foreign_map();
 
+       complete(&cpu_running);
+       synchronise_count_slave(cpu);
+
        /*
         * irq will be enabled in ->smp_finish(), enabling it too early
         * is dangerous.
index 3f74f6c1f065fb6b5d2b930dc1d164e1624e63f0..9fea6c6bbf49e3768e81ad3ffb8b8af2bbdef61f 100644 (file)
@@ -48,7 +48,7 @@
 
 #include "uasm.c"
 
-static const struct insn const insn_table[insn_invalid] = {
+static const struct insn insn_table[insn_invalid] = {
        [insn_addiu]    = {M(addiu_op, 0, 0, 0, 0, 0), RS | RT | SIMM},
        [insn_addu]     = {M(spec_op, 0, 0, 0, 0, addu_op), RS | RT | RD},
        [insn_and]      = {M(spec_op, 0, 0, 0, 0, and_op), RS | RT | RD},
index bd67ac74fe2d3420509a03647a20856341bb70c8..9632436d74d7a74b3d584ab6e87a1fc7e55827cc 100644 (file)
@@ -28,16 +28,15 @@ EXPORT_SYMBOL(PCIBIOS_MIN_MEM);
 
 static int __init pcibios_set_cache_line_size(void)
 {
-       struct cpuinfo_mips *c = &current_cpu_data;
        unsigned int lsize;
 
        /*
         * Set PCI cacheline size to that of the highest level in the
         * cache hierarchy.
         */
-       lsize = c->dcache.linesz;
-       lsize = c->scache.linesz ? : lsize;
-       lsize = c->tcache.linesz ? : lsize;
+       lsize = cpu_dcache_line_size();
+       lsize = cpu_scache_line_size() ? : lsize;
+       lsize = cpu_tcache_line_size() ? : lsize;
 
        BUG_ON(!lsize);
 
index 974276e828b2cdd3eecb07493f480a8779da2d41..e2690d7ca4ddd992e69fffe802cac6355ad14de8 100644 (file)
@@ -35,7 +35,8 @@ static __always_inline long gettimeofday_fallback(struct timeval *_tv,
        "       syscall\n"
        : "=r" (ret), "=r" (error)
        : "r" (tv), "r" (tz), "r" (nr)
-       : "memory");
+       : "$1", "$3", "$8", "$9", "$10", "$11", "$12", "$13",
+         "$14", "$15", "$24", "$25", "hi", "lo", "memory");
 
        return error ? -ret : ret;
 }
@@ -55,7 +56,8 @@ static __always_inline long clock_gettime_fallback(clockid_t _clkid,
        "       syscall\n"
        : "=r" (ret), "=r" (error)
        : "r" (clkid), "r" (ts), "r" (nr)
-       : "memory");
+       : "$1", "$3", "$8", "$9", "$10", "$11", "$12", "$13",
+         "$14", "$15", "$24", "$25", "hi", "lo", "memory");
 
        return error ? -ret : ret;
 }
index 36f858c37ca70b576e851a52ada48e2400de86a1..81b0031f909f6d65855ed3051505a4baa9469a64 100644 (file)
@@ -199,7 +199,7 @@ config PPC
        select HAVE_OPTPROBES                   if PPC64
        select HAVE_PERF_EVENTS
        select HAVE_PERF_EVENTS_NMI             if PPC64
-       select HAVE_HARDLOCKUP_DETECTOR_PERF    if HAVE_PERF_EVENTS_NMI && !HAVE_HARDLOCKUP_DETECTOR_ARCH
+       select HAVE_HARDLOCKUP_DETECTOR_PERF    if PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !HAVE_HARDLOCKUP_DETECTOR_ARCH
        select HAVE_PERF_REGS
        select HAVE_PERF_USER_STACK_DUMP
        select HAVE_RCU_TABLE_FREE              if SMP
index 0695ce047d565199e4501333fa41ece48cdf9e45..34fc9bbfca9e68d6372e1d34b79ebf95d978e685 100644 (file)
@@ -293,7 +293,8 @@ CONFIG_MAGIC_SYSRQ=y
 CONFIG_DEBUG_KERNEL=y
 CONFIG_DEBUG_STACK_USAGE=y
 CONFIG_DEBUG_STACKOVERFLOW=y
-CONFIG_LOCKUP_DETECTOR=y
+CONFIG_SOFTLOCKUP_DETECTOR=y
+CONFIG_HARDLOCKUP_DETECTOR=y
 CONFIG_LATENCYTOP=y
 CONFIG_SCHED_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
index 5175028c56ce74e3e50a2b30eabccf7b87ed8f0e..c5246d29f3859965316bd4d48e4e816283439bf0 100644 (file)
@@ -324,7 +324,8 @@ CONFIG_MAGIC_SYSRQ=y
 CONFIG_DEBUG_KERNEL=y
 CONFIG_DEBUG_STACK_USAGE=y
 CONFIG_DEBUG_STACKOVERFLOW=y
-CONFIG_LOCKUP_DETECTOR=y
+CONFIG_SOFTLOCKUP_DETECTOR=y
+CONFIG_HARDLOCKUP_DETECTOR=y
 CONFIG_DEBUG_MUTEXES=y
 CONFIG_LATENCYTOP=y
 CONFIG_SCHED_TRACER=y
index 1a61aa20dfbac9d5072ae83ef90640b8be380bd3..fd5d98a0b95c7b1ae5fda56892c2ecd43ea29f3a 100644 (file)
@@ -291,7 +291,8 @@ CONFIG_MAGIC_SYSRQ=y
 CONFIG_DEBUG_KERNEL=y
 CONFIG_DEBUG_STACK_USAGE=y
 CONFIG_DEBUG_STACKOVERFLOW=y
-CONFIG_LOCKUP_DETECTOR=y
+CONFIG_SOFTLOCKUP_DETECTOR=y
+CONFIG_HARDLOCKUP_DETECTOR=y
 CONFIG_LATENCYTOP=y
 CONFIG_SCHED_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
index 49d8422767b4de686ec0ee64fbf69ac415f05003..e925c1c99c71cab982967e7f7df6325e3135506f 100644 (file)
@@ -223,17 +223,27 @@ system_call_exit:
        andi.   r0,r9,(_TIF_SYSCALL_DOTRACE|_TIF_SINGLESTEP|_TIF_USER_WORK_MASK|_TIF_PERSYSCALL_MASK)
        bne-    .Lsyscall_exit_work
 
-       /* If MSR_FP and MSR_VEC are set in user msr, then no need to restore */
-       li      r7,MSR_FP
+       andi.   r0,r8,MSR_FP
+       beq 2f
 #ifdef CONFIG_ALTIVEC
-       oris    r7,r7,MSR_VEC@h
+       andis.  r0,r8,MSR_VEC@h
+       bne     3f
 #endif
-       and     r0,r8,r7
-       cmpd    r0,r7
-       bne     .Lsyscall_restore_math
-.Lsyscall_restore_math_cont:
+2:     addi    r3,r1,STACK_FRAME_OVERHEAD
+#ifdef CONFIG_PPC_BOOK3S
+       li      r10,MSR_RI
+       mtmsrd  r10,1           /* Restore RI */
+#endif
+       bl      restore_math
+#ifdef CONFIG_PPC_BOOK3S
+       li      r11,0
+       mtmsrd  r11,1
+#endif
+       ld      r8,_MSR(r1)
+       ld      r3,RESULT(r1)
+       li      r11,-MAX_ERRNO
 
-       cmpld   r3,r11
+3:     cmpld   r3,r11
        ld      r5,_CCR(r1)
        bge-    .Lsyscall_error
 .Lsyscall_error_cont:
@@ -267,40 +277,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
        std     r5,_CCR(r1)
        b       .Lsyscall_error_cont
 
-.Lsyscall_restore_math:
-       /*
-        * Some initial tests from restore_math to avoid the heavyweight
-        * C code entry and MSR manipulations.
-        */
-       LOAD_REG_IMMEDIATE(r0, MSR_TS_MASK)
-       and.    r0,r0,r8
-       bne     1f
-
-       ld      r7,PACACURRENT(r13)
-       lbz     r0,THREAD+THREAD_LOAD_FP(r7)
-#ifdef CONFIG_ALTIVEC
-       lbz     r6,THREAD+THREAD_LOAD_VEC(r7)
-       add     r0,r0,r6
-#endif
-       cmpdi   r0,0
-       beq     .Lsyscall_restore_math_cont
-
-1:     addi    r3,r1,STACK_FRAME_OVERHEAD
-#ifdef CONFIG_PPC_BOOK3S
-       li      r10,MSR_RI
-       mtmsrd  r10,1           /* Restore RI */
-#endif
-       bl      restore_math
-#ifdef CONFIG_PPC_BOOK3S
-       li      r11,0
-       mtmsrd  r11,1
-#endif
-       /* Restore volatiles, reload MSR from updated one */
-       ld      r8,_MSR(r1)
-       ld      r3,RESULT(r1)
-       li      r11,-MAX_ERRNO
-       b       .Lsyscall_restore_math_cont
-
 /* Traced system call support */
 .Lsyscall_dotrace:
        bl      save_nvgprs
index 9f3e2c932dccc1c3a1158fc174a8cf57e63dd75d..1f0fd361e09b9415f81242d67d9b939f36fd63ba 100644 (file)
@@ -362,7 +362,8 @@ void enable_kernel_vsx(void)
 
        cpumsr = msr_check_and_set(MSR_FP|MSR_VEC|MSR_VSX);
 
-       if (current->thread.regs && (current->thread.regs->msr & MSR_VSX)) {
+       if (current->thread.regs &&
+           (current->thread.regs->msr & (MSR_VSX|MSR_VEC|MSR_FP))) {
                check_if_tm_restore_required(current);
                /*
                 * If a thread has already been reclaimed then the
@@ -386,7 +387,7 @@ void flush_vsx_to_thread(struct task_struct *tsk)
 {
        if (tsk->thread.regs) {
                preempt_disable();
-               if (tsk->thread.regs->msr & MSR_VSX) {
+               if (tsk->thread.regs->msr & (MSR_VSX|MSR_VEC|MSR_FP)) {
                        BUG_ON(tsk != current);
                        giveup_vsx(tsk);
                }
@@ -511,10 +512,6 @@ void restore_math(struct pt_regs *regs)
 {
        unsigned long msr;
 
-       /*
-        * Syscall exit makes a similar initial check before branching
-        * to restore_math. Keep them in synch.
-        */
        if (!msr_tm_active(regs->msr) &&
                !current->thread.load_fp && !loadvec(current->thread))
                return;
index cf0e1245b8cc1c78948a4004be2d20c5b5ac0b78..8d3320562c70f3ef7308645fb7b805fc14794e42 100644 (file)
@@ -351,7 +351,7 @@ static void nmi_ipi_lock_start(unsigned long *flags)
        hard_irq_disable();
        while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1) {
                raw_local_irq_restore(*flags);
-               cpu_relax();
+               spin_until_cond(atomic_read(&__nmi_ipi_lock) == 0);
                raw_local_irq_save(*flags);
                hard_irq_disable();
        }
@@ -360,7 +360,7 @@ static void nmi_ipi_lock_start(unsigned long *flags)
 static void nmi_ipi_lock(void)
 {
        while (atomic_cmpxchg(&__nmi_ipi_lock, 0, 1) == 1)
-               cpu_relax();
+               spin_until_cond(atomic_read(&__nmi_ipi_lock) == 0);
 }
 
 static void nmi_ipi_unlock(void)
@@ -475,7 +475,7 @@ int smp_send_nmi_ipi(int cpu, void (*fn)(struct pt_regs *), u64 delay_us)
        nmi_ipi_lock_start(&flags);
        while (nmi_ipi_busy_count) {
                nmi_ipi_unlock_end(&flags);
-               cpu_relax();
+               spin_until_cond(nmi_ipi_busy_count == 0);
                nmi_ipi_lock_start(&flags);
        }
 
index b67f8b03a32d0f12ce29eeb4ac3be0a97384fe72..34721a257a770c450baac0288f8006c94ff1975b 100644 (file)
@@ -71,15 +71,20 @@ static inline void wd_smp_lock(unsigned long *flags)
         * This may be called from low level interrupt handlers at some
         * point in future.
         */
-       local_irq_save(*flags);
-       while (unlikely(test_and_set_bit_lock(0, &__wd_smp_lock)))
-               cpu_relax();
+       raw_local_irq_save(*flags);
+       hard_irq_disable(); /* Make it soft-NMI safe */
+       while (unlikely(test_and_set_bit_lock(0, &__wd_smp_lock))) {
+               raw_local_irq_restore(*flags);
+               spin_until_cond(!test_bit(0, &__wd_smp_lock));
+               raw_local_irq_save(*flags);
+               hard_irq_disable();
+       }
 }
 
 static inline void wd_smp_unlock(unsigned long *flags)
 {
        clear_bit_unlock(0, &__wd_smp_lock);
-       local_irq_restore(*flags);
+       raw_local_irq_restore(*flags);
 }
 
 static void wd_lockup_ipi(struct pt_regs *regs)
@@ -96,10 +101,10 @@ static void wd_lockup_ipi(struct pt_regs *regs)
                nmi_panic(regs, "Hard LOCKUP");
 }
 
-static void set_cpu_stuck(int cpu, u64 tb)
+static void set_cpumask_stuck(const struct cpumask *cpumask, u64 tb)
 {
-       cpumask_set_cpu(cpu, &wd_smp_cpus_stuck);
-       cpumask_clear_cpu(cpu, &wd_smp_cpus_pending);
+       cpumask_or(&wd_smp_cpus_stuck, &wd_smp_cpus_stuck, cpumask);
+       cpumask_andnot(&wd_smp_cpus_pending, &wd_smp_cpus_pending, cpumask);
        if (cpumask_empty(&wd_smp_cpus_pending)) {
                wd_smp_last_reset_tb = tb;
                cpumask_andnot(&wd_smp_cpus_pending,
@@ -107,6 +112,10 @@ static void set_cpu_stuck(int cpu, u64 tb)
                                &wd_smp_cpus_stuck);
        }
 }
+static void set_cpu_stuck(int cpu, u64 tb)
+{
+       set_cpumask_stuck(cpumask_of(cpu), tb);
+}
 
 static void watchdog_smp_panic(int cpu, u64 tb)
 {
@@ -135,11 +144,9 @@ static void watchdog_smp_panic(int cpu, u64 tb)
        }
        smp_flush_nmi_ipi(1000000);
 
-       /* Take the stuck CPU out of the watch group */
-       for_each_cpu(c, &wd_smp_cpus_pending)
-               set_cpu_stuck(c, tb);
+       /* Take the stuck CPUs out of the watch group */
+       set_cpumask_stuck(&wd_smp_cpus_pending, tb);
 
-out:
        wd_smp_unlock(&flags);
 
        printk_safe_flush();
@@ -152,6 +159,11 @@ out:
 
        if (hardlockup_panic)
                nmi_panic(NULL, "Hard LOCKUP");
+
+       return;
+
+out:
+       wd_smp_unlock(&flags);
 }
 
 static void wd_smp_clear_cpu_pending(int cpu, u64 tb)
@@ -258,9 +270,11 @@ static void wd_timer_fn(unsigned long data)
 
 void arch_touch_nmi_watchdog(void)
 {
+       unsigned long ticks = tb_ticks_per_usec * wd_timer_period_ms * 1000;
        int cpu = smp_processor_id();
 
-       watchdog_timer_interrupt(cpu);
+       if (get_tb() - per_cpu(wd_timer_tb, cpu) >= ticks)
+               watchdog_timer_interrupt(cpu);
 }
 EXPORT_SYMBOL(arch_touch_nmi_watchdog);
 
@@ -283,6 +297,8 @@ static void stop_watchdog_timer_on(unsigned int cpu)
 
 static int start_wd_on_cpu(unsigned int cpu)
 {
+       unsigned long flags;
+
        if (cpumask_test_cpu(cpu, &wd_cpus_enabled)) {
                WARN_ON(1);
                return 0;
@@ -297,12 +313,14 @@ static int start_wd_on_cpu(unsigned int cpu)
        if (!cpumask_test_cpu(cpu, &watchdog_cpumask))
                return 0;
 
+       wd_smp_lock(&flags);
        cpumask_set_cpu(cpu, &wd_cpus_enabled);
        if (cpumask_weight(&wd_cpus_enabled) == 1) {
                cpumask_set_cpu(cpu, &wd_smp_cpus_pending);
                wd_smp_last_reset_tb = get_tb();
        }
-       smp_wmb();
+       wd_smp_unlock(&flags);
+
        start_watchdog_timer_on(cpu);
 
        return 0;
@@ -310,12 +328,17 @@ static int start_wd_on_cpu(unsigned int cpu)
 
 static int stop_wd_on_cpu(unsigned int cpu)
 {
+       unsigned long flags;
+
        if (!cpumask_test_cpu(cpu, &wd_cpus_enabled))
                return 0; /* Can happen in CPU unplug case */
 
        stop_watchdog_timer_on(cpu);
 
+       wd_smp_lock(&flags);
        cpumask_clear_cpu(cpu, &wd_cpus_enabled);
+       wd_smp_unlock(&flags);
+
        wd_smp_clear_cpu_pending(cpu, get_tb());
 
        return 0;
index 2abee070373fb3a8b757b8d3cb269e5d0b89dff6..a553aeea7af683812ba2f5a80d65e97cda163919 100644 (file)
@@ -56,6 +56,7 @@ u64 pnv_first_deep_stop_state = MAX_STOP_STATE;
  */
 static u64 pnv_deepest_stop_psscr_val;
 static u64 pnv_deepest_stop_psscr_mask;
+static u64 pnv_deepest_stop_flag;
 static bool deepest_stop_found;
 
 static int pnv_save_sprs_for_deep_states(void)
@@ -185,8 +186,40 @@ static void pnv_alloc_idle_core_states(void)
 
        update_subcore_sibling_mask();
 
-       if (supported_cpuidle_states & OPAL_PM_LOSE_FULL_CONTEXT)
-               pnv_save_sprs_for_deep_states();
+       if (supported_cpuidle_states & OPAL_PM_LOSE_FULL_CONTEXT) {
+               int rc = pnv_save_sprs_for_deep_states();
+
+               if (likely(!rc))
+                       return;
+
+               /*
+                * The stop-api is unable to restore hypervisor
+                * resources on wakeup from platform idle states which
+                * lose full context. So disable such states.
+                */
+               supported_cpuidle_states &= ~OPAL_PM_LOSE_FULL_CONTEXT;
+               pr_warn("cpuidle-powernv: Disabling idle states that lose full context\n");
+               pr_warn("cpuidle-powernv: Idle power-savings, CPU-Hotplug affected\n");
+
+               if (cpu_has_feature(CPU_FTR_ARCH_300) &&
+                   (pnv_deepest_stop_flag & OPAL_PM_LOSE_FULL_CONTEXT)) {
+                       /*
+                        * Use the default stop state for CPU-Hotplug
+                        * if available.
+                        */
+                       if (default_stop_found) {
+                               pnv_deepest_stop_psscr_val =
+                                       pnv_default_stop_val;
+                               pnv_deepest_stop_psscr_mask =
+                                       pnv_default_stop_mask;
+                               pr_warn("cpuidle-powernv: Offlined CPUs will stop with psscr = 0x%016llx\n",
+                                       pnv_deepest_stop_psscr_val);
+                       } else { /* Fallback to snooze loop for CPU-Hotplug */
+                               deepest_stop_found = false;
+                               pr_warn("cpuidle-powernv: Offlined CPUs will busy wait\n");
+                       }
+               }
+       }
 }
 
 u32 pnv_get_supported_cpuidle_states(void)
@@ -375,7 +408,8 @@ unsigned long pnv_cpu_offline(unsigned int cpu)
                                                pnv_deepest_stop_psscr_val;
                srr1 = power9_idle_stop(psscr);
 
-       } else if (idle_states & OPAL_PM_WINKLE_ENABLED) {
+       } else if ((idle_states & OPAL_PM_WINKLE_ENABLED) &&
+                  (idle_states & OPAL_PM_LOSE_FULL_CONTEXT)) {
                srr1 = power7_idle_insn(PNV_THREAD_WINKLE);
        } else if ((idle_states & OPAL_PM_SLEEP_ENABLED) ||
                   (idle_states & OPAL_PM_SLEEP_ENABLED_ER1)) {
@@ -553,6 +587,7 @@ static int __init pnv_power9_idle_init(struct device_node *np, u32 *flags,
                        max_residency_ns = residency_ns[i];
                        pnv_deepest_stop_psscr_val = psscr_val[i];
                        pnv_deepest_stop_psscr_mask = psscr_mask[i];
+                       pnv_deepest_stop_flag = flags[i];
                        deepest_stop_found = true;
                }
 
index 0efd0583a8c9da1afc2acc508f2e2e1a5388441d..6249214148c24bb44f2203117200768a410cee2d 100644 (file)
@@ -68,6 +68,7 @@ typedef struct { unsigned long iopgprot; } iopgprot_t;
 #define iopgprot_val(x)        ((x).iopgprot)
 
 #define __pte(x)       ((pte_t) { (x) } )
+#define __pmd(x)       ((pmd_t) { { (x) }, })
 #define __iopte(x)     ((iopte_t) { (x) } )
 #define __pgd(x)       ((pgd_t) { (x) } )
 #define __ctxd(x)      ((ctxd_t) { (x) } )
@@ -95,6 +96,7 @@ typedef unsigned long iopgprot_t;
 #define iopgprot_val(x)        (x)
 
 #define __pte(x)       (x)
+#define __pmd(x)       ((pmd_t) { { (x) }, })
 #define __iopte(x)     (x)
 #define __pgd(x)       (x)
 #define __ctxd(x)      (x)
index f10e2f7123949dedc8904ea37895d402ed097d2b..9ebebf1fd93d2f5db8c09804c336b006041b28af 100644 (file)
@@ -1266,8 +1266,6 @@ static int pci_sun4v_probe(struct platform_device *op)
                         * ATU group, but ATU hcalls won't be available.
                         */
                        hv_atu = false;
-                       pr_err(PFX "Could not register hvapi ATU err=%d\n",
-                              err);
                } else {
                        pr_info(PFX "Registered hvapi ATU major[%lu] minor[%lu]\n",
                                vatu_major, vatu_minor);
index a38787b843220807d0f5527f864981fc75e8486f..732af9a9f6ddef510e39da7d6eca5547e210296e 100644 (file)
@@ -602,7 +602,7 @@ void pcibios_fixup_bus(struct pci_bus *bus)
 {
        struct pci_dev *dev;
        int i, has_io, has_mem;
-       unsigned int cmd;
+       unsigned int cmd = 0;
        struct linux_pcic *pcic;
        /* struct linux_pbm_info* pbm = &pcic->pbm; */
        int node;
index d6b6c97fe3c73061f911737a5da23ab41a3eab77..703127aaf4a5015ee4a1099e6129735f7162222b 100644 (file)
@@ -5,26 +5,26 @@
        .align  4
 ENTRY(__multi3) /* %o0 = u, %o1 = v */
        mov     %o1, %g1
-       srl     %o3, 0, %g4
-       mulx    %g4, %g1, %o1
+       srl     %o3, 0, %o4
+       mulx    %o4, %g1, %o1
        srlx    %g1, 0x20, %g3
-       mulx    %g3, %g4, %g5
-       sllx    %g5, 0x20, %o5
-       srl     %g1, 0, %g4
+       mulx    %g3, %o4, %g7
+       sllx    %g7, 0x20, %o5
+       srl     %g1, 0, %o4
        sub     %o1, %o5, %o5
        srlx    %o5, 0x20, %o5
-       addcc   %g5, %o5, %g5
+       addcc   %g7, %o5, %g7
        srlx    %o3, 0x20, %o5
-       mulx    %g4, %o5, %g4
+       mulx    %o4, %o5, %o4
        mulx    %g3, %o5, %o5
        sethi   %hi(0x80000000), %g3
-       addcc   %g5, %g4, %g5
-       srlx    %g5, 0x20, %g5
+       addcc   %g7, %o4, %g7
+       srlx    %g7, 0x20, %g7
        add     %g3, %g3, %g3
        movcc   %xcc, %g0, %g3
-       addcc   %o5, %g5, %o5
-       sllx    %g4, 0x20, %g4
-       add     %o1, %g4, %o1
+       addcc   %o5, %g7, %o5
+       sllx    %o4, 0x20, %o4
+       add     %o1, %o4, %o1
        add     %o5, %g3, %g2
        mulx    %g1, %o2, %g1
        add     %g1, %g2, %g1
index 73574c91e857876ea82378164cc7e1959f40ca46..6e01f585d57c191a53bb2cb6feaa90823cc8cc1b 100644 (file)
@@ -101,6 +101,7 @@ config X86
        select GENERIC_STRNCPY_FROM_USER
        select GENERIC_STRNLEN_USER
        select GENERIC_TIME_VSYSCALL
+       select HARDLOCKUP_CHECK_TIMESTAMP       if X86_64
        select HAVE_ACPI_APEI                   if ACPI
        select HAVE_ACPI_APEI_NMI               if ACPI
        select HAVE_ALIGNED_STRUCT_PAGE         if SLUB
@@ -164,7 +165,7 @@ config X86
        select HAVE_PCSPKR_PLATFORM
        select HAVE_PERF_EVENTS
        select HAVE_PERF_EVENTS_NMI
-       select HAVE_HARDLOCKUP_DETECTOR_PERF    if HAVE_PERF_EVENTS_NMI
+       select HAVE_HARDLOCKUP_DETECTOR_PERF    if PERF_EVENTS && HAVE_PERF_EVENTS_NMI
        select HAVE_PERF_REGS
        select HAVE_PERF_USER_STACK_DUMP
        select HAVE_REGS_AND_STACK_ACCESS_API
index 1cd792db15efe760e3a6fc8b17b9a4c4e6f35233..1eab79c9ac484172a63d9cc0c4a409b7fffe7e8b 100644 (file)
        .set T1, REG_T1
 .endm
 
-#define K_BASE         %r8
 #define HASH_PTR       %r9
+#define BLOCKS_CTR     %r8
 #define BUFFER_PTR     %r10
 #define BUFFER_PTR2    %r13
-#define BUFFER_END     %r11
 
 #define PRECALC_BUF    %r14
 #define WK_BUF         %r15
                 * blended AVX2 and ALU instruction scheduling
                 * 1 vector iteration per 8 rounds
                 */
-               vmovdqu ((i * 2) + PRECALC_OFFSET)(BUFFER_PTR), W_TMP
+               vmovdqu (i * 2)(BUFFER_PTR), W_TMP
        .elseif ((i & 7) == 1)
-               vinsertf128 $1, (((i-1) * 2)+PRECALC_OFFSET)(BUFFER_PTR2),\
+               vinsertf128 $1, ((i-1) * 2)(BUFFER_PTR2),\
                         WY_TMP, WY_TMP
        .elseif ((i & 7) == 2)
                vpshufb YMM_SHUFB_BSWAP, WY_TMP, WY
        .elseif ((i & 7) == 4)
-               vpaddd  K_XMM(K_BASE), WY, WY_TMP
+               vpaddd  K_XMM + K_XMM_AR(%rip), WY, WY_TMP
        .elseif ((i & 7) == 7)
                vmovdqu  WY_TMP, PRECALC_WK(i&~7)
 
                vpxor   WY, WY_TMP, WY_TMP
        .elseif ((i & 7) == 7)
                vpxor   WY_TMP2, WY_TMP, WY
-               vpaddd  K_XMM(K_BASE), WY, WY_TMP
+               vpaddd  K_XMM + K_XMM_AR(%rip), WY, WY_TMP
                vmovdqu WY_TMP, PRECALC_WK(i&~7)
 
                PRECALC_ROTATE_WY
                vpsrld  $30, WY, WY
                vpor    WY, WY_TMP, WY
        .elseif ((i & 7) == 7)
-               vpaddd  K_XMM(K_BASE), WY, WY_TMP
+               vpaddd  K_XMM + K_XMM_AR(%rip), WY, WY_TMP
                vmovdqu WY_TMP, PRECALC_WK(i&~7)
 
                PRECALC_ROTATE_WY
 
 .endm
 
+/* Add constant only if (%2 > %3) condition met (uses RTA as temp)
+ * %1 + %2 >= %3 ? %4 : 0
+ */
+.macro ADD_IF_GE a, b, c, d
+       mov     \a, RTA
+       add     $\d, RTA
+       cmp     $\c, \b
+       cmovge  RTA, \a
+.endm
+
 /*
  * macro implements 80 rounds of SHA-1, for multiple blocks with s/w pipelining
  */
        lea     (2*4*80+32)(%rsp), WK_BUF
 
        # Precalc WK for first 2 blocks
-       PRECALC_OFFSET = 0
+       ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 2, 64
        .set i, 0
        .rept    160
                PRECALC i
                .set i, i + 1
        .endr
-       PRECALC_OFFSET = 128
+
+       /* Go to next block if needed */
+       ADD_IF_GE BUFFER_PTR, BLOCKS_CTR, 3, 128
+       ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 4, 128
        xchg    WK_BUF, PRECALC_BUF
 
        .align 32
@@ -479,8 +491,8 @@ _loop:
         * we use K_BASE value as a signal of a last block,
         * it is set below by: cmovae BUFFER_PTR, K_BASE
         */
-       cmp     K_BASE, BUFFER_PTR
-       jne     _begin
+       test BLOCKS_CTR, BLOCKS_CTR
+       jnz _begin
        .align 32
        jmp     _end
        .align 32
@@ -512,10 +524,10 @@ _loop0:
                .set j, j+2
        .endr
 
-       add     $(2*64), BUFFER_PTR       /* move to next odd-64-byte block */
-       cmp     BUFFER_END, BUFFER_PTR    /* is current block the last one? */
-       cmovae  K_BASE, BUFFER_PTR      /* signal the last iteration smartly */
-
+       /* Update Counter */
+       sub $1, BLOCKS_CTR
+       /* Move to the next block only if needed*/
+       ADD_IF_GE BUFFER_PTR, BLOCKS_CTR, 4, 128
        /*
         * rounds
         * 60,62,64,66,68
@@ -532,8 +544,8 @@ _loop0:
        UPDATE_HASH     12(HASH_PTR), D
        UPDATE_HASH     16(HASH_PTR), E
 
-       cmp     K_BASE, BUFFER_PTR      /* is current block the last one? */
-       je      _loop
+       test    BLOCKS_CTR, BLOCKS_CTR
+       jz      _loop
 
        mov     TB, B
 
@@ -575,10 +587,10 @@ _loop2:
                .set j, j+2
        .endr
 
-       add     $(2*64), BUFFER_PTR2      /* move to next even-64-byte block */
-
-       cmp     BUFFER_END, BUFFER_PTR2   /* is current block the last one */
-       cmovae  K_BASE, BUFFER_PTR       /* signal the last iteration smartly */
+       /* update counter */
+       sub     $1, BLOCKS_CTR
+       /* Move to the next block only if needed*/
+       ADD_IF_GE BUFFER_PTR2, BLOCKS_CTR, 4, 128
 
        jmp     _loop3
 _loop3:
@@ -641,19 +653,12 @@ _loop3:
 
        avx2_zeroupper
 
-       lea     K_XMM_AR(%rip), K_BASE
-
+       /* Setup initial values */
        mov     CTX, HASH_PTR
        mov     BUF, BUFFER_PTR
-       lea     64(BUF), BUFFER_PTR2
-
-       shl     $6, CNT                 /* mul by 64 */
-       add     BUF, CNT
-       add     $64, CNT
-       mov     CNT, BUFFER_END
 
-       cmp     BUFFER_END, BUFFER_PTR2
-       cmovae  K_BASE, BUFFER_PTR2
+       mov     BUF, BUFFER_PTR2
+       mov     CNT, BLOCKS_CTR
 
        xmm_mov BSWAP_SHUFB_CTL(%rip), YMM_SHUFB_BSWAP
 
index f960a043cdeba4a36f4ba20fddf81369f8ced38f..fc61739150e7c2c5a484c020b14441551fe5403a 100644 (file)
@@ -201,7 +201,7 @@ asmlinkage void sha1_transform_avx2(u32 *digest, const char *data,
 
 static bool avx2_usable(void)
 {
-       if (false && avx_usable() && boot_cpu_has(X86_FEATURE_AVX2)
+       if (avx_usable() && boot_cpu_has(X86_FEATURE_AVX2)
                && boot_cpu_has(X86_FEATURE_BMI1)
                && boot_cpu_has(X86_FEATURE_BMI2))
                return true;
index daf8936d0628bc7240d2c8dc6df1ec37d2d22b80..64b233ab7cad98b613df5b47de975fb8a77b75a0 100644 (file)
@@ -1313,6 +1313,8 @@ ENTRY(nmi)
         * other IST entries.
         */
 
+       ASM_CLAC
+
        /* Use %rdx as our temp variable throughout */
        pushq   %rdx
 
index 8e3db8f642a7a02dd8f99db3a25dedbfd1deb01a..af12e294caeda5f518cf8b80ce5765996f1d8c56 100644 (file)
@@ -2114,7 +2114,7 @@ static void refresh_pce(void *ignored)
        load_mm_cr4(this_cpu_read(cpu_tlbstate.loaded_mm));
 }
 
-static void x86_pmu_event_mapped(struct perf_event *event)
+static void x86_pmu_event_mapped(struct perf_event *event, struct mm_struct *mm)
 {
        if (!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED))
                return;
@@ -2129,22 +2129,20 @@ static void x86_pmu_event_mapped(struct perf_event *event)
         * For now, this can't happen because all callers hold mmap_sem
         * for write.  If this changes, we'll need a different solution.
         */
-       lockdep_assert_held_exclusive(&current->mm->mmap_sem);
+       lockdep_assert_held_exclusive(&mm->mmap_sem);
 
-       if (atomic_inc_return(&current->mm->context.perf_rdpmc_allowed) == 1)
-               on_each_cpu_mask(mm_cpumask(current->mm), refresh_pce, NULL, 1);
+       if (atomic_inc_return(&mm->context.perf_rdpmc_allowed) == 1)
+               on_each_cpu_mask(mm_cpumask(mm), refresh_pce, NULL, 1);
 }
 
-static void x86_pmu_event_unmapped(struct perf_event *event)
+static void x86_pmu_event_unmapped(struct perf_event *event, struct mm_struct *mm)
 {
-       if (!current->mm)
-               return;
 
        if (!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED))
                return;
 
-       if (atomic_dec_and_test(&current->mm->context.perf_rdpmc_allowed))
-               on_each_cpu_mask(mm_cpumask(current->mm), refresh_pce, NULL, 1);
+       if (atomic_dec_and_test(&mm->context.perf_rdpmc_allowed))
+               on_each_cpu_mask(mm_cpumask(mm), refresh_pce, NULL, 1);
 }
 
 static int x86_pmu_event_idx(struct perf_event *event)
index 8ae8c5ce3a1f94debb2fe81f5652cbd7b35f2c4a..ddd8d3516bfcfa58fc3baf7b3f4f70b7231190ab 100644 (file)
@@ -69,7 +69,7 @@ struct bts_buffer {
        struct bts_phys buf[0];
 };
 
-struct pmu bts_pmu;
+static struct pmu bts_pmu;
 
 static size_t buf_size(struct page *page)
 {
index eb0533558c2b705957a30704218b983eda65e61e..d32c0eed38ca92665d378fa2d0792eed12aad818 100644 (file)
@@ -587,7 +587,7 @@ static __initconst const u64 p4_hw_cache_event_ids
  * P4_CONFIG_ALIASABLE or bits for P4_PEBS_METRIC, they are
  * either up to date automatically or not applicable at all.
  */
-struct p4_event_alias {
+static struct p4_event_alias {
        u64 original;
        u64 alternative;
 } p4_event_aliases[] = {
index a45e2114a8460925b44bd6e0983f3ee37e83d861..8e2457cb6b4a416e1c84d9baa990ea512ec74ba0 100644 (file)
@@ -559,7 +559,7 @@ static struct attribute_group rapl_pmu_format_group = {
        .attrs = rapl_formats_attr,
 };
 
-const struct attribute_group *rapl_attr_groups[] = {
+static const struct attribute_group *rapl_attr_groups[] = {
        &rapl_pmu_attr_group,
        &rapl_pmu_format_group,
        &rapl_pmu_events_group,
index 44ec523287f670dff8eee3e3c9eb1bb1606489b6..1c5390f1cf0992787afa8d34bb361f622b00dcb2 100644 (file)
@@ -721,7 +721,7 @@ static struct attribute *uncore_pmu_attrs[] = {
        NULL,
 };
 
-static struct attribute_group uncore_pmu_attr_group = {
+static const struct attribute_group uncore_pmu_attr_group = {
        .attrs = uncore_pmu_attrs,
 };
 
index cda56933200503662d133050fc47bdae050e8d61..6a5cbe90f8593eb23a7a403132240fc9afc87c9d 100644 (file)
@@ -272,7 +272,7 @@ static struct attribute *nhmex_uncore_ubox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_ubox_format_group = {
+static const struct attribute_group nhmex_uncore_ubox_format_group = {
        .name           = "format",
        .attrs          = nhmex_uncore_ubox_formats_attr,
 };
@@ -299,7 +299,7 @@ static struct attribute *nhmex_uncore_cbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_cbox_format_group = {
+static const struct attribute_group nhmex_uncore_cbox_format_group = {
        .name = "format",
        .attrs = nhmex_uncore_cbox_formats_attr,
 };
@@ -407,7 +407,7 @@ static struct attribute *nhmex_uncore_bbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_bbox_format_group = {
+static const struct attribute_group nhmex_uncore_bbox_format_group = {
        .name = "format",
        .attrs = nhmex_uncore_bbox_formats_attr,
 };
@@ -484,7 +484,7 @@ static struct attribute *nhmex_uncore_sbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_sbox_format_group = {
+static const struct attribute_group nhmex_uncore_sbox_format_group = {
        .name                   = "format",
        .attrs                  = nhmex_uncore_sbox_formats_attr,
 };
@@ -898,7 +898,7 @@ static struct attribute *nhmex_uncore_mbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_mbox_format_group = {
+static const struct attribute_group nhmex_uncore_mbox_format_group = {
        .name           = "format",
        .attrs          = nhmex_uncore_mbox_formats_attr,
 };
@@ -1163,7 +1163,7 @@ static struct attribute *nhmex_uncore_rbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhmex_uncore_rbox_format_group = {
+static const struct attribute_group nhmex_uncore_rbox_format_group = {
        .name = "format",
        .attrs = nhmex_uncore_rbox_formats_attr,
 };
index a3dcc12bef4ab3a67aab29c6acf3480920e06952..db1127ce685eb8ea687bb4e0279f13ad5706a59f 100644 (file)
@@ -130,7 +130,7 @@ static struct attribute *snb_uncore_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group snb_uncore_format_group = {
+static const struct attribute_group snb_uncore_format_group = {
        .name           = "format",
        .attrs          = snb_uncore_formats_attr,
 };
@@ -289,7 +289,7 @@ static struct attribute *snb_uncore_imc_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group snb_uncore_imc_format_group = {
+static const struct attribute_group snb_uncore_imc_format_group = {
        .name = "format",
        .attrs = snb_uncore_imc_formats_attr,
 };
@@ -769,7 +769,7 @@ static struct attribute *nhm_uncore_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group nhm_uncore_format_group = {
+static const struct attribute_group nhm_uncore_format_group = {
        .name = "format",
        .attrs = nhm_uncore_formats_attr,
 };
index 4f9127644b80abd87d4c49204dde197a9ade7ac2..db1fe377e6dd9ddfcfbc006a7346879f263d408a 100644 (file)
@@ -602,27 +602,27 @@ static struct uncore_event_desc snbep_uncore_qpi_events[] = {
        { /* end: all zeroes */ },
 };
 
-static struct attribute_group snbep_uncore_format_group = {
+static const struct attribute_group snbep_uncore_format_group = {
        .name = "format",
        .attrs = snbep_uncore_formats_attr,
 };
 
-static struct attribute_group snbep_uncore_ubox_format_group = {
+static const struct attribute_group snbep_uncore_ubox_format_group = {
        .name = "format",
        .attrs = snbep_uncore_ubox_formats_attr,
 };
 
-static struct attribute_group snbep_uncore_cbox_format_group = {
+static const struct attribute_group snbep_uncore_cbox_format_group = {
        .name = "format",
        .attrs = snbep_uncore_cbox_formats_attr,
 };
 
-static struct attribute_group snbep_uncore_pcu_format_group = {
+static const struct attribute_group snbep_uncore_pcu_format_group = {
        .name = "format",
        .attrs = snbep_uncore_pcu_formats_attr,
 };
 
-static struct attribute_group snbep_uncore_qpi_format_group = {
+static const struct attribute_group snbep_uncore_qpi_format_group = {
        .name = "format",
        .attrs = snbep_uncore_qpi_formats_attr,
 };
@@ -1431,27 +1431,27 @@ static struct attribute *ivbep_uncore_qpi_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group ivbep_uncore_format_group = {
+static const struct attribute_group ivbep_uncore_format_group = {
        .name = "format",
        .attrs = ivbep_uncore_formats_attr,
 };
 
-static struct attribute_group ivbep_uncore_ubox_format_group = {
+static const struct attribute_group ivbep_uncore_ubox_format_group = {
        .name = "format",
        .attrs = ivbep_uncore_ubox_formats_attr,
 };
 
-static struct attribute_group ivbep_uncore_cbox_format_group = {
+static const struct attribute_group ivbep_uncore_cbox_format_group = {
        .name = "format",
        .attrs = ivbep_uncore_cbox_formats_attr,
 };
 
-static struct attribute_group ivbep_uncore_pcu_format_group = {
+static const struct attribute_group ivbep_uncore_pcu_format_group = {
        .name = "format",
        .attrs = ivbep_uncore_pcu_formats_attr,
 };
 
-static struct attribute_group ivbep_uncore_qpi_format_group = {
+static const struct attribute_group ivbep_uncore_qpi_format_group = {
        .name = "format",
        .attrs = ivbep_uncore_qpi_formats_attr,
 };
@@ -1887,7 +1887,7 @@ static struct attribute *knl_uncore_ubox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group knl_uncore_ubox_format_group = {
+static const struct attribute_group knl_uncore_ubox_format_group = {
        .name = "format",
        .attrs = knl_uncore_ubox_formats_attr,
 };
@@ -1927,7 +1927,7 @@ static struct attribute *knl_uncore_cha_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group knl_uncore_cha_format_group = {
+static const struct attribute_group knl_uncore_cha_format_group = {
        .name = "format",
        .attrs = knl_uncore_cha_formats_attr,
 };
@@ -2037,7 +2037,7 @@ static struct attribute *knl_uncore_pcu_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group knl_uncore_pcu_format_group = {
+static const struct attribute_group knl_uncore_pcu_format_group = {
        .name = "format",
        .attrs = knl_uncore_pcu_formats_attr,
 };
@@ -2187,7 +2187,7 @@ static struct attribute *knl_uncore_irp_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group knl_uncore_irp_format_group = {
+static const struct attribute_group knl_uncore_irp_format_group = {
        .name = "format",
        .attrs = knl_uncore_irp_formats_attr,
 };
@@ -2385,7 +2385,7 @@ static struct attribute *hswep_uncore_ubox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group hswep_uncore_ubox_format_group = {
+static const struct attribute_group hswep_uncore_ubox_format_group = {
        .name = "format",
        .attrs = hswep_uncore_ubox_formats_attr,
 };
@@ -2439,7 +2439,7 @@ static struct attribute *hswep_uncore_cbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group hswep_uncore_cbox_format_group = {
+static const struct attribute_group hswep_uncore_cbox_format_group = {
        .name = "format",
        .attrs = hswep_uncore_cbox_formats_attr,
 };
@@ -2621,7 +2621,7 @@ static struct attribute *hswep_uncore_sbox_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group hswep_uncore_sbox_format_group = {
+static const struct attribute_group hswep_uncore_sbox_format_group = {
        .name = "format",
        .attrs = hswep_uncore_sbox_formats_attr,
 };
@@ -3314,7 +3314,7 @@ static struct attribute *skx_uncore_cha_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group skx_uncore_chabox_format_group = {
+static const struct attribute_group skx_uncore_chabox_format_group = {
        .name = "format",
        .attrs = skx_uncore_cha_formats_attr,
 };
@@ -3427,7 +3427,7 @@ static struct attribute *skx_uncore_iio_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group skx_uncore_iio_format_group = {
+static const struct attribute_group skx_uncore_iio_format_group = {
        .name = "format",
        .attrs = skx_uncore_iio_formats_attr,
 };
@@ -3484,7 +3484,7 @@ static struct attribute *skx_uncore_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group skx_uncore_format_group = {
+static const struct attribute_group skx_uncore_format_group = {
        .name = "format",
        .attrs = skx_uncore_formats_attr,
 };
@@ -3605,7 +3605,7 @@ static struct attribute *skx_upi_uncore_formats_attr[] = {
        NULL,
 };
 
-static struct attribute_group skx_upi_uncore_format_group = {
+static const struct attribute_group skx_upi_uncore_format_group = {
        .name = "format",
        .attrs = skx_upi_uncore_formats_attr,
 };
index ca3c48c0872f4beba6b03ac92c50eb6a352b3006..5a28e8e55e36fd2164c0cda175288a7fcc534c4b 100644 (file)
 #define X86_FEATURE_PAUSEFILTER (15*32+10) /* filtered pause intercept */
 #define X86_FEATURE_PFTHRESHOLD (15*32+12) /* pause filter threshold */
 #define X86_FEATURE_AVIC       (15*32+13) /* Virtual Interrupt Controller */
-#define X86_FEATURE_VIRTUAL_VMLOAD_VMSAVE (15*32+15) /* Virtual VMLOAD VMSAVE */
+#define X86_FEATURE_V_VMSAVE_VMLOAD (15*32+15) /* Virtual VMSAVE VMLOAD */
 
 /* Intel-defined CPU features, CPUID level 0x00000007:0 (ecx), word 16 */
 #define X86_FEATURE_AVX512VBMI  (16*32+ 1) /* AVX512 Vector Bit Manipulation instructions*/
index 1c18d83d3f094d1f5abfa09a08b67e9cb6b7376d..9aeb91935ce02387d8dae5e2f51bf1750f420a43 100644 (file)
@@ -247,11 +247,11 @@ extern int force_personality32;
 
 /*
  * This is the base location for PIE (ET_DYN with INTERP) loads. On
- * 64-bit, this is raised to 4GB to leave the entire 32-bit address
+ * 64-bit, this is above 4GB to leave the entire 32-bit address
  * space open for things that want to use the area for 32-bit pointers.
  */
 #define ELF_ET_DYN_BASE                (mmap_is_ia32() ? 0x000400000UL : \
-                                                 0x100000000UL)
+                                                 (TASK_SIZE / 3 * 2))
 
 /* This yields a mask that user programs can use to figure out what
    instruction set this CPU supports.  This could be done in user space,
index 21126155a739f4a3495499f052abcd3eab5a0bb7..0ead9dbb91301d0f7f8923dcf33f25515bd182b8 100644 (file)
@@ -43,6 +43,9 @@ struct hypervisor_x86 {
 
        /* pin current vcpu to specified physical cpu (run rarely) */
        void            (*pin_vcpu)(int);
+
+       /* called during init_mem_mapping() to setup early mappings. */
+       void            (*init_mem_mapping)(void);
 };
 
 extern const struct hypervisor_x86 *x86_hyper;
@@ -57,8 +60,15 @@ extern const struct hypervisor_x86 x86_hyper_kvm;
 extern void init_hypervisor_platform(void);
 extern bool hypervisor_x2apic_available(void);
 extern void hypervisor_pin_vcpu(int cpu);
+
+static inline void hypervisor_init_mem_mapping(void)
+{
+       if (x86_hyper && x86_hyper->init_mem_mapping)
+               x86_hyper->init_mem_mapping();
+}
 #else
 static inline void init_hypervisor_platform(void) { }
 static inline bool hypervisor_x2apic_available(void) { return false; }
+static inline void hypervisor_init_mem_mapping(void) { }
 #endif /* CONFIG_HYPERVISOR_GUEST */
 #endif /* _ASM_X86_HYPERVISOR_H */
index 7cf7c70b6ef2a20483361fb8333a85ace25bf1d7..0ee83321a3136fcca7a00a3b7e6c375e7a51e13f 100644 (file)
@@ -40,13 +40,16 @@ static void aperfmperf_snapshot_khz(void *dummy)
        struct aperfmperf_sample *s = this_cpu_ptr(&samples);
        ktime_t now = ktime_get();
        s64 time_delta = ktime_ms_delta(now, s->time);
+       unsigned long flags;
 
        /* Don't bother re-computing within the cache threshold time. */
        if (time_delta < APERFMPERF_CACHE_THRESHOLD_MS)
                return;
 
+       local_irq_save(flags);
        rdmsrl(MSR_IA32_APERF, aperf);
        rdmsrl(MSR_IA32_MPERF, mperf);
+       local_irq_restore(flags);
 
        aperf_delta = aperf - s->aperf;
        mperf_delta = mperf - s->mperf;
index d7cc190ae45719bf84d721e781dc7fde00c85e14..f7370abd33c67570ad713a03a20f9114e6ad6399 100644 (file)
@@ -122,7 +122,7 @@ static struct attribute *thermal_throttle_attrs[] = {
        NULL
 };
 
-static struct attribute_group thermal_attr_group = {
+static const struct attribute_group thermal_attr_group = {
        .attrs  = thermal_throttle_attrs,
        .name   = "thermal_throttle"
 };
index 9cb98ee103db1bf0cdc2349614fe0d4e59cfb21f..86e8f0b2537b3eaedd2da204d67c94947a1b16f1 100644 (file)
@@ -561,7 +561,7 @@ static struct attribute *mc_default_attrs[] = {
        NULL
 };
 
-static struct attribute_group mc_attr_group = {
+static const struct attribute_group mc_attr_group = {
        .attrs                  = mc_default_attrs,
        .name                   = "microcode",
 };
@@ -707,7 +707,7 @@ static struct attribute *cpu_root_microcode_attrs[] = {
        NULL
 };
 
-static struct attribute_group cpu_root_microcode_group = {
+static const struct attribute_group cpu_root_microcode_group = {
        .name  = "microcode",
        .attrs = cpu_root_microcode_attrs,
 };
index c5bb63be4ba1e6a2df1203fe73079faad9fc62e7..40d5a8a752125ed5d26a7605d5eabad572879bfc 100644 (file)
@@ -237,6 +237,18 @@ set_mtrr(unsigned int reg, unsigned long base, unsigned long size, mtrr_type typ
        stop_machine(mtrr_rendezvous_handler, &data, cpu_online_mask);
 }
 
+static void set_mtrr_cpuslocked(unsigned int reg, unsigned long base,
+                               unsigned long size, mtrr_type type)
+{
+       struct set_mtrr_data data = { .smp_reg = reg,
+                                     .smp_base = base,
+                                     .smp_size = size,
+                                     .smp_type = type
+                                   };
+
+       stop_machine_cpuslocked(mtrr_rendezvous_handler, &data, cpu_online_mask);
+}
+
 static void set_mtrr_from_inactive_cpu(unsigned int reg, unsigned long base,
                                      unsigned long size, mtrr_type type)
 {
@@ -370,7 +382,7 @@ int mtrr_add_page(unsigned long base, unsigned long size,
        /* Search for an empty MTRR */
        i = mtrr_if->get_free_region(base, size, replace);
        if (i >= 0) {
-               set_mtrr(i, base, size, type);
+               set_mtrr_cpuslocked(i, base, size, type);
                if (likely(replace < 0)) {
                        mtrr_usage_table[i] = 1;
                } else {
@@ -378,7 +390,7 @@ int mtrr_add_page(unsigned long base, unsigned long size,
                        if (increment)
                                mtrr_usage_table[i]++;
                        if (unlikely(replace != i)) {
-                               set_mtrr(replace, 0, 0, 0);
+                               set_mtrr_cpuslocked(replace, 0, 0, 0);
                                mtrr_usage_table[replace] = 0;
                        }
                }
@@ -506,7 +518,7 @@ int mtrr_del_page(int reg, unsigned long base, unsigned long size)
                goto out;
        }
        if (--mtrr_usage_table[reg] < 1)
-               set_mtrr(reg, 0, 0, 0);
+               set_mtrr_cpuslocked(reg, 0, 0, 0);
        error = reg;
  out:
        mutex_unlock(&mtrr_mutex);
index 46c3c73e7f43f5fbb56fe0028d8aafb64a3f8a9e..9ba79543d9ee9f1ee19bce38955467afa04b125f 100644 (file)
@@ -53,6 +53,7 @@ void __head __startup_64(unsigned long physaddr)
        pudval_t *pud;
        pmdval_t *pmd, pmd_entry;
        int i;
+       unsigned int *next_pgt_ptr;
 
        /* Is the address too large? */
        if (physaddr >> MAX_PHYSMEM_BITS)
@@ -91,9 +92,9 @@ void __head __startup_64(unsigned long physaddr)
         * creates a bunch of nonsense entries but that is fine --
         * it avoids problems around wraparound.
         */
-
-       pud = fixup_pointer(early_dynamic_pgts[next_early_pgt++], physaddr);
-       pmd = fixup_pointer(early_dynamic_pgts[next_early_pgt++], physaddr);
+       next_pgt_ptr = fixup_pointer(&next_early_pgt, physaddr);
+       pud = fixup_pointer(early_dynamic_pgts[(*next_pgt_ptr)++], physaddr);
+       pmd = fixup_pointer(early_dynamic_pgts[(*next_pgt_ptr)++], physaddr);
 
        if (IS_ENABLED(CONFIG_X86_5LEVEL)) {
                p4d = fixup_pointer(early_dynamic_pgts[next_early_pgt++], physaddr);
index 4afc67f5facc497606e2fb898ba7b7d402f04caa..06e1ff5562c0b4a5a28f3c052ed74b545c7ccca0 100644 (file)
@@ -55,7 +55,7 @@ static struct bin_attribute *boot_params_data_attrs[] = {
        NULL,
 };
 
-static struct attribute_group boot_params_attr_group = {
+static const struct attribute_group boot_params_attr_group = {
        .attrs = boot_params_version_attrs,
        .bin_attrs = boot_params_data_attrs,
 };
@@ -202,7 +202,7 @@ static struct bin_attribute *setup_data_data_attrs[] = {
        NULL,
 };
 
-static struct attribute_group setup_data_attr_group = {
+static const struct attribute_group setup_data_attr_group = {
        .attrs = setup_data_type_attrs,
        .bin_attrs = setup_data_data_attrs,
 };
index b474c8de7fba09cfe27d1e0124f694d152654570..54b9e89d4d6be3844b8b3c310433467d0e5f1481 100644 (file)
@@ -971,7 +971,8 @@ void common_cpu_up(unsigned int cpu, struct task_struct *idle)
  * Returns zero if CPU booted OK, else error code from
  * ->wakeup_secondary_cpu.
  */
-static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
+static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle,
+                      int *cpu0_nmi_registered)
 {
        volatile u32 *trampoline_status =
                (volatile u32 *) __va(real_mode_header->trampoline_status);
@@ -979,7 +980,6 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
        unsigned long start_ip = real_mode_header->trampoline_start;
 
        unsigned long boot_error = 0;
-       int cpu0_nmi_registered = 0;
        unsigned long timeout;
 
        idle->thread.sp = (unsigned long)task_pt_regs(idle);
@@ -1035,7 +1035,7 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
                boot_error = apic->wakeup_secondary_cpu(apicid, start_ip);
        else
                boot_error = wakeup_cpu_via_init_nmi(cpu, start_ip, apicid,
-                                                    &cpu0_nmi_registered);
+                                                    cpu0_nmi_registered);
 
        if (!boot_error) {
                /*
@@ -1080,12 +1080,6 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
                 */
                smpboot_restore_warm_reset_vector();
        }
-       /*
-        * Clean up the nmi handler. Do this after the callin and callout sync
-        * to avoid impact of possible long unregister time.
-        */
-       if (cpu0_nmi_registered)
-               unregister_nmi_handler(NMI_LOCAL, "wake_cpu0");
 
        return boot_error;
 }
@@ -1093,8 +1087,9 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
 int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
 {
        int apicid = apic->cpu_present_to_apicid(cpu);
+       int cpu0_nmi_registered = 0;
        unsigned long flags;
-       int err;
+       int err, ret = 0;
 
        WARN_ON(irqs_disabled());
 
@@ -1131,10 +1126,11 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
 
        common_cpu_up(cpu, tidle);
 
-       err = do_boot_cpu(apicid, cpu, tidle);
+       err = do_boot_cpu(apicid, cpu, tidle, &cpu0_nmi_registered);
        if (err) {
                pr_err("do_boot_cpu failed(%d) to wakeup CPU#%u\n", err, cpu);
-               return -EIO;
+               ret = -EIO;
+               goto unreg_nmi;
        }
 
        /*
@@ -1150,7 +1146,15 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
                touch_nmi_watchdog();
        }
 
-       return 0;
+unreg_nmi:
+       /*
+        * Clean up the nmi handler. Do this after the callin and callout sync
+        * to avoid impact of possible long unregister time.
+        */
+       if (cpu0_nmi_registered)
+               unregister_nmi_handler(NMI_LOCAL, "wake_cpu0");
+
+       return ret;
 }
 
 /**
index 1107626938ccff4f26da130f8b570928ca7c8efb..56ba05312759d3ed4568e546492d9d8bfad05b71 100644 (file)
@@ -1100,7 +1100,7 @@ static __init int svm_hardware_setup(void)
 
        if (vls) {
                if (!npt_enabled ||
-                   !boot_cpu_has(X86_FEATURE_VIRTUAL_VMLOAD_VMSAVE) ||
+                   !boot_cpu_has(X86_FEATURE_V_VMSAVE_VMLOAD) ||
                    !IS_ENABLED(CONFIG_X86_64)) {
                        vls = false;
                } else {
index 673541eb3b3f16c8c029349d597d67f4bb83a77a..bf3f1065d6addb88b898ba3a86089cccff6ed15e 100644 (file)
@@ -18,6 +18,7 @@
 #include <asm/dma.h>           /* for MAX_DMA_PFN */
 #include <asm/microcode.h>
 #include <asm/kaslr.h>
+#include <asm/hypervisor.h>
 
 /*
  * We need to define the tracepoints somewhere, and tlb.c
@@ -636,6 +637,8 @@ void __init init_mem_mapping(void)
        load_cr3(swapper_pg_dir);
        __flush_tlb_all();
 
+       hypervisor_init_mem_mapping();
+
        early_memtest(0, max_pfn_mapped << PAGE_SHIFT);
 }
 
index 229d04a83f8561ec08e394f57d8cb5d671bd56ff..a88cfbfbd0781a4d20c398d7ca1b30f50f05be43 100644 (file)
@@ -50,8 +50,7 @@ unsigned long tasksize_64bit(void)
 static unsigned long stack_maxrandom_size(unsigned long task_size)
 {
        unsigned long max = 0;
-       if ((current->flags & PF_RANDOMIZE) &&
-               !(current->personality & ADDR_NO_RANDOMIZE)) {
+       if (current->flags & PF_RANDOMIZE) {
                max = (-1UL) & __STACK_RND_MASK(task_size == tasksize_32bit());
                max <<= PAGE_SHIFT;
        }
@@ -79,13 +78,13 @@ static int mmap_is_legacy(void)
 
 static unsigned long arch_rnd(unsigned int rndbits)
 {
+       if (!(current->flags & PF_RANDOMIZE))
+               return 0;
        return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
 }
 
 unsigned long arch_mmap_rnd(void)
 {
-       if (!(current->flags & PF_RANDOMIZE))
-               return 0;
        return arch_rnd(mmap_is_ia32() ? mmap32_rnd_bits : mmap64_rnd_bits);
 }
 
index 3e4bdb442fbcfc783059be96efe4d4e2d6b58b47..f44c0bc95aa2f45ad42462a5f23f4db4672d1257 100644 (file)
@@ -26,7 +26,7 @@
 static struct bau_operations ops __ro_after_init;
 
 /* timeouts in nanoseconds (indexed by UVH_AGING_PRESCALE_SEL urgency7 30:28) */
-static int timeout_base_ns[] = {
+static const int timeout_base_ns[] = {
                20,
                160,
                1280,
@@ -1216,7 +1216,7 @@ static struct bau_pq_entry *find_another_by_swack(struct bau_pq_entry *msg,
  * set a bit in the UVH_LB_BAU_INTD_SOFTWARE_ACKNOWLEDGE register.
  * Such a message must be ignored.
  */
-void process_uv2_message(struct msg_desc *mdp, struct bau_control *bcp)
+static void process_uv2_message(struct msg_desc *mdp, struct bau_control *bcp)
 {
        unsigned long mmr_image;
        unsigned char swack_vec;
index 87d791356ea9052f79f23e00f5d0024ff8cf0b7f..de503c225ae1f194b10c71b44528ad2a2a7a4c0d 100644 (file)
@@ -12,6 +12,7 @@
 #include <asm/setup.h>
 #include <asm/hypervisor.h>
 #include <asm/e820/api.h>
+#include <asm/early_ioremap.h>
 
 #include <asm/xen/cpuid.h>
 #include <asm/xen/hypervisor.h>
 #include "mmu.h"
 #include "smp.h"
 
-void __ref xen_hvm_init_shared_info(void)
+static unsigned long shared_info_pfn;
+
+void xen_hvm_init_shared_info(void)
 {
        struct xen_add_to_physmap xatp;
-       u64 pa;
-
-       if (HYPERVISOR_shared_info == &xen_dummy_shared_info) {
-               /*
-                * Search for a free page starting at 4kB physical address.
-                * Low memory is preferred to avoid an EPT large page split up
-                * by the mapping.
-                * Starting below X86_RESERVE_LOW (usually 64kB) is fine as
-                * the BIOS used for HVM guests is well behaved and won't
-                * clobber memory other than the first 4kB.
-                */
-               for (pa = PAGE_SIZE;
-                    !e820__mapped_all(pa, pa + PAGE_SIZE, E820_TYPE_RAM) ||
-                    memblock_is_reserved(pa);
-                    pa += PAGE_SIZE)
-                       ;
-
-               memblock_reserve(pa, PAGE_SIZE);
-               HYPERVISOR_shared_info = __va(pa);
-       }
 
        xatp.domid = DOMID_SELF;
        xatp.idx = 0;
        xatp.space = XENMAPSPACE_shared_info;
-       xatp.gpfn = virt_to_pfn(HYPERVISOR_shared_info);
+       xatp.gpfn = shared_info_pfn;
        if (HYPERVISOR_memory_op(XENMEM_add_to_physmap, &xatp))
                BUG();
 }
 
+static void __init reserve_shared_info(void)
+{
+       u64 pa;
+
+       /*
+        * Search for a free page starting at 4kB physical address.
+        * Low memory is preferred to avoid an EPT large page split up
+        * by the mapping.
+        * Starting below X86_RESERVE_LOW (usually 64kB) is fine as
+        * the BIOS used for HVM guests is well behaved and won't
+        * clobber memory other than the first 4kB.
+        */
+       for (pa = PAGE_SIZE;
+            !e820__mapped_all(pa, pa + PAGE_SIZE, E820_TYPE_RAM) ||
+            memblock_is_reserved(pa);
+            pa += PAGE_SIZE)
+               ;
+
+       shared_info_pfn = PHYS_PFN(pa);
+
+       memblock_reserve(pa, PAGE_SIZE);
+       HYPERVISOR_shared_info = early_memremap(pa, PAGE_SIZE);
+}
+
+static void __init xen_hvm_init_mem_mapping(void)
+{
+       early_memunmap(HYPERVISOR_shared_info, PAGE_SIZE);
+       HYPERVISOR_shared_info = __va(PFN_PHYS(shared_info_pfn));
+}
+
 static void __init init_hvm_pv_info(void)
 {
        int major, minor;
@@ -153,6 +166,7 @@ static void __init xen_hvm_guest_init(void)
 
        init_hvm_pv_info();
 
+       reserve_shared_info();
        xen_hvm_init_shared_info();
 
        /*
@@ -218,5 +232,6 @@ const struct hypervisor_x86 x86_hyper_xen_hvm = {
        .init_platform          = xen_hvm_guest_init,
        .pin_vcpu               = xen_pin_vcpu,
        .x2apic_available       = xen_x2apic_para_available,
+       .init_mem_mapping       = xen_hvm_init_mem_mapping,
 };
 EXPORT_SYMBOL(x86_hyper_xen_hvm);
index 83e92beb3c9feb25f3be917e2b675f90b1544b5f..9b1ea478577b033195f34b457c85ad16eb759c6a 100644 (file)
@@ -387,9 +387,11 @@ static void bio_integrity_verify_fn(struct work_struct *work)
  */
 bool __bio_integrity_endio(struct bio *bio)
 {
-       if (bio_op(bio) == REQ_OP_READ && !bio->bi_status) {
-               struct bio_integrity_payload *bip = bio_integrity(bio);
+       struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
+       struct bio_integrity_payload *bip = bio_integrity(bio);
 
+       if (bio_op(bio) == REQ_OP_READ && !bio->bi_status &&
+           (bip->bip_flags & BIP_BLOCK_INTEGRITY) && bi->profile->verify_fn) {
                INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
                queue_work(kintegrityd_wq, &bip->bip_work);
                return false;
index 0c3354cf3552877d7542cdb32f0ce4102a79fcce..76944e3271bf34a730e62fa0b27266abce86942e 100644 (file)
@@ -36,12 +36,18 @@ int blk_mq_pci_map_queues(struct blk_mq_tag_set *set, struct pci_dev *pdev)
        for (queue = 0; queue < set->nr_hw_queues; queue++) {
                mask = pci_irq_get_affinity(pdev, queue);
                if (!mask)
-                       return -EINVAL;
+                       goto fallback;
 
                for_each_cpu(cpu, mask)
                        set->mq_map[cpu] = queue;
        }
 
        return 0;
+
+fallback:
+       WARN_ON_ONCE(set->nr_hw_queues > 1);
+       for_each_possible_cpu(cpu)
+               set->mq_map[cpu] = 0;
+       return 0;
 }
 EXPORT_SYMBOL_GPL(blk_mq_pci_map_queues);
index 211ef367345f270f300cb5cdcb9a63c0a50d2420..4603b115e234887860cbb28520c36ba9e5e7efd8 100644 (file)
@@ -360,12 +360,12 @@ struct request *blk_mq_alloc_request(struct request_queue *q, unsigned int op,
                return ERR_PTR(ret);
 
        rq = blk_mq_get_request(q, NULL, op, &alloc_data);
+       blk_queue_exit(q);
 
        if (!rq)
                return ERR_PTR(-EWOULDBLOCK);
 
        blk_mq_put_ctx(alloc_data.ctx);
-       blk_queue_exit(q);
 
        rq->__data_len = 0;
        rq->__sector = (sector_t) -1;
@@ -411,12 +411,11 @@ struct request *blk_mq_alloc_request_hctx(struct request_queue *q,
        alloc_data.ctx = __blk_mq_get_ctx(q, cpu);
 
        rq = blk_mq_get_request(q, NULL, op, &alloc_data);
+       blk_queue_exit(q);
 
        if (!rq)
                return ERR_PTR(-EWOULDBLOCK);
 
-       blk_queue_exit(q);
-
        return rq;
 }
 EXPORT_SYMBOL_GPL(blk_mq_alloc_request_hctx);
@@ -684,8 +683,8 @@ EXPORT_SYMBOL(blk_mq_kick_requeue_list);
 void blk_mq_delay_kick_requeue_list(struct request_queue *q,
                                    unsigned long msecs)
 {
-       kblockd_schedule_delayed_work(&q->requeue_work,
-                                     msecs_to_jiffies(msecs));
+       kblockd_mod_delayed_work_on(WORK_CPU_UNBOUND, &q->requeue_work,
+                                   msecs_to_jiffies(msecs));
 }
 EXPORT_SYMBOL(blk_mq_delay_kick_requeue_list);
 
index 538c61677c100e9361bc556154719b7b2368740e..783f4c838aee4c1daabdea8ac8d6f21e75060492 100644 (file)
@@ -100,9 +100,13 @@ acpi_evaluate_object_typed(acpi_handle handle,
                free_buffer_on_error = TRUE;
        }
 
-       status = acpi_get_handle(handle, pathname, &target_handle);
-       if (ACPI_FAILURE(status)) {
-               return_ACPI_STATUS(status);
+       if (pathname) {
+               status = acpi_get_handle(handle, pathname, &target_handle);
+               if (ACPI_FAILURE(status)) {
+                       return_ACPI_STATUS(status);
+               }
+       } else {
+               target_handle = handle;
        }
 
        full_pathname = acpi_ns_get_external_pathname(target_handle);
index 62068a5e814f026a44e212bd2b93941590a3fe53..ae3d6d15263307c2c159fd459d09bc93ef66fa03 100644 (file)
@@ -1741,7 +1741,7 @@ error:
  * functioning ECDT EC first in order to handle the events.
  * https://bugzilla.kernel.org/show_bug.cgi?id=115021
  */
-int __init acpi_ec_ecdt_start(void)
+static int __init acpi_ec_ecdt_start(void)
 {
        acpi_handle handle;
 
@@ -2003,20 +2003,17 @@ static inline void acpi_ec_query_exit(void)
 int __init acpi_ec_init(void)
 {
        int result;
+       int ecdt_fail, dsdt_fail;
 
        /* register workqueue for _Qxx evaluations */
        result = acpi_ec_query_init();
        if (result)
-               goto err_exit;
-       /* Now register the driver for the EC */
-       result = acpi_bus_register_driver(&acpi_ec_driver);
-       if (result)
-               goto err_exit;
+               return result;
 
-err_exit:
-       if (result)
-               acpi_ec_query_exit();
-       return result;
+       /* Drivers must be started after acpi_ec_query_init() */
+       ecdt_fail = acpi_ec_ecdt_start();
+       dsdt_fail = acpi_bus_register_driver(&acpi_ec_driver);
+       return ecdt_fail && dsdt_fail ? -ENODEV : 0;
 }
 
 /* EC driver currently not unloadable */
index 58dd7ab3c6538bd490b45e2e420c3d822ecf8c3e..3f5af4d7a73989f86cdcc80c1e8e7b949a975702 100644 (file)
@@ -185,7 +185,6 @@ typedef int (*acpi_ec_query_func) (void *data);
 int acpi_ec_init(void);
 int acpi_ec_ecdt_probe(void);
 int acpi_ec_dsdt_probe(void);
-int acpi_ec_ecdt_start(void);
 void acpi_ec_block_transactions(void);
 void acpi_ec_unblock_transactions(void);
 int acpi_ec_add_query_handler(struct acpi_ec *ec, u8 query_bit,
index 917c789f953dfb280f6f00895ce1fcf7ae2484ec..476a52c60cf3de98b18abe2c35a681c2be3ae058 100644 (file)
@@ -1047,7 +1047,7 @@ static struct fwnode_handle *acpi_graph_get_child_prop_value(
        fwnode_for_each_child_node(fwnode, child) {
                u32 nr;
 
-               if (!fwnode_property_read_u32(fwnode, prop_name, &nr))
+               if (fwnode_property_read_u32(child, prop_name, &nr))
                        continue;
 
                if (val == nr)
index 33897298f03e3ed680272ba6f4109077d434a0d8..70fd5502c2843a7cb0338245767a3c7dc2aeb308 100644 (file)
@@ -2084,7 +2084,6 @@ int __init acpi_scan_init(void)
 
        acpi_gpe_apply_masked_gpes();
        acpi_update_all_gpes();
-       acpi_ec_ecdt_start();
 
        acpi_scan_initialized = true;
 
index 4ac3e06b41d846440d35079dfa54885111eec0df..98aa8c808a3346dc559e2f419ae4378dec075ffc 100644 (file)
 #include <linux/kernel.h>
 #include <linux/serial_core.h>
 
+/*
+ * Erratum 44 for QDF2432v1 and QDF2400v1 SoCs describes the BUSY bit as
+ * occasionally getting stuck as 1. To avoid the potential for a hang, check
+ * TXFE == 0 instead of BUSY == 1. This may not be suitable for all UART
+ * implementations, so only do so if an affected platform is detected in
+ * parse_spcr().
+ */
+bool qdf2400_e44_present;
+EXPORT_SYMBOL(qdf2400_e44_present);
+
 /*
  * Some Qualcomm Datacenter Technologies SoCs have a defective UART BUSY bit.
  * Detect them by examining the OEM fields in the SPCR header, similiar to PCI
@@ -147,8 +157,30 @@ int __init parse_spcr(bool earlycon)
                goto done;
        }
 
-       if (qdf2400_erratum_44_present(&table->header))
-               uart = "qdf2400_e44";
+       /*
+        * If the E44 erratum is required, then we need to tell the pl011
+        * driver to implement the work-around.
+        *
+        * The global variable is used by the probe function when it
+        * creates the UARTs, whether or not they're used as a console.
+        *
+        * If the user specifies "traditional" earlycon, the qdf2400_e44
+        * console name matches the EARLYCON_DECLARE() statement, and
+        * SPCR is not used.  Parameter "earlycon" is false.
+        *
+        * If the user specifies "SPCR" earlycon, then we need to update
+        * the console name so that it also says "qdf2400_e44".  Parameter
+        * "earlycon" is true.
+        *
+        * For consistency, if we change the console name, then we do it
+        * for everyone, not just earlycon.
+        */
+       if (qdf2400_erratum_44_present(&table->header)) {
+               qdf2400_e44_present = true;
+               if (earlycon)
+                       uart = "qdf2400_e44";
+       }
+
        if (xgene_8250_erratum_present(table))
                iotype = "mmio32";
 
index b9f907eedbf770ee32359468d2b8d07e57bde667..bfbe1e15412889dfb40e699af7c3623a06d83253 100644 (file)
@@ -30,7 +30,6 @@
 #include <linux/syscore_ops.h>
 #include <linux/reboot.h>
 #include <linux/security.h>
-#include <linux/swait.h>
 
 #include <generated/utsrelease.h>
 
@@ -112,13 +111,13 @@ static inline long firmware_loading_timeout(void)
  * state of the firmware loading.
  */
 struct fw_state {
-       struct swait_queue_head wq;
+       struct completion completion;
        enum fw_status status;
 };
 
 static void fw_state_init(struct fw_state *fw_st)
 {
-       init_swait_queue_head(&fw_st->wq);
+       init_completion(&fw_st->completion);
        fw_st->status = FW_STATUS_UNKNOWN;
 }
 
@@ -131,9 +130,7 @@ static int __fw_state_wait_common(struct fw_state *fw_st, long timeout)
 {
        long ret;
 
-       ret = swait_event_interruptible_timeout(fw_st->wq,
-                               __fw_state_is_done(READ_ONCE(fw_st->status)),
-                               timeout);
+       ret = wait_for_completion_killable_timeout(&fw_st->completion, timeout);
        if (ret != 0 && fw_st->status == FW_STATUS_ABORTED)
                return -ENOENT;
        if (!ret)
@@ -148,35 +145,34 @@ static void __fw_state_set(struct fw_state *fw_st,
        WRITE_ONCE(fw_st->status, status);
 
        if (status == FW_STATUS_DONE || status == FW_STATUS_ABORTED)
-               swake_up(&fw_st->wq);
+               complete_all(&fw_st->completion);
 }
 
 #define fw_state_start(fw_st)                                  \
        __fw_state_set(fw_st, FW_STATUS_LOADING)
 #define fw_state_done(fw_st)                                   \
        __fw_state_set(fw_st, FW_STATUS_DONE)
+#define fw_state_aborted(fw_st)                                        \
+       __fw_state_set(fw_st, FW_STATUS_ABORTED)
 #define fw_state_wait(fw_st)                                   \
        __fw_state_wait_common(fw_st, MAX_SCHEDULE_TIMEOUT)
 
-#ifndef CONFIG_FW_LOADER_USER_HELPER
-
-#define fw_state_is_aborted(fw_st)     false
-
-#else /* CONFIG_FW_LOADER_USER_HELPER */
-
 static int __fw_state_check(struct fw_state *fw_st, enum fw_status status)
 {
        return fw_st->status == status;
 }
 
+#define fw_state_is_aborted(fw_st)                             \
+       __fw_state_check(fw_st, FW_STATUS_ABORTED)
+
+#ifdef CONFIG_FW_LOADER_USER_HELPER
+
 #define fw_state_aborted(fw_st)                                        \
        __fw_state_set(fw_st, FW_STATUS_ABORTED)
 #define fw_state_is_done(fw_st)                                        \
        __fw_state_check(fw_st, FW_STATUS_DONE)
 #define fw_state_is_loading(fw_st)                             \
        __fw_state_check(fw_st, FW_STATUS_LOADING)
-#define fw_state_is_aborted(fw_st)                             \
-       __fw_state_check(fw_st, FW_STATUS_ABORTED)
 #define fw_state_wait_timeout(fw_st, timeout)                  \
        __fw_state_wait_common(fw_st, timeout)
 
@@ -1200,6 +1196,28 @@ _request_firmware_prepare(struct firmware **firmware_p, const char *name,
        return 1; /* need to load */
 }
 
+/*
+ * Batched requests need only one wake, we need to do this step last due to the
+ * fallback mechanism. The buf is protected with kref_get(), and it won't be
+ * released until the last user calls release_firmware().
+ *
+ * Failed batched requests are possible as well, in such cases we just share
+ * the struct firmware_buf and won't release it until all requests are woken
+ * and have gone through this same path.
+ */
+static void fw_abort_batch_reqs(struct firmware *fw)
+{
+       struct firmware_buf *buf;
+
+       /* Loaded directly? */
+       if (!fw || !fw->priv)
+               return;
+
+       buf = fw->priv;
+       if (!fw_state_is_aborted(&buf->fw_st))
+               fw_state_aborted(&buf->fw_st);
+}
+
 /* called from request_firmware() and request_firmware_work_func() */
 static int
 _request_firmware(const struct firmware **firmware_p, const char *name,
@@ -1243,6 +1261,7 @@ _request_firmware(const struct firmware **firmware_p, const char *name,
 
  out:
        if (ret < 0) {
+               fw_abort_batch_reqs(fw);
                release_firmware(fw);
                fw = NULL;
        }
index 98e34e4c62b8b228a6ee6d37256e860ee7d3c2b7..2468c28d477110e6bcb2cbf5225dd481c6ec5b95 100644 (file)
@@ -2075,9 +2075,9 @@ static int blkfront_resume(struct xenbus_device *dev)
                        /*
                         * Get the bios in the request so we can re-queue them.
                         */
-                       if (req_op(shadow[i].request) == REQ_OP_FLUSH ||
-                           req_op(shadow[i].request) == REQ_OP_DISCARD ||
-                           req_op(shadow[i].request) == REQ_OP_SECURE_ERASE ||
+                       if (req_op(shadow[j].request) == REQ_OP_FLUSH ||
+                           req_op(shadow[j].request) == REQ_OP_DISCARD ||
+                           req_op(shadow[j].request) == REQ_OP_SECURE_ERASE ||
                            shadow[j].request->cmd_flags & REQ_FUA) {
                                /*
                                 * Flush operations don't contain bios, so
index fcae5ca6ac9234debecf5a4f84521405aa64a74c..54a67f8a28ebfb929610ea07697d7c7e4eb33de1 100644 (file)
@@ -262,7 +262,7 @@ config CLKSRC_LPC32XX
 
 config CLKSRC_PISTACHIO
        bool "Clocksource for Pistachio SoC" if COMPILE_TEST
-       depends on HAS_IOMEM
+       depends on GENERIC_CLOCKEVENTS && HAS_IOMEM
        select TIMER_OF
        help
          Enables the clocksource for the Pistachio SoC.
index c62e71614c75a0bf9ed9766aafb6033e31821ddd..fd4b7f684bd030151d388071e05a3c2cfcb5e2c4 100644 (file)
@@ -1444,7 +1444,7 @@ static int __init arch_timer_mem_acpi_init(int platform_timer_count)
         * While unlikely, it's theoretically possible that none of the frames
         * in a timer expose the combination of feature we want.
         */
-       for (i = i; i < timer_count; i++) {
+       for (i = 0; i < timer_count; i++) {
                timer = &timers[i];
 
                frame = arch_timer_mem_find_best_frame(timer);
index bc48cbf6a7957198d4ecd6bf91840df350843d48..269db74a065815e2211194fbda9774dcf1e80041 100644 (file)
@@ -305,7 +305,7 @@ static int em_sti_probe(struct platform_device *pdev)
        irq = platform_get_irq(pdev, 0);
        if (irq < 0) {
                dev_err(&pdev->dev, "failed to get irq\n");
-               return -EINVAL;
+               return irq;
        }
 
        /* map memory, let base point to the STI instance */
@@ -314,11 +314,12 @@ static int em_sti_probe(struct platform_device *pdev)
        if (IS_ERR(p->base))
                return PTR_ERR(p->base);
 
-       if (devm_request_irq(&pdev->dev, irq, em_sti_interrupt,
-                            IRQF_TIMER | IRQF_IRQPOLL | IRQF_NOBALANCING,
-                            dev_name(&pdev->dev), p)) {
+       ret = devm_request_irq(&pdev->dev, irq, em_sti_interrupt,
+                              IRQF_TIMER | IRQF_IRQPOLL | IRQF_NOBALANCING,
+                              dev_name(&pdev->dev), p);
+       if (ret) {
                dev_err(&pdev->dev, "failed to request low IRQ\n");
-               return -ENOENT;
+               return ret;
        }
 
        /* get hold of clock */
index d509b500a7b5f8565514352a0313cb4704976042..4d7aef9d9c15422df0e9973e72379b2e21256f47 100644 (file)
@@ -128,9 +128,9 @@ static __init int timer_base_init(struct device_node *np,
        const char *name = of_base->name ? of_base->name : np->full_name;
 
        of_base->base = of_io_request_and_map(np, of_base->index, name);
-       if (!of_base->base) {
+       if (IS_ERR(of_base->base)) {
                pr_err("Failed to iomap (%s)\n", name);
-               return -ENXIO;
+               return PTR_ERR(of_base->base);
        }
 
        return 0;
index 0566455f233ed3cd663506321f08f7ae546862be..65ee4fcace1f260632cf53c150c3260cc27240ad 100644 (file)
@@ -1613,8 +1613,7 @@ static inline bool intel_pstate_sample(struct cpudata *cpu, u64 time)
 
 static inline int32_t get_avg_frequency(struct cpudata *cpu)
 {
-       return mul_ext_fp(cpu->sample.core_avg_perf,
-                         cpu->pstate.max_pstate_physical * cpu->pstate.scaling);
+       return mul_ext_fp(cpu->sample.core_avg_perf, cpu_khz);
 }
 
 static inline int32_t get_avg_pstate(struct cpudata *cpu)
index 37b0698b7193e60be4107a8be107285e55c65877..42896a67aeae38325cbda2acb7ca655c1b43915c 100644 (file)
@@ -235,6 +235,7 @@ static inline int validate_dt_prop_sizes(const char *prop1, int prop1_len,
        return -1;
 }
 
+extern u32 pnv_get_supported_cpuidle_states(void);
 static int powernv_add_idle_states(void)
 {
        struct device_node *power_mgt;
@@ -248,6 +249,8 @@ static int powernv_add_idle_states(void)
        const char *names[CPUIDLE_STATE_MAX];
        u32 has_stop_states = 0;
        int i, rc;
+       u32 supported_flags = pnv_get_supported_cpuidle_states();
+
 
        /* Currently we have snooze statically defined */
 
@@ -362,6 +365,13 @@ static int powernv_add_idle_states(void)
        for (i = 0; i < dt_idle_states; i++) {
                unsigned int exit_latency, target_residency;
                bool stops_timebase = false;
+
+               /*
+                * Skip the platform idle state whose flag isn't in
+                * the supported_cpuidle_states flag mask.
+                */
+               if ((flags[i] & supported_flags) != flags[i])
+                       continue;
                /*
                 * If an idle state has exit latency beyond
                 * POWERNV_THRESHOLD_LATENCY_NS then don't use it
index 427cbe01272926acd38eba211dceb28fec61db64..dadc4a808df5a37764adf45df9f90dc1fce5b03a 100644 (file)
@@ -1073,7 +1073,7 @@ static int aead_perform(struct aead_request *req, int encrypt,
                req_ctx->hmac_virt = dma_pool_alloc(buffer_pool, flags,
                                &crypt->icv_rev_aes);
                if (unlikely(!req_ctx->hmac_virt))
-                       goto free_buf_src;
+                       goto free_buf_dst;
                if (!encrypt) {
                        scatterwalk_map_and_copy(req_ctx->hmac_virt,
                                req->src, cryptlen, authsize, 0);
@@ -1088,10 +1088,10 @@ static int aead_perform(struct aead_request *req, int encrypt,
        BUG_ON(qmgr_stat_overflow(SEND_QID));
        return -EINPROGRESS;
 
-free_buf_src:
-       free_buf_chain(dev, req_ctx->src, crypt->src_buf);
 free_buf_dst:
        free_buf_chain(dev, req_ctx->dst, crypt->dst_buf);
+free_buf_src:
+       free_buf_chain(dev, req_ctx->src, crypt->src_buf);
        crypt->ctl_flags = CTL_FLAG_UNUSED;
        return -ENOMEM;
 }
index e338c374356294f97c1b331c4c04a4977d495290..45c65f805fd6ba5a3899b5c29a9651fa2d6dcc43 100644 (file)
@@ -557,7 +557,7 @@ static void mvebu_gpio_irq_handler(struct irq_desc *desc)
        edge_cause = mvebu_gpio_read_edge_cause(mvchip);
        edge_mask  = mvebu_gpio_read_edge_mask(mvchip);
 
-       cause = (data_in ^ level_mask) | (edge_cause & edge_mask);
+       cause = (data_in & level_mask) | (edge_cause & edge_mask);
 
        for (i = 0; i < mvchip->chip.ngpio; i++) {
                int irq;
index 16fe9742597b540ff4d82e869dc13d66326bbd2a..fc80add5fedb57dc07b6f567eef138494452316e 100644 (file)
@@ -2,6 +2,7 @@
 #include <linux/mutex.h>
 #include <linux/device.h>
 #include <linux/sysfs.h>
+#include <linux/gpio.h>
 #include <linux/gpio/consumer.h>
 #include <linux/gpio/driver.h>
 #include <linux/interrupt.h>
@@ -432,6 +433,11 @@ static struct attribute *gpiochip_attrs[] = {
 };
 ATTRIBUTE_GROUPS(gpiochip);
 
+static struct gpio_desc *gpio_to_valid_desc(int gpio)
+{
+       return gpio_is_valid(gpio) ? gpio_to_desc(gpio) : NULL;
+}
+
 /*
  * /sys/class/gpio/export ... write-only
  *     integer N ... number of GPIO to export (full access)
@@ -450,7 +456,7 @@ static ssize_t export_store(struct class *class,
        if (status < 0)
                goto done;
 
-       desc = gpio_to_desc(gpio);
+       desc = gpio_to_valid_desc(gpio);
        /* reject invalid GPIOs */
        if (!desc) {
                pr_warn("%s: invalid GPIO %ld\n", __func__, gpio);
@@ -493,7 +499,7 @@ static ssize_t unexport_store(struct class *class,
        if (status < 0)
                goto done;
 
-       desc = gpio_to_desc(gpio);
+       desc = gpio_to_valid_desc(gpio);
        /* reject bogus commands (gpio_unexport ignores them) */
        if (!desc) {
                pr_warn("%s: invalid GPIO %ld\n", __func__, gpio);
index a6899180b265721831837282e6c66d12a8bccf48..c586f44312f9772fb93f8b1442827c842d610594 100644 (file)
@@ -244,6 +244,12 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
                struct dma_fence *f = e->fence;
                struct amd_sched_fence *s_fence = to_amd_sched_fence(f);
 
+               if (dma_fence_is_signaled(f)) {
+                       hash_del(&e->node);
+                       dma_fence_put(f);
+                       kmem_cache_free(amdgpu_sync_slab, e);
+                       continue;
+               }
                if (ring && s_fence) {
                        /* For fences from the same ring it is sufficient
                         * when they are scheduled.
@@ -256,13 +262,6 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
                        }
                }
 
-               if (dma_fence_is_signaled(f)) {
-                       hash_del(&e->node);
-                       dma_fence_put(f);
-                       kmem_cache_free(amdgpu_sync_slab, e);
-                       continue;
-               }
-
                return f;
        }
 
index 0511fce5c9478da035a150bace5b94767a988661..e3a81ed66bc23555a624d7b235fcade3b8ff2d21 100644 (file)
@@ -4581,7 +4581,7 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
 
                sseu->slice_mask |= BIT(s);
 
-               if (IS_GEN9_BC(dev_priv))
+               if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv))
                        sseu->subslice_mask =
                                INTEL_INFO(dev_priv)->sseu.subslice_mask;
 
index 39ed58a21fc1f517f56d179d48b3b9bef00e9410..e1e971ee2ed57aae59dd505a49e8a6ee546e58ff 100644 (file)
@@ -688,19 +688,19 @@ static inline bool skip_rcs_switch(struct i915_hw_ppgtt *ppgtt,
 }
 
 static bool
-needs_pd_load_pre(struct i915_hw_ppgtt *ppgtt,
-                 struct intel_engine_cs *engine,
-                 struct i915_gem_context *to)
+needs_pd_load_pre(struct i915_hw_ppgtt *ppgtt, struct intel_engine_cs *engine)
 {
+       struct i915_gem_context *from = engine->legacy_active_context;
+
        if (!ppgtt)
                return false;
 
        /* Always load the ppgtt on first use */
-       if (!engine->legacy_active_context)
+       if (!from)
                return true;
 
        /* Same context without new entries, skip */
-       if (engine->legacy_active_context == to &&
+       if ((!from->ppgtt || from->ppgtt == ppgtt) &&
            !(intel_engine_flag(engine) & ppgtt->pd_dirty_rings))
                return false;
 
@@ -744,7 +744,7 @@ static int do_rcs_switch(struct drm_i915_gem_request *req)
        if (skip_rcs_switch(ppgtt, engine, to))
                return 0;
 
-       if (needs_pd_load_pre(ppgtt, engine, to)) {
+       if (needs_pd_load_pre(ppgtt, engine)) {
                /* Older GENs and non render rings still want the load first,
                 * "PP_DCLV followed by PP_DIR_BASE register through Load
                 * Register Immediate commands in Ring Buffer before submitting
@@ -841,7 +841,7 @@ int i915_switch_context(struct drm_i915_gem_request *req)
                struct i915_hw_ppgtt *ppgtt =
                        to->ppgtt ?: req->i915->mm.aliasing_ppgtt;
 
-               if (needs_pd_load_pre(ppgtt, engine, to)) {
+               if (needs_pd_load_pre(ppgtt, engine)) {
                        int ret;
 
                        trace_switch_mm(engine, to);
@@ -852,6 +852,7 @@ int i915_switch_context(struct drm_i915_gem_request *req)
                        ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
                }
 
+               engine->legacy_active_context = to;
                return 0;
        }
 
index 7032c542a9b1d007c4bc69e328db040b88efc15d..4dd4c2159a92e26d91fd98e560c587a91b964e48 100644 (file)
@@ -242,6 +242,10 @@ int i915_gem_render_state_emit(struct drm_i915_gem_request *req)
                        goto err_unpin;
        }
 
+       ret = req->engine->emit_flush(req, EMIT_INVALIDATE);
+       if (ret)
+               goto err_unpin;
+
        ret = req->engine->emit_bb_start(req,
                                         so->batch_offset, so->batch_size,
                                         I915_DISPATCH_SECURE);
index 9edeaaef77adf6c38e23733769e271147698f02a..d3b3252a874252641af555d8797a72e20056a34b 100644 (file)
@@ -1762,7 +1762,7 @@ cnl_get_buf_trans_edp(struct drm_i915_private *dev_priv,
        if (dev_priv->vbt.edp.low_vswing) {
                if (voltage == VOLTAGE_INFO_0_85V) {
                        *n_entries = ARRAY_SIZE(cnl_ddi_translations_edp_0_85V);
-                       return cnl_ddi_translations_dp_0_85V;
+                       return cnl_ddi_translations_edp_0_85V;
                } else if (voltage == VOLTAGE_INFO_0_95V) {
                        *n_entries = ARRAY_SIZE(cnl_ddi_translations_edp_0_95V);
                        return cnl_ddi_translations_edp_0_95V;
index 9471c88d449eaf84adf3411d61135bf090be9b96..cc484b56eeaa33213a0832f43770132a1b165df1 100644 (file)
@@ -3485,6 +3485,13 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv)
            !gpu_reset_clobbers_display(dev_priv))
                return;
 
+       /* We have a modeset vs reset deadlock, defensively unbreak it.
+        *
+        * FIXME: We can do a _lot_ better, this is just a first iteration.
+        */
+       i915_gem_set_wedged(dev_priv);
+       DRM_DEBUG_DRIVER("Wedging GPU to avoid deadlocks with pending modeset updates\n");
+
        /*
         * Need mode_config.mutex so that we don't
         * trample ongoing ->detect() and whatnot.
index 52b3a1fd4059bab91a898f7f498a79af2c117a4b..57ef5833c4274958f3eed9f39b66a598dcfd1cd7 100644 (file)
@@ -63,7 +63,6 @@ enum {
 };
 
 /* Logical Rings */
-void intel_logical_ring_stop(struct intel_engine_cs *engine);
 void intel_logical_ring_cleanup(struct intel_engine_cs *engine);
 int logical_render_ring_init(struct intel_engine_cs *engine);
 int logical_xcs_ring_init(struct intel_engine_cs *engine);
index 6b5d3be283c4e7e00f72bb32e550b52ab06e1526..807299dd45ebf0663fbc97b3831bd8f92148ec20 100644 (file)
@@ -193,7 +193,6 @@ struct bmc150_accel_data {
        struct regmap *regmap;
        int irq;
        struct bmc150_accel_interrupt interrupts[BMC150_ACCEL_INTERRUPTS];
-       atomic_t active_intr;
        struct bmc150_accel_trigger triggers[BMC150_ACCEL_TRIGGERS];
        struct mutex mutex;
        u8 fifo_mode, watermark;
@@ -493,11 +492,6 @@ static int bmc150_accel_set_interrupt(struct bmc150_accel_data *data, int i,
                goto out_fix_power_state;
        }
 
-       if (state)
-               atomic_inc(&data->active_intr);
-       else
-               atomic_dec(&data->active_intr);
-
        return 0;
 
 out_fix_power_state:
@@ -1710,8 +1704,7 @@ static int bmc150_accel_resume(struct device *dev)
        struct bmc150_accel_data *data = iio_priv(indio_dev);
 
        mutex_lock(&data->mutex);
-       if (atomic_read(&data->active_intr))
-               bmc150_accel_set_mode(data, BMC150_ACCEL_SLEEP_MODE_NORMAL, 0);
+       bmc150_accel_set_mode(data, BMC150_ACCEL_SLEEP_MODE_NORMAL, 0);
        bmc150_accel_fifo_set_mode(data);
        mutex_unlock(&data->mutex);
 
index 07d1489cd457a6b5445b8b3ba35dad95b1792acc..e44f62bf9caa9f1a45c6699875b885d497bb0a7a 100644 (file)
@@ -166,6 +166,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_ihl = 0x02,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x23,
+                       .value = BIT(0),
+               },
                .multi_read_bit = true,
                .bootime = 2,
        },
@@ -234,6 +238,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_od = 0x40,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x23,
+                       .value = BIT(0),
+               },
                .multi_read_bit = true,
                .bootime = 2,
        },
@@ -316,6 +324,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                                .en_mask = 0x08,
                        },
                },
+               .sim = {
+                       .addr = 0x24,
+                       .value = BIT(0),
+               },
                .multi_read_bit = false,
                .bootime = 2,
        },
@@ -379,6 +391,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_int1 = 0x04,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x21,
+                       .value = BIT(1),
+               },
                .multi_read_bit = true,
                .bootime = 2, /* guess */
        },
@@ -437,6 +453,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_od = 0x40,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x21,
+                       .value = BIT(7),
+               },
                .multi_read_bit = false,
                .bootime = 2, /* guess */
        },
@@ -499,6 +519,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .addr_ihl = 0x22,
                        .mask_ihl = 0x80,
                },
+               .sim = {
+                       .addr = 0x23,
+                       .value = BIT(0),
+               },
                .multi_read_bit = true,
                .bootime = 2,
        },
@@ -547,6 +571,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_int1 = 0x04,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x21,
+                       .value = BIT(1),
+               },
                .multi_read_bit = false,
                .bootime = 2,
        },
@@ -614,6 +642,10 @@ static const struct st_sensor_settings st_accel_sensors_settings[] = {
                        .mask_ihl = 0x02,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
+               .sim = {
+                       .addr = 0x23,
+                       .value = BIT(0),
+               },
                .multi_read_bit = true,
                .bootime = 2,
        },
index e0ea411a0b2df9563085c70552086946843ba2ca..c02b23d675cbc1540ec47769515da714c784cf67 100644 (file)
@@ -22,6 +22,7 @@
 
 #include <linux/iio/iio.h>
 #include <linux/iio/driver.h>
+#include <linux/iopoll.h>
 
 #define ASPEED_RESOLUTION_BITS         10
 #define ASPEED_CLOCKS_PER_SAMPLE       12
 
 #define ASPEED_ENGINE_ENABLE           BIT(0)
 
+#define ASPEED_ADC_CTRL_INIT_RDY       BIT(8)
+
+#define ASPEED_ADC_INIT_POLLING_TIME   500
+#define ASPEED_ADC_INIT_TIMEOUT                500000
+
 struct aspeed_adc_model_data {
        const char *model_name;
        unsigned int min_sampling_rate; // Hz
        unsigned int max_sampling_rate; // Hz
        unsigned int vref_voltage;      // mV
+       bool wait_init_sequence;
 };
 
 struct aspeed_adc_data {
@@ -211,6 +218,24 @@ static int aspeed_adc_probe(struct platform_device *pdev)
                goto scaler_error;
        }
 
+       model_data = of_device_get_match_data(&pdev->dev);
+
+       if (model_data->wait_init_sequence) {
+               /* Enable engine in normal mode. */
+               writel(ASPEED_OPERATION_MODE_NORMAL | ASPEED_ENGINE_ENABLE,
+                      data->base + ASPEED_REG_ENGINE_CONTROL);
+
+               /* Wait for initial sequence complete. */
+               ret = readl_poll_timeout(data->base + ASPEED_REG_ENGINE_CONTROL,
+                                        adc_engine_control_reg_val,
+                                        adc_engine_control_reg_val &
+                                        ASPEED_ADC_CTRL_INIT_RDY,
+                                        ASPEED_ADC_INIT_POLLING_TIME,
+                                        ASPEED_ADC_INIT_TIMEOUT);
+               if (ret)
+                       goto scaler_error;
+       }
+
        /* Start all channels in normal mode. */
        ret = clk_prepare_enable(data->clk_scaler->clk);
        if (ret)
@@ -274,6 +299,7 @@ static const struct aspeed_adc_model_data ast2500_model_data = {
        .vref_voltage = 1800, // mV
        .min_sampling_rate = 1,
        .max_sampling_rate = 1000000,
+       .wait_init_sequence = true,
 };
 
 static const struct of_device_id aspeed_adc_matches[] = {
index 64799ad7ebad02a797607470678aa500842ed99f..462a99c13e7a210a74f1d9b2a9c87de7ebc3fad6 100644 (file)
@@ -28,6 +28,8 @@
 #include <linux/iio/driver.h>
 
 #define AXP288_ADC_EN_MASK             0xF1
+#define AXP288_ADC_TS_PIN_GPADC                0xF2
+#define AXP288_ADC_TS_PIN_ON           0xF3
 
 enum axp288_adc_id {
        AXP288_ADC_TS,
@@ -121,6 +123,26 @@ static int axp288_adc_read_channel(int *val, unsigned long address,
        return IIO_VAL_INT;
 }
 
+static int axp288_adc_set_ts(struct regmap *regmap, unsigned int mode,
+                               unsigned long address)
+{
+       int ret;
+
+       /* channels other than GPADC do not need to switch TS pin */
+       if (address != AXP288_GP_ADC_H)
+               return 0;
+
+       ret = regmap_write(regmap, AXP288_ADC_TS_PIN_CTRL, mode);
+       if (ret)
+               return ret;
+
+       /* When switching to the GPADC pin give things some time to settle */
+       if (mode == AXP288_ADC_TS_PIN_GPADC)
+               usleep_range(6000, 10000);
+
+       return 0;
+}
+
 static int axp288_adc_read_raw(struct iio_dev *indio_dev,
                        struct iio_chan_spec const *chan,
                        int *val, int *val2, long mask)
@@ -131,7 +153,16 @@ static int axp288_adc_read_raw(struct iio_dev *indio_dev,
        mutex_lock(&indio_dev->mlock);
        switch (mask) {
        case IIO_CHAN_INFO_RAW:
+               if (axp288_adc_set_ts(info->regmap, AXP288_ADC_TS_PIN_GPADC,
+                                       chan->address)) {
+                       dev_err(&indio_dev->dev, "GPADC mode\n");
+                       ret = -EINVAL;
+                       break;
+               }
                ret = axp288_adc_read_channel(val, chan->address, info->regmap);
+               if (axp288_adc_set_ts(info->regmap, AXP288_ADC_TS_PIN_ON,
+                                               chan->address))
+                       dev_err(&indio_dev->dev, "TS pin restore\n");
                break;
        default:
                ret = -EINVAL;
@@ -141,6 +172,15 @@ static int axp288_adc_read_raw(struct iio_dev *indio_dev,
        return ret;
 }
 
+static int axp288_adc_set_state(struct regmap *regmap)
+{
+       /* ADC should be always enabled for internal FG to function */
+       if (regmap_write(regmap, AXP288_ADC_TS_PIN_CTRL, AXP288_ADC_TS_PIN_ON))
+               return -EIO;
+
+       return regmap_write(regmap, AXP20X_ADC_EN1, AXP288_ADC_EN_MASK);
+}
+
 static const struct iio_info axp288_adc_iio_info = {
        .read_raw = &axp288_adc_read_raw,
        .driver_module = THIS_MODULE,
@@ -169,7 +209,7 @@ static int axp288_adc_probe(struct platform_device *pdev)
         * Set ADC to enabled state at all time, including system suspend.
         * otherwise internal fuel gauge functionality may be affected.
         */
-       ret = regmap_write(info->regmap, AXP20X_ADC_EN1, AXP288_ADC_EN_MASK);
+       ret = axp288_adc_set_state(axp20x->regmap);
        if (ret) {
                dev_err(&pdev->dev, "unable to enable ADC device\n");
                return ret;
index 81d4c39e414a4da6b0f8df0ebd8922d371a909d7..137f577d94326a299e0f944edfd108bdc2db31fe 100644 (file)
@@ -256,6 +256,7 @@ static int sun4i_gpadc_read(struct iio_dev *indio_dev, int channel, int *val,
 
 err:
        pm_runtime_put_autosuspend(indio_dev->dev.parent);
+       disable_irq(irq);
        mutex_unlock(&info->mutex);
 
        return ret;
@@ -365,7 +366,6 @@ static irqreturn_t sun4i_gpadc_temp_data_irq_handler(int irq, void *dev_id)
                complete(&info->completion);
 
 out:
-       disable_irq_nosync(info->temp_data_irq);
        return IRQ_HANDLED;
 }
 
@@ -380,7 +380,6 @@ static irqreturn_t sun4i_gpadc_fifo_data_irq_handler(int irq, void *dev_id)
                complete(&info->completion);
 
 out:
-       disable_irq_nosync(info->fifo_data_irq);
        return IRQ_HANDLED;
 }
 
index 01fc76f7d6602c7b14c5f0cd4ea0c01090278087..c168e0db329ab49b6b59d720cc62cb2b37352088 100644 (file)
@@ -77,7 +77,7 @@
 #define VF610_ADC_ADSTS_MASK           0x300
 #define VF610_ADC_ADLPC_EN             0x80
 #define VF610_ADC_ADHSC_EN             0x400
-#define VF610_ADC_REFSEL_VALT          0x100
+#define VF610_ADC_REFSEL_VALT          0x800
 #define VF610_ADC_REFSEL_VBG           0x1000
 #define VF610_ADC_ADTRG_HARD           0x2000
 #define VF610_ADC_AVGS_8               0x4000
index 79c8c7cd70d5c6d74fc2e32cad372f8644233c6b..6e6a1ecc99ddf4b69252b6b4232c9cda6f00d4be 100644 (file)
@@ -550,6 +550,31 @@ out:
 }
 EXPORT_SYMBOL(st_sensors_read_info_raw);
 
+static int st_sensors_init_interface_mode(struct iio_dev *indio_dev,
+                       const struct st_sensor_settings *sensor_settings)
+{
+       struct st_sensor_data *sdata = iio_priv(indio_dev);
+       struct device_node *np = sdata->dev->of_node;
+       struct st_sensors_platform_data *pdata;
+
+       pdata = (struct st_sensors_platform_data *)sdata->dev->platform_data;
+       if (((np && of_property_read_bool(np, "spi-3wire")) ||
+            (pdata && pdata->spi_3wire)) && sensor_settings->sim.addr) {
+               int err;
+
+               err = sdata->tf->write_byte(&sdata->tb, sdata->dev,
+                                           sensor_settings->sim.addr,
+                                           sensor_settings->sim.value);
+               if (err < 0) {
+                       dev_err(&indio_dev->dev,
+                               "failed to init interface mode\n");
+                       return err;
+               }
+       }
+
+       return 0;
+}
+
 int st_sensors_check_device_support(struct iio_dev *indio_dev,
                        int num_sensors_list,
                        const struct st_sensor_settings *sensor_settings)
@@ -574,6 +599,10 @@ int st_sensors_check_device_support(struct iio_dev *indio_dev,
                return -ENODEV;
        }
 
+       err = st_sensors_init_interface_mode(indio_dev, &sensor_settings[i]);
+       if (err < 0)
+               return err;
+
        if (sensor_settings[i].wai_addr) {
                err = sdata->tf->read_byte(&sdata->tb, sdata->dev,
                                           sensor_settings[i].wai_addr, &wai);
index e7d4ea75e007c0bd82ef087812524258ec69f490..7599693f7fe9597cb750319ccb2635362c1dc9f0 100644 (file)
@@ -626,7 +626,7 @@ static irqreturn_t tsl2563_event_handler(int irq, void *private)
        struct tsl2563_chip *chip = iio_priv(dev_info);
 
        iio_push_event(dev_info,
-                      IIO_UNMOD_EVENT_CODE(IIO_LIGHT,
+                      IIO_UNMOD_EVENT_CODE(IIO_INTENSITY,
                                            0,
                                            IIO_EV_TYPE_THRESH,
                                            IIO_EV_DIR_EITHER),
index aa61ec15c1396ca3925ecf1a099fbf91a302dae7..f1bce05ffa135703792f24317db81375d3e4ebb3 100644 (file)
@@ -456,7 +456,7 @@ static const struct st_sensor_settings st_press_sensors_settings[] = {
                        .mask_od = 0x40,
                        .addr_stat_drdy = ST_SENSORS_DEFAULT_STAT_ADDR,
                },
-               .multi_read_bit = true,
+               .multi_read_bit = false,
                .bootime = 2,
        },
 };
index a5dfab6adf495b88e846ee03b815536a7917bf48..221468f77128440acb2f5956d2ecf25bac7cc80d 100644 (file)
@@ -537,10 +537,11 @@ void ib_unregister_device(struct ib_device *device)
        }
        up_read(&lists_rwsem);
 
-       mutex_unlock(&device_mutex);
-
        ib_device_unregister_rdmacg(device);
        ib_device_unregister_sysfs(device);
+
+       mutex_unlock(&device_mutex);
+
        ib_cache_cleanup_one(device);
 
        ib_security_destroy_port_pkey_list(device);
index c551d2b275fdf339310a087bef9c6e821d7c7e09..739bd69ef1d47fa163ea9f57cf86cbbf09f5fba7 100644 (file)
@@ -1015,7 +1015,7 @@ static struct ib_ucq_object *create_cq(struct ib_uverbs_file *file,
        cq->uobject       = &obj->uobject;
        cq->comp_handler  = ib_uverbs_comp_handler;
        cq->event_handler = ib_uverbs_cq_event_handler;
-       cq->cq_context    = &ev_file->ev_queue;
+       cq->cq_context    = ev_file ? &ev_file->ev_queue : NULL;
        atomic_set(&cq->usecnt, 0);
 
        obj->uobject.object = cq;
@@ -1522,6 +1522,7 @@ static int create_qp(struct ib_uverbs_file *file,
                qp->qp_type       = attr.qp_type;
                atomic_set(&qp->usecnt, 0);
                atomic_inc(&pd->usecnt);
+               qp->port = 0;
                if (attr.send_cq)
                        atomic_inc(&attr.send_cq->usecnt);
                if (attr.recv_cq)
@@ -1962,8 +1963,9 @@ static int modify_qp(struct ib_uverbs_file *file,
        attr->alt_timeout         = cmd->base.alt_timeout;
        attr->rate_limit          = cmd->rate_limit;
 
-       attr->ah_attr.type = rdma_ah_find_type(qp->device,
-                                              cmd->base.dest.port_num);
+       if (cmd->base.attr_mask & IB_QP_AV)
+               attr->ah_attr.type = rdma_ah_find_type(qp->device,
+                                                      cmd->base.dest.port_num);
        if (cmd->base.dest.is_global) {
                rdma_ah_set_grh(&attr->ah_attr, NULL,
                                cmd->base.dest.flow_label,
@@ -1981,8 +1983,9 @@ static int modify_qp(struct ib_uverbs_file *file,
        rdma_ah_set_port_num(&attr->ah_attr,
                             cmd->base.dest.port_num);
 
-       attr->alt_ah_attr.type = rdma_ah_find_type(qp->device,
-                                                  cmd->base.dest.port_num);
+       if (cmd->base.attr_mask & IB_QP_ALT_PATH)
+               attr->alt_ah_attr.type =
+                       rdma_ah_find_type(qp->device, cmd->base.dest.port_num);
        if (cmd->base.alt_dest.is_global) {
                rdma_ah_set_grh(&attr->alt_ah_attr, NULL,
                                cmd->base.alt_dest.flow_label,
index c023e2c81b8f2b06443452f91edcc506b46b6d17..5e530d2bee4448ddcb68724199a5c93d36eec32a 100644 (file)
@@ -1153,7 +1153,6 @@ static void ib_uverbs_free_hw_resources(struct ib_uverbs_device *uverbs_dev,
                kref_get(&file->ref);
                mutex_unlock(&uverbs_dev->lists_mutex);
 
-               ib_uverbs_event_handler(&file->event_handler, &event);
 
                mutex_lock(&file->cleanup_mutex);
                ucontext = file->ucontext;
@@ -1170,6 +1169,7 @@ static void ib_uverbs_free_hw_resources(struct ib_uverbs_device *uverbs_dev,
                         * for example due to freeing the resources
                         * (e.g mmput).
                         */
+                       ib_uverbs_event_handler(&file->event_handler, &event);
                        ib_dev->disassociate_ucontext(ucontext);
                        mutex_lock(&file->cleanup_mutex);
                        ib_uverbs_cleanup_ucontext(file, ucontext, true);
index 7f8fe443df46f5b562ac3b2561e19226e3ab6b68..b456e3ca1876ae8db6bb5aeeb34dc8e7e48eab88 100644 (file)
@@ -838,6 +838,7 @@ struct ib_qp *ib_create_qp(struct ib_pd *pd,
        spin_lock_init(&qp->mr_lock);
        INIT_LIST_HEAD(&qp->rdma_mrs);
        INIT_LIST_HEAD(&qp->sig_mrs);
+       qp->port = 0;
 
        if (qp_init_attr->qp_type == IB_QPT_XRC_TGT)
                return ib_create_xrc_qp(qp, qp_init_attr);
@@ -1297,7 +1298,11 @@ int ib_modify_qp_with_udata(struct ib_qp *qp, struct ib_qp_attr *attr,
                if (ret)
                        return ret;
        }
-       return ib_security_modify_qp(qp, attr, attr_mask, udata);
+       ret = ib_security_modify_qp(qp, attr, attr_mask, udata);
+       if (!ret && (attr_mask & IB_QP_PORT))
+               qp->port = attr->port_num;
+
+       return ret;
 }
 EXPORT_SYMBOL(ib_modify_qp_with_udata);
 
index 5332f06b99ba4cfa3ef404ec816d09fb68ec0ddf..c2fba76becd4e985a35b9ed3d1ec1314c955706b 100644 (file)
@@ -661,7 +661,7 @@ struct ib_mr *c4iw_alloc_mr(struct ib_pd *pd,
        rhp = php->rhp;
 
        if (mr_type != IB_MR_TYPE_MEM_REG ||
-           max_num_sg > t4_max_fr_depth(&rhp->rdev.lldi.ulptx_memwrite_dsgl &&
+           max_num_sg > t4_max_fr_depth(rhp->rdev.lldi.ulptx_memwrite_dsgl &&
                                         use_dsgl))
                return ERR_PTR(-EINVAL);
 
index f78a733a63ec7e1f884930a72e53ba2100ee888e..d545302b8ef8c520d4eceb630a39db06b7a584a4 100644 (file)
@@ -64,8 +64,10 @@ struct ib_ah *hns_roce_create_ah(struct ib_pd *ibpd,
        } else {
                u8 *dmac = rdma_ah_retrieve_dmac(ah_attr);
 
-               if (!dmac)
+               if (!dmac) {
+                       kfree(ah);
                        return ERR_PTR(-EINVAL);
+               }
                memcpy(ah->av.mac, dmac, ETH_ALEN);
        }
 
index 9ec1ae9a82c9843878ea10266b0d5338d8862308..a49ff2eb6fb3bca54f40bc460238eeb4a5b4ba52 100644 (file)
@@ -130,20 +130,32 @@ static enum i40iw_status_code i40iw_sc_parse_fpm_commit_buf(
        u64 base = 0;
        u32 i, j;
        u32 k = 0;
-       u32 low;
 
        /* copy base values in obj_info */
-       for (i = I40IW_HMC_IW_QP, j = 0;
-                       i <= I40IW_HMC_IW_PBLE; i++, j += 8) {
+       for (i = I40IW_HMC_IW_QP, j = 0; i <= I40IW_HMC_IW_PBLE; i++, j += 8) {
+               if ((i == I40IW_HMC_IW_SRQ) ||
+                       (i == I40IW_HMC_IW_FSIMC) ||
+                       (i == I40IW_HMC_IW_FSIAV)) {
+                       info[i].base = 0;
+                       info[i].cnt = 0;
+                       continue;
+               }
                get_64bit_val(buf, j, &temp);
                info[i].base = RS_64_1(temp, 32) * 512;
                if (info[i].base > base) {
                        base = info[i].base;
                        k = i;
                }
-               low = (u32)(temp);
-               if (low)
-                       info[i].cnt = low;
+               if (i == I40IW_HMC_IW_APBVT_ENTRY) {
+                       info[i].cnt = 1;
+                       continue;
+               }
+               if (i == I40IW_HMC_IW_QP)
+                       info[i].cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_QPS);
+               else if (i == I40IW_HMC_IW_CQ)
+                       info[i].cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_CQS);
+               else
+                       info[i].cnt = (u32)(temp);
        }
        size = info[k].cnt * info[k].size + info[k].base;
        if (size & 0x1FFFFF)
@@ -154,6 +166,31 @@ static enum i40iw_status_code i40iw_sc_parse_fpm_commit_buf(
        return 0;
 }
 
+/**
+ * i40iw_sc_decode_fpm_query() - Decode a 64 bit value into max count and size
+ * @buf: ptr to fpm query buffer
+ * @buf_idx: index into buf
+ * @info: ptr to i40iw_hmc_obj_info struct
+ * @rsrc_idx: resource index into info
+ *
+ * Decode a 64 bit value from fpm query buffer into max count and size
+ */
+static u64 i40iw_sc_decode_fpm_query(u64 *buf,
+                                           u32 buf_idx,
+                                           struct i40iw_hmc_obj_info *obj_info,
+                                           u32 rsrc_idx)
+{
+       u64 temp;
+       u32 size;
+
+       get_64bit_val(buf, buf_idx, &temp);
+       obj_info[rsrc_idx].max_cnt = (u32)temp;
+       size = (u32)RS_64_1(temp, 32);
+       obj_info[rsrc_idx].size = LS_64_1(1, size);
+
+       return temp;
+}
+
 /**
  * i40iw_sc_parse_fpm_query_buf() - parses fpm query buffer
  * @buf: ptr to fpm query buffer
@@ -168,9 +205,9 @@ static enum i40iw_status_code i40iw_sc_parse_fpm_query_buf(
                                struct i40iw_hmc_info *hmc_info,
                                struct i40iw_hmc_fpm_misc *hmc_fpm_misc)
 {
-       u64 temp;
        struct i40iw_hmc_obj_info *obj_info;
-       u32 i, j, size;
+       u64 temp;
+       u32 size;
        u16 max_pe_sds;
 
        obj_info = hmc_info->hmc_obj;
@@ -185,41 +222,52 @@ static enum i40iw_status_code i40iw_sc_parse_fpm_query_buf(
        hmc_fpm_misc->max_sds = max_pe_sds;
        hmc_info->sd_table.sd_cnt = max_pe_sds + hmc_info->first_sd_index;
 
-       for (i = I40IW_HMC_IW_QP, j = 8;
-            i <= I40IW_HMC_IW_ARP; i++, j += 8) {
-               get_64bit_val(buf, j, &temp);
-               if (i == I40IW_HMC_IW_QP)
-                       obj_info[i].max_cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_QPS);
-               else if (i == I40IW_HMC_IW_CQ)
-                       obj_info[i].max_cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_CQS);
-               else
-                       obj_info[i].max_cnt = (u32)temp;
+       get_64bit_val(buf, 8, &temp);
+       obj_info[I40IW_HMC_IW_QP].max_cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_QPS);
+       size = (u32)RS_64_1(temp, 32);
+       obj_info[I40IW_HMC_IW_QP].size = LS_64_1(1, size);
 
-               size = (u32)RS_64_1(temp, 32);
-               obj_info[i].size = ((u64)1 << size);
-       }
-       for (i = I40IW_HMC_IW_MR, j = 48;
-                       i <= I40IW_HMC_IW_PBLE; i++, j += 8) {
-               get_64bit_val(buf, j, &temp);
-               obj_info[i].max_cnt = (u32)temp;
-               size = (u32)RS_64_1(temp, 32);
-               obj_info[i].size = LS_64_1(1, size);
-       }
+       get_64bit_val(buf, 16, &temp);
+       obj_info[I40IW_HMC_IW_CQ].max_cnt = (u32)RS_64(temp, I40IW_QUERY_FPM_MAX_CQS);
+       size = (u32)RS_64_1(temp, 32);
+       obj_info[I40IW_HMC_IW_CQ].size = LS_64_1(1, size);
+
+       i40iw_sc_decode_fpm_query(buf, 32, obj_info, I40IW_HMC_IW_HTE);
+       i40iw_sc_decode_fpm_query(buf, 40, obj_info, I40IW_HMC_IW_ARP);
+
+       obj_info[I40IW_HMC_IW_APBVT_ENTRY].size = 8192;
+       obj_info[I40IW_HMC_IW_APBVT_ENTRY].max_cnt = 1;
+
+       i40iw_sc_decode_fpm_query(buf, 48, obj_info, I40IW_HMC_IW_MR);
+       i40iw_sc_decode_fpm_query(buf, 56, obj_info, I40IW_HMC_IW_XF);
 
-       get_64bit_val(buf, 120, &temp);
-       hmc_fpm_misc->max_ceqs = (u8)RS_64(temp, I40IW_QUERY_FPM_MAX_CEQS);
-       get_64bit_val(buf, 120, &temp);
-       hmc_fpm_misc->ht_multiplier = RS_64(temp, I40IW_QUERY_FPM_HTMULTIPLIER);
-       get_64bit_val(buf, 120, &temp);
-       hmc_fpm_misc->timer_bucket = RS_64(temp, I40IW_QUERY_FPM_TIMERBUCKET);
        get_64bit_val(buf, 64, &temp);
+       obj_info[I40IW_HMC_IW_XFFL].max_cnt = (u32)temp;
+       obj_info[I40IW_HMC_IW_XFFL].size = 4;
        hmc_fpm_misc->xf_block_size = RS_64(temp, I40IW_QUERY_FPM_XFBLOCKSIZE);
        if (!hmc_fpm_misc->xf_block_size)
                return I40IW_ERR_INVALID_SIZE;
+
+       i40iw_sc_decode_fpm_query(buf, 72, obj_info, I40IW_HMC_IW_Q1);
+
        get_64bit_val(buf, 80, &temp);
+       obj_info[I40IW_HMC_IW_Q1FL].max_cnt = (u32)temp;
+       obj_info[I40IW_HMC_IW_Q1FL].size = 4;
        hmc_fpm_misc->q1_block_size = RS_64(temp, I40IW_QUERY_FPM_Q1BLOCKSIZE);
        if (!hmc_fpm_misc->q1_block_size)
                return I40IW_ERR_INVALID_SIZE;
+
+       i40iw_sc_decode_fpm_query(buf, 88, obj_info, I40IW_HMC_IW_TIMER);
+
+       get_64bit_val(buf, 112, &temp);
+       obj_info[I40IW_HMC_IW_PBLE].max_cnt = (u32)temp;
+       obj_info[I40IW_HMC_IW_PBLE].size = 8;
+
+       get_64bit_val(buf, 120, &temp);
+       hmc_fpm_misc->max_ceqs = (u8)RS_64(temp, I40IW_QUERY_FPM_MAX_CEQS);
+       hmc_fpm_misc->ht_multiplier = RS_64(temp, I40IW_QUERY_FPM_HTMULTIPLIER);
+       hmc_fpm_misc->timer_bucket = RS_64(temp, I40IW_QUERY_FPM_TIMERBUCKET);
+
        return 0;
 }
 
@@ -3392,13 +3440,6 @@ enum i40iw_status_code i40iw_sc_init_iw_hmc(struct i40iw_sc_dev *dev, u8 hmc_fn_
                hmc_info->sd_table.sd_entry = virt_mem.va;
        }
 
-       /* fill size of objects which are fixed */
-       hmc_info->hmc_obj[I40IW_HMC_IW_XFFL].size = 4;
-       hmc_info->hmc_obj[I40IW_HMC_IW_Q1FL].size = 4;
-       hmc_info->hmc_obj[I40IW_HMC_IW_PBLE].size = 8;
-       hmc_info->hmc_obj[I40IW_HMC_IW_APBVT_ENTRY].size = 8192;
-       hmc_info->hmc_obj[I40IW_HMC_IW_APBVT_ENTRY].max_cnt = 1;
-
        return ret_code;
 }
 
@@ -4840,7 +4881,7 @@ void i40iw_vsi_stats_free(struct i40iw_sc_vsi *vsi)
 {
        u8 fcn_id = vsi->fcn_id;
 
-       if ((vsi->stats_fcn_id_alloc) && (fcn_id != I40IW_INVALID_FCN_ID))
+       if (vsi->stats_fcn_id_alloc && fcn_id < I40IW_MAX_STATS_COUNT)
                vsi->dev->fcn_id_array[fcn_id] = false;
        i40iw_hw_stats_stop_timer(vsi);
 }
index a39ac12b6a7e8430b80e42d9a38d7abd3b48d3b8..2ebaadbed379406e9d6c10f01fe0577ab515ebc9 100644 (file)
@@ -1507,8 +1507,8 @@ enum {
        I40IW_CQ0_ALIGNMENT_MASK =              (256 - 1),
        I40IW_HOST_CTX_ALIGNMENT_MASK =         (4 - 1),
        I40IW_SHADOWAREA_MASK =                 (128 - 1),
-       I40IW_FPM_QUERY_BUF_ALIGNMENT_MASK =    0,
-       I40IW_FPM_COMMIT_BUF_ALIGNMENT_MASK =   0
+       I40IW_FPM_QUERY_BUF_ALIGNMENT_MASK =    (4 - 1),
+       I40IW_FPM_COMMIT_BUF_ALIGNMENT_MASK =   (4 - 1)
 };
 
 enum i40iw_alignment {
index 71050c5d29a05f3f6cb9433aafd7e09fbb5695aa..7f5583d83622a57821d1d4718c6bbc49d84410d0 100644 (file)
@@ -685,7 +685,7 @@ static enum i40iw_status_code i40iw_puda_cq_create(struct i40iw_puda_rsrc *rsrc)
        cqsize = rsrc->cq_size * (sizeof(struct i40iw_cqe));
        tsize = cqsize + sizeof(struct i40iw_cq_shadow_area);
        ret = i40iw_allocate_dma_mem(dev->hw, &rsrc->cqmem, tsize,
-                                    I40IW_CQ0_ALIGNMENT_MASK);
+                                    I40IW_CQ0_ALIGNMENT);
        if (ret)
                return ret;
 
index 91c421762f06797be844f0834d7ff09ff36bb5f9..f7013f11d8085966bf6d9dab310f0f8d0f35422a 100644 (file)
@@ -62,7 +62,7 @@ enum i40iw_status_code {
        I40IW_ERR_INVALID_ALIGNMENT = -23,
        I40IW_ERR_FLUSHED_QUEUE = -24,
        I40IW_ERR_INVALID_PUSH_PAGE_INDEX = -25,
-       I40IW_ERR_INVALID_IMM_DATA_SIZE = -26,
+       I40IW_ERR_INVALID_INLINE_DATA_SIZE = -26,
        I40IW_ERR_TIMEOUT = -27,
        I40IW_ERR_OPCODE_MISMATCH = -28,
        I40IW_ERR_CQP_COMPL_ERROR = -29,
index b0d3a0e8a9b522a3549d63593a0af38f682e71ba..1060725d18bce852a108d6ce9e0b55cb460bfb0d 100644 (file)
@@ -435,7 +435,7 @@ static enum i40iw_status_code i40iw_inline_rdma_write(struct i40iw_qp_uk *qp,
 
        op_info = &info->op.inline_rdma_write;
        if (op_info->len > I40IW_MAX_INLINE_DATA_SIZE)
-               return I40IW_ERR_INVALID_IMM_DATA_SIZE;
+               return I40IW_ERR_INVALID_INLINE_DATA_SIZE;
 
        ret_code = i40iw_inline_data_size_to_wqesize(op_info->len, &wqe_size);
        if (ret_code)
@@ -511,7 +511,7 @@ static enum i40iw_status_code i40iw_inline_send(struct i40iw_qp_uk *qp,
 
        op_info = &info->op.inline_send;
        if (op_info->len > I40IW_MAX_INLINE_DATA_SIZE)
-               return I40IW_ERR_INVALID_IMM_DATA_SIZE;
+               return I40IW_ERR_INVALID_INLINE_DATA_SIZE;
 
        ret_code = i40iw_inline_data_size_to_wqesize(op_info->len, &wqe_size);
        if (ret_code)
@@ -784,7 +784,7 @@ static enum i40iw_status_code i40iw_cq_poll_completion(struct i40iw_cq_uk *cq,
        get_64bit_val(cqe, 0, &qword0);
        get_64bit_val(cqe, 16, &qword2);
 
-       info->tcp_seq_num = (u8)RS_64(qword0, I40IWCQ_TCPSEQNUM);
+       info->tcp_seq_num = (u32)RS_64(qword0, I40IWCQ_TCPSEQNUM);
 
        info->qp_id = (u32)RS_64(qword2, I40IWCQ_QPID);
 
@@ -1187,7 +1187,7 @@ enum i40iw_status_code i40iw_inline_data_size_to_wqesize(u32 data_size,
                                                         u8 *wqe_size)
 {
        if (data_size > I40IW_MAX_INLINE_DATA_SIZE)
-               return I40IW_ERR_INVALID_IMM_DATA_SIZE;
+               return I40IW_ERR_INVALID_INLINE_DATA_SIZE;
 
        if (data_size <= 16)
                *wqe_size = I40IW_QP_WQE_MIN_SIZE;
index a7f2e60085c46c2300e3695029fbe1373cc0d480..f7fcde1ff0aae66fef09d6eaccf8f9de6b008ba1 100644 (file)
@@ -1085,6 +1085,12 @@ static int mlx5_ib_modify_port(struct ib_device *ibdev, u8 port, int mask,
        bool is_ib = (mlx5_ib_port_link_layer(ibdev, port) ==
                      IB_LINK_LAYER_INFINIBAND);
 
+       /* CM layer calls ib_modify_port() regardless of the link layer. For
+        * Ethernet ports, qkey violation and Port capabilities are meaningless.
+        */
+       if (!is_ib)
+               return 0;
+
        if (MLX5_CAP_GEN(dev->mdev, ib_virt) && is_ib) {
                change_mask = props->clr_port_cap_mask | props->set_port_cap_mask;
                value = ~props->clr_port_cap_mask | props->set_port_cap_mask;
index 0889ff367c8625f694346136c22bd510d2dea2a7..f58f8f5f3ebe10e8270606327b58440efd70aae3 100644 (file)
@@ -1238,6 +1238,7 @@ static int create_raw_packet_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
                        goto err_destroy_tis;
 
                sq->base.container_mibqp = qp;
+               sq->base.mqp.event = mlx5_ib_qp_event;
        }
 
        if (qp->rq.wqe_cnt) {
index 69bda611d31385044766915410ac72eb70cffac5..90aa326fd7c0974e0c5a70eb5d2b46f50d03c490 100644 (file)
@@ -65,13 +65,28 @@ int pvrdma_req_notify_cq(struct ib_cq *ibcq,
        struct pvrdma_dev *dev = to_vdev(ibcq->device);
        struct pvrdma_cq *cq = to_vcq(ibcq);
        u32 val = cq->cq_handle;
+       unsigned long flags;
+       int has_data = 0;
 
        val |= (notify_flags & IB_CQ_SOLICITED_MASK) == IB_CQ_SOLICITED ?
                PVRDMA_UAR_CQ_ARM_SOL : PVRDMA_UAR_CQ_ARM;
 
+       spin_lock_irqsave(&cq->cq_lock, flags);
+
        pvrdma_write_uar_cq(dev, val);
 
-       return 0;
+       if (notify_flags & IB_CQ_REPORT_MISSED_EVENTS) {
+               unsigned int head;
+
+               has_data = pvrdma_idx_ring_has_data(&cq->ring_state->rx,
+                                                   cq->ibcq.cqe, &head);
+               if (unlikely(has_data == PVRDMA_INVALID_IDX))
+                       dev_err(&dev->pdev->dev, "CQ ring state invalid\n");
+       }
+
+       spin_unlock_irqrestore(&cq->cq_lock, flags);
+
+       return has_data;
 }
 
 /**
index 3b616cb7c67f88512e82dbc8ba5f61e565cbd2a5..714cf7f9b13859988ca7df9f1fc98a10e0e6fbae 100644 (file)
@@ -1248,6 +1248,10 @@ static const struct acpi_device_id elan_acpi_id[] = {
        { "ELAN0100", 0 },
        { "ELAN0600", 0 },
        { "ELAN0605", 0 },
+       { "ELAN0608", 0 },
+       { "ELAN0605", 0 },
+       { "ELAN0609", 0 },
+       { "ELAN060B", 0 },
        { "ELAN1000", 0 },
        { }
 };
index 922ea02edcc3ef6c091b572275245e7d5fabd3f0..20b5b21c1bba8892f37aab4a0190dfd6352b72d1 100644 (file)
@@ -380,8 +380,8 @@ int trackpoint_detect(struct psmouse *psmouse, bool set_properties)
                return 0;
 
        if (trackpoint_read(ps2dev, TP_EXT_BTN, &button_info)) {
-               psmouse_warn(psmouse, "failed to get extended button data\n");
-               button_info = 0;
+               psmouse_warn(psmouse, "failed to get extended button data, assuming 3 buttons\n");
+               button_info = 0x33;
        }
 
        psmouse->private = kzalloc(sizeof(struct trackpoint_data), GFP_KERNEL);
index b97188acc4f1006185a5f8cb4fb3ee0e5cf735cf..2d80fa8a0634aba34b366609d8bcc50f432bb31c 100644 (file)
@@ -1519,6 +1519,13 @@ static int arm_smmu_add_device(struct device *dev)
 
        if (using_legacy_binding) {
                ret = arm_smmu_register_legacy_master(dev, &smmu);
+
+               /*
+                * If dev->iommu_fwspec is initally NULL, arm_smmu_register_legacy_master()
+                * will allocate/initialise a new one. Thus we need to update fwspec for
+                * later use.
+                */
+               fwspec = dev->iommu_fwspec;
                if (ret)
                        goto out_free;
        } else if (fwspec && fwspec->ops == &arm_smmu_ops) {
index 28b26c80f4cf937b8547328b5d724a69e51b7d05..072bd227b6c677b40fee80f99e1364d733e681a3 100644 (file)
@@ -137,14 +137,14 @@ static void __init aic_common_ext_irq_of_init(struct irq_domain *domain)
 #define AT91_RTC_IMR           0x28
 #define AT91_RTC_IRQ_MASK      0x1f
 
-void __init aic_common_rtc_irq_fixup(struct device_node *root)
+void __init aic_common_rtc_irq_fixup(void)
 {
        struct device_node *np;
        void __iomem *regs;
 
-       np = of_find_compatible_node(root, NULL, "atmel,at91rm9200-rtc");
+       np = of_find_compatible_node(NULL, NULL, "atmel,at91rm9200-rtc");
        if (!np)
-               np = of_find_compatible_node(root, NULL,
+               np = of_find_compatible_node(NULL, NULL,
                                             "atmel,at91sam9x5-rtc");
 
        if (!np)
@@ -165,7 +165,7 @@ void __init aic_common_rtc_irq_fixup(struct device_node *root)
 #define AT91_RTT_ALMIEN                (1 << 16)               /* Alarm Interrupt Enable */
 #define AT91_RTT_RTTINCIEN     (1 << 17)               /* Real Time Timer Increment Interrupt Enable */
 
-void __init aic_common_rtt_irq_fixup(struct device_node *root)
+void __init aic_common_rtt_irq_fixup(void)
 {
        struct device_node *np;
        void __iomem *regs;
@@ -196,11 +196,10 @@ static void __init aic_common_irq_fixup(const struct of_device_id *matches)
                return;
 
        match = of_match_node(matches, root);
-       of_node_put(root);
 
        if (match) {
-               void (*fixup)(struct device_node *) = match->data;
-               fixup(root);
+               void (*fixup)(void) = match->data;
+               fixup();
        }
 
        of_node_put(root);
index af60376d50debe30132acd00c58254c4d1a7ab9a..242e62c1851ead9744442e0236fa84198f745d7f 100644 (file)
@@ -33,8 +33,8 @@ struct irq_domain *__init aic_common_of_init(struct device_node *node,
                                             const char *name, int nirqs,
                                             const struct of_device_id *matches);
 
-void __init aic_common_rtc_irq_fixup(struct device_node *root);
+void __init aic_common_rtc_irq_fixup(void);
 
-void __init aic_common_rtt_irq_fixup(struct device_node *root);
+void __init aic_common_rtt_irq_fixup(void);
 
 #endif /* __IRQ_ATMEL_AIC_COMMON_H */
index 37f952dd9fc94bdc5faa6c2721822b8780dd46e1..bb1ad451392fd8b7a315a6b2c5bea523af04a62a 100644 (file)
@@ -209,20 +209,20 @@ static const struct irq_domain_ops aic_irq_ops = {
        .xlate  = aic_irq_domain_xlate,
 };
 
-static void __init at91rm9200_aic_irq_fixup(struct device_node *root)
+static void __init at91rm9200_aic_irq_fixup(void)
 {
-       aic_common_rtc_irq_fixup(root);
+       aic_common_rtc_irq_fixup();
 }
 
-static void __init at91sam9260_aic_irq_fixup(struct device_node *root)
+static void __init at91sam9260_aic_irq_fixup(void)
 {
-       aic_common_rtt_irq_fixup(root);
+       aic_common_rtt_irq_fixup();
 }
 
-static void __init at91sam9g45_aic_irq_fixup(struct device_node *root)
+static void __init at91sam9g45_aic_irq_fixup(void)
 {
-       aic_common_rtc_irq_fixup(root);
-       aic_common_rtt_irq_fixup(root);
+       aic_common_rtc_irq_fixup();
+       aic_common_rtt_irq_fixup();
 }
 
 static const struct of_device_id aic_irq_fixups[] __initconst = {
index c04ee9a23d094f9d6a9e0c95a7451cf73626727a..6acad2ea0fb3565a2604ecbb3fb5235395d62fad 100644 (file)
@@ -305,9 +305,9 @@ static const struct irq_domain_ops aic5_irq_ops = {
        .xlate  = aic5_irq_domain_xlate,
 };
 
-static void __init sama5d3_aic_irq_fixup(struct device_node *root)
+static void __init sama5d3_aic_irq_fixup(void)
 {
-       aic_common_rtc_irq_fixup(root);
+       aic_common_rtc_irq_fixup();
 }
 
 static const struct of_device_id aic5_irq_fixups[] __initconst = {
index bddf169c4b37b7c9d2a91518aa233480dbb52f5d..b009b916a2923504414aafe6961b404614770da5 100644 (file)
@@ -189,6 +189,7 @@ static int __init brcmstb_l2_intc_of_init(struct device_node *np,
 
        ct->chip.irq_suspend = brcmstb_l2_intc_suspend;
        ct->chip.irq_resume = brcmstb_l2_intc_resume;
+       ct->chip.irq_pm_shutdown = brcmstb_l2_intc_suspend;
 
        if (data->can_wake) {
                /* This IRQ chip can wake the system, set all child interrupts
index 249240d9a4259eb72b7afb68c2e2723a72e7c9c5..833a90fe33aed839a81b781831027e677eef5581 100644 (file)
@@ -43,6 +43,7 @@ static int of_pmsi_get_dev_id(struct irq_domain *domain, struct device *dev,
                        *dev_id = args.args[0];
                        break;
                }
+               index++;
        } while (!ret);
 
        return ret;
index 68932873eebc0c3d14caa00b048d0c961e765a98..284738add89b3b94f77ad257fd314bc89a423cae 100644 (file)
@@ -1835,7 +1835,7 @@ static int __init its_of_probe(struct device_node *node)
 
 #define ACPI_GICV3_ITS_MEM_SIZE (SZ_128K)
 
-#if defined(CONFIG_ACPI_NUMA) && (ACPI_CA_VERSION >= 0x20170531)
+#ifdef CONFIG_ACPI_NUMA
 struct its_srat_map {
        /* numa node id */
        u32     numa_node;
@@ -1843,7 +1843,7 @@ struct its_srat_map {
        u32     its_id;
 };
 
-static struct its_srat_map its_srat_maps[MAX_NUMNODES] __initdata;
+static struct its_srat_map *its_srat_maps __initdata;
 static int its_in_srat __initdata;
 
 static int __init acpi_get_its_numa_node(u32 its_id)
@@ -1857,6 +1857,12 @@ static int __init acpi_get_its_numa_node(u32 its_id)
        return NUMA_NO_NODE;
 }
 
+static int __init gic_acpi_match_srat_its(struct acpi_subtable_header *header,
+                                         const unsigned long end)
+{
+       return 0;
+}
+
 static int __init gic_acpi_parse_srat_its(struct acpi_subtable_header *header,
                         const unsigned long end)
 {
@@ -1873,12 +1879,6 @@ static int __init gic_acpi_parse_srat_its(struct acpi_subtable_header *header,
                return -EINVAL;
        }
 
-       if (its_in_srat >= MAX_NUMNODES) {
-               pr_err("SRAT: ITS affinity exceeding max count[%d]\n",
-                               MAX_NUMNODES);
-               return -EINVAL;
-       }
-
        node = acpi_map_pxm_to_node(its_affinity->proximity_domain);
 
        if (node == NUMA_NO_NODE || node >= MAX_NUMNODES) {
@@ -1897,14 +1897,37 @@ static int __init gic_acpi_parse_srat_its(struct acpi_subtable_header *header,
 
 static void __init acpi_table_parse_srat_its(void)
 {
+       int count;
+
+       count = acpi_table_parse_entries(ACPI_SIG_SRAT,
+                       sizeof(struct acpi_table_srat),
+                       ACPI_SRAT_TYPE_GIC_ITS_AFFINITY,
+                       gic_acpi_match_srat_its, 0);
+       if (count <= 0)
+               return;
+
+       its_srat_maps = kmalloc(count * sizeof(struct its_srat_map),
+                               GFP_KERNEL);
+       if (!its_srat_maps) {
+               pr_warn("SRAT: Failed to allocate memory for its_srat_maps!\n");
+               return;
+       }
+
        acpi_table_parse_entries(ACPI_SIG_SRAT,
                        sizeof(struct acpi_table_srat),
                        ACPI_SRAT_TYPE_GIC_ITS_AFFINITY,
                        gic_acpi_parse_srat_its, 0);
 }
+
+/* free the its_srat_maps after ITS probing */
+static void __init acpi_its_srat_maps_free(void)
+{
+       kfree(its_srat_maps);
+}
 #else
 static void __init acpi_table_parse_srat_its(void)     { }
 static int __init acpi_get_its_numa_node(u32 its_id) { return NUMA_NO_NODE; }
+static void __init acpi_its_srat_maps_free(void) { }
 #endif
 
 static int __init gic_acpi_parse_madt_its(struct acpi_subtable_header *header,
@@ -1951,6 +1974,7 @@ static void __init its_acpi_probe(void)
        acpi_table_parse_srat_its();
        acpi_table_parse_madt(ACPI_MADT_TYPE_GENERIC_TRANSLATOR,
                              gic_acpi_parse_madt_its, 0);
+       acpi_its_srat_maps_free();
 }
 #else
 static void __init its_acpi_probe(void) { }
index dbffb7ab62033b346ca7fc3b3468d6bcb3a19a63..984c3ecfd22c21aff59b3cad177c676348d86dd4 100644 (file)
@@ -353,6 +353,8 @@ static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs
 
                        if (static_key_true(&supports_deactivate))
                                gic_write_eoir(irqnr);
+                       else
+                               isb();
 
                        err = handle_domain_irq(gic_data.domain, irqnr, regs);
                        if (err) {
@@ -640,11 +642,16 @@ static void gic_smp_init(void)
 static int gic_set_affinity(struct irq_data *d, const struct cpumask *mask_val,
                            bool force)
 {
-       unsigned int cpu = cpumask_any_and(mask_val, cpu_online_mask);
+       unsigned int cpu;
        void __iomem *reg;
        int enabled;
        u64 val;
 
+       if (force)
+               cpu = cpumask_first(mask_val);
+       else
+               cpu = cpumask_any_and(mask_val, cpu_online_mask);
+
        if (cpu >= nr_cpu_ids)
                return -EINVAL;
 
@@ -831,8 +838,11 @@ static int gic_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
        if (ret)
                return ret;
 
-       for (i = 0; i < nr_irqs; i++)
-               gic_irq_domain_map(domain, virq + i, hwirq + i);
+       for (i = 0; i < nr_irqs; i++) {
+               ret = gic_irq_domain_map(domain, virq + i, hwirq + i);
+               if (ret)
+                       return ret;
+       }
 
        return 0;
 }
index 1b1df4f770bdefe0a16e0109624801f8af90f1f7..d3e7c43718b82b9e7fb2191dedadab1950c3ac7d 100644 (file)
@@ -361,6 +361,7 @@ static void __exception_irq_entry gic_handle_irq(struct pt_regs *regs)
                if (likely(irqnr > 15 && irqnr < 1020)) {
                        if (static_key_true(&supports_deactivate))
                                writel_relaxed(irqstat, cpu_base + GIC_CPU_EOI);
+                       isb();
                        handle_domain_irq(gic->domain, irqnr, regs);
                        continue;
                }
@@ -401,10 +402,12 @@ static void gic_handle_cascade_irq(struct irq_desc *desc)
                goto out;
 
        cascade_irq = irq_find_mapping(chip_data->domain, gic_irq);
-       if (unlikely(gic_irq < 32 || gic_irq > 1020))
+       if (unlikely(gic_irq < 32 || gic_irq > 1020)) {
                handle_bad_irq(desc);
-       else
+       } else {
+               isb();
                generic_handle_irq(cascade_irq);
+       }
 
  out:
        chained_irq_exit(chip, desc);
@@ -1027,8 +1030,11 @@ static int gic_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
        if (ret)
                return ret;
 
-       for (i = 0; i < nr_irqs; i++)
-               gic_irq_domain_map(domain, virq + i, hwirq + i);
+       for (i = 0; i < nr_irqs; i++) {
+               ret = gic_irq_domain_map(domain, virq + i, hwirq + i);
+               if (ret)
+                       return ret;
+       }
 
        return 0;
 }
index 78fc5d5e90514353b258658da350a85c408b89db..92e6570b11435e7e4dade7aa433231008c82b294 100644 (file)
@@ -26,7 +26,7 @@
 
 #define FSM_TIMER_DEBUG 0
 
-void
+int
 mISDN_FsmNew(struct Fsm *fsm,
             struct FsmNode *fnlist, int fncount)
 {
@@ -34,6 +34,8 @@ mISDN_FsmNew(struct Fsm *fsm,
 
        fsm->jumpmatrix = kzalloc(sizeof(FSMFNPTR) * fsm->state_count *
                                  fsm->event_count, GFP_KERNEL);
+       if (fsm->jumpmatrix == NULL)
+               return -ENOMEM;
 
        for (i = 0; i < fncount; i++)
                if ((fnlist[i].state >= fsm->state_count) ||
@@ -45,6 +47,7 @@ mISDN_FsmNew(struct Fsm *fsm,
                } else
                        fsm->jumpmatrix[fsm->state_count * fnlist[i].event +
                                        fnlist[i].state] = (FSMFNPTR) fnlist[i].routine;
+       return 0;
 }
 EXPORT_SYMBOL(mISDN_FsmNew);
 
index 928f5be192c1fd4a5c8197160efb9dad21e1141d..e1def84902212e1637cf96f0c9cefb8c3950edbd 100644 (file)
@@ -55,7 +55,7 @@ struct FsmTimer {
        void *arg;
 };
 
-extern void mISDN_FsmNew(struct Fsm *, struct FsmNode *, int);
+extern int mISDN_FsmNew(struct Fsm *, struct FsmNode *, int);
 extern void mISDN_FsmFree(struct Fsm *);
 extern int mISDN_FsmEvent(struct FsmInst *, int , void *);
 extern void mISDN_FsmChangeState(struct FsmInst *, int);
index bebc57b72138e721aa5a290482ae46e748aac2d2..3192b0eb39445435d7b38c0f914bded7c93b2fac 100644 (file)
@@ -414,8 +414,7 @@ l1_init(u_int *deb)
        l1fsm_s.event_count = L1_EVENT_COUNT;
        l1fsm_s.strEvent = strL1Event;
        l1fsm_s.strState = strL1SState;
-       mISDN_FsmNew(&l1fsm_s, L1SFnList, ARRAY_SIZE(L1SFnList));
-       return 0;
+       return mISDN_FsmNew(&l1fsm_s, L1SFnList, ARRAY_SIZE(L1SFnList));
 }
 
 void
index 7243a6746f8b099d79d3221def6dcfcf138a6f51..9ff0903a0e89fb2f4f7373e1ef4d1d1663934b36 100644 (file)
@@ -2247,15 +2247,26 @@ static struct Bprotocol X75SLP = {
 int
 Isdnl2_Init(u_int *deb)
 {
+       int res;
        debug = deb;
        mISDN_register_Bprotocol(&X75SLP);
        l2fsm.state_count = L2_STATE_COUNT;
        l2fsm.event_count = L2_EVENT_COUNT;
        l2fsm.strEvent = strL2Event;
        l2fsm.strState = strL2State;
-       mISDN_FsmNew(&l2fsm, L2FnList, ARRAY_SIZE(L2FnList));
-       TEIInit(deb);
+       res = mISDN_FsmNew(&l2fsm, L2FnList, ARRAY_SIZE(L2FnList));
+       if (res)
+               goto error;
+       res = TEIInit(deb);
+       if (res)
+               goto error_fsm;
        return 0;
+
+error_fsm:
+       mISDN_FsmFree(&l2fsm);
+error:
+       mISDN_unregister_Bprotocol(&X75SLP);
+       return res;
 }
 
 void
index 908127efccf8ceb94ef2ab1ed7df4b17b26ce5b9..12d9e5f4beb1f81c5aa5e5af81bc9aca61c21668 100644 (file)
@@ -1387,23 +1387,37 @@ create_teimanager(struct mISDNdevice *dev)
 
 int TEIInit(u_int *deb)
 {
+       int res;
        debug = deb;
        teifsmu.state_count = TEI_STATE_COUNT;
        teifsmu.event_count = TEI_EVENT_COUNT;
        teifsmu.strEvent = strTeiEvent;
        teifsmu.strState = strTeiState;
-       mISDN_FsmNew(&teifsmu, TeiFnListUser, ARRAY_SIZE(TeiFnListUser));
+       res = mISDN_FsmNew(&teifsmu, TeiFnListUser, ARRAY_SIZE(TeiFnListUser));
+       if (res)
+               goto error;
        teifsmn.state_count = TEI_STATE_COUNT;
        teifsmn.event_count = TEI_EVENT_COUNT;
        teifsmn.strEvent = strTeiEvent;
        teifsmn.strState = strTeiState;
-       mISDN_FsmNew(&teifsmn, TeiFnListNet, ARRAY_SIZE(TeiFnListNet));
+       res = mISDN_FsmNew(&teifsmn, TeiFnListNet, ARRAY_SIZE(TeiFnListNet));
+       if (res)
+               goto error_smn;
        deactfsm.state_count =  DEACT_STATE_COUNT;
        deactfsm.event_count = DEACT_EVENT_COUNT;
        deactfsm.strEvent = strDeactEvent;
        deactfsm.strState = strDeactState;
-       mISDN_FsmNew(&deactfsm, DeactFnList, ARRAY_SIZE(DeactFnList));
+       res = mISDN_FsmNew(&deactfsm, DeactFnList, ARRAY_SIZE(DeactFnList));
+       if (res)
+               goto error_deact;
        return 0;
+
+error_deact:
+       mISDN_FsmFree(&teifsmn);
+error_smn:
+       mISDN_FsmFree(&teifsmu);
+error:
+       return res;
 }
 
 void TEIFree(void)
index c99634612fc408fbc97df9c7c8ef8e028efe2fad..b01e458d31e94ce9eba98b9300366121de550619 100644 (file)
@@ -7996,7 +7996,7 @@ bool md_write_start(struct mddev *mddev, struct bio *bi)
        if (mddev->safemode == 1)
                mddev->safemode = 0;
        /* sync_checkers is always 0 when writes_pending is in per-cpu mode */
-       if (mddev->in_sync || !mddev->sync_checkers) {
+       if (mddev->in_sync || mddev->sync_checkers) {
                spin_lock(&mddev->lock);
                if (mddev->in_sync) {
                        mddev->in_sync = 0;
@@ -8656,6 +8656,9 @@ void md_check_recovery(struct mddev *mddev)
        if (mddev_trylock(mddev)) {
                int spares = 0;
 
+               if (!mddev->external && mddev->safemode == 1)
+                       mddev->safemode = 0;
+
                if (mddev->ro) {
                        struct md_rdev *rdev;
                        if (!mddev->external && mddev->in_sync)
index bfa1e907c472e49855f9e0abb4c307cd45514d4d..2dcbafa8e66ca7f5418fce8434fe9674977f0832 100644 (file)
@@ -236,9 +236,10 @@ struct r5l_io_unit {
        bool need_split_bio;
        struct bio *split_bio;
 
-       unsigned int has_flush:1;      /* include flush request */
-       unsigned int has_fua:1;        /* include fua request */
-       unsigned int has_null_flush:1; /* include empty flush request */
+       unsigned int has_flush:1;               /* include flush request */
+       unsigned int has_fua:1;                 /* include fua request */
+       unsigned int has_null_flush:1;          /* include null flush request */
+       unsigned int has_flush_payload:1;       /* include flush payload  */
        /*
         * io isn't sent yet, flush/fua request can only be submitted till it's
         * the first IO in running_ios list
@@ -571,6 +572,8 @@ static void r5l_log_endio(struct bio *bio)
        struct r5l_io_unit *io_deferred;
        struct r5l_log *log = io->log;
        unsigned long flags;
+       bool has_null_flush;
+       bool has_flush_payload;
 
        if (bio->bi_status)
                md_error(log->rdev->mddev, log->rdev);
@@ -580,6 +583,16 @@ static void r5l_log_endio(struct bio *bio)
 
        spin_lock_irqsave(&log->io_list_lock, flags);
        __r5l_set_io_unit_state(io, IO_UNIT_IO_END);
+
+       /*
+        * if the io doesn't not have null_flush or flush payload,
+        * it is not safe to access it after releasing io_list_lock.
+        * Therefore, it is necessary to check the condition with
+        * the lock held.
+        */
+       has_null_flush = io->has_null_flush;
+       has_flush_payload = io->has_flush_payload;
+
        if (log->need_cache_flush && !list_empty(&io->stripe_list))
                r5l_move_to_end_ios(log);
        else
@@ -600,19 +613,23 @@ static void r5l_log_endio(struct bio *bio)
        if (log->need_cache_flush)
                md_wakeup_thread(log->rdev->mddev->thread);
 
-       if (io->has_null_flush) {
+       /* finish flush only io_unit and PAYLOAD_FLUSH only io_unit */
+       if (has_null_flush) {
                struct bio *bi;
 
                WARN_ON(bio_list_empty(&io->flush_barriers));
                while ((bi = bio_list_pop(&io->flush_barriers)) != NULL) {
                        bio_endio(bi);
-                       atomic_dec(&io->pending_stripe);
+                       if (atomic_dec_and_test(&io->pending_stripe)) {
+                               __r5l_stripe_write_finished(io);
+                               return;
+                       }
                }
        }
-
-       /* finish flush only io_unit and PAYLOAD_FLUSH only io_unit */
-       if (atomic_read(&io->pending_stripe) == 0)
-               __r5l_stripe_write_finished(io);
+       /* decrease pending_stripe for flush payload */
+       if (has_flush_payload)
+               if (atomic_dec_and_test(&io->pending_stripe))
+                       __r5l_stripe_write_finished(io);
 }
 
 static void r5l_do_submit_io(struct r5l_log *log, struct r5l_io_unit *io)
@@ -881,6 +898,11 @@ static void r5l_append_flush_payload(struct r5l_log *log, sector_t sect)
        payload->size = cpu_to_le32(sizeof(__le64));
        payload->flush_stripes[0] = cpu_to_le64(sect);
        io->meta_offset += meta_size;
+       /* multiple flush payloads count as one pending_stripe */
+       if (!io->has_flush_payload) {
+               io->has_flush_payload = 1;
+               atomic_inc(&io->pending_stripe);
+       }
        mutex_unlock(&log->io_mutex);
 }
 
@@ -2540,23 +2562,32 @@ static ssize_t r5c_journal_mode_show(struct mddev *mddev, char *page)
  */
 int r5c_journal_mode_set(struct mddev *mddev, int mode)
 {
-       struct r5conf *conf = mddev->private;
-       struct r5l_log *log = conf->log;
-
-       if (!log)
-               return -ENODEV;
+       struct r5conf *conf;
+       int err;
 
        if (mode < R5C_JOURNAL_MODE_WRITE_THROUGH ||
            mode > R5C_JOURNAL_MODE_WRITE_BACK)
                return -EINVAL;
 
+       err = mddev_lock(mddev);
+       if (err)
+               return err;
+       conf = mddev->private;
+       if (!conf || !conf->log) {
+               mddev_unlock(mddev);
+               return -ENODEV;
+       }
+
        if (raid5_calc_degraded(conf) > 0 &&
-           mode == R5C_JOURNAL_MODE_WRITE_BACK)
+           mode == R5C_JOURNAL_MODE_WRITE_BACK) {
+               mddev_unlock(mddev);
                return -EINVAL;
+       }
 
        mddev_suspend(mddev);
        conf->log->r5c_journal_mode = mode;
        mddev_resume(mddev);
+       mddev_unlock(mddev);
 
        pr_debug("md/raid:%s: setting r5c cache mode to %d: %s\n",
                 mdname(mddev), mode, r5c_journal_mode_str[mode]);
index 99e644cda4d13db301b713a5752788c0f646dfa1..ebf69ff48ae2656594b0840fb7bd8a917e159406 100644 (file)
@@ -72,7 +72,7 @@ struct atmel_smc_timing_xlate {
        { .name = nm, .converter = atmel_smc_cs_conf_set_pulse, .shift = pos}
 
 #define ATMEL_SMC_CYCLE_XLATE(nm, pos) \
-       { .name = nm, .converter = atmel_smc_cs_conf_set_setup, .shift = pos}
+       { .name = nm, .converter = atmel_smc_cs_conf_set_cycle, .shift = pos}
 
 static void at91sam9_ebi_get_config(struct atmel_ebi_dev *ebid,
                                    struct atmel_ebi_dev_config *conf)
@@ -120,12 +120,14 @@ static int atmel_ebi_xslate_smc_timings(struct atmel_ebi_dev *ebid,
        if (!ret) {
                required = true;
                ncycles = DIV_ROUND_UP(val, clk_period_ns);
-               if (ncycles > ATMEL_SMC_MODE_TDF_MAX ||
-                   ncycles < ATMEL_SMC_MODE_TDF_MIN) {
+               if (ncycles > ATMEL_SMC_MODE_TDF_MAX) {
                        ret = -EINVAL;
                        goto out;
                }
 
+               if (ncycles < ATMEL_SMC_MODE_TDF_MIN)
+                       ncycles = ATMEL_SMC_MODE_TDF_MIN;
+
                smcconf->mode |= ATMEL_SMC_MODE_TDF(ncycles);
        }
 
@@ -263,7 +265,7 @@ static int atmel_ebi_xslate_smc_config(struct atmel_ebi_dev *ebid,
        }
 
        ret = atmel_ebi_xslate_smc_timings(ebid, np, &conf->smcconf);
-       if (ret)
+       if (ret < 0)
                return -EINVAL;
 
        if ((ret > 0 && !required) || (!ret && required)) {
index 954cf0f66a31fa87b5be4e8cf4c4f81691980c0a..20cc0ea470fae22053584554e096ce31fad06275 100644 (file)
@@ -206,7 +206,7 @@ EXPORT_SYMBOL_GPL(atmel_smc_cs_conf_set_pulse);
  *          parameter
  *
  * This function encodes the @ncycles value as described in the datasheet
- * (section "SMC Pulse Register"), and then stores the result in the
+ * (section "SMC Cycle Register"), and then stores the result in the
  * @conf->setup field at @shift position.
  *
  * Returns -EINVAL if @shift is invalid, -ERANGE if @ncycles does not fit in
index fbe0f245ce8ee50943340de1c05d6490a3a13ced..fe1811523e4a7aac70e0498e7cd70fac55e2046b 100644 (file)
@@ -644,6 +644,9 @@ static const struct regmap_range da9062_aa_readable_ranges[] = {
        }, {
                .range_min = DA9062AA_VLDO1_B,
                .range_max = DA9062AA_VLDO4_B,
+       }, {
+               .range_min = DA9062AA_BBAT_CONT,
+               .range_max = DA9062AA_BBAT_CONT,
        }, {
                .range_min = DA9062AA_INTERFACE,
                .range_max = DA9062AA_CONFIG_E,
@@ -720,6 +723,9 @@ static const struct regmap_range da9062_aa_writeable_ranges[] = {
        }, {
                .range_min = DA9062AA_VLDO1_B,
                .range_max = DA9062AA_VLDO4_B,
+       }, {
+               .range_min = DA9062AA_BBAT_CONT,
+               .range_max = DA9062AA_BBAT_CONT,
        }, {
                .range_min = DA9062AA_GP_ID_0,
                .range_max = DA9062AA_GP_ID_19,
index 8621a198a2ce3eae317e4007cd6c533b56cfecc4..bac33311f55a6d7a6699b362c39730fae9c4f107 100644 (file)
@@ -215,6 +215,12 @@ static int mei_me_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 
        pci_set_drvdata(pdev, dev);
 
+       /*
+        * MEI requires to resume from runtime suspend mode
+        * in order to perform link reset flow upon system suspend.
+        */
+       pdev->dev_flags |= PCI_DEV_FLAGS_NEEDS_RESUME;
+
        /*
        * For not wake-able HW runtime pm framework
        * can't be used on pci device level.
index f811cd52446852beecfedf02b7100f4bb6789169..e38a5f144373451fc87007ffc1cf4292059c5408 100644 (file)
@@ -137,6 +137,12 @@ static int mei_txe_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 
        pci_set_drvdata(pdev, dev);
 
+       /*
+        * MEI requires to resume from runtime suspend mode
+        * in order to perform link reset flow upon system suspend.
+        */
+       pdev->dev_flags |= PCI_DEV_FLAGS_NEEDS_RESUME;
+
        /*
        * For not wake-able HW runtime pm framework
        * can't be used on pci device level.
index e5938c791330c9be1203c2698bc511e4f98aaea4..f1bbfd389367ff4530137be199c4063c65f97f5c 100644 (file)
@@ -2170,7 +2170,9 @@ static void mmc_blk_remove_req(struct mmc_blk_data *md)
                 * from being accepted.
                 */
                card = md->queue.card;
+               spin_lock_irq(md->queue.queue->queue_lock);
                queue_flag_set(QUEUE_FLAG_BYPASS, md->queue.queue);
+               spin_unlock_irq(md->queue.queue->queue_lock);
                blk_set_queue_dying(md->queue.queue);
                mmc_cleanup_queue(&md->queue);
                if (md->disk->flags & GENHD_FL_UP) {
index 4ffea14b7eb645d92a91d62907d64c97cf8a9998..2bae69e39544452dc323a9a3bd15ae9b3e7de1b9 100644 (file)
@@ -1289,7 +1289,7 @@ out_err:
 static int mmc_select_hs400es(struct mmc_card *card)
 {
        struct mmc_host *host = card->host;
-       int err = 0;
+       int err = -EINVAL;
        u8 val;
 
        if (!(host->caps & MMC_CAP_8_BIT_DATA)) {
index 04ff3c97a535143933acec93caa8483a43f633c6..2ab4788d021f0512082c6bd67d1edaa57ba61379 100644 (file)
@@ -2086,7 +2086,7 @@ static int omap_hsmmc_probe(struct platform_device *pdev)
        mmc->max_seg_size = mmc->max_req_size;
 
        mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
-                    MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_ERASE;
+                    MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_ERASE | MMC_CAP_CMD23;
 
        mmc->caps |= mmc_pdata(host)->caps;
        if (mmc->caps & MMC_CAP_8_BIT_DATA)
index f336a9b855765e5ea236fddf9564f5f084793597..9ec8f033ac5f077b05fdfc32a7f97f1747339ca8 100644 (file)
@@ -113,6 +113,7 @@ static blk_status_t do_blktrans_request(struct mtd_blktrans_ops *tr,
                for (; nsect > 0; nsect--, block++, buf += tr->blksize)
                        if (tr->writesect(dev, block, buf))
                                return BLK_STS_IOERR;
+               return BLK_STS_OK;
        default:
                return BLK_STS_IOERR;
        }
index 9bee6c1c70cca33941ae8db8002d69e760693bac..fc63992ab0e0ad3300aa9a8d991fc107788ae780 100644 (file)
@@ -1569,7 +1569,8 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
        new_slave->delay = 0;
        new_slave->link_failure_count = 0;
 
-       if (bond_update_speed_duplex(new_slave))
+       if (bond_update_speed_duplex(new_slave) &&
+           bond_needs_speed_duplex(bond))
                new_slave->link = BOND_LINK_DOWN;
 
        new_slave->last_rx = jiffies -
@@ -2140,11 +2141,13 @@ static void bond_miimon_commit(struct bonding *bond)
                        continue;
 
                case BOND_LINK_UP:
-                       if (bond_update_speed_duplex(slave)) {
+                       if (bond_update_speed_duplex(slave) &&
+                           bond_needs_speed_duplex(bond)) {
                                slave->link = BOND_LINK_DOWN;
-                               netdev_warn(bond->dev,
-                                           "failed to get link speed/duplex for %s\n",
-                                           slave->dev->name);
+                               if (net_ratelimit())
+                                       netdev_warn(bond->dev,
+                                                   "failed to get link speed/duplex for %s\n",
+                                                   slave->dev->name);
                                continue;
                        }
                        bond_set_slave_link_state(slave, BOND_LINK_UP,
index ef4be781fd054696edac7c494cf713f0dd5f3d4e..09ea62ee96d38b3d99bb48cbd0f72dcc715f8315 100644 (file)
@@ -529,6 +529,7 @@ enum {                                 /* adapter flags */
        USING_SOFT_PARAMS  = (1 << 6),
        MASTER_PF          = (1 << 7),
        FW_OFLD_CONN       = (1 << 9),
+       ROOT_NO_RELAXED_ORDERING = (1 << 10),
 };
 
 enum {
index e403fa18f1b15e570748b2136ae1bf8b54d14afb..33bb8678833adc6f83551d992f201f6314e68762 100644 (file)
@@ -4654,11 +4654,6 @@ static void print_port_info(const struct net_device *dev)
                    dev->name, adap->params.vpd.id, adap->name, buf);
 }
 
-static void enable_pcie_relaxed_ordering(struct pci_dev *dev)
-{
-       pcie_capability_set_word(dev, PCI_EXP_DEVCTL, PCI_EXP_DEVCTL_RELAX_EN);
-}
-
 /*
  * Free the following resources:
  * - memory used for tables
@@ -4908,7 +4903,6 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
        }
 
        pci_enable_pcie_error_reporting(pdev);
-       enable_pcie_relaxed_ordering(pdev);
        pci_set_master(pdev);
        pci_save_state(pdev);
 
@@ -4947,6 +4941,23 @@ static int init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
        adapter->msg_enable = DFLT_MSG_ENABLE;
        memset(adapter->chan_map, 0xff, sizeof(adapter->chan_map));
 
+       /* If possible, we use PCIe Relaxed Ordering Attribute to deliver
+        * Ingress Packet Data to Free List Buffers in order to allow for
+        * chipset performance optimizations between the Root Complex and
+        * Memory Controllers.  (Messages to the associated Ingress Queue
+        * notifying new Packet Placement in the Free Lists Buffers will be
+        * send without the Relaxed Ordering Attribute thus guaranteeing that
+        * all preceding PCIe Transaction Layer Packets will be processed
+        * first.)  But some Root Complexes have various issues with Upstream
+        * Transaction Layer Packets with the Relaxed Ordering Attribute set.
+        * The PCIe devices which under the Root Complexes will be cleared the
+        * Relaxed Ordering bit in the configuration space, So we check our
+        * PCIe configuration space to see if it's flagged with advice against
+        * using Relaxed Ordering.
+        */
+       if (!pcie_relaxed_ordering_enabled(pdev))
+               adapter->flags |= ROOT_NO_RELAXED_ORDERING;
+
        spin_lock_init(&adapter->stats_lock);
        spin_lock_init(&adapter->tid_release_lock);
        spin_lock_init(&adapter->win0_lock);
index ede12209f20be4573a5464de9e7f7a0da0088c31..4ef68f69b58c45322d65f414e06d068ade4ab22d 100644 (file)
@@ -2719,6 +2719,7 @@ int t4_sge_alloc_rxq(struct adapter *adap, struct sge_rspq *iq, bool fwevtq,
        struct fw_iq_cmd c;
        struct sge *s = &adap->sge;
        struct port_info *pi = netdev_priv(dev);
+       int relaxed = !(adap->flags & ROOT_NO_RELAXED_ORDERING);
 
        /* Size needs to be multiple of 16, including status entry. */
        iq->size = roundup(iq->size, 16);
@@ -2772,8 +2773,8 @@ int t4_sge_alloc_rxq(struct adapter *adap, struct sge_rspq *iq, bool fwevtq,
 
                flsz = fl->size / 8 + s->stat_len / sizeof(struct tx_desc);
                c.iqns_to_fl0congen |= htonl(FW_IQ_CMD_FL0PACKEN_F |
-                                            FW_IQ_CMD_FL0FETCHRO_F |
-                                            FW_IQ_CMD_FL0DATARO_F |
+                                            FW_IQ_CMD_FL0FETCHRO_V(relaxed) |
+                                            FW_IQ_CMD_FL0DATARO_V(relaxed) |
                                             FW_IQ_CMD_FL0PADEN_F);
                if (cong >= 0)
                        c.iqns_to_fl0congen |=
index 109bc630408b65b1dc67cedbaab00ce0601d2574..08c6ddb84a049ecc75f414a968dba78974d5837f 100644 (file)
@@ -408,6 +408,7 @@ enum { /* adapter flags */
        USING_MSI          = (1UL << 1),
        USING_MSIX         = (1UL << 2),
        QUEUES_BOUND       = (1UL << 3),
+       ROOT_NO_RELAXED_ORDERING = (1UL << 4),
 };
 
 /*
index ac7a150c54e9b4e93b751821eb09d80669eecc69..2b85b874fd0d2c96fadb4888b0bd4695628bf611 100644 (file)
@@ -2888,6 +2888,24 @@ static int cxgb4vf_pci_probe(struct pci_dev *pdev,
         */
        adapter->name = pci_name(pdev);
        adapter->msg_enable = DFLT_MSG_ENABLE;
+
+       /* If possible, we use PCIe Relaxed Ordering Attribute to deliver
+        * Ingress Packet Data to Free List Buffers in order to allow for
+        * chipset performance optimizations between the Root Complex and
+        * Memory Controllers.  (Messages to the associated Ingress Queue
+        * notifying new Packet Placement in the Free Lists Buffers will be
+        * send without the Relaxed Ordering Attribute thus guaranteeing that
+        * all preceding PCIe Transaction Layer Packets will be processed
+        * first.)  But some Root Complexes have various issues with Upstream
+        * Transaction Layer Packets with the Relaxed Ordering Attribute set.
+        * The PCIe devices which under the Root Complexes will be cleared the
+        * Relaxed Ordering bit in the configuration space, So we check our
+        * PCIe configuration space to see if it's flagged with advice against
+        * using Relaxed Ordering.
+        */
+       if (!pcie_relaxed_ordering_enabled(pdev))
+               adapter->flags |= ROOT_NO_RELAXED_ORDERING;
+
        err = adap_init0(adapter);
        if (err)
                goto err_unmap_bar;
index e37dde2ba97f6d529177d475be2a6d001071de0b..05498e7f284034d86a8f8531de95cf7f59ec4890 100644 (file)
@@ -2205,6 +2205,7 @@ int t4vf_sge_alloc_rxq(struct adapter *adapter, struct sge_rspq *rspq,
        struct port_info *pi = netdev_priv(dev);
        struct fw_iq_cmd cmd, rpl;
        int ret, iqandst, flsz = 0;
+       int relaxed = !(adapter->flags & ROOT_NO_RELAXED_ORDERING);
 
        /*
         * If we're using MSI interrupts and we're not initializing the
@@ -2300,6 +2301,8 @@ int t4vf_sge_alloc_rxq(struct adapter *adapter, struct sge_rspq *rspq,
                        cpu_to_be32(
                                FW_IQ_CMD_FL0HOSTFCMODE_V(SGE_HOSTFCMODE_NONE) |
                                FW_IQ_CMD_FL0PACKEN_F |
+                               FW_IQ_CMD_FL0FETCHRO_V(relaxed) |
+                               FW_IQ_CMD_FL0DATARO_V(relaxed) |
                                FW_IQ_CMD_FL0PADEN_F);
 
                /* In T6, for egress queue type FL there is internal overhead
index 09b9bc17bce998a99f360577a92a0211b552bf38..5fe5cdc5135776abb8fd4df748b0948520504b48 100644 (file)
@@ -432,7 +432,7 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
                /* Virtual PCI function needs to determine UAR page size from
                 * firmware. Only master PCI function can set the uar page size
                 */
-               if (enable_4k_uar)
+               if (enable_4k_uar || !dev->persist->num_vfs)
                        dev->uar_page_shift = DEFAULT_UAR_PAGE_SHIFT;
                else
                        dev->uar_page_shift = PAGE_SHIFT;
@@ -2277,7 +2277,7 @@ static int mlx4_init_hca(struct mlx4_dev *dev)
 
                dev->caps.max_fmr_maps = (1 << (32 - ilog2(dev->caps.num_mpts))) - 1;
 
-               if (enable_4k_uar) {
+               if (enable_4k_uar || !dev->persist->num_vfs) {
                        init_hca.log_uar_sz = ilog2(dev->caps.num_uars) +
                                                    PAGE_SHIFT - DEFAULT_UAR_PAGE_SHIFT;
                        init_hca.uar_page_sz = DEFAULT_UAR_PAGE_SHIFT - 12;
index dd7fa9cf225ff3c9e60d6f62e3f7f17317f2ea52..b0837b58c3a1084e8261ff3bf6d6057c4d8e7f54 100644 (file)
@@ -115,14 +115,10 @@ nfp_flower_cmsg_portmod_rx(struct nfp_app *app, struct sk_buff *skb)
                return;
        }
 
-       if (link) {
+       if (link)
                netif_carrier_on(netdev);
-               rtnl_lock();
-               dev_set_mtu(netdev, be16_to_cpu(msg->mtu));
-               rtnl_unlock();
-       } else {
+       else
                netif_carrier_off(netdev);
-       }
        rcu_read_unlock();
 }
 
index 4631ca8b8eb2780865bc2eefae51ce5c5a61f5cf..9f77ce038a4a339260e0118bb6a9f61e60547870 100644 (file)
@@ -908,8 +908,7 @@ static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev)
        return NETDEV_TX_OK;
 
 err_unmap:
-       --f;
-       while (f >= 0) {
+       while (--f >= 0) {
                frag = &skb_shinfo(skb)->frags[f];
                dma_unmap_page(dp->dev, tx_ring->txbufs[wr_idx].dma_addr,
                               skb_frag_size(frag), DMA_TO_DEVICE);
index 66ff15d08bad4eb0a12e30e69547905789d15336..0a66389c06c2522a95ed8d6d8fdd04ce1762cf4d 100644 (file)
@@ -2311,7 +2311,7 @@ netxen_md_rdqueue(struct netxen_adapter *adapter,
                                 loop_cnt++) {
                NX_WR_DUMP_REG(select_addr, adapter->ahw.pci_base0, queue_id);
                read_addr = queueEntry->read_addr;
-               for (k = 0; k < read_cnt; k--) {
+               for (k = 0; k < read_cnt; k++) {
                        NX_RD_DUMP_REG(read_addr, adapter->ahw.pci_base0,
                                                        &read_value);
                        *data_buff++ = read_value;
index c905971c5f3a2849262dcf3ce220de93e32e4bd2..990a63d7fcb7213fa5c5d6e40a10ec5102af892a 100644 (file)
@@ -938,7 +938,6 @@ enum efx_stats_action {
 static int efx_mcdi_mac_stats(struct efx_nic *efx,
                              enum efx_stats_action action, int clear)
 {
-       struct efx_ef10_nic_data *nic_data = efx->nic_data;
        MCDI_DECLARE_BUF(inbuf, MC_CMD_MAC_STATS_IN_LEN);
        int rc;
        int change = action == EFX_STATS_PULL ? 0 : 1;
@@ -960,7 +959,12 @@ static int efx_mcdi_mac_stats(struct efx_nic *efx,
                              MAC_STATS_IN_PERIODIC_NOEVENT, 1,
                              MAC_STATS_IN_PERIOD_MS, period);
        MCDI_SET_DWORD(inbuf, MAC_STATS_IN_DMA_LEN, dma_len);
-       MCDI_SET_DWORD(inbuf, MAC_STATS_IN_PORT_ID, nic_data->vport_id);
+
+       if (efx_nic_rev(efx) >= EFX_REV_HUNT_A0) {
+               struct efx_ef10_nic_data *nic_data = efx->nic_data;
+
+               MCDI_SET_DWORD(inbuf, MAC_STATS_IN_PORT_ID, nic_data->vport_id);
+       }
 
        rc = efx_mcdi_rpc_quiet(efx, MC_CMD_MAC_STATS, inbuf, sizeof(inbuf),
                                NULL, 0, NULL);
index db157a47000c65fb588d3f1c5dfe21bbd20983fb..72ec711fcba242775feea3064bdcb69de6c1812e 100644 (file)
@@ -204,6 +204,7 @@ int stmmac_mdio_register(struct net_device *ndev)
        struct stmmac_priv *priv = netdev_priv(ndev);
        struct stmmac_mdio_bus_data *mdio_bus_data = priv->plat->mdio_bus_data;
        struct device_node *mdio_node = priv->plat->mdio_node;
+       struct device *dev = ndev->dev.parent;
        int addr, found;
 
        if (!mdio_bus_data)
@@ -237,7 +238,7 @@ int stmmac_mdio_register(struct net_device *ndev)
        else
                err = mdiobus_register(new_bus);
        if (err != 0) {
-               netdev_err(ndev, "Cannot register the MDIO bus\n");
+               dev_err(dev, "Cannot register the MDIO bus\n");
                goto bus_register_fail;
        }
 
@@ -285,14 +286,12 @@ int stmmac_mdio_register(struct net_device *ndev)
                        irq_str = irq_num;
                        break;
                }
-               netdev_info(ndev, "PHY ID %08x at %d IRQ %s (%s)%s\n",
-                           phydev->phy_id, addr, irq_str, phydev_name(phydev),
-                           act ? " active" : "");
+               phy_attached_info(phydev);
                found = 1;
        }
 
        if (!found && !mdio_node) {
-               netdev_warn(ndev, "No PHY found\n");
+               dev_warn(dev, "No PHY found\n");
                mdiobus_unregister(new_bus);
                mdiobus_free(new_bus);
                return -ENODEV;
index 32ad87345f5798498584d8dcfbda2f9ac993e619..0a2c0a42283f780b947fcf52fcb4220e73f0fa2f 100644 (file)
@@ -1879,6 +1879,9 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
 
 err_detach:
        tun_detach_all(dev);
+       /* register_netdevice() already called tun_free_netdev() */
+       goto err_free_dev;
+
 err_free_flow:
        tun_flow_uninit(tun);
        security_tun_dev_free_security(tun->security);
index d21258d277ce4bca965e588e537df507d4e23892..f1b60740e02080261d998686a28e27c9ee4ea8a2 100644 (file)
@@ -159,8 +159,10 @@ void brcmf_feat_attach(struct brcmf_pub *drvr)
 
        brcmf_feat_firmware_capabilities(ifp);
        memset(&gscan_cfg, 0, sizeof(gscan_cfg));
-       brcmf_feat_iovar_data_set(ifp, BRCMF_FEAT_GSCAN, "pfn_gscan_cfg",
-                                 &gscan_cfg, sizeof(gscan_cfg));
+       if (drvr->bus_if->chip != BRCM_CC_43430_CHIP_ID)
+               brcmf_feat_iovar_data_set(ifp, BRCMF_FEAT_GSCAN,
+                                         "pfn_gscan_cfg",
+                                         &gscan_cfg, sizeof(gscan_cfg));
        brcmf_feat_iovar_int_get(ifp, BRCMF_FEAT_PNO, "pfn");
        if (drvr->bus_if->wowl_supported)
                brcmf_feat_iovar_int_get(ifp, BRCMF_FEAT_WOWL, "wowl");
index b4ecd1fe137489617ab477608e0152712db0f784..97208ce19f927685ce79b2b2321ee5b3086f7a3c 100644 (file)
@@ -154,7 +154,7 @@ static const struct iwl_tt_params iwl9000_tt_params = {
 const struct iwl_cfg iwl9160_2ac_cfg = {
        .name = "Intel(R) Dual Band Wireless AC 9160",
        .fw_name_pre = IWL9260A_FW_PRE,
-       .fw_name_pre_next_step = IWL9260B_FW_PRE,
+       .fw_name_pre_b_or_c_step = IWL9260B_FW_PRE,
        IWL_DEVICE_9000,
        .ht_params = &iwl9000_ht_params,
        .nvm_ver = IWL9000_NVM_VERSION,
@@ -165,7 +165,7 @@ const struct iwl_cfg iwl9160_2ac_cfg = {
 const struct iwl_cfg iwl9260_2ac_cfg = {
        .name = "Intel(R) Dual Band Wireless AC 9260",
        .fw_name_pre = IWL9260A_FW_PRE,
-       .fw_name_pre_next_step = IWL9260B_FW_PRE,
+       .fw_name_pre_b_or_c_step = IWL9260B_FW_PRE,
        IWL_DEVICE_9000,
        .ht_params = &iwl9000_ht_params,
        .nvm_ver = IWL9000_NVM_VERSION,
@@ -176,7 +176,7 @@ const struct iwl_cfg iwl9260_2ac_cfg = {
 const struct iwl_cfg iwl9270_2ac_cfg = {
        .name = "Intel(R) Dual Band Wireless AC 9270",
        .fw_name_pre = IWL9260A_FW_PRE,
-       .fw_name_pre_next_step = IWL9260B_FW_PRE,
+       .fw_name_pre_b_or_c_step = IWL9260B_FW_PRE,
        IWL_DEVICE_9000,
        .ht_params = &iwl9000_ht_params,
        .nvm_ver = IWL9000_NVM_VERSION,
@@ -186,8 +186,8 @@ const struct iwl_cfg iwl9270_2ac_cfg = {
 
 const struct iwl_cfg iwl9460_2ac_cfg = {
        .name = "Intel(R) Dual Band Wireless AC 9460",
-       .fw_name_pre = IWL9000_FW_PRE,
-       .fw_name_pre_rf_next_step = IWL9000RFB_FW_PRE,
+       .fw_name_pre = IWL9260A_FW_PRE,
+       .fw_name_pre_b_or_c_step = IWL9260B_FW_PRE,
        IWL_DEVICE_9000,
        .ht_params = &iwl9000_ht_params,
        .nvm_ver = IWL9000_NVM_VERSION,
@@ -198,8 +198,8 @@ const struct iwl_cfg iwl9460_2ac_cfg = {
 
 const struct iwl_cfg iwl9560_2ac_cfg = {
        .name = "Intel(R) Dual Band Wireless AC 9560",
-       .fw_name_pre = IWL9000_FW_PRE,
-       .fw_name_pre_rf_next_step = IWL9000RFB_FW_PRE,
+       .fw_name_pre = IWL9260A_FW_PRE,
+       .fw_name_pre_b_or_c_step = IWL9260B_FW_PRE,
        IWL_DEVICE_9000,
        .ht_params = &iwl9000_ht_params,
        .nvm_ver = IWL9000_NVM_VERSION,
index 0fa8c473f1e2738b8490e2555af1ae95319d96dd..c73a6438ce8fbcd12a8e12e359c5f7d25e66a76f 100644 (file)
@@ -328,6 +328,7 @@ typedef unsigned int __bitwise iwl_ucode_tlv_capa_t;
  * @IWL_UCODE_TLV_CAPA_TX_POWER_ACK: reduced TX power API has larger
  *     command size (command version 4) that supports toggling ACK TX
  *     power reduction.
+ * @IWL_UCODE_TLV_CAPA_MLME_OFFLOAD: supports MLME offload
  *
  * @NUM_IWL_UCODE_TLV_CAPA: number of bits used
  */
@@ -373,6 +374,7 @@ enum iwl_ucode_tlv_capa {
        IWL_UCODE_TLV_CAPA_EXTEND_SHARED_MEM_CFG        = (__force iwl_ucode_tlv_capa_t)80,
        IWL_UCODE_TLV_CAPA_LQM_SUPPORT                  = (__force iwl_ucode_tlv_capa_t)81,
        IWL_UCODE_TLV_CAPA_TX_POWER_ACK                 = (__force iwl_ucode_tlv_capa_t)84,
+       IWL_UCODE_TLV_CAPA_MLME_OFFLOAD                 = (__force iwl_ucode_tlv_capa_t)96,
 
        NUM_IWL_UCODE_TLV_CAPA
 #ifdef __CHECKER__
index c52623cb7c2a1bc89d635f4406df70bc3101f340..d19c74827fbb6094147ab48fc266cd6dbbb15e28 100644 (file)
@@ -276,10 +276,10 @@ struct iwl_pwr_tx_backoff {
  * @fw_name_pre: Firmware filename prefix. The api version and extension
  *     (.ucode) will be added to filename before loading from disk. The
  *     filename is constructed as fw_name_pre<api>.ucode.
- * @fw_name_pre_next_step: same as @fw_name_pre, only for next step
+ * @fw_name_pre_b_or_c_step: same as @fw_name_pre, only for b or c steps
  *     (if supported)
- * @fw_name_pre_rf_next_step: same as @fw_name_pre_next_step, only for rf next
- *     step. Supported only in integrated solutions.
+ * @fw_name_pre_rf_next_step: same as @fw_name_pre_b_or_c_step, only for rf
+ *     next step. Supported only in integrated solutions.
  * @ucode_api_max: Highest version of uCode API supported by driver.
  * @ucode_api_min: Lowest version of uCode API supported by driver.
  * @max_inst_size: The maximal length of the fw inst section
@@ -330,7 +330,7 @@ struct iwl_cfg {
        /* params specific to an individual device within a device family */
        const char *name;
        const char *fw_name_pre;
-       const char *fw_name_pre_next_step;
+       const char *fw_name_pre_b_or_c_step;
        const char *fw_name_pre_rf_next_step;
        /* params not likely to change within a device family */
        const struct iwl_base_params *base_params;
index 6fdb5921e17f41846db6a4c7302d9248247d1c03..4e0f86fe0a6f0874a96d41ae257c9621a2f035ee 100644 (file)
@@ -216,8 +216,9 @@ static int iwl_request_firmware(struct iwl_drv *drv, bool first)
        const char *fw_pre_name;
 
        if (drv->trans->cfg->device_family == IWL_DEVICE_FAMILY_9000 &&
-           CSR_HW_REV_STEP(drv->trans->hw_rev) == SILICON_B_STEP)
-               fw_pre_name = cfg->fw_name_pre_next_step;
+           (CSR_HW_REV_STEP(drv->trans->hw_rev) == SILICON_B_STEP ||
+            CSR_HW_REV_STEP(drv->trans->hw_rev) == SILICON_C_STEP))
+               fw_pre_name = cfg->fw_name_pre_b_or_c_step;
        else if (drv->trans->cfg->integrated &&
                 CSR_HW_RFID_STEP(drv->trans->hw_rf_id) == SILICON_B_STEP &&
                 cfg->fw_name_pre_rf_next_step)
index 5c08f4d40f6ac78fbdeb575e152742c9c0a3cef5..3ee6767392b61151efc774610223e2f6216d22f3 100644 (file)
@@ -785,7 +785,8 @@ iwl_parse_nvm_mcc_info(struct device *dev, const struct iwl_cfg *cfg,
                       int num_of_ch, __le32 *channels, u16 fw_mcc)
 {
        int ch_idx;
-       u16 ch_flags, prev_ch_flags = 0;
+       u16 ch_flags;
+       u32 reg_rule_flags, prev_reg_rule_flags = 0;
        const u8 *nvm_chan = cfg->ext_nvm ?
                             iwl_ext_nvm_channels : iwl_nvm_channels;
        struct ieee80211_regdomain *regd;
@@ -834,8 +835,11 @@ iwl_parse_nvm_mcc_info(struct device *dev, const struct iwl_cfg *cfg,
                        continue;
                }
 
+               reg_rule_flags = iwl_nvm_get_regdom_bw_flags(nvm_chan, ch_idx,
+                                                            ch_flags, cfg);
+
                /* we can't continue the same rule */
-               if (ch_idx == 0 || prev_ch_flags != ch_flags ||
+               if (ch_idx == 0 || prev_reg_rule_flags != reg_rule_flags ||
                    center_freq - prev_center_freq > 20) {
                        valid_rules++;
                        new_rule = true;
@@ -854,18 +858,17 @@ iwl_parse_nvm_mcc_info(struct device *dev, const struct iwl_cfg *cfg,
                rule->power_rule.max_eirp =
                        DBM_TO_MBM(IWL_DEFAULT_MAX_TX_POWER);
 
-               rule->flags = iwl_nvm_get_regdom_bw_flags(nvm_chan, ch_idx,
-                                                         ch_flags, cfg);
+               rule->flags = reg_rule_flags;
 
                /* rely on auto-calculation to merge BW of contiguous chans */
                rule->flags |= NL80211_RRF_AUTO_BW;
                rule->freq_range.max_bandwidth_khz = 0;
 
-               prev_ch_flags = ch_flags;
                prev_center_freq = center_freq;
+               prev_reg_rule_flags = reg_rule_flags;
 
                IWL_DEBUG_DEV(dev, IWL_DL_LAR,
-                             "Ch. %d [%sGHz] %s%s%s%s%s%s%s%s%s(0x%02x): Ad-Hoc %ssupported\n",
+                             "Ch. %d [%sGHz] %s%s%s%s%s%s%s%s%s(0x%02x) reg_flags 0x%x: %s\n",
                              center_freq,
                              band == NL80211_BAND_5GHZ ? "5.2" : "2.4",
                              CHECK_AND_PRINT_I(VALID),
@@ -877,10 +880,10 @@ iwl_parse_nvm_mcc_info(struct device *dev, const struct iwl_cfg *cfg,
                              CHECK_AND_PRINT_I(160MHZ),
                              CHECK_AND_PRINT_I(INDOOR_ONLY),
                              CHECK_AND_PRINT_I(GO_CONCURRENT),
-                             ch_flags,
+                             ch_flags, reg_rule_flags,
                              ((ch_flags & NVM_CHANNEL_ACTIVE) &&
                               !(ch_flags & NVM_CHANNEL_RADAR))
-                                        ? "" : "not ");
+                                        ? "Ad-Hoc" : "");
        }
 
        regd->n_reg_rules = valid_rules;
index 79e7a7a285dc960597e945ce7b80f96e729a86ad..82863e9273eb66ce9058a5974d2f35988290e1dc 100644 (file)
@@ -1275,8 +1275,10 @@ static int iwl_mvm_sar_get_wgds_table(struct iwl_mvm *mvm)
 
                        entry = &wifi_pkg->package.elements[idx++];
                        if ((entry->type != ACPI_TYPE_INTEGER) ||
-                           (entry->integer.value > U8_MAX))
-                               return -EINVAL;
+                           (entry->integer.value > U8_MAX)) {
+                               ret = -EINVAL;
+                               goto out_free;
+                       }
 
                        mvm->geo_profiles[i].values[j] = entry->integer.value;
                }
index c7b1e58e33847a8250ab708695373a0817f85874..ce901be5fba87e3674f06b030cfae8759e063c3b 100644 (file)
@@ -2597,8 +2597,18 @@ static void iwl_mvm_purge_deferred_tx_frames(struct iwl_mvm *mvm,
        spin_lock_bh(&mvm_sta->lock);
        for (i = 0; i <= IWL_MAX_TID_COUNT; i++) {
                tid_data = &mvm_sta->tid_data[i];
-               while ((skb = __skb_dequeue(&tid_data->deferred_tx_frames)))
+
+               while ((skb = __skb_dequeue(&tid_data->deferred_tx_frames))) {
+                       struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+
+                       /*
+                        * The first deferred frame should've stopped the MAC
+                        * queues, so we should never get a second deferred
+                        * frame for the RA/TID.
+                        */
+                       iwl_mvm_start_mac_queues(mvm, info->hw_queue);
                        ieee80211_free_txskb(mvm->hw, skb);
+               }
        }
        spin_unlock_bh(&mvm_sta->lock);
 }
index 65beca3a457a19d2dcd804d01a590f7404e04f65..8999a1199d60d27bdec88d6c638feb85403c3ced 100644 (file)
@@ -1291,7 +1291,7 @@ void iwl_mvm_rs_tx_status(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
         * first index into rate scale table.
         */
        if (info->flags & IEEE80211_TX_STAT_AMPDU) {
-               rs_collect_tpc_data(mvm, lq_sta, curr_tbl, lq_rate.index,
+               rs_collect_tpc_data(mvm, lq_sta, curr_tbl, tx_resp_rate.index,
                                    info->status.ampdu_len,
                                    info->status.ampdu_ack_len,
                                    reduced_txp);
@@ -1312,7 +1312,7 @@ void iwl_mvm_rs_tx_status(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
                if (info->status.ampdu_ack_len == 0)
                        info->status.ampdu_len = 1;
 
-               rs_collect_tlc_data(mvm, lq_sta, curr_tbl, lq_rate.index,
+               rs_collect_tlc_data(mvm, lq_sta, curr_tbl, tx_resp_rate.index,
                                    info->status.ampdu_len,
                                    info->status.ampdu_ack_len);
 
@@ -1348,11 +1348,11 @@ void iwl_mvm_rs_tx_status(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
                                continue;
 
                        rs_collect_tpc_data(mvm, lq_sta, tmp_tbl,
-                                           lq_rate.index, 1,
+                                           tx_resp_rate.index, 1,
                                            i < retries ? 0 : legacy_success,
                                            reduced_txp);
                        rs_collect_tlc_data(mvm, lq_sta, tmp_tbl,
-                                           lq_rate.index, 1,
+                                           tx_resp_rate.index, 1,
                                            i < retries ? 0 : legacy_success);
                }
 
index f3e608196369a3885804b42331f442a16623e388..71c8b800ffa99874bd4120e59e2aa7f51212d2da 100644 (file)
@@ -636,9 +636,9 @@ static bool iwl_mvm_reorder(struct iwl_mvm *mvm,
 
        baid_data = rcu_dereference(mvm->baid_map[baid]);
        if (!baid_data) {
-               WARN(!(reorder & IWL_RX_MPDU_REORDER_BA_OLD_SN),
-                    "Received baid %d, but no data exists for this BAID\n",
-                    baid);
+               IWL_DEBUG_RX(mvm,
+                            "Got valid BAID but no baid allocated, bypass the re-ordering buffer. Baid %d reorder 0x%x\n",
+                             baid, reorder);
                return false;
        }
 
@@ -759,7 +759,9 @@ static void iwl_mvm_agg_rx_received(struct iwl_mvm *mvm,
 
        data = rcu_dereference(mvm->baid_map[baid]);
        if (!data) {
-               WARN_ON(!(reorder_data & IWL_RX_MPDU_REORDER_BA_OLD_SN));
+               IWL_DEBUG_RX(mvm,
+                            "Got valid BAID but no baid allocated, bypass the re-ordering buffer. Baid %d reorder 0x%x\n",
+                             baid, reorder_data);
                goto out;
        }
 
index ab66b4394dfc8ca2afc0cf1321f93f28f186d1cb..027ee5e72172c85f9eaa98f1fd1a26489ab74820 100644 (file)
@@ -121,7 +121,8 @@ int iwl_mvm_sta_send_to_fw(struct iwl_mvm *mvm, struct ieee80211_sta *sta,
                .mac_id_n_color = cpu_to_le32(mvm_sta->mac_id_n_color),
                .add_modify = update ? 1 : 0,
                .station_flags_msk = cpu_to_le32(STA_FLG_FAT_EN_MSK |
-                                                STA_FLG_MIMO_EN_MSK),
+                                                STA_FLG_MIMO_EN_MSK |
+                                                STA_FLG_RTS_MIMO_PROT),
                .tid_disable_tx = cpu_to_le16(mvm_sta->tid_disable_agg),
        };
        int ret;
@@ -290,8 +291,8 @@ static void iwl_mvm_rx_agg_session_expired(unsigned long data)
                goto unlock;
 
        mvm_sta = iwl_mvm_sta_from_mac80211(sta);
-       ieee80211_stop_rx_ba_session_offl(mvm_sta->vif,
-                                         sta->addr, ba_data->tid);
+       ieee80211_rx_ba_timer_expired(mvm_sta->vif,
+                                     sta->addr, ba_data->tid);
 unlock:
        rcu_read_unlock();
 }
index 60360ed73f26165c890d0af3bac742b5669fb8d9..5fcc9dd6be56de52fa0a063969cf58011b8eb3d8 100644 (file)
@@ -185,8 +185,14 @@ static u16 iwl_mvm_tx_csum(struct iwl_mvm *mvm, struct sk_buff *skb,
        else
                udp_hdr(skb)->check = 0;
 
-       /* mac header len should include IV, size is in words */
-       if (info->control.hw_key)
+       /*
+        * mac header len should include IV, size is in words unless
+        * the IV is added by the firmware like in WEP.
+        * In new Tx API, the IV is always added by the firmware.
+        */
+       if (!iwl_mvm_has_new_tx_api(mvm) && info->control.hw_key &&
+           info->control.hw_key->cipher != WLAN_CIPHER_SUITE_WEP40 &&
+           info->control.hw_key->cipher != WLAN_CIPHER_SUITE_WEP104)
                mh_len += info->control.hw_key->iv_len;
        mh_len /= 2;
        offload_assist |= mh_len << TX_CMD_OFFLD_MH_SIZE;
@@ -1815,6 +1821,8 @@ void iwl_mvm_rx_ba_notif(struct iwl_mvm *mvm, struct iwl_rx_cmd_buffer *rxb)
        struct iwl_mvm_tid_data *tid_data;
        struct iwl_mvm_sta *mvmsta;
 
+       ba_info.flags = IEEE80211_TX_STAT_AMPDU;
+
        if (iwl_mvm_has_new_tx_api(mvm)) {
                struct iwl_mvm_compressed_ba_notif *ba_res =
                        (void *)pkt->data;
index f16c1bb9bf94b6bef6bfb6ce2ff36c3faf771c00..84f4ba01e14fa2e84878dc75fda5d050e3500880 100644 (file)
@@ -510,9 +510,17 @@ static const struct pci_device_id iwl_hw_card_ids[] = {
 
 /* 9000 Series */
        {IWL_PCI_DEVICE(0x271B, 0x0010, iwl9160_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x271B, 0x0014, iwl9160_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x271B, 0x0210, iwl9160_2ac_cfg)},
        {IWL_PCI_DEVICE(0x2526, 0x0000, iwl9260_2ac_cfg)},
        {IWL_PCI_DEVICE(0x2526, 0x0010, iwl9260_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0014, iwl9260_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0xA014, iwl9260_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x4010, iwl9260_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0210, iwl9260_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0214, iwl9260_2ac_cfg)},
        {IWL_PCI_DEVICE(0x2526, 0x1410, iwl9270_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x1610, iwl9270_2ac_cfg)},
        {IWL_PCI_DEVICE(0x9DF0, 0x0A10, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x9DF0, 0x0010, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x9DF0, 0x0210, iwl9460_2ac_cfg)},
@@ -527,10 +535,22 @@ static const struct pci_device_id iwl_hw_card_ids[] = {
        {IWL_PCI_DEVICE(0x9DF0, 0x2A10, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x30DC, 0x0060, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x2526, 0x0060, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0260, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0064, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x00A4, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x40A4, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x02A4, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x00A0, iwl9460_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x02A0, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x9DF0, 0x0060, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0xA370, 0x0060, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x31DC, 0x0060, iwl9460_2ac_cfg)},
        {IWL_PCI_DEVICE(0x2526, 0x0030, iwl9560_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x4030, iwl9560_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0230, iwl9560_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0234, iwl9560_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x0238, iwl9560_2ac_cfg)},
+       {IWL_PCI_DEVICE(0x2526, 0x023C, iwl9560_2ac_cfg)},
        {IWL_PCI_DEVICE(0x9DF0, 0x0030, iwl9560_2ac_cfg)},
        {IWL_PCI_DEVICE(0xA370, 0x0030, iwl9560_2ac_cfg)},
        {IWL_PCI_DEVICE(0x31DC, 0x0030, iwl9560_2ac_cfg)},
index c49f1f8b2e57459deb2605f242c9324e8960a95b..37046ac2c4413a51b9a77864cdf9e434a6ac7bd6 100644 (file)
@@ -336,7 +336,7 @@ static int nvme_get_stream_params(struct nvme_ctrl *ctrl,
 
        c.directive.opcode = nvme_admin_directive_recv;
        c.directive.nsid = cpu_to_le32(nsid);
-       c.directive.numd = cpu_to_le32(sizeof(*s));
+       c.directive.numd = cpu_to_le32((sizeof(*s) >> 2) - 1);
        c.directive.doper = NVME_DIR_RCV_ST_OP_PARAM;
        c.directive.dtype = NVME_DIR_STREAMS;
 
@@ -1509,7 +1509,7 @@ static void nvme_set_queue_limits(struct nvme_ctrl *ctrl,
        blk_queue_write_cache(q, vwc, vwc);
 }
 
-static void nvme_configure_apst(struct nvme_ctrl *ctrl)
+static int nvme_configure_apst(struct nvme_ctrl *ctrl)
 {
        /*
         * APST (Autonomous Power State Transition) lets us program a
@@ -1538,16 +1538,16 @@ static void nvme_configure_apst(struct nvme_ctrl *ctrl)
         * then don't do anything.
         */
        if (!ctrl->apsta)
-               return;
+               return 0;
 
        if (ctrl->npss > 31) {
                dev_warn(ctrl->device, "NPSS is invalid; not using APST\n");
-               return;
+               return 0;
        }
 
        table = kzalloc(sizeof(*table), GFP_KERNEL);
        if (!table)
-               return;
+               return 0;
 
        if (!ctrl->apst_enabled || ctrl->ps_max_latency_us == 0) {
                /* Turn off APST. */
@@ -1629,6 +1629,7 @@ static void nvme_configure_apst(struct nvme_ctrl *ctrl)
                dev_err(ctrl->device, "failed to set APST feature (%d)\n", ret);
 
        kfree(table);
+       return ret;
 }
 
 static void nvme_set_latency_tolerance(struct device *dev, s32 val)
@@ -1835,13 +1836,16 @@ int nvme_init_identify(struct nvme_ctrl *ctrl)
                 * In fabrics we need to verify the cntlid matches the
                 * admin connect
                 */
-               if (ctrl->cntlid != le16_to_cpu(id->cntlid))
+               if (ctrl->cntlid != le16_to_cpu(id->cntlid)) {
                        ret = -EINVAL;
+                       goto out_free;
+               }
 
                if (!ctrl->opts->discovery_nqn && !ctrl->kas) {
                        dev_err(ctrl->device,
                                "keep-alive support is mandatory for fabrics\n");
                        ret = -EINVAL;
+                       goto out_free;
                }
        } else {
                ctrl->cntlid = le16_to_cpu(id->cntlid);
@@ -1856,11 +1860,20 @@ int nvme_init_identify(struct nvme_ctrl *ctrl)
        else if (!ctrl->apst_enabled && prev_apst_enabled)
                dev_pm_qos_hide_latency_tolerance(ctrl->device);
 
-       nvme_configure_apst(ctrl);
-       nvme_configure_directives(ctrl);
+       ret = nvme_configure_apst(ctrl);
+       if (ret < 0)
+               return ret;
+
+       ret = nvme_configure_directives(ctrl);
+       if (ret < 0)
+               return ret;
 
        ctrl->identified = true;
 
+       return 0;
+
+out_free:
+       kfree(id);
        return ret;
 }
 EXPORT_SYMBOL_GPL(nvme_init_identify);
@@ -2004,9 +2017,11 @@ static ssize_t wwid_show(struct device *dev, struct device_attribute *attr,
        if (memchr_inv(ns->eui, 0, sizeof(ns->eui)))
                return sprintf(buf, "eui.%8phN\n", ns->eui);
 
-       while (ctrl->serial[serial_len - 1] == ' ')
+       while (serial_len > 0 && (ctrl->serial[serial_len - 1] == ' ' ||
+                                 ctrl->serial[serial_len - 1] == '\0'))
                serial_len--;
-       while (ctrl->model[model_len - 1] == ' ')
+       while (model_len > 0 && (ctrl->model[model_len - 1] == ' ' ||
+                                ctrl->model[model_len - 1] == '\0'))
                model_len--;
 
        return sprintf(buf, "nvme.%04x-%*phN-%*phN-%08x\n", ctrl->vid,
index 2e582a2409437bca7d603598290d023bd988c577..5f5cd306f76d05e8cb7351cb91403d7f9a554d9a 100644 (file)
@@ -794,7 +794,8 @@ static int nvmf_check_allowed_opts(struct nvmf_ctrl_options *opts,
                int i;
 
                for (i = 0; i < ARRAY_SIZE(opt_tokens); i++) {
-                       if (opt_tokens[i].token & ~allowed_opts) {
+                       if ((opt_tokens[i].token & opts->mask) &&
+                           (opt_tokens[i].token & ~allowed_opts)) {
                                pr_warn("invalid parameter '%s'\n",
                                        opt_tokens[i].pattern);
                        }
index cd888a47d0fccb728b155f75aa85f0836ec4eac3..925467b31a333940dc62d62c4c5f3376316e4120 100644 (file)
@@ -801,6 +801,7 @@ static inline void nvme_handle_cqe(struct nvme_queue *nvmeq,
                return;
        }
 
+       nvmeq->cqe_seen = 1;
        req = blk_mq_tag_to_rq(*nvmeq->tags, cqe->command_id);
        nvme_end_request(req, cqe->status, cqe->result);
 }
@@ -830,10 +831,8 @@ static void nvme_process_cq(struct nvme_queue *nvmeq)
                consumed++;
        }
 
-       if (consumed) {
+       if (consumed)
                nvme_ring_cq_doorbell(nvmeq);
-               nvmeq->cqe_seen = 1;
-       }
 }
 
 static irqreturn_t nvme_irq(int irq, void *data)
@@ -1558,11 +1557,9 @@ static inline void nvme_release_cmb(struct nvme_dev *dev)
        if (dev->cmb) {
                iounmap(dev->cmb);
                dev->cmb = NULL;
-               if (dev->cmbsz) {
-                       sysfs_remove_file_from_group(&dev->ctrl.device->kobj,
-                                                    &dev_attr_cmb.attr, NULL);
-                       dev->cmbsz = 0;
-               }
+               sysfs_remove_file_from_group(&dev->ctrl.device->kobj,
+                                            &dev_attr_cmb.attr, NULL);
+               dev->cmbsz = 0;
        }
 }
 
@@ -1953,16 +1950,14 @@ static int nvme_pci_enable(struct nvme_dev *dev)
 
        /*
         * CMBs can currently only exist on >=1.2 PCIe devices. We only
-        * populate sysfs if a CMB is implemented. Note that we add the
-        * CMB attribute to the nvme_ctrl kobj which removes the need to remove
-        * it on exit. Since nvme_dev_attrs_group has no name we can pass
-        * NULL as final argument to sysfs_add_file_to_group.
+        * populate sysfs if a CMB is implemented. Since nvme_dev_attrs_group
+        * has no name we can pass NULL as final argument to
+        * sysfs_add_file_to_group.
         */
 
        if (readl(dev->bar + NVME_REG_VS) >= NVME_VS(1, 2, 0)) {
                dev->cmb = nvme_map_cmb(dev);
-
-               if (dev->cmbsz) {
+               if (dev->cmb) {
                        if (sysfs_add_file_to_group(&dev->ctrl.device->kobj,
                                                    &dev_attr_cmb.attr, NULL))
                                dev_warn(dev->ctrl.device,
index 2d7a98ab53fbf2de131990b753b929fe31cd154b..a53bb6635b8378d00f7fb4d367a16170d81a846e 100644 (file)
@@ -199,12 +199,6 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
        copy_and_pad(id->mn, sizeof(id->mn), model, sizeof(model) - 1);
        copy_and_pad(id->fr, sizeof(id->fr), UTS_RELEASE, strlen(UTS_RELEASE));
 
-       memset(id->mn, ' ', sizeof(id->mn));
-       strncpy((char *)id->mn, "Linux", sizeof(id->mn));
-
-       memset(id->fr, ' ', sizeof(id->fr));
-       strncpy((char *)id->fr, UTS_RELEASE, sizeof(id->fr));
-
        id->rab = 6;
 
        /*
index 31ca55dfcb1d49f3a1d88f7c6f5d0e7f0ee1e1ea..309c84aa7595b9b2ffab72c8a632c97f29def6d1 100644 (file)
@@ -114,6 +114,11 @@ struct nvmet_fc_tgtport {
        struct kref                     ref;
 };
 
+struct nvmet_fc_defer_fcp_req {
+       struct list_head                req_list;
+       struct nvmefc_tgt_fcp_req       *fcp_req;
+};
+
 struct nvmet_fc_tgt_queue {
        bool                            ninetypercent;
        u16                             qid;
@@ -132,6 +137,8 @@ struct nvmet_fc_tgt_queue {
        struct nvmet_fc_tgt_assoc       *assoc;
        struct nvmet_fc_fcp_iod         *fod;           /* array of fcp_iods */
        struct list_head                fod_list;
+       struct list_head                pending_cmd_list;
+       struct list_head                avail_defer_list;
        struct workqueue_struct         *work_q;
        struct kref                     ref;
 } __aligned(sizeof(unsigned long long));
@@ -223,6 +230,8 @@ static void nvmet_fc_tgt_q_put(struct nvmet_fc_tgt_queue *queue);
 static int nvmet_fc_tgt_q_get(struct nvmet_fc_tgt_queue *queue);
 static void nvmet_fc_tgtport_put(struct nvmet_fc_tgtport *tgtport);
 static int nvmet_fc_tgtport_get(struct nvmet_fc_tgtport *tgtport);
+static void nvmet_fc_handle_fcp_rqst(struct nvmet_fc_tgtport *tgtport,
+                                       struct nvmet_fc_fcp_iod *fod);
 
 
 /* *********************** FC-NVME DMA Handling **************************** */
@@ -385,7 +394,7 @@ nvmet_fc_free_ls_iodlist(struct nvmet_fc_tgtport *tgtport)
 static struct nvmet_fc_ls_iod *
 nvmet_fc_alloc_ls_iod(struct nvmet_fc_tgtport *tgtport)
 {
-       static struct nvmet_fc_ls_iod *iod;
+       struct nvmet_fc_ls_iod *iod;
        unsigned long flags;
 
        spin_lock_irqsave(&tgtport->lock, flags);
@@ -462,10 +471,10 @@ nvmet_fc_destroy_fcp_iodlist(struct nvmet_fc_tgtport *tgtport,
 static struct nvmet_fc_fcp_iod *
 nvmet_fc_alloc_fcp_iod(struct nvmet_fc_tgt_queue *queue)
 {
-       static struct nvmet_fc_fcp_iod *fod;
-       unsigned long flags;
+       struct nvmet_fc_fcp_iod *fod;
+
+       lockdep_assert_held(&queue->qlock);
 
-       spin_lock_irqsave(&queue->qlock, flags);
        fod = list_first_entry_or_null(&queue->fod_list,
                                        struct nvmet_fc_fcp_iod, fcp_list);
        if (fod) {
@@ -477,17 +486,37 @@ nvmet_fc_alloc_fcp_iod(struct nvmet_fc_tgt_queue *queue)
                 * will "inherit" that reference.
                 */
        }
-       spin_unlock_irqrestore(&queue->qlock, flags);
        return fod;
 }
 
 
+static void
+nvmet_fc_queue_fcp_req(struct nvmet_fc_tgtport *tgtport,
+                      struct nvmet_fc_tgt_queue *queue,
+                      struct nvmefc_tgt_fcp_req *fcpreq)
+{
+       struct nvmet_fc_fcp_iod *fod = fcpreq->nvmet_fc_private;
+
+       /*
+        * put all admin cmds on hw queue id 0. All io commands go to
+        * the respective hw queue based on a modulo basis
+        */
+       fcpreq->hwqid = queue->qid ?
+                       ((queue->qid - 1) % tgtport->ops->max_hw_queues) : 0;
+
+       if (tgtport->ops->target_features & NVMET_FCTGTFEAT_CMD_IN_ISR)
+               queue_work_on(queue->cpu, queue->work_q, &fod->work);
+       else
+               nvmet_fc_handle_fcp_rqst(tgtport, fod);
+}
+
 static void
 nvmet_fc_free_fcp_iod(struct nvmet_fc_tgt_queue *queue,
                        struct nvmet_fc_fcp_iod *fod)
 {
        struct nvmefc_tgt_fcp_req *fcpreq = fod->fcpreq;
        struct nvmet_fc_tgtport *tgtport = fod->tgtport;
+       struct nvmet_fc_defer_fcp_req *deferfcp;
        unsigned long flags;
 
        fc_dma_sync_single_for_cpu(tgtport->dev, fod->rspdma,
@@ -495,21 +524,56 @@ nvmet_fc_free_fcp_iod(struct nvmet_fc_tgt_queue *queue,
 
        fcpreq->nvmet_fc_private = NULL;
 
-       spin_lock_irqsave(&queue->qlock, flags);
-       list_add_tail(&fod->fcp_list, &fod->queue->fod_list);
        fod->active = false;
        fod->abort = false;
        fod->aborted = false;
        fod->writedataactive = false;
        fod->fcpreq = NULL;
+
+       tgtport->ops->fcp_req_release(&tgtport->fc_target_port, fcpreq);
+
+       spin_lock_irqsave(&queue->qlock, flags);
+       deferfcp = list_first_entry_or_null(&queue->pending_cmd_list,
+                               struct nvmet_fc_defer_fcp_req, req_list);
+       if (!deferfcp) {
+               list_add_tail(&fod->fcp_list, &fod->queue->fod_list);
+               spin_unlock_irqrestore(&queue->qlock, flags);
+
+               /* Release reference taken at queue lookup and fod allocation */
+               nvmet_fc_tgt_q_put(queue);
+               return;
+       }
+
+       /* Re-use the fod for the next pending cmd that was deferred */
+       list_del(&deferfcp->req_list);
+
+       fcpreq = deferfcp->fcp_req;
+
+       /* deferfcp can be reused for another IO at a later date */
+       list_add_tail(&deferfcp->req_list, &queue->avail_defer_list);
+
        spin_unlock_irqrestore(&queue->qlock, flags);
 
+       /* Save NVME CMD IO in fod */
+       memcpy(&fod->cmdiubuf, fcpreq->rspaddr, fcpreq->rsplen);
+
+       /* Setup new fcpreq to be processed */
+       fcpreq->rspaddr = NULL;
+       fcpreq->rsplen  = 0;
+       fcpreq->nvmet_fc_private = fod;
+       fod->fcpreq = fcpreq;
+       fod->active = true;
+
+       /* inform LLDD IO is now being processed */
+       tgtport->ops->defer_rcv(&tgtport->fc_target_port, fcpreq);
+
+       /* Submit deferred IO for processing */
+       nvmet_fc_queue_fcp_req(tgtport, queue, fcpreq);
+
        /*
-        * release the reference taken at queue lookup and fod allocation
+        * Leave the queue lookup get reference taken when
+        * fod was originally allocated.
         */
-       nvmet_fc_tgt_q_put(queue);
-
-       tgtport->ops->fcp_req_release(&tgtport->fc_target_port, fcpreq);
 }
 
 static int
@@ -569,6 +633,8 @@ nvmet_fc_alloc_target_queue(struct nvmet_fc_tgt_assoc *assoc,
        queue->port = assoc->tgtport->port;
        queue->cpu = nvmet_fc_queue_to_cpu(assoc->tgtport, qid);
        INIT_LIST_HEAD(&queue->fod_list);
+       INIT_LIST_HEAD(&queue->avail_defer_list);
+       INIT_LIST_HEAD(&queue->pending_cmd_list);
        atomic_set(&queue->connected, 0);
        atomic_set(&queue->sqtail, 0);
        atomic_set(&queue->rsn, 1);
@@ -638,6 +704,7 @@ nvmet_fc_delete_target_queue(struct nvmet_fc_tgt_queue *queue)
 {
        struct nvmet_fc_tgtport *tgtport = queue->assoc->tgtport;
        struct nvmet_fc_fcp_iod *fod = queue->fod;
+       struct nvmet_fc_defer_fcp_req *deferfcp, *tempptr;
        unsigned long flags;
        int i, writedataactive;
        bool disconnect;
@@ -666,6 +733,36 @@ nvmet_fc_delete_target_queue(struct nvmet_fc_tgt_queue *queue)
                        }
                }
        }
+
+       /* Cleanup defer'ed IOs in queue */
+       list_for_each_entry_safe(deferfcp, tempptr, &queue->avail_defer_list,
+                               req_list) {
+               list_del(&deferfcp->req_list);
+               kfree(deferfcp);
+       }
+
+       for (;;) {
+               deferfcp = list_first_entry_or_null(&queue->pending_cmd_list,
+                               struct nvmet_fc_defer_fcp_req, req_list);
+               if (!deferfcp)
+                       break;
+
+               list_del(&deferfcp->req_list);
+               spin_unlock_irqrestore(&queue->qlock, flags);
+
+               tgtport->ops->defer_rcv(&tgtport->fc_target_port,
+                               deferfcp->fcp_req);
+
+               tgtport->ops->fcp_abort(&tgtport->fc_target_port,
+                               deferfcp->fcp_req);
+
+               tgtport->ops->fcp_req_release(&tgtport->fc_target_port,
+                               deferfcp->fcp_req);
+
+               kfree(deferfcp);
+
+               spin_lock_irqsave(&queue->qlock, flags);
+       }
        spin_unlock_irqrestore(&queue->qlock, flags);
 
        flush_workqueue(queue->work_q);
@@ -2172,11 +2269,38 @@ nvmet_fc_handle_fcp_rqst_work(struct work_struct *work)
  * Pass a FC-NVME FCP CMD IU received from the FC link to the nvmet-fc
  * layer for processing.
  *
- * The nvmet-fc layer will copy cmd payload to an internal structure for
- * processing.  As such, upon completion of the routine, the LLDD may
- * immediately free/reuse the CMD IU buffer passed in the call.
+ * The nvmet_fc layer allocates a local job structure (struct
+ * nvmet_fc_fcp_iod) from the queue for the io and copies the
+ * CMD IU buffer to the job structure. As such, on a successful
+ * completion (returns 0), the LLDD may immediately free/reuse
+ * the CMD IU buffer passed in the call.
  *
- * If this routine returns error, the lldd should abort the exchange.
+ * However, in some circumstances, due to the packetized nature of FC
+ * and the api of the FC LLDD which may issue a hw command to send the
+ * response, but the LLDD may not get the hw completion for that command
+ * and upcall the nvmet_fc layer before a new command may be
+ * asynchronously received - its possible for a command to be received
+ * before the LLDD and nvmet_fc have recycled the job structure. It gives
+ * the appearance of more commands received than fits in the sq.
+ * To alleviate this scenario, a temporary queue is maintained in the
+ * transport for pending LLDD requests waiting for a queue job structure.
+ * In these "overrun" cases, a temporary queue element is allocated
+ * the LLDD request and CMD iu buffer information remembered, and the
+ * routine returns a -EOVERFLOW status. Subsequently, when a queue job
+ * structure is freed, it is immediately reallocated for anything on the
+ * pending request list. The LLDDs defer_rcv() callback is called,
+ * informing the LLDD that it may reuse the CMD IU buffer, and the io
+ * is then started normally with the transport.
+ *
+ * The LLDD, when receiving an -EOVERFLOW completion status, is to treat
+ * the completion as successful but must not reuse the CMD IU buffer
+ * until the LLDD's defer_rcv() callback has been called for the
+ * corresponding struct nvmefc_tgt_fcp_req pointer.
+ *
+ * If there is any other condition in which an error occurs, the
+ * transport will return a non-zero status indicating the error.
+ * In all cases other than -EOVERFLOW, the transport has not accepted the
+ * request and the LLDD should abort the exchange.
  *
  * @target_port: pointer to the (registered) target port the FCP CMD IU
  *              was received on.
@@ -2194,6 +2318,8 @@ nvmet_fc_rcv_fcp_req(struct nvmet_fc_target_port *target_port,
        struct nvme_fc_cmd_iu *cmdiu = cmdiubuf;
        struct nvmet_fc_tgt_queue *queue;
        struct nvmet_fc_fcp_iod *fod;
+       struct nvmet_fc_defer_fcp_req *deferfcp;
+       unsigned long flags;
 
        /* validate iu, so the connection id can be used to find the queue */
        if ((cmdiubuf_len != sizeof(*cmdiu)) ||
@@ -2214,29 +2340,60 @@ nvmet_fc_rcv_fcp_req(struct nvmet_fc_target_port *target_port,
         * when the fod is freed.
         */
 
+       spin_lock_irqsave(&queue->qlock, flags);
+
        fod = nvmet_fc_alloc_fcp_iod(queue);
-       if (!fod) {
+       if (fod) {
+               spin_unlock_irqrestore(&queue->qlock, flags);
+
+               fcpreq->nvmet_fc_private = fod;
+               fod->fcpreq = fcpreq;
+
+               memcpy(&fod->cmdiubuf, cmdiubuf, cmdiubuf_len);
+
+               nvmet_fc_queue_fcp_req(tgtport, queue, fcpreq);
+
+               return 0;
+       }
+
+       if (!tgtport->ops->defer_rcv) {
+               spin_unlock_irqrestore(&queue->qlock, flags);
                /* release the queue lookup reference */
                nvmet_fc_tgt_q_put(queue);
                return -ENOENT;
        }
 
-       fcpreq->nvmet_fc_private = fod;
-       fod->fcpreq = fcpreq;
-       /*
-        * put all admin cmds on hw queue id 0. All io commands go to
-        * the respective hw queue based on a modulo basis
-        */
-       fcpreq->hwqid = queue->qid ?
-                       ((queue->qid - 1) % tgtport->ops->max_hw_queues) : 0;
-       memcpy(&fod->cmdiubuf, cmdiubuf, cmdiubuf_len);
+       deferfcp = list_first_entry_or_null(&queue->avail_defer_list,
+                       struct nvmet_fc_defer_fcp_req, req_list);
+       if (deferfcp) {
+               /* Just re-use one that was previously allocated */
+               list_del(&deferfcp->req_list);
+       } else {
+               spin_unlock_irqrestore(&queue->qlock, flags);
 
-       if (tgtport->ops->target_features & NVMET_FCTGTFEAT_CMD_IN_ISR)
-               queue_work_on(queue->cpu, queue->work_q, &fod->work);
-       else
-               nvmet_fc_handle_fcp_rqst(tgtport, fod);
+               /* Now we need to dynamically allocate one */
+               deferfcp = kmalloc(sizeof(*deferfcp), GFP_KERNEL);
+               if (!deferfcp) {
+                       /* release the queue lookup reference */
+                       nvmet_fc_tgt_q_put(queue);
+                       return -ENOMEM;
+               }
+               spin_lock_irqsave(&queue->qlock, flags);
+       }
 
-       return 0;
+       /* For now, use rspaddr / rsplen to save payload information */
+       fcpreq->rspaddr = cmdiubuf;
+       fcpreq->rsplen  = cmdiubuf_len;
+       deferfcp->fcp_req = fcpreq;
+
+       /* defer processing till a fod becomes available */
+       list_add_tail(&deferfcp->req_list, &queue->pending_cmd_list);
+
+       /* NOTE: the queue lookup reference is still valid */
+
+       spin_unlock_irqrestore(&queue->qlock, flags);
+
+       return -EOVERFLOW;
 }
 EXPORT_SYMBOL_GPL(nvmet_fc_rcv_fcp_req);
 
index 28c38c756f92858906ca2aee54a9c922522f9253..e0a28ea341fe95e84a866559dad8e317c8056401 100644 (file)
@@ -89,6 +89,7 @@ int of_dma_configure(struct device *dev, struct device_node *np)
        bool coherent;
        unsigned long offset;
        const struct iommu_ops *iommu;
+       u64 mask;
 
        /*
         * Set default coherent_dma_mask to 32 bit.  Drivers are expected to
@@ -134,10 +135,9 @@ int of_dma_configure(struct device *dev, struct device_node *np)
         * Limit coherent and dma mask based on size and default mask
         * set by the driver.
         */
-       dev->coherent_dma_mask = min(dev->coherent_dma_mask,
-                                    DMA_BIT_MASK(ilog2(dma_addr + size)));
-       *dev->dma_mask = min((*dev->dma_mask),
-                            DMA_BIT_MASK(ilog2(dma_addr + size)));
+       mask = DMA_BIT_MASK(ilog2(dma_addr + size - 1) + 1);
+       dev->coherent_dma_mask &= mask;
+       *dev->dma_mask &= mask;
 
        coherent = of_dma_is_coherent(np);
        dev_dbg(dev, "device is%sdma coherent\n",
index 5c63b920b4713eaf41954c4e8f0d041472c78b6d..ed92c1254cff473113f9b24e25a598a72871891d 100644 (file)
@@ -956,7 +956,7 @@ static int __init dino_probe(struct parisc_device *dev)
 
        dino_dev->hba.dev = dev;
        dino_dev->hba.base_addr = ioremap_nocache(hpa, 4096);
-       dino_dev->hba.lmmio_space_offset = 0;   /* CPU addrs == bus addrs */
+       dino_dev->hba.lmmio_space_offset = PCI_F_EXTEND;
        spin_lock_init(&dino_dev->dinosaur_pen);
        dino_dev->hba.iommu = ccio_get_iommu(dev);
 
index b4b7eab2940024024c46ead23d6b1c415fa146f7..fdf65a6c13f62558e887cd001f4c61ca2daef989 100644 (file)
@@ -514,7 +514,7 @@ EXPORT_SYMBOL(pci_find_resource);
  */
 struct pci_dev *pci_find_pcie_root_port(struct pci_dev *dev)
 {
-       struct pci_dev *bridge, *highest_pcie_bridge = NULL;
+       struct pci_dev *bridge, *highest_pcie_bridge = dev;
 
        bridge = pci_upstream_bridge(dev);
        while (bridge && pci_is_pcie(bridge)) {
index c31310db04047367c44b96ff6afa41a6b991bc28..e6a917b4acd3f3c63dff5d57b905ab093d7d0487 100644 (file)
@@ -1762,6 +1762,48 @@ static void pci_configure_extended_tags(struct pci_dev *dev)
                                         PCI_EXP_DEVCTL_EXT_TAG);
 }
 
+/**
+ * pcie_relaxed_ordering_enabled - Probe for PCIe relaxed ordering enable
+ * @dev: PCI device to query
+ *
+ * Returns true if the device has enabled relaxed ordering attribute.
+ */
+bool pcie_relaxed_ordering_enabled(struct pci_dev *dev)
+{
+       u16 v;
+
+       pcie_capability_read_word(dev, PCI_EXP_DEVCTL, &v);
+
+       return !!(v & PCI_EXP_DEVCTL_RELAX_EN);
+}
+EXPORT_SYMBOL(pcie_relaxed_ordering_enabled);
+
+static void pci_configure_relaxed_ordering(struct pci_dev *dev)
+{
+       struct pci_dev *root;
+
+       /* PCI_EXP_DEVICE_RELAX_EN is RsvdP in VFs */
+       if (dev->is_virtfn)
+               return;
+
+       if (!pcie_relaxed_ordering_enabled(dev))
+               return;
+
+       /*
+        * For now, we only deal with Relaxed Ordering issues with Root
+        * Ports. Peer-to-Peer DMA is another can of worms.
+        */
+       root = pci_find_pcie_root_port(dev);
+       if (!root)
+               return;
+
+       if (root->dev_flags & PCI_DEV_FLAGS_NO_RELAXED_ORDERING) {
+               pcie_capability_clear_word(dev, PCI_EXP_DEVCTL,
+                                          PCI_EXP_DEVCTL_RELAX_EN);
+               dev_info(&dev->dev, "Disable Relaxed Ordering because the Root Port didn't support it\n");
+       }
+}
+
 static void pci_configure_device(struct pci_dev *dev)
 {
        struct hotplug_params hpp;
@@ -1769,6 +1811,7 @@ static void pci_configure_device(struct pci_dev *dev)
 
        pci_configure_mps(dev);
        pci_configure_extended_tags(dev);
+       pci_configure_relaxed_ordering(dev);
 
        memset(&hpp, 0, sizeof(hpp));
        ret = pci_get_hp_params(dev, &hpp);
index 6967c6b4cf6b017170619c285ee58feaa5975a5a..140760403f36a932b7d39eb73ae3eb96d76cab37 100644 (file)
@@ -4015,6 +4015,95 @@ DECLARE_PCI_FIXUP_CLASS_EARLY(0x1797, 0x6868, PCI_CLASS_NOT_DEFINED, 8,
 DECLARE_PCI_FIXUP_CLASS_EARLY(0x1797, 0x6869, PCI_CLASS_NOT_DEFINED, 8,
                              quirk_tw686x_class);
 
+/*
+ * Some devices have problems with Transaction Layer Packets with the Relaxed
+ * Ordering Attribute set.  Such devices should mark themselves and other
+ * Device Drivers should check before sending TLPs with RO set.
+ */
+static void quirk_relaxedordering_disable(struct pci_dev *dev)
+{
+       dev->dev_flags |= PCI_DEV_FLAGS_NO_RELAXED_ORDERING;
+       dev_info(&dev->dev, "Disable Relaxed Ordering Attributes to avoid PCIe Completion erratum\n");
+}
+
+/*
+ * Intel Xeon processors based on Broadwell/Haswell microarchitecture Root
+ * Complex has a Flow Control Credit issue which can cause performance
+ * problems with Upstream Transaction Layer Packets with Relaxed Ordering set.
+ */
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f01, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f02, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f03, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f04, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f05, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f06, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f07, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f08, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f09, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f0a, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f0b, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f0c, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f0d, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x6f0e, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f01, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f02, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f03, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f04, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f05, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f06, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f07, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f08, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f09, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f0a, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f0b, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f0c, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f0d, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_INTEL, 0x2f0e, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+
+/*
+ * The AMD ARM A1100 (AKA "SEATTLE") SoC has a bug in its PCIe Root Complex
+ * where Upstream Transaction Layer Packets with the Relaxed Ordering
+ * Attribute clear are allowed to bypass earlier TLPs with Relaxed Ordering
+ * set.  This is a violation of the PCIe 3.0 Transaction Ordering Rules
+ * outlined in Section 2.4.1 (PCI Express(r) Base Specification Revision 3.0
+ * November 10, 2010).  As a result, on this platform we can't use Relaxed
+ * Ordering for Upstream TLPs.
+ */
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_AMD, 0x1a00, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_AMD, 0x1a01, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+DECLARE_PCI_FIXUP_CLASS_EARLY(PCI_VENDOR_ID_AMD, 0x1a02, PCI_CLASS_NOT_DEFINED, 8,
+                             quirk_relaxedordering_disable);
+
 /*
  * Per PCIe r3.0, sec 2.2.9, "Completion headers must supply the same
  * values for the Attribute as were supplied in the header of the
index 4fac49e55d473e80c87eb932698a33383ffca04b..4b43aa62fbc789b049ecf05b3be7ec393dfa868d 100644 (file)
@@ -1301,7 +1301,6 @@ static void ds1307_clks_register(struct ds1307 *ds1307)
 static const struct regmap_config regmap_config = {
        .reg_bits = 8,
        .val_bits = 8,
-       .max_register = 0x12,
 };
 
 static int ds1307_probe(struct i2c_client *client,
index f4538d7a3016e2b1514df38c9e6e9b48939d6ce7..d145e0d9022755d476f8f48c22ea9c3a62377e28 100644 (file)
@@ -47,6 +47,17 @@ config SCSI_NETLINK
        default n
        depends on NET
 
+config SCSI_MQ_DEFAULT
+       bool "SCSI: use blk-mq I/O path by default"
+       depends on SCSI
+       ---help---
+         This option enables the new blk-mq based I/O path for SCSI
+         devices by default.  With the option the scsi_mod.use_blk_mq
+         module/boot option defaults to Y, without it to N, but it can
+         still be overridden either way.
+
+         If unsure say N.
+
 config SCSI_PROC_FS
        bool "legacy /proc/scsi/ support"
        depends on SCSI && PROC_FS
index 4591113c49de3af951908ed2257f6f5e88663b96..a1a2c71e162651f93d499c56e024e7073cc7fc79 100644 (file)
@@ -549,7 +549,9 @@ static void get_container_name_callback(void *context, struct fib * fibptr)
        if ((le32_to_cpu(get_name_reply->status) == CT_OK)
         && (get_name_reply->data[0] != '\0')) {
                char *sp = get_name_reply->data;
-               sp[sizeof(((struct aac_get_name_resp *)NULL)->data)] = '\0';
+               int data_size = FIELD_SIZEOF(struct aac_get_name_resp, data);
+
+               sp[data_size - 1] = '\0';
                while (*sp == ' ')
                        ++sp;
                if (*sp) {
@@ -579,12 +581,15 @@ static void get_container_name_callback(void *context, struct fib * fibptr)
 static int aac_get_container_name(struct scsi_cmnd * scsicmd)
 {
        int status;
+       int data_size;
        struct aac_get_name *dinfo;
        struct fib * cmd_fibcontext;
        struct aac_dev * dev;
 
        dev = (struct aac_dev *)scsicmd->device->host->hostdata;
 
+       data_size = FIELD_SIZEOF(struct aac_get_name_resp, data);
+
        cmd_fibcontext = aac_fib_alloc_tag(dev, scsicmd);
 
        aac_fib_init(cmd_fibcontext);
@@ -593,7 +598,7 @@ static int aac_get_container_name(struct scsi_cmnd * scsicmd)
        dinfo->command = cpu_to_le32(VM_ContainerConfig);
        dinfo->type = cpu_to_le32(CT_READ_NAME);
        dinfo->cid = cpu_to_le32(scmd_id(scsicmd));
-       dinfo->count = cpu_to_le32(sizeof(((struct aac_get_name_resp *)NULL)->data));
+       dinfo->count = cpu_to_le32(data_size - 1);
 
        status = aac_fib_send(ContainerCommand,
                  cmd_fibcontext,
index d31a9bc2ba69abdcf6cd6c6c8e6d99229ae50e4a..ee2667e20e4239f9129ebe21e7983a2428065b88 100644 (file)
@@ -2274,7 +2274,7 @@ struct aac_get_name_resp {
        __le32          parm3;
        __le32          parm4;
        __le32          parm5;
-       u8              data[16];
+       u8              data[17];
 };
 
 #define CT_CID_TO_32BITS_UID 165
index 2029ad225121162bddfec4ffcce7e38a940385f3..5be0086142cac6991598cd9099522212709a216c 100644 (file)
@@ -3845,8 +3845,10 @@ csio_hw_start(struct csio_hw *hw)
 
        if (csio_is_hw_ready(hw))
                return 0;
-       else
+       else if (csio_match_state(hw, csio_hws_uninit))
                return -EINVAL;
+       else
+               return -ENODEV;
 }
 
 int
index ea0c31086cc6b8b44517c8fcd34a44fb6c93fad9..dcd074169aa9b7794850c7034a1208353ba17266 100644 (file)
@@ -969,10 +969,14 @@ static int csio_probe_one(struct pci_dev *pdev, const struct pci_device_id *id)
 
        pci_set_drvdata(pdev, hw);
 
-       if (csio_hw_start(hw) != 0) {
-               dev_err(&pdev->dev,
-                       "Failed to start FW, continuing in debug mode.\n");
-               return 0;
+       rv = csio_hw_start(hw);
+       if (rv) {
+               if (rv == -EINVAL) {
+                       dev_err(&pdev->dev,
+                               "Failed to start FW, continuing in debug mode.\n");
+                       return 0;
+               }
+               goto err_lnode_exit;
        }
 
        sprintf(hw->fwrev_str, "%u.%u.%u.%u\n",
index a69a9ac836f5d13c37d5220ecbb2711d588c4c35..1d02cf9fe06c5e941e5d1a76254f86ce658fc04b 100644 (file)
@@ -1635,6 +1635,9 @@ static int init_act_open(struct cxgbi_sock *csk)
                goto rel_resource;
        }
 
+       if (!(n->nud_state & NUD_VALID))
+               neigh_event_send(n, NULL);
+
        csk->atid = cxgb4_alloc_atid(lldi->tids, csk);
        if (csk->atid < 0) {
                pr_err("%s, NO atid available.\n", ndev->name);
index b0c68d24db011ba9465dd31a41c05acd4f13c283..da5bdbdcce527262489cae939761b9cf1834eeb0 100644 (file)
@@ -3351,6 +3351,16 @@ static void ipr_worker_thread(struct work_struct *work)
                return;
        }
 
+       if (ioa_cfg->scsi_unblock) {
+               ioa_cfg->scsi_unblock = 0;
+               ioa_cfg->scsi_blocked = 0;
+               spin_unlock_irqrestore(ioa_cfg->host->host_lock, lock_flags);
+               scsi_unblock_requests(ioa_cfg->host);
+               spin_lock_irqsave(ioa_cfg->host->host_lock, lock_flags);
+               if (ioa_cfg->scsi_blocked)
+                       scsi_block_requests(ioa_cfg->host);
+       }
+
        if (!ioa_cfg->scan_enabled) {
                spin_unlock_irqrestore(ioa_cfg->host->host_lock, lock_flags);
                return;
@@ -7211,9 +7221,8 @@ static int ipr_ioa_bringdown_done(struct ipr_cmnd *ipr_cmd)
        ENTER;
        if (!ioa_cfg->hrrq[IPR_INIT_HRRQ].removing_ioa) {
                ipr_trace;
-               spin_unlock_irq(ioa_cfg->host->host_lock);
-               scsi_unblock_requests(ioa_cfg->host);
-               spin_lock_irq(ioa_cfg->host->host_lock);
+               ioa_cfg->scsi_unblock = 1;
+               schedule_work(&ioa_cfg->work_q);
        }
 
        ioa_cfg->in_reset_reload = 0;
@@ -7287,13 +7296,7 @@ static int ipr_ioa_reset_done(struct ipr_cmnd *ipr_cmd)
        list_add_tail(&ipr_cmd->queue, &ipr_cmd->hrrq->hrrq_free_q);
        wake_up_all(&ioa_cfg->reset_wait_q);
 
-       spin_unlock(ioa_cfg->host->host_lock);
-       scsi_unblock_requests(ioa_cfg->host);
-       spin_lock(ioa_cfg->host->host_lock);
-
-       if (!ioa_cfg->hrrq[IPR_INIT_HRRQ].allow_cmds)
-               scsi_block_requests(ioa_cfg->host);
-
+       ioa_cfg->scsi_unblock = 1;
        schedule_work(&ioa_cfg->work_q);
        LEAVE;
        return IPR_RC_JOB_RETURN;
@@ -9249,8 +9252,11 @@ static void _ipr_initiate_ioa_reset(struct ipr_ioa_cfg *ioa_cfg,
                spin_unlock(&ioa_cfg->hrrq[i]._lock);
        }
        wmb();
-       if (!ioa_cfg->hrrq[IPR_INIT_HRRQ].removing_ioa)
+       if (!ioa_cfg->hrrq[IPR_INIT_HRRQ].removing_ioa) {
+               ioa_cfg->scsi_unblock = 0;
+               ioa_cfg->scsi_blocked = 1;
                scsi_block_requests(ioa_cfg->host);
+       }
 
        ipr_cmd = ipr_get_free_ipr_cmnd(ioa_cfg);
        ioa_cfg->reset_cmd = ipr_cmd;
@@ -9306,9 +9312,8 @@ static void ipr_initiate_ioa_reset(struct ipr_ioa_cfg *ioa_cfg,
                        wake_up_all(&ioa_cfg->reset_wait_q);
 
                        if (!ioa_cfg->hrrq[IPR_INIT_HRRQ].removing_ioa) {
-                               spin_unlock_irq(ioa_cfg->host->host_lock);
-                               scsi_unblock_requests(ioa_cfg->host);
-                               spin_lock_irq(ioa_cfg->host->host_lock);
+                               ioa_cfg->scsi_unblock = 1;
+                               schedule_work(&ioa_cfg->work_q);
                        }
                        return;
                } else {
index e98a87a653357b54996326d6f422d2ffa7923bee..c7f0e9e3cd7d4986c1341c3b7e115e65de551723 100644 (file)
@@ -1488,6 +1488,8 @@ struct ipr_ioa_cfg {
        u8 cfg_locked:1;
        u8 clear_isr:1;
        u8 probe_done:1;
+       u8 scsi_unblock:1;
+       u8 scsi_blocked:1;
 
        u8 revid;
 
index 4ed48ed38e79316f02ca1e299e56f66eea84ba8e..7ee1a94c0b33eefd57a6889df66649477ad4713b 100644 (file)
@@ -205,8 +205,10 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
                                atomic_read(&tgtp->xmt_ls_rsp_error));
 
                len += snprintf(buf+len, PAGE_SIZE-len,
-                               "FCP: Rcv %08x Release %08x Drop %08x\n",
+                               "FCP: Rcv %08x Defer %08x Release %08x "
+                               "Drop %08x\n",
                                atomic_read(&tgtp->rcv_fcp_cmd_in),
+                               atomic_read(&tgtp->rcv_fcp_cmd_defer),
                                atomic_read(&tgtp->xmt_fcp_release),
                                atomic_read(&tgtp->rcv_fcp_cmd_drop));
 
index 5cc8b0f7d885fb0dfd5a00342f6ca72c72a40129..744f3f395b64852a294a9300adb64a496087aed7 100644 (file)
@@ -782,8 +782,11 @@ lpfc_debugfs_nvmestat_data(struct lpfc_vport *vport, char *buf, int size)
                                atomic_read(&tgtp->xmt_ls_rsp_error));
 
                len += snprintf(buf + len, size - len,
-                               "FCP: Rcv %08x Drop %08x\n",
+                               "FCP: Rcv %08x Defer %08x Release %08x "
+                               "Drop %08x\n",
                                atomic_read(&tgtp->rcv_fcp_cmd_in),
+                               atomic_read(&tgtp->rcv_fcp_cmd_defer),
+                               atomic_read(&tgtp->xmt_fcp_release),
                                atomic_read(&tgtp->rcv_fcp_cmd_drop));
 
                if (atomic_read(&tgtp->rcv_fcp_cmd_in) !=
index fbeec344c6cc3be0bdd878db6bafda7353dcf901..bbbd0f84160d36563008a212afd8252f86ef15c8 100644 (file)
@@ -841,12 +841,31 @@ lpfc_nvmet_xmt_fcp_release(struct nvmet_fc_target_port *tgtport,
        lpfc_nvmet_ctxbuf_post(phba, ctxp->ctxbuf);
 }
 
+static void
+lpfc_nvmet_defer_rcv(struct nvmet_fc_target_port *tgtport,
+                    struct nvmefc_tgt_fcp_req *rsp)
+{
+       struct lpfc_nvmet_tgtport *tgtp;
+       struct lpfc_nvmet_rcv_ctx *ctxp =
+               container_of(rsp, struct lpfc_nvmet_rcv_ctx, ctx.fcp_req);
+       struct rqb_dmabuf *nvmebuf = ctxp->rqb_buffer;
+       struct lpfc_hba *phba = ctxp->phba;
+
+       lpfc_nvmeio_data(phba, "NVMET DEFERRCV: xri x%x sz %d CPU %02x\n",
+                        ctxp->oxid, ctxp->size, smp_processor_id());
+
+       tgtp = phba->targetport->private;
+       atomic_inc(&tgtp->rcv_fcp_cmd_defer);
+       lpfc_rq_buf_free(phba, &nvmebuf->hbuf); /* repost */
+}
+
 static struct nvmet_fc_target_template lpfc_tgttemplate = {
        .targetport_delete = lpfc_nvmet_targetport_delete,
        .xmt_ls_rsp     = lpfc_nvmet_xmt_ls_rsp,
        .fcp_op         = lpfc_nvmet_xmt_fcp_op,
        .fcp_abort      = lpfc_nvmet_xmt_fcp_abort,
        .fcp_req_release = lpfc_nvmet_xmt_fcp_release,
+       .defer_rcv      = lpfc_nvmet_defer_rcv,
 
        .max_hw_queues  = 1,
        .max_sgl_segments = LPFC_NVMET_DEFAULT_SEGS,
@@ -1504,6 +1523,17 @@ lpfc_nvmet_unsol_fcp_buffer(struct lpfc_hba *phba,
                return;
        }
 
+       /* Processing of FCP command is deferred */
+       if (rc == -EOVERFLOW) {
+               lpfc_nvmeio_data(phba,
+                                "NVMET RCV BUSY: xri x%x sz %d from %06x\n",
+                                oxid, size, sid);
+               /* defer reposting rcv buffer till .defer_rcv callback */
+               ctxp->rqb_buffer = nvmebuf;
+               atomic_inc(&tgtp->rcv_fcp_cmd_out);
+               return;
+       }
+
        atomic_inc(&tgtp->rcv_fcp_cmd_drop);
        lpfc_printf_log(phba, KERN_ERR, LOG_NVME_IOERR,
                        "6159 FCP Drop IO x%x: err x%x: x%x x%x x%x\n",
index e675ef17be08a0f67dd76f9d33a2ab831f961ce8..48a76788b003cb746afa45376272af362c4446ce 100644 (file)
@@ -49,6 +49,7 @@ struct lpfc_nvmet_tgtport {
        atomic_t rcv_fcp_cmd_in;
        atomic_t rcv_fcp_cmd_out;
        atomic_t rcv_fcp_cmd_drop;
+       atomic_t rcv_fcp_cmd_defer;
        atomic_t xmt_fcp_release;
 
        /* Stats counters - lpfc_nvmet_xmt_fcp_op */
index 316c3df0c3fd80f9689499f0351082703772707f..71c4746341ea379e3881d32cf3078f135d15f999 100644 (file)
@@ -6228,8 +6228,8 @@ static int megasas_probe_one(struct pci_dev *pdev,
 fail_start_aen:
 fail_io_attach:
        megasas_mgmt_info.count--;
-       megasas_mgmt_info.instance[megasas_mgmt_info.max_index] = NULL;
        megasas_mgmt_info.max_index--;
+       megasas_mgmt_info.instance[megasas_mgmt_info.max_index] = NULL;
 
        instance->instancet->disable_intr(instance);
        megasas_destroy_irqs(instance);
index 33142610882f4d9bedce93076160180513ca5f82..b18646d6057f476e70d59137a0450e5f90d4b3fc 100644 (file)
@@ -401,9 +401,6 @@ qla27xx_fwdt_entry_t263(struct scsi_qla_host *vha,
                for (i = 0; i < vha->hw->max_req_queues; i++) {
                        struct req_que *req = vha->hw->req_q_map[i];
 
-                       if (!test_bit(i, vha->hw->req_qid_map))
-                               continue;
-
                        if (req || !buf) {
                                length = req ?
                                    req->length : REQUEST_ENTRY_CNT_24XX;
@@ -418,9 +415,6 @@ qla27xx_fwdt_entry_t263(struct scsi_qla_host *vha,
                for (i = 0; i < vha->hw->max_rsp_queues; i++) {
                        struct rsp_que *rsp = vha->hw->rsp_q_map[i];
 
-                       if (!test_bit(i, vha->hw->rsp_qid_map))
-                               continue;
-
                        if (rsp || !buf) {
                                length = rsp ?
                                    rsp->length : RESPONSE_ENTRY_CNT_MQ;
@@ -660,9 +654,6 @@ qla27xx_fwdt_entry_t274(struct scsi_qla_host *vha,
                for (i = 0; i < vha->hw->max_req_queues; i++) {
                        struct req_que *req = vha->hw->req_q_map[i];
 
-                       if (!test_bit(i, vha->hw->req_qid_map))
-                               continue;
-
                        if (req || !buf) {
                                qla27xx_insert16(i, buf, len);
                                qla27xx_insert16(1, buf, len);
@@ -675,9 +666,6 @@ qla27xx_fwdt_entry_t274(struct scsi_qla_host *vha,
                for (i = 0; i < vha->hw->max_rsp_queues; i++) {
                        struct rsp_que *rsp = vha->hw->rsp_q_map[i];
 
-                       if (!test_bit(i, vha->hw->rsp_qid_map))
-                               continue;
-
                        if (rsp || !buf) {
                                qla27xx_insert16(i, buf, len);
                                qla27xx_insert16(1, buf, len);
index b20da0d27ad78494bc9a0172d428dc176cbcb852..3f82ea1b72dc8739283992b4f425d77692b51e7b 100644 (file)
@@ -500,7 +500,6 @@ static int tcm_qla2xxx_handle_cmd(scsi_qla_host_t *vha, struct qla_tgt_cmd *cmd,
 static void tcm_qla2xxx_handle_data_work(struct work_struct *work)
 {
        struct qla_tgt_cmd *cmd = container_of(work, struct qla_tgt_cmd, work);
-       unsigned long flags;
 
        /*
         * Ensure that the complete FCP WRITE payload has been received.
@@ -508,17 +507,6 @@ static void tcm_qla2xxx_handle_data_work(struct work_struct *work)
         */
        cmd->cmd_in_wq = 0;
 
-       spin_lock_irqsave(&cmd->cmd_lock, flags);
-       cmd->data_work = 1;
-       if (cmd->aborted) {
-               cmd->data_work_free = 1;
-               spin_unlock_irqrestore(&cmd->cmd_lock, flags);
-
-               tcm_qla2xxx_free_cmd(cmd);
-               return;
-       }
-       spin_unlock_irqrestore(&cmd->cmd_lock, flags);
-
        cmd->qpair->tgt_counters.qla_core_ret_ctio++;
        if (!cmd->write_data_transferred) {
                /*
@@ -765,31 +753,13 @@ static void tcm_qla2xxx_queue_tm_rsp(struct se_cmd *se_cmd)
        qlt_xmit_tm_rsp(mcmd);
 }
 
-#define DATA_WORK_NOT_FREE(_cmd) (_cmd->data_work && !_cmd->data_work_free)
 static void tcm_qla2xxx_aborted_task(struct se_cmd *se_cmd)
 {
        struct qla_tgt_cmd *cmd = container_of(se_cmd,
                                struct qla_tgt_cmd, se_cmd);
-       unsigned long flags;
 
        if (qlt_abort_cmd(cmd))
                return;
-
-       spin_lock_irqsave(&cmd->cmd_lock, flags);
-       if ((cmd->state == QLA_TGT_STATE_NEW)||
-           ((cmd->state == QLA_TGT_STATE_DATA_IN) &&
-               DATA_WORK_NOT_FREE(cmd))) {
-               cmd->data_work_free = 1;
-               spin_unlock_irqrestore(&cmd->cmd_lock, flags);
-               /*
-                * cmd has not reached fw, Use this trigger to free it.
-                */
-               tcm_qla2xxx_free_cmd(cmd);
-               return;
-       }
-       spin_unlock_irqrestore(&cmd->cmd_lock, flags);
-       return;
-
 }
 
 static void tcm_qla2xxx_clear_sess_lookup(struct tcm_qla2xxx_lport *,
index 3d38c6d463b810f19fe2cbdd3dc69801465cd05a..1bf274e3b2b614b5331f4a082668e660ceba74e7 100644 (file)
@@ -800,7 +800,11 @@ MODULE_LICENSE("GPL");
 module_param(scsi_logging_level, int, S_IRUGO|S_IWUSR);
 MODULE_PARM_DESC(scsi_logging_level, "a bit mask of logging levels");
 
+#ifdef CONFIG_SCSI_MQ_DEFAULT
 bool scsi_use_blk_mq = true;
+#else
+bool scsi_use_blk_mq = false;
+#endif
 module_param_named(use_blk_mq, scsi_use_blk_mq, bool, S_IWUSR | S_IRUGO);
 
 static int __init init_scsi(void)
index bea36adeee178ab57bced3591b0969bc2d973264..e2647f2d44304b0ded65ebb0f55d629c7289d0b7 100644 (file)
@@ -1277,6 +1277,9 @@ static void sd_uninit_command(struct scsi_cmnd *SCpnt)
 {
        struct request *rq = SCpnt->request;
 
+       if (SCpnt->flags & SCMD_ZONE_WRITE_LOCK)
+               sd_zbc_write_unlock_zone(SCpnt);
+
        if (rq->rq_flags & RQF_SPECIAL_PAYLOAD)
                __free_page(rq->special_vec.bv_page);
 
index 96855df9f49ddf3756b8fc6ee98b127727d9bbeb..8aa54779aac1b4e6112d86dc536dd14b1522d6eb 100644 (file)
@@ -294,6 +294,9 @@ int sd_zbc_write_lock_zone(struct scsi_cmnd *cmd)
            test_and_set_bit(zno, sdkp->zones_wlock))
                return BLKPREP_DEFER;
 
+       WARN_ON_ONCE(cmd->flags & SCMD_ZONE_WRITE_LOCK);
+       cmd->flags |= SCMD_ZONE_WRITE_LOCK;
+
        return BLKPREP_OK;
 }
 
@@ -302,9 +305,10 @@ void sd_zbc_write_unlock_zone(struct scsi_cmnd *cmd)
        struct request *rq = cmd->request;
        struct scsi_disk *sdkp = scsi_disk(rq->rq_disk);
 
-       if (sdkp->zones_wlock) {
+       if (sdkp->zones_wlock && cmd->flags & SCMD_ZONE_WRITE_LOCK) {
                unsigned int zno = sd_zbc_zone_no(sdkp, blk_rq_pos(rq));
                WARN_ON_ONCE(!test_bit(zno, sdkp->zones_wlock));
+               cmd->flags &= ~SCMD_ZONE_WRITE_LOCK;
                clear_bit_unlock(zno, sdkp->zones_wlock);
                smp_mb__after_atomic();
        }
@@ -335,9 +339,6 @@ void sd_zbc_complete(struct scsi_cmnd *cmd,
        case REQ_OP_WRITE_ZEROES:
        case REQ_OP_WRITE_SAME:
 
-               /* Unlock the zone */
-               sd_zbc_write_unlock_zone(cmd);
-
                if (result &&
                    sshdr->sense_key == ILLEGAL_REQUEST &&
                    sshdr->asc == 0x21)
index f1cdf32d7514120b31515ddf2901960d4f766694..8927f9f54ad926f5b50439e23685a35b24c50edf 100644 (file)
@@ -99,7 +99,7 @@ static int ses_recv_diag(struct scsi_device *sdev, int page_code,
 
        ret =  scsi_execute_req(sdev, cmd, DMA_FROM_DEVICE, buf, bufflen,
                                NULL, SES_TIMEOUT, SES_RETRIES, NULL);
-       if (unlikely(!ret))
+       if (unlikely(ret))
                return ret;
 
        recv_page_code = ((unsigned char *)buf)[0];
index 8e5013d9cad445178a0461edd8eaf9abe4e84c38..94e402ed30f6ae54ecb32160480d554e6e118827 100644 (file)
@@ -4299,11 +4299,11 @@ static int st_probe(struct device *dev)
        kref_init(&tpnt->kref);
        tpnt->disk = disk;
        disk->private_data = &tpnt->driver;
-       disk->queue = SDp->request_queue;
        /* SCSI tape doesn't register this gendisk via add_disk().  Manually
         * take queue reference that release_disk() expects. */
-       if (!blk_get_queue(disk->queue))
+       if (!blk_get_queue(SDp->request_queue))
                goto out_put_disk;
+       disk->queue = SDp->request_queue;
        tpnt->driver = &st_template;
 
        tpnt->device = SDp;
index 3039072911a5bce09d375c04b9faf65071894c11..afc7ecc3c1876158d33e45933458fc035593c39f 100644 (file)
@@ -200,16 +200,11 @@ static int imx7_pgc_domain_probe(struct platform_device *pdev)
 
        domain->dev = &pdev->dev;
 
-       ret = pm_genpd_init(&domain->genpd, NULL, true);
-       if (ret) {
-               dev_err(domain->dev, "Failed to init power domain\n");
-               return ret;
-       }
-
        domain->regulator = devm_regulator_get_optional(domain->dev, "power");
        if (IS_ERR(domain->regulator)) {
                if (PTR_ERR(domain->regulator) != -ENODEV) {
-                       dev_err(domain->dev, "Failed to get domain's regulator\n");
+                       if (PTR_ERR(domain->regulator) != -EPROBE_DEFER)
+                               dev_err(domain->dev, "Failed to get domain's regulator\n");
                        return PTR_ERR(domain->regulator);
                }
        } else {
@@ -217,6 +212,12 @@ static int imx7_pgc_domain_probe(struct platform_device *pdev)
                                      domain->voltage, domain->voltage);
        }
 
+       ret = pm_genpd_init(&domain->genpd, NULL, true);
+       if (ret) {
+               dev_err(domain->dev, "Failed to init power domain\n");
+               return ret;
+       }
+
        ret = of_genpd_add_provider_simple(domain->dev->of_node,
                                           &domain->genpd);
        if (ret) {
index 279e7c5551dd55588fa95fee7547ea8a945ab58d..39225de9d7f1455e43d2ac6279193a68e5518562 100644 (file)
@@ -745,6 +745,9 @@ void *knav_pool_create(const char *name,
        bool slot_found;
        int ret;
 
+       if (!kdev)
+               return ERR_PTR(-EPROBE_DEFER);
+
        if (!kdev->dev)
                return ERR_PTR(-ENODEV);
 
index b0b283810e72245cb24d0f5ab1efc468d55e52c0..de31b9389e2ee710ed48677bcddbe64e88a26284 100644 (file)
@@ -176,6 +176,8 @@ static int ti_sci_pm_domain_probe(struct platform_device *pdev)
 
        ti_sci_pd->dev = dev;
 
+       ti_sci_pd->pd.name = "ti_sci_pd";
+
        ti_sci_pd->pd.attach_dev = ti_sci_pd_attach_dev;
        ti_sci_pd->pd.detach_dev = ti_sci_pd_detach_dev;
 
index ca11be21f64b606091d5e8e0e62044ae263dc3ab..34ca7823255d692d05aa753e7c5bc28a385c5583 100644 (file)
@@ -2396,6 +2396,7 @@ static ssize_t comedi_write(struct file *file, const char __user *buf,
                        continue;
                }
 
+               set_current_state(TASK_RUNNING);
                wp = async->buf_write_ptr;
                n1 = min(n, async->prealloc_bufsz - wp);
                n2 = n - n1;
@@ -2528,6 +2529,8 @@ static ssize_t comedi_read(struct file *file, char __user *buf, size_t nbytes,
                        }
                        continue;
                }
+
+               set_current_state(TASK_RUNNING);
                rp = async->buf_read_ptr;
                n1 = min(n, async->prealloc_bufsz - rp);
                n2 = n - n1;
index a6a8393d66645e75c13ecf8f67ca9c71ef48a6e5..3e00df74b18c883d4e19f0bac7ec087988b58821 100644 (file)
@@ -472,7 +472,7 @@ static int ad2s1210_read_raw(struct iio_dev *indio_dev,
                             long m)
 {
        struct ad2s1210_state *st = iio_priv(indio_dev);
-       bool negative;
+       u16 negative;
        int ret = 0;
        u16 pos;
        s16 vel;
index e583dd8a418b537eda69a9606fa8a3fc2c6a6207..d4fa41be80f9a1719574af28c8981ef8e8d287ca 100644 (file)
@@ -1510,11 +1510,13 @@ cxgbit_pass_open_rpl(struct cxgbit_device *cdev, struct sk_buff *skb)
 
        if (!cnp) {
                pr_info("%s stid %d lookup failure\n", __func__, stid);
-               return;
+               goto rel_skb;
        }
 
        cxgbit_wake_up(&cnp->com.wr_wait, __func__, rpl->status);
        cxgbit_put_cnp(cnp);
+rel_skb:
+       __kfree_skb(skb);
 }
 
 static void
@@ -1530,11 +1532,13 @@ cxgbit_close_listsrv_rpl(struct cxgbit_device *cdev, struct sk_buff *skb)
 
        if (!cnp) {
                pr_info("%s stid %d lookup failure\n", __func__, stid);
-               return;
+               goto rel_skb;
        }
 
        cxgbit_wake_up(&cnp->com.wr_wait, __func__, rpl->status);
        cxgbit_put_cnp(cnp);
+rel_skb:
+       __kfree_skb(skb);
 }
 
 static void
@@ -1819,12 +1823,16 @@ static void cxgbit_set_tcb_rpl(struct cxgbit_device *cdev, struct sk_buff *skb)
        struct tid_info *t = lldi->tids;
 
        csk = lookup_tid(t, tid);
-       if (unlikely(!csk))
+       if (unlikely(!csk)) {
                pr_err("can't find connection for tid %u.\n", tid);
-       else
+               goto rel_skb;
+       } else {
                cxgbit_wake_up(&csk->com.wr_wait, __func__, rpl->status);
+       }
 
        cxgbit_put_csk(csk);
+rel_skb:
+       __kfree_skb(skb);
 }
 
 static void cxgbit_rx_data(struct cxgbit_device *cdev, struct sk_buff *skb)
index dda13f1af38e581f746e4528590510c07efb916f..514986b57c2d60ce19c1074f4d19d65dd550be2e 100644 (file)
@@ -827,7 +827,7 @@ cxgbit_put_login_tx(struct iscsi_conn *conn, struct iscsi_login *login,
 
 static void
 cxgbit_skb_copy_to_sg(struct sk_buff *skb, struct scatterlist *sg,
-                     unsigned int nents)
+                     unsigned int nents, u32 skip)
 {
        struct skb_seq_state st;
        const u8 *buf;
@@ -846,7 +846,7 @@ cxgbit_skb_copy_to_sg(struct sk_buff *skb, struct scatterlist *sg,
                }
 
                consumed += sg_pcopy_from_buffer(sg, nents, (void *)buf,
-                                                buf_len, consumed);
+                                                buf_len, skip + consumed);
        }
 }
 
@@ -912,7 +912,7 @@ cxgbit_handle_immediate_data(struct iscsi_cmd *cmd, struct iscsi_scsi_req *hdr,
                struct scatterlist *sg = &cmd->se_cmd.t_data_sg[0];
                u32 sg_nents = max(1UL, DIV_ROUND_UP(pdu_cb->dlen, PAGE_SIZE));
 
-               cxgbit_skb_copy_to_sg(csk->skb, sg, sg_nents);
+               cxgbit_skb_copy_to_sg(csk->skb, sg, sg_nents, 0);
        }
 
        cmd->write_data_done += pdu_cb->dlen;
@@ -1069,11 +1069,13 @@ static int cxgbit_handle_iscsi_dataout(struct cxgbit_sock *csk)
                  cmd->se_cmd.data_length);
 
        if (!(pdu_cb->flags & PDUCBF_RX_DATA_DDPD)) {
+               u32 skip = data_offset % PAGE_SIZE;
+
                sg_off = data_offset / PAGE_SIZE;
                sg_start = &cmd->se_cmd.t_data_sg[sg_off];
-               sg_nents = max(1UL, DIV_ROUND_UP(data_len, PAGE_SIZE));
+               sg_nents = max(1UL, DIV_ROUND_UP(skip + data_len, PAGE_SIZE));
 
-               cxgbit_skb_copy_to_sg(csk->skb, sg_start, sg_nents);
+               cxgbit_skb_copy_to_sg(csk->skb, sg_start, sg_nents, skip);
        }
 
 check_payload:
index 74e4975dd1b1e74d6c39517be102ccdc92614e80..5001261f5d69d759dd25161401510551d35c3da7 100644 (file)
@@ -418,6 +418,7 @@ int iscsit_reset_np_thread(
                return 0;
        }
        np->np_thread_state = ISCSI_NP_THREAD_RESET;
+       atomic_inc(&np->np_reset_count);
 
        if (np->np_thread) {
                spin_unlock_bh(&np->np_thread_lock);
@@ -2167,6 +2168,7 @@ iscsit_setup_text_cmd(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
        cmd->cmd_sn             = be32_to_cpu(hdr->cmdsn);
        cmd->exp_stat_sn        = be32_to_cpu(hdr->exp_statsn);
        cmd->data_direction     = DMA_NONE;
+       kfree(cmd->text_in_ptr);
        cmd->text_in_ptr        = NULL;
 
        return 0;
@@ -3487,9 +3489,9 @@ iscsit_build_text_rsp(struct iscsi_cmd *cmd, struct iscsi_conn *conn,
                return text_length;
 
        if (completed) {
-               hdr->flags |= ISCSI_FLAG_CMD_FINAL;
+               hdr->flags = ISCSI_FLAG_CMD_FINAL;
        } else {
-               hdr->flags |= ISCSI_FLAG_TEXT_CONTINUE;
+               hdr->flags = ISCSI_FLAG_TEXT_CONTINUE;
                cmd->read_data_done += text_length;
                if (cmd->targ_xfer_tag == 0xFFFFFFFF)
                        cmd->targ_xfer_tag = session_get_next_ttt(conn->sess);
index e9bdc8b86e7d1d71d77cf4388370af6fcf3fded7..dc13afbd4c88dec2390ca8e65b3332d2f78acd73 100644 (file)
@@ -1243,9 +1243,11 @@ static int __iscsi_target_login_thread(struct iscsi_np *np)
        flush_signals(current);
 
        spin_lock_bh(&np->np_thread_lock);
-       if (np->np_thread_state == ISCSI_NP_THREAD_RESET) {
+       if (atomic_dec_if_positive(&np->np_reset_count) >= 0) {
                np->np_thread_state = ISCSI_NP_THREAD_ACTIVE;
+               spin_unlock_bh(&np->np_thread_lock);
                complete(&np->np_restart_comp);
+               return 1;
        } else if (np->np_thread_state == ISCSI_NP_THREAD_SHUTDOWN) {
                spin_unlock_bh(&np->np_thread_lock);
                goto exit;
@@ -1278,7 +1280,8 @@ static int __iscsi_target_login_thread(struct iscsi_np *np)
                goto exit;
        } else if (rc < 0) {
                spin_lock_bh(&np->np_thread_lock);
-               if (np->np_thread_state == ISCSI_NP_THREAD_RESET) {
+               if (atomic_dec_if_positive(&np->np_reset_count) >= 0) {
+                       np->np_thread_state = ISCSI_NP_THREAD_ACTIVE;
                        spin_unlock_bh(&np->np_thread_lock);
                        complete(&np->np_restart_comp);
                        iscsit_put_transport(conn->conn_transport);
index 36913734c6bc58ed326ac793dc19d3d265ec878d..02e8a5d8665837f415ac7b2d78e51067d1ef5b62 100644 (file)
@@ -364,7 +364,7 @@ void core_tpg_del_initiator_node_acl(struct se_node_acl *acl)
        mutex_lock(&tpg->acl_node_mutex);
        if (acl->dynamic_node_acl)
                acl->dynamic_node_acl = 0;
-       list_del(&acl->acl_list);
+       list_del_init(&acl->acl_list);
        mutex_unlock(&tpg->acl_node_mutex);
 
        target_shutdown_sessions(acl);
@@ -548,7 +548,7 @@ int core_tpg_deregister(struct se_portal_group *se_tpg)
         * in transport_deregister_session().
         */
        list_for_each_entry_safe(nacl, nacl_tmp, &node_list, acl_list) {
-               list_del(&nacl->acl_list);
+               list_del_init(&nacl->acl_list);
 
                core_tpg_wait_for_nacl_pr_ref(nacl);
                core_free_device_list_for_node(nacl, se_tpg);
index 97fed9a298bdc29a19d184ba49df458e04e4f9f1..836d552b0385e978bc1a0b98c59a3379c262fd61 100644 (file)
@@ -466,7 +466,7 @@ static void target_complete_nacl(struct kref *kref)
        }
 
        mutex_lock(&se_tpg->acl_node_mutex);
-       list_del(&nacl->acl_list);
+       list_del_init(&nacl->acl_list);
        mutex_unlock(&se_tpg->acl_node_mutex);
 
        core_tpg_wait_for_nacl_pr_ref(nacl);
@@ -538,7 +538,7 @@ void transport_free_session(struct se_session *se_sess)
                        spin_unlock_irqrestore(&se_nacl->nacl_sess_lock, flags);
 
                        if (se_nacl->dynamic_stop)
-                               list_del(&se_nacl->acl_list);
+                               list_del_init(&se_nacl->acl_list);
                }
                mutex_unlock(&se_tpg->acl_node_mutex);
 
index 80ee130f8253ec44324ac2ba616cfd93beb1ba76..942d094269fba5db66ff7e791dcfaab1c6acec15 100644 (file)
@@ -563,8 +563,6 @@ static int scatter_data_area(struct tcmu_dev *udev,
                                        block_remaining);
                        to_offset = get_block_offset_user(udev, dbi,
                                        block_remaining);
-                       offset = DATA_BLOCK_SIZE - block_remaining;
-                       to += offset;
 
                        if (*iov_cnt != 0 &&
                            to_offset == iov_tail(*iov)) {
@@ -575,8 +573,10 @@ static int scatter_data_area(struct tcmu_dev *udev,
                                (*iov)->iov_len = copy_bytes;
                        }
                        if (copy_data) {
-                               memcpy(to, from + sg->length - sg_remaining,
-                                       copy_bytes);
+                               offset = DATA_BLOCK_SIZE - block_remaining;
+                               memcpy(to + offset,
+                                      from + sg->length - sg_remaining,
+                                      copy_bytes);
                                tcmu_flush_dcache_range(to, copy_bytes);
                        }
                        sg_remaining -= copy_bytes;
@@ -637,9 +637,8 @@ static void gather_data_area(struct tcmu_dev *udev, struct tcmu_cmd *cmd,
                        copy_bytes = min_t(size_t, sg_remaining,
                                        block_remaining);
                        offset = DATA_BLOCK_SIZE - block_remaining;
-                       from += offset;
                        tcmu_flush_dcache_range(from, copy_bytes);
-                       memcpy(to + sg->length - sg_remaining, from,
+                       memcpy(to + sg->length - sg_remaining, from + offset,
                                        copy_bytes);
 
                        sg_remaining -= copy_bytes;
@@ -1433,6 +1432,8 @@ static int tcmu_update_uio_info(struct tcmu_dev *udev)
        if (udev->dev_config[0])
                snprintf(str + used, size - used, "/%s", udev->dev_config);
 
+       /* If the old string exists, free it */
+       kfree(info->name);
        info->name = str;
 
        return 0;
index 308b6e17c88aace0775b4ed0a5460097559d66ca..fe2f00ceafc5d7e1a3bbafea4224c838dc2158a4 100644 (file)
@@ -333,6 +333,15 @@ static int tb_drom_parse_entry_port(struct tb_switch *sw,
        int res;
        enum tb_port_type type;
 
+       /*
+        * Some DROMs list more ports than the controller actually has
+        * so we skip those but allow the parser to continue.
+        */
+       if (header->index > sw->config.max_port_number) {
+               dev_info_once(&sw->dev, "ignoring unnecessary extra entries in DROM\n");
+               return 0;
+       }
+
        port = &sw->ports[header->index];
        port->disabled = header->port_disabled;
        if (port->disabled)
index 284749fb0f6b96d3d6d667332e569ba0d745e048..a6d5164c33a9ca7e3d02bbda7fc94ff783d75929 100644 (file)
@@ -69,13 +69,8 @@ static void pty_close(struct tty_struct *tty, struct file *filp)
 #ifdef CONFIG_UNIX98_PTYS
                if (tty->driver == ptm_driver) {
                        mutex_lock(&devpts_mutex);
-                       if (tty->link->driver_data) {
-                               struct path *path = tty->link->driver_data;
-
-                               devpts_pty_kill(path->dentry);
-                               path_put(path);
-                               kfree(path);
-                       }
+                       if (tty->link->driver_data)
+                               devpts_pty_kill(tty->link->driver_data);
                        mutex_unlock(&devpts_mutex);
                }
 #endif
@@ -607,25 +602,24 @@ static inline void legacy_pty_init(void) { }
 static struct cdev ptmx_cdev;
 
 /**
- *     pty_open_peer - open the peer of a pty
- *     @tty: the peer of the pty being opened
+ *     ptm_open_peer - open the peer of a pty
+ *     @master: the open struct file of the ptmx device node
+ *     @tty: the master of the pty being opened
+ *     @flags: the flags for open
  *
- *     Open the cached dentry in tty->link, providing a safe way for userspace
- *     to get the slave end of a pty (where they have the master fd and cannot
- *     access or trust the mount namespace /dev/pts was mounted inside).
+ *     Provide a race free way for userspace to open the slave end of a pty
+ *     (where they have the master fd and cannot access or trust the mount
+ *     namespace /dev/pts was mounted inside).
  */
-static struct file *pty_open_peer(struct tty_struct *tty, int flags)
-{
-       if (tty->driver->subtype != PTY_TYPE_MASTER)
-               return ERR_PTR(-EIO);
-       return dentry_open(tty->link->driver_data, flags, current_cred());
-}
-
-static int pty_get_peer(struct tty_struct *tty, int flags)
+int ptm_open_peer(struct file *master, struct tty_struct *tty, int flags)
 {
        int fd = -1;
-       struct file *filp = NULL;
+       struct file *filp;
        int retval = -EINVAL;
+       struct path path;
+
+       if (tty->driver != ptm_driver)
+               return -EIO;
 
        fd = get_unused_fd_flags(0);
        if (fd < 0) {
@@ -633,7 +627,16 @@ static int pty_get_peer(struct tty_struct *tty, int flags)
                goto err;
        }
 
-       filp = pty_open_peer(tty, flags);
+       /* Compute the slave's path */
+       path.mnt = devpts_mntget(master, tty->driver_data);
+       if (IS_ERR(path.mnt)) {
+               retval = PTR_ERR(path.mnt);
+               goto err_put;
+       }
+       path.dentry = tty->link->driver_data;
+
+       filp = dentry_open(&path, flags, current_cred());
+       mntput(path.mnt);
        if (IS_ERR(filp)) {
                retval = PTR_ERR(filp);
                goto err_put;
@@ -662,8 +665,6 @@ static int pty_unix98_ioctl(struct tty_struct *tty,
                return pty_get_pktmode(tty, (int __user *)arg);
        case TIOCGPTN: /* Get PT Number */
                return put_user(tty->index, (unsigned int __user *)arg);
-       case TIOCGPTPEER: /* Open the other end */
-               return pty_get_peer(tty, (int) arg);
        case TIOCSIG:    /* Send signal to other side of pty */
                return pty_signal(tty, (int) arg);
        }
@@ -791,7 +792,6 @@ static int ptmx_open(struct inode *inode, struct file *filp)
 {
        struct pts_fs_info *fsi;
        struct tty_struct *tty;
-       struct path *pts_path;
        struct dentry *dentry;
        int retval;
        int index;
@@ -845,26 +845,16 @@ static int ptmx_open(struct inode *inode, struct file *filp)
                retval = PTR_ERR(dentry);
                goto err_release;
        }
-       /* We need to cache a fake path for TIOCGPTPEER. */
-       pts_path = kmalloc(sizeof(struct path), GFP_KERNEL);
-       if (!pts_path)
-               goto err_release;
-       pts_path->mnt = filp->f_path.mnt;
-       pts_path->dentry = dentry;
-       path_get(pts_path);
-       tty->link->driver_data = pts_path;
+       tty->link->driver_data = dentry;
 
        retval = ptm_driver->ops->open(tty, filp);
        if (retval)
-               goto err_path_put;
+               goto err_release;
 
        tty_debug_hangup(tty, "opening (count=%d)\n", tty->count);
 
        tty_unlock(tty);
        return 0;
-err_path_put:
-       path_put(pts_path);
-       kfree(pts_path);
 err_release:
        tty_unlock(tty);
        // This will also put-ref the fsi
index b5def356af63b70e3ebc2e23a48da0753e41b47e..1aab3010fbfae76e2c25cb60085f21051a1f24cf 100644 (file)
@@ -1043,13 +1043,24 @@ int serial8250_register_8250_port(struct uart_8250_port *up)
                if (up->dl_write)
                        uart->dl_write = up->dl_write;
 
-               if (serial8250_isa_config != NULL)
-                       serial8250_isa_config(0, &uart->port,
-                                       &uart->capabilities);
+               if (uart->port.type != PORT_8250_CIR) {
+                       if (serial8250_isa_config != NULL)
+                               serial8250_isa_config(0, &uart->port,
+                                               &uart->capabilities);
+
+                       ret = uart_add_one_port(&serial8250_reg,
+                                               &uart->port);
+                       if (ret == 0)
+                               ret = uart->port.line;
+               } else {
+                       dev_info(uart->port.dev,
+                               "skipping CIR port at 0x%lx / 0x%llx, IRQ %d\n",
+                               uart->port.iobase,
+                               (unsigned long long)uart->port.mapbase,
+                               uart->port.irq);
 
-               ret = uart_add_one_port(&serial8250_reg, &uart->port);
-               if (ret == 0)
-                       ret = uart->port.line;
+                       ret = 0;
+               }
        }
        mutex_unlock(&serial_mutex);
 
index 8a857bb34fbb26c6d60784d3fe7576730a9aa5b3..1888d168a41c87c605962da2605df8ab1c02bd20 100644 (file)
@@ -142,15 +142,7 @@ static struct vendor_data vendor_sbsa = {
        .fixed_options          = true,
 };
 
-/*
- * Erratum 44 for QDF2432v1 and QDF2400v1 SoCs describes the BUSY bit as
- * occasionally getting stuck as 1. To avoid the potential for a hang, check
- * TXFE == 0 instead of BUSY == 1. This may not be suitable for all UART
- * implementations, so only do so if an affected platform is detected in
- * parse_spcr().
- */
-static bool qdf2400_e44_present = false;
-
+#ifdef CONFIG_ACPI_SPCR_TABLE
 static struct vendor_data vendor_qdt_qdf2400_e44 = {
        .reg_offset             = pl011_std_offsets,
        .fr_busy                = UART011_FR_TXFE,
@@ -165,6 +157,7 @@ static struct vendor_data vendor_qdt_qdf2400_e44 = {
        .always_enabled         = true,
        .fixed_options          = true,
 };
+#endif
 
 static u16 pl011_st_offsets[REG_ARRAY_SIZE] = {
        [REG_DR] = UART01x_DR,
@@ -2375,12 +2368,14 @@ static int __init pl011_console_match(struct console *co, char *name, int idx,
        resource_size_t addr;
        int i;
 
-       if (strcmp(name, "qdf2400_e44") == 0) {
-               pr_info_once("UART: Working around QDF2400 SoC erratum 44");
-               qdf2400_e44_present = true;
-       } else if (strcmp(name, "pl011") != 0) {
+       /*
+        * Systems affected by the Qualcomm Technologies QDF2400 E44 erratum
+        * have a distinct console name, so make sure we check for that.
+        * The actual implementation of the erratum occurs in the probe
+        * function.
+        */
+       if ((strcmp(name, "qdf2400_e44") != 0) && (strcmp(name, "pl011") != 0))
                return -ENODEV;
-       }
 
        if (uart_parse_earlycon(options, &iotype, &addr, &options))
                return -ENODEV;
@@ -2734,11 +2729,17 @@ static int sbsa_uart_probe(struct platform_device *pdev)
        }
        uap->port.irq   = ret;
 
-       uap->reg_offset = vendor_sbsa.reg_offset;
-       uap->vendor     = qdf2400_e44_present ?
-                                       &vendor_qdt_qdf2400_e44 : &vendor_sbsa;
+#ifdef CONFIG_ACPI_SPCR_TABLE
+       if (qdf2400_e44_present) {
+               dev_info(&pdev->dev, "working around QDF2400 SoC erratum 44\n");
+               uap->vendor = &vendor_qdt_qdf2400_e44;
+       } else
+#endif
+               uap->vendor = &vendor_sbsa;
+
+       uap->reg_offset = uap->vendor->reg_offset;
        uap->fifosize   = 32;
-       uap->port.iotype = vendor_sbsa.access_32b ? UPIO_MEM32 : UPIO_MEM;
+       uap->port.iotype = uap->vendor->access_32b ? UPIO_MEM32 : UPIO_MEM;
        uap->port.ops   = &sbsa_uart_pops;
        uap->fixed_baud = baudrate;
 
index 974b13d244010de234d0350b9c45425251ff99c6..10c4038c0e8dfe9d07b7fd5f02732de03cc4ed3e 100644 (file)
@@ -2518,6 +2518,9 @@ long tty_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
        case TIOCSSERIAL:
                tty_warn_deprecated_flags(p);
                break;
+       case TIOCGPTPEER:
+               /* Special because the struct file is needed */
+               return ptm_open_peer(file, tty, (int)arg);
        default:
                retval = tty_jobctrl_ioctl(tty, real_tty, file, cmd, arg);
                if (retval != -ENOIOCTLCMD)
index ab1bb3b538ac6175dd1fbe5babb686a6a8f4bc49..7f277b092b5bf070c21d9c321d15b2e0d630792e 100644 (file)
@@ -1888,7 +1888,7 @@ void usb_hcd_flush_endpoint(struct usb_device *udev,
        /* No more submits can occur */
        spin_lock_irq(&hcd_urb_list_lock);
 rescan:
-       list_for_each_entry (urb, &ep->urb_list, urb_list) {
+       list_for_each_entry_reverse(urb, &ep->urb_list, urb_list) {
                int     is_in;
 
                if (urb->unlinked)
@@ -2485,6 +2485,8 @@ void usb_hc_died (struct usb_hcd *hcd)
        }
        if (usb_hcd_is_primary_hcd(hcd) && hcd->shared_hcd) {
                hcd = hcd->shared_hcd;
+               clear_bit(HCD_FLAG_RH_RUNNING, &hcd->flags);
+               set_bit(HCD_FLAG_DEAD, &hcd->flags);
                if (hcd->rh_registered) {
                        clear_bit(HCD_FLAG_POLL_RH, &hcd->flags);
 
index 6e6797d145dd80136c413129641bdbbae2f1e63b..822f8c50e4233c70d159a4e374ad66b49502c0c1 100644 (file)
@@ -4725,7 +4725,8 @@ hub_power_remaining(struct usb_hub *hub)
 static void hub_port_connect(struct usb_hub *hub, int port1, u16 portstatus,
                u16 portchange)
 {
-       int status, i;
+       int status = -ENODEV;
+       int i;
        unsigned unit_load;
        struct usb_device *hdev = hub->hdev;
        struct usb_hcd *hcd = bus_to_hcd(hdev->bus);
@@ -4929,9 +4930,10 @@ loop:
 
 done:
        hub_port_disable(hub, port1, 1);
-       if (hcd->driver->relinquish_port && !hub->hdev->parent)
-               hcd->driver->relinquish_port(hcd, port1);
-
+       if (hcd->driver->relinquish_port && !hub->hdev->parent) {
+               if (status != -ENOTCONN && status != -ENODEV)
+                       hcd->driver->relinquish_port(hcd, port1);
+       }
 }
 
 /* Handle physical or logical connection change events.
index 3116edfcdc18558aa768d248f1ff1881448ced09..574da2b4529cc26cc0a97ee7146e42889ca55bbe 100644 (file)
@@ -150,6 +150,9 @@ static const struct usb_device_id usb_quirk_list[] = {
        /* appletouch */
        { USB_DEVICE(0x05ac, 0x021a), .driver_info = USB_QUIRK_RESET_RESUME },
 
+       /* Genesys Logic hub, internally used by Moshi USB to Ethernet Adapter */
+       { USB_DEVICE(0x05e3, 0x0616), .driver_info = USB_QUIRK_NO_LPM },
+
        /* Avision AV600U */
        { USB_DEVICE(0x0638, 0x0a13), .driver_info =
          USB_QUIRK_STRING_FETCH_255 },
@@ -249,6 +252,7 @@ static const struct usb_device_id usb_amd_resume_quirk_list[] = {
        { USB_DEVICE(0x093a, 0x2500), .driver_info = USB_QUIRK_RESET_RESUME },
        { USB_DEVICE(0x093a, 0x2510), .driver_info = USB_QUIRK_RESET_RESUME },
        { USB_DEVICE(0x093a, 0x2521), .driver_info = USB_QUIRK_RESET_RESUME },
+       { USB_DEVICE(0x03f0, 0x2b4a), .driver_info = USB_QUIRK_RESET_RESUME },
 
        /* Logitech Optical Mouse M90/M100 */
        { USB_DEVICE(0x046d, 0xc05a), .driver_info = USB_QUIRK_RESET_RESUME },
index 6b299c7b765611e0cb6c0d30e5de50f8ef279cb7..f064f1549333dcd7dab10fd491e43cf31bac887e 100644 (file)
@@ -896,9 +896,40 @@ static void __dwc3_prepare_one_trb(struct dwc3_ep *dep, struct dwc3_trb *trb,
                if (!node) {
                        trb->ctrl = DWC3_TRBCTL_ISOCHRONOUS_FIRST;
 
+                       /*
+                        * USB Specification 2.0 Section 5.9.2 states that: "If
+                        * there is only a single transaction in the microframe,
+                        * only a DATA0 data packet PID is used.  If there are
+                        * two transactions per microframe, DATA1 is used for
+                        * the first transaction data packet and DATA0 is used
+                        * for the second transaction data packet.  If there are
+                        * three transactions per microframe, DATA2 is used for
+                        * the first transaction data packet, DATA1 is used for
+                        * the second, and DATA0 is used for the third."
+                        *
+                        * IOW, we should satisfy the following cases:
+                        *
+                        * 1) length <= maxpacket
+                        *      - DATA0
+                        *
+                        * 2) maxpacket < length <= (2 * maxpacket)
+                        *      - DATA1, DATA0
+                        *
+                        * 3) (2 * maxpacket) < length <= (3 * maxpacket)
+                        *      - DATA2, DATA1, DATA0
+                        */
                        if (speed == USB_SPEED_HIGH) {
                                struct usb_ep *ep = &dep->endpoint;
-                               trb->size |= DWC3_TRB_SIZE_PCM1(ep->mult - 1);
+                               unsigned int mult = ep->mult - 1;
+                               unsigned int maxp = usb_endpoint_maxp(ep->desc);
+
+                               if (length <= (2 * maxp))
+                                       mult--;
+
+                               if (length <= maxp)
+                                       mult--;
+
+                               trb->size |= DWC3_TRB_SIZE_PCM1(mult);
                        }
                } else {
                        trb->ctrl = DWC3_TRBCTL_ISOCHRONOUS;
index 62dc9c7798e78a034a9e63f8c3163fcebec9e57a..e1de8fe599a35695eceda22d3138492af8995b20 100644 (file)
@@ -838,21 +838,32 @@ static struct renesas_usb3_request *usb3_get_request(struct renesas_usb3_ep
        return usb3_req;
 }
 
-static void usb3_request_done(struct renesas_usb3_ep *usb3_ep,
-                             struct renesas_usb3_request *usb3_req, int status)
+static void __usb3_request_done(struct renesas_usb3_ep *usb3_ep,
+                               struct renesas_usb3_request *usb3_req,
+                               int status)
 {
        struct renesas_usb3 *usb3 = usb3_ep_to_usb3(usb3_ep);
-       unsigned long flags;
 
        dev_dbg(usb3_to_dev(usb3), "giveback: ep%2d, %u, %u, %d\n",
                usb3_ep->num, usb3_req->req.length, usb3_req->req.actual,
                status);
        usb3_req->req.status = status;
-       spin_lock_irqsave(&usb3->lock, flags);
        usb3_ep->started = false;
        list_del_init(&usb3_req->queue);
-       spin_unlock_irqrestore(&usb3->lock, flags);
+       spin_unlock(&usb3->lock);
        usb_gadget_giveback_request(&usb3_ep->ep, &usb3_req->req);
+       spin_lock(&usb3->lock);
+}
+
+static void usb3_request_done(struct renesas_usb3_ep *usb3_ep,
+                             struct renesas_usb3_request *usb3_req, int status)
+{
+       struct renesas_usb3 *usb3 = usb3_ep_to_usb3(usb3_ep);
+       unsigned long flags;
+
+       spin_lock_irqsave(&usb3->lock, flags);
+       __usb3_request_done(usb3_ep, usb3_req, status);
+       spin_unlock_irqrestore(&usb3->lock, flags);
 }
 
 static void usb3_irq_epc_pipe0_status_end(struct renesas_usb3 *usb3)
index 858fefd67ebed72dbe59cc4e07806be1b8895577..c8f38649f749311a81ff63aa194c1b97cb71718f 100644 (file)
@@ -98,6 +98,7 @@ enum amd_chipset_gen {
        AMD_CHIPSET_HUDSON2,
        AMD_CHIPSET_BOLTON,
        AMD_CHIPSET_YANGTZE,
+       AMD_CHIPSET_TAISHAN,
        AMD_CHIPSET_UNKNOWN,
 };
 
@@ -141,6 +142,11 @@ static int amd_chipset_sb_type_init(struct amd_chipset_info *pinfo)
                        pinfo->sb_type.gen = AMD_CHIPSET_SB700;
                else if (rev >= 0x40 && rev <= 0x4f)
                        pinfo->sb_type.gen = AMD_CHIPSET_SB800;
+       }
+       pinfo->smbus_dev = pci_get_device(PCI_VENDOR_ID_AMD,
+                                         0x145c, NULL);
+       if (pinfo->smbus_dev) {
+               pinfo->sb_type.gen = AMD_CHIPSET_TAISHAN;
        } else {
                pinfo->smbus_dev = pci_get_device(PCI_VENDOR_ID_AMD,
                                PCI_DEVICE_ID_AMD_HUDSON2_SMBUS, NULL);
@@ -260,11 +266,12 @@ int usb_hcd_amd_remote_wakeup_quirk(struct pci_dev *pdev)
 {
        /* Make sure amd chipset type has already been initialized */
        usb_amd_find_chipset_info();
-       if (amd_chipset.sb_type.gen != AMD_CHIPSET_YANGTZE)
-               return 0;
-
-       dev_dbg(&pdev->dev, "QUIRK: Enable AMD remote wakeup fix\n");
-       return 1;
+       if (amd_chipset.sb_type.gen == AMD_CHIPSET_YANGTZE ||
+           amd_chipset.sb_type.gen == AMD_CHIPSET_TAISHAN) {
+               dev_dbg(&pdev->dev, "QUIRK: Enable AMD remote wakeup fix\n");
+               return 1;
+       }
+       return 0;
 }
 EXPORT_SYMBOL_GPL(usb_hcd_amd_remote_wakeup_quirk);
 
index 76decb8011ebc2d37d3e7e64a394bd2bb6e5c8ba..3344ffd5bb13743812bab737f52b8e81eb77b9db 100644 (file)
@@ -139,6 +139,7 @@ static void musb_h_tx_flush_fifo(struct musb_hw_ep *ep)
                                "Could not flush host TX%d fifo: csr: %04x\n",
                                ep->epnum, csr))
                        return;
+               mdelay(1);
        }
 }
 
index 8fb86a5f458e01275186dc83ff86e38b38fff35b..3d0dd2f9741571423522ee80c972d8db2ffec853 100644 (file)
@@ -197,6 +197,7 @@ struct msm_otg {
        struct regulator *v3p3;
        struct regulator *v1p8;
        struct regulator *vddcx;
+       struct regulator_bulk_data supplies[3];
 
        struct reset_control *phy_rst;
        struct reset_control *link_rst;
@@ -1731,7 +1732,6 @@ static int msm_otg_reboot_notify(struct notifier_block *this,
 
 static int msm_otg_probe(struct platform_device *pdev)
 {
-       struct regulator_bulk_data regs[3];
        int ret = 0;
        struct device_node *np = pdev->dev.of_node;
        struct msm_otg_platform_data *pdata;
@@ -1817,17 +1817,18 @@ static int msm_otg_probe(struct platform_device *pdev)
                return motg->irq;
        }
 
-       regs[0].supply = "vddcx";
-       regs[1].supply = "v3p3";
-       regs[2].supply = "v1p8";
+       motg->supplies[0].supply = "vddcx";
+       motg->supplies[1].supply = "v3p3";
+       motg->supplies[2].supply = "v1p8";
 
-       ret = devm_regulator_bulk_get(motg->phy.dev, ARRAY_SIZE(regs), regs);
+       ret = devm_regulator_bulk_get(motg->phy.dev, ARRAY_SIZE(motg->supplies),
+                                     motg->supplies);
        if (ret)
                return ret;
 
-       motg->vddcx = regs[0].consumer;
-       motg->v3p3  = regs[1].consumer;
-       motg->v1p8  = regs[2].consumer;
+       motg->vddcx = motg->supplies[0].consumer;
+       motg->v3p3  = motg->supplies[1].consumer;
+       motg->v1p8  = motg->supplies[2].consumer;
 
        clk_set_rate(motg->clk, 60000000);
 
index 93fba9033b00a7136b85a9995d7da32a140c61e9..2c8161bcf5b5e22ee8ed2eb413d29fdfde504e95 100644 (file)
@@ -639,14 +639,11 @@ static int usbhsg_ep_disable(struct usb_ep *ep)
        struct usbhsg_uep *uep = usbhsg_ep_to_uep(ep);
        struct usbhs_pipe *pipe;
        unsigned long flags;
-       int ret = 0;
 
        spin_lock_irqsave(&uep->lock, flags);
        pipe = usbhsg_uep_to_pipe(uep);
-       if (!pipe) {
-               ret = -EINVAL;
+       if (!pipe)
                goto out;
-       }
 
        usbhsg_pipe_disable(uep);
        usbhs_pipe_free(pipe);
index d544b331c9f2ce80d83095f30184e2102eda6ea6..02b67abfc2a16139a3230d1120a17a4a9ab1cc09 100644 (file)
 /* Low Power Status register (LPSTS) */
 #define LPSTS_SUSPM    0x4000
 
-/* USB General control register 2 (UGCTRL2), bit[31:6] should be 0 */
+/*
+ * USB General control register 2 (UGCTRL2)
+ * Remarks: bit[31:11] and bit[9:6] should be 0
+ */
 #define UGCTRL2_RESERVED_3     0x00000001      /* bit[3:0] should be B'0001 */
 #define UGCTRL2_USB0SEL_OTG    0x00000030
+#define UGCTRL2_VBUSSEL                0x00000400
 
 static void usbhs_write32(struct usbhs_priv *priv, u32 reg, u32 data)
 {
@@ -34,7 +38,8 @@ static int usbhs_rcar3_power_ctrl(struct platform_device *pdev,
 {
        struct usbhs_priv *priv = usbhs_pdev_to_priv(pdev);
 
-       usbhs_write32(priv, UGCTRL2, UGCTRL2_RESERVED_3 | UGCTRL2_USB0SEL_OTG);
+       usbhs_write32(priv, UGCTRL2, UGCTRL2_RESERVED_3 | UGCTRL2_USB0SEL_OTG |
+                     UGCTRL2_VBUSSEL);
 
        if (enable) {
                usbhs_bset(priv, LPSTS, LPSTS_SUSPM, LPSTS_SUSPM);
index f64e914a8985495bf6ef51bb3de4772c40ffbf49..2d945c9f975c04d5cd7909e00a37017a97e8062e 100644 (file)
@@ -142,6 +142,7 @@ static const struct usb_device_id id_table[] = {
        { USB_DEVICE(0x10C4, 0x8998) }, /* KCF Technologies PRN */
        { USB_DEVICE(0x10C4, 0x8A2A) }, /* HubZ dual ZigBee and Z-Wave dongle */
        { USB_DEVICE(0x10C4, 0x8A5E) }, /* CEL EM3588 ZigBee USB Stick Long Range */
+       { USB_DEVICE(0x10C4, 0x8B34) }, /* Qivicon ZigBee USB Radio Stick */
        { USB_DEVICE(0x10C4, 0xEA60) }, /* Silicon Labs factory default */
        { USB_DEVICE(0x10C4, 0xEA61) }, /* Silicon Labs factory default */
        { USB_DEVICE(0x10C4, 0xEA70) }, /* Silicon Labs factory default */
index ebe51f11105d8889d6144a72bbf719f14e4b1103..fe123153b1a5439f016f2637d4f9d19f0fde3f14 100644 (file)
@@ -2025,6 +2025,8 @@ static const struct usb_device_id option_ids[] = {
        { USB_DEVICE_INTERFACE_CLASS(0x2001, 0x7d04, 0xff) },                   /* D-Link DWM-158 */
        { USB_DEVICE_INTERFACE_CLASS(0x2001, 0x7e19, 0xff),                     /* D-Link DWM-221 B1 */
          .driver_info = (kernel_ulong_t)&net_intf4_blacklist },
+       { USB_DEVICE_INTERFACE_CLASS(0x2001, 0x7e35, 0xff),                     /* D-Link DWM-222 */
+         .driver_info = (kernel_ulong_t)&net_intf4_blacklist },
        { USB_DEVICE_AND_INTERFACE_INFO(0x07d1, 0x3e01, 0xff, 0xff, 0xff) }, /* D-Link DWM-152/C1 */
        { USB_DEVICE_AND_INTERFACE_INFO(0x07d1, 0x3e02, 0xff, 0xff, 0xff) }, /* D-Link DWM-156/C1 */
        { USB_DEVICE_AND_INTERFACE_INFO(0x07d1, 0x7e11, 0xff, 0xff, 0xff) }, /* D-Link DWM-156/A3 */
index c9ebefd8f35fdbe5491627f4df89b80c87aeabad..a585b477415dde58b38c9e2defe0d26c24552757 100644 (file)
@@ -52,6 +52,8 @@ static const struct usb_device_id id_table[] = {
        { USB_DEVICE(IODATA_VENDOR_ID, IODATA_PRODUCT_ID_RSAQ5) },
        { USB_DEVICE(ATEN_VENDOR_ID, ATEN_PRODUCT_ID),
                .driver_info = PL2303_QUIRK_ENDPOINT_HACK },
+       { USB_DEVICE(ATEN_VENDOR_ID, ATEN_PRODUCT_UC485),
+               .driver_info = PL2303_QUIRK_ENDPOINT_HACK },
        { USB_DEVICE(ATEN_VENDOR_ID, ATEN_PRODUCT_ID2) },
        { USB_DEVICE(ATEN_VENDOR_ID2, ATEN_PRODUCT_ID) },
        { USB_DEVICE(ELCOM_VENDOR_ID, ELCOM_PRODUCT_ID) },
index 09d9be88209e1ce6b1f53dc052a53c5e4c491336..3b5a15d1dc0dd50a0ebf9d925e33ef4cf8db0090 100644 (file)
@@ -27,6 +27,7 @@
 #define ATEN_VENDOR_ID         0x0557
 #define ATEN_VENDOR_ID2                0x0547
 #define ATEN_PRODUCT_ID                0x2008
+#define ATEN_PRODUCT_UC485     0x2021
 #define ATEN_PRODUCT_ID2       0x2118
 
 #define IODATA_VENDOR_ID       0x04bb
index cbea9f329e715aad97b23928d79579cdcf5e8c92..cde115359793001dcf18b8884c3c9e22eec0fa73 100644 (file)
@@ -124,9 +124,9 @@ UNUSUAL_DEV(0x0bc2, 0xab2a, 0x0000, 0x9999,
 /* Reported-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> */
 UNUSUAL_DEV(0x13fd, 0x3940, 0x0000, 0x9999,
                "Initio Corporation",
-               "",
+               "INIC-3069",
                USB_SC_DEVICE, USB_PR_DEVICE, NULL,
-               US_FL_NO_ATA_1X),
+               US_FL_NO_ATA_1X | US_FL_IGNORE_RESIDUE),
 
 /* Reported-by: Tom Arild Naess <tanaess@gmail.com> */
 UNUSUAL_DEV(0x152d, 0x0539, 0x0000, 0x9999,
index 06615934fed1cc537694b03d023b613c95aea9f1..0dceb9fa3a0629af3bd1a1990508d2c32bb42db2 100644 (file)
@@ -315,6 +315,7 @@ static int usb_stor_control_thread(void * __us)
 {
        struct us_data *us = (struct us_data *)__us;
        struct Scsi_Host *host = us_to_host(us);
+       struct scsi_cmnd *srb;
 
        for (;;) {
                usb_stor_dbg(us, "*** thread sleeping\n");
@@ -330,6 +331,7 @@ static int usb_stor_control_thread(void * __us)
                scsi_lock(host);
 
                /* When we are called with no command pending, we're done */
+               srb = us->srb;
                if (us->srb == NULL) {
                        scsi_unlock(host);
                        mutex_unlock(&us->dev_mutex);
@@ -398,14 +400,11 @@ static int usb_stor_control_thread(void * __us)
                /* lock access to the state */
                scsi_lock(host);
 
-               /* indicate that the command is done */
-               if (us->srb->result != DID_ABORT << 16) {
-                       usb_stor_dbg(us, "scsi cmd done, result=0x%x\n",
-                                    us->srb->result);
-                       us->srb->scsi_done(us->srb);
-               } else {
+               /* was the command aborted? */
+               if (us->srb->result == DID_ABORT << 16) {
 SkipForAbort:
                        usb_stor_dbg(us, "scsi command aborted\n");
+                       srb = NULL;     /* Don't call srb->scsi_done() */
                }
 
                /*
@@ -429,6 +428,13 @@ SkipForAbort:
 
                /* unlock the device pointers */
                mutex_unlock(&us->dev_mutex);
+
+               /* now that the locks are released, notify the SCSI core */
+               if (srb) {
+                       usb_stor_dbg(us, "scsi cmd done, result=0x%x\n",
+                                       srb->result);
+                       srb->scsi_done(srb);
+               }
        } /* for (;;) */
 
        /* Wait until we are told to stop */
index ff01bed7112f1566ca13b3e17330851bf02fec05..1e784adb89b17534ce31751b546ef20801b2427f 100644 (file)
@@ -17,6 +17,7 @@
 #include <asm/efi.h>
 
 static bool request_mem_succeeded = false;
+static bool nowc = false;
 
 static struct fb_var_screeninfo efifb_defined = {
        .activate               = FB_ACTIVATE_NOW,
@@ -99,6 +100,8 @@ static int efifb_setup(char *options)
                                screen_info.lfb_height = simple_strtoul(this_opt+7, NULL, 0);
                        else if (!strncmp(this_opt, "width:", 6))
                                screen_info.lfb_width = simple_strtoul(this_opt+6, NULL, 0);
+                       else if (!strcmp(this_opt, "nowc"))
+                               nowc = true;
                }
        }
 
@@ -255,7 +258,10 @@ static int efifb_probe(struct platform_device *dev)
        info->apertures->ranges[0].base = efifb_fix.smem_start;
        info->apertures->ranges[0].size = size_remap;
 
-       info->screen_base = ioremap_wc(efifb_fix.smem_start, efifb_fix.smem_len);
+       if (nowc)
+               info->screen_base = ioremap(efifb_fix.smem_start, efifb_fix.smem_len);
+       else
+               info->screen_base = ioremap_wc(efifb_fix.smem_start, efifb_fix.smem_len);
        if (!info->screen_base) {
                pr_err("efifb: abort, cannot ioremap video memory 0x%x @ 0x%lx\n",
                        efifb_fix.smem_len, efifb_fix.smem_start);
index c166e0725be5dab13e9a685a93ea7ea9c23a3351..ba82f97fb42b2d10fdbebd227fcb7e5eb19dcbdc 100644 (file)
@@ -1073,20 +1073,16 @@ static int imxfb_remove(struct platform_device *pdev)
        imxfb_disable_controller(fbi);
 
        unregister_framebuffer(info);
-
+       fb_dealloc_cmap(&info->cmap);
        pdata = dev_get_platdata(&pdev->dev);
        if (pdata && pdata->exit)
                pdata->exit(fbi->pdev);
-
-       fb_dealloc_cmap(&info->cmap);
-       kfree(info->pseudo_palette);
-       framebuffer_release(info);
-
        dma_free_wc(&pdev->dev, fbi->map_size, info->screen_base,
                    fbi->map_dma);
-
        iounmap(fbi->regs);
        release_mem_region(res->start, resource_size(res));
+       kfree(info->pseudo_palette);
+       framebuffer_release(info);
 
        return 0;
 }
index eecf695c16f41b6996520e47b8fa7e1b71efbabc..09e5bb013d28071c69b63b1968ef6eca44252f1e 100644 (file)
@@ -193,7 +193,6 @@ static struct notifier_block omap_dss_pm_notif_block = {
 
 static int __init omap_dss_probe(struct platform_device *pdev)
 {
-       struct omap_dss_board_info *pdata = pdev->dev.platform_data;
        int r;
 
        core.pdev = pdev;
index 8feab810aed9222b97bd5959b42e6f09ffaaf83d..7f188b8d0c6703dda0ed4ae7647d7b082c429933 100644 (file)
@@ -7,9 +7,6 @@ obj-y   += xenbus/
 nostackp := $(call cc-option, -fno-stack-protector)
 CFLAGS_features.o                      := $(nostackp)
 
-CFLAGS_efi.o                           += -fshort-wchar
-LDFLAGS                                        += $(call ld-option, --no-wchar-size-warning)
-
 dom0-$(CONFIG_ARM64) += arm-device.o
 dom0-$(CONFIG_PCI) += pci.o
 dom0-$(CONFIG_USB_SUPPORT) += dbgp.o
index 4da69dbf7dcad7f2e2898dd807a118596f86058a..1bdd02a6d6ac757c5a500db192b3d7923c316753 100644 (file)
@@ -10,8 +10,7 @@ bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
        unsigned long bfn1 = pfn_to_bfn(page_to_pfn(vec1->bv_page));
        unsigned long bfn2 = pfn_to_bfn(page_to_pfn(vec2->bv_page));
 
-       return __BIOVEC_PHYS_MERGEABLE(vec1, vec2) &&
-               ((bfn1 == bfn2) || ((bfn1+1) == bfn2));
+       return bfn1 + PFN_DOWN(vec1->bv_offset + vec1->bv_len) == bfn2;
 #else
        /*
         * XXX: Add support for merging bio_vec when using different page
index bae1f5d36c26e8eac1a7ce9473a3ca7999b90642..2d43118077e4eecc12f68bee605fb068def11c95 100644 (file)
@@ -574,7 +574,7 @@ static void shutdown_pirq(struct irq_data *data)
 
 static void enable_pirq(struct irq_data *data)
 {
-       startup_pirq(data);
+       enable_dynirq(data);
 }
 
 static void disable_pirq(struct irq_data *data)
index e460802149555b6f0d5def7659d8e8207828eb53..3e59590c7254ddc8f1a08f4232262a74a29e3711 100644 (file)
@@ -857,6 +857,8 @@ static int xenwatch_thread(void *unused)
        struct list_head *ent;
        struct xs_watch_event *event;
 
+       xenwatch_pid = current->pid;
+
        for (;;) {
                wait_event_interruptible(watch_events_waitq,
                                         !list_empty(&watch_events));
@@ -925,7 +927,6 @@ int xs_init(void)
        task = kthread_run(xenwatch_thread, NULL, "xenwatch");
        if (IS_ERR(task))
                return PTR_ERR(task);
-       xenwatch_pid = task->pid;
 
        /* shutdown watches for kexec boot */
        xs_reset_watches();
index 879ff9c7ffd01a0f1f2d6a7e92c32926af5ef4c2..6466153f2bf099d357166192710167667abefa09 100644 (file)
@@ -664,8 +664,7 @@ static unsigned long randomize_stack_top(unsigned long stack_top)
 {
        unsigned long random_variable = 0;
 
-       if ((current->flags & PF_RANDOMIZE) &&
-               !(current->personality & ADDR_NO_RANDOMIZE)) {
+       if (current->flags & PF_RANDOMIZE) {
                random_variable = get_random_long();
                random_variable &= STACK_RND_MASK;
                random_variable <<= PAGE_SHIFT;
index 080e2ebb8aa0137baef69edda45aa895bf8b7c7c..f45b61fe9a9ab8d54d87e98ab494e29cedd2e0ab 100644 (file)
@@ -3516,7 +3516,7 @@ static blk_status_t wait_dev_flush(struct btrfs_device *device)
        struct bio *bio = device->flush_bio;
 
        if (!device->flush_bio_sent)
-               return 0;
+               return BLK_STS_OK;
 
        device->flush_bio_sent = 0;
        wait_for_completion_io(&device->flush_wait);
@@ -3563,7 +3563,7 @@ static int barrier_all_devices(struct btrfs_fs_info *info)
                        continue;
 
                write_dev_flush(dev);
-               dev->last_flush_error = 0;
+               dev->last_flush_error = BLK_STS_OK;
        }
 
        /* wait for all the barriers */
index 95c212037095fea727f4a35ea19d9d7e54ad8cfe..24bcd5cd9cf2fc680cc7ed32fb58cd254bb055b5 100644 (file)
@@ -7924,11 +7924,12 @@ err:
        return ret;
 }
 
-static inline int submit_dio_repair_bio(struct inode *inode, struct bio *bio,
-                                       int mirror_num)
+static inline blk_status_t submit_dio_repair_bio(struct inode *inode,
+                                                struct bio *bio,
+                                                int mirror_num)
 {
        struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
-       int ret;
+       blk_status_t ret;
 
        BUG_ON(bio_op(bio) == REQ_OP_WRITE);
 
@@ -7980,10 +7981,10 @@ static int btrfs_check_dio_repairable(struct inode *inode,
        return 1;
 }
 
-static int dio_read_error(struct inode *inode, struct bio *failed_bio,
-                       struct page *page, unsigned int pgoff,
-                       u64 start, u64 end, int failed_mirror,
-                       bio_end_io_t *repair_endio, void *repair_arg)
+static blk_status_t dio_read_error(struct inode *inode, struct bio *failed_bio,
+                                  struct page *page, unsigned int pgoff,
+                                  u64 start, u64 end, int failed_mirror,
+                                  bio_end_io_t *repair_endio, void *repair_arg)
 {
        struct io_failure_record *failrec;
        struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
@@ -7993,18 +7994,19 @@ static int dio_read_error(struct inode *inode, struct bio *failed_bio,
        int read_mode = 0;
        int segs;
        int ret;
+       blk_status_t status;
 
        BUG_ON(bio_op(failed_bio) == REQ_OP_WRITE);
 
        ret = btrfs_get_io_failure_record(inode, start, end, &failrec);
        if (ret)
-               return ret;
+               return errno_to_blk_status(ret);
 
        ret = btrfs_check_dio_repairable(inode, failed_bio, failrec,
                                         failed_mirror);
        if (!ret) {
                free_io_failure(failure_tree, io_tree, failrec);
-               return -EIO;
+               return BLK_STS_IOERR;
        }
 
        segs = bio_segments(failed_bio);
@@ -8022,13 +8024,13 @@ static int dio_read_error(struct inode *inode, struct bio *failed_bio,
                    "Repair DIO Read Error: submitting new dio read[%#x] to this_mirror=%d, in_validation=%d\n",
                    read_mode, failrec->this_mirror, failrec->in_validation);
 
-       ret = submit_dio_repair_bio(inode, bio, failrec->this_mirror);
-       if (ret) {
+       status = submit_dio_repair_bio(inode, bio, failrec->this_mirror);
+       if (status) {
                free_io_failure(failure_tree, io_tree, failrec);
                bio_put(bio);
        }
 
-       return ret;
+       return status;
 }
 
 struct btrfs_retry_complete {
@@ -8065,8 +8067,8 @@ end:
        bio_put(bio);
 }
 
-static int __btrfs_correct_data_nocsum(struct inode *inode,
-                                      struct btrfs_io_bio *io_bio)
+static blk_status_t __btrfs_correct_data_nocsum(struct inode *inode,
+                                               struct btrfs_io_bio *io_bio)
 {
        struct btrfs_fs_info *fs_info;
        struct bio_vec bvec;
@@ -8076,8 +8078,8 @@ static int __btrfs_correct_data_nocsum(struct inode *inode,
        unsigned int pgoff;
        u32 sectorsize;
        int nr_sectors;
-       int ret;
-       int err = 0;
+       blk_status_t ret;
+       blk_status_t err = BLK_STS_OK;
 
        fs_info = BTRFS_I(inode)->root->fs_info;
        sectorsize = fs_info->sectorsize;
@@ -8183,11 +8185,12 @@ static blk_status_t __btrfs_subio_endio_read(struct inode *inode,
        int csum_pos;
        bool uptodate = (err == 0);
        int ret;
+       blk_status_t status;
 
        fs_info = BTRFS_I(inode)->root->fs_info;
        sectorsize = fs_info->sectorsize;
 
-       err = 0;
+       err = BLK_STS_OK;
        start = io_bio->logical;
        done.inode = inode;
        io_bio->bio.bi_iter = io_bio->iter;
@@ -8209,12 +8212,12 @@ try_again:
                done.start = start;
                init_completion(&done.done);
 
-               ret = dio_read_error(inode, &io_bio->bio, bvec.bv_page,
-                               pgoff, start, start + sectorsize - 1,
-                               io_bio->mirror_num,
-                               btrfs_retry_endio, &done);
-               if (ret) {
-                       err = errno_to_blk_status(ret);
+               status = dio_read_error(inode, &io_bio->bio, bvec.bv_page,
+                                       pgoff, start, start + sectorsize - 1,
+                                       io_bio->mirror_num, btrfs_retry_endio,
+                                       &done);
+               if (status) {
+                       err = status;
                        goto next;
                }
 
@@ -8250,7 +8253,7 @@ static blk_status_t btrfs_subio_endio_read(struct inode *inode,
                if (unlikely(err))
                        return __btrfs_correct_data_nocsum(inode, io_bio);
                else
-                       return 0;
+                       return BLK_STS_OK;
        } else {
                return __btrfs_subio_endio_read(inode, io_bio, err);
        }
@@ -8423,9 +8426,9 @@ static inline blk_status_t btrfs_lookup_and_bind_dio_csum(struct inode *inode,
        return 0;
 }
 
-static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode,
-                                        u64 file_offset, int skip_sum,
-                                        int async_submit)
+static inline blk_status_t
+__btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, u64 file_offset,
+                      int skip_sum, int async_submit)
 {
        struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
        struct btrfs_dio_private *dip = bio->bi_private;
@@ -8488,6 +8491,7 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip,
        int clone_offset = 0;
        int clone_len;
        int ret;
+       blk_status_t status;
 
        map_length = orig_bio->bi_iter.bi_size;
        submit_len = map_length;
@@ -8537,9 +8541,9 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip,
                 */
                atomic_inc(&dip->pending_bios);
 
-               ret = __btrfs_submit_dio_bio(bio, inode, file_offset, skip_sum,
-                                            async_submit);
-               if (ret) {
+               status = __btrfs_submit_dio_bio(bio, inode, file_offset, skip_sum,
+                                               async_submit);
+               if (status) {
                        bio_put(bio);
                        atomic_dec(&dip->pending_bios);
                        goto out_err;
@@ -8557,9 +8561,9 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip,
        } while (submit_len > 0);
 
 submit:
-       ret = __btrfs_submit_dio_bio(bio, inode, file_offset, skip_sum,
-                                    async_submit);
-       if (!ret)
+       status = __btrfs_submit_dio_bio(bio, inode, file_offset, skip_sum,
+                                       async_submit);
+       if (!status)
                return 0;
 
        bio_put(bio);
index 208638384cd2abfb1206b2f5927b5763a6330283..2cf6ba40f7c441bb3483f9c543e1e17bbcea1971 100644 (file)
@@ -905,7 +905,7 @@ static void raid_write_end_io(struct bio *bio)
        if (!atomic_dec_and_test(&rbio->stripes_pending))
                return;
 
-       err = 0;
+       err = BLK_STS_OK;
 
        /* OK, we have read all the stripes we need to. */
        max_errors = (rbio->operation == BTRFS_RBIO_PARITY_SCRUB) ?
@@ -1324,7 +1324,7 @@ write_data:
        return;
 
 cleanup:
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
 }
 
 /*
@@ -1475,7 +1475,7 @@ static void raid_rmw_end_io(struct bio *bio)
 
 cleanup:
 
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
 }
 
 static void async_rmw_stripe(struct btrfs_raid_bio *rbio)
@@ -1579,7 +1579,7 @@ static int raid56_rmw_stripe(struct btrfs_raid_bio *rbio)
        return 0;
 
 cleanup:
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
        return -EIO;
 
 finish:
@@ -1795,12 +1795,12 @@ static void __raid_recover_end_io(struct btrfs_raid_bio *rbio)
        void **pointers;
        int faila = -1, failb = -1;
        struct page *page;
-       int err;
+       blk_status_t err;
        int i;
 
        pointers = kcalloc(rbio->real_stripes, sizeof(void *), GFP_NOFS);
        if (!pointers) {
-               err = -ENOMEM;
+               err = BLK_STS_RESOURCE;
                goto cleanup_io;
        }
 
@@ -1856,7 +1856,7 @@ static void __raid_recover_end_io(struct btrfs_raid_bio *rbio)
                                         * a bad data or Q stripe.
                                         * TODO, we should redo the xor here.
                                         */
-                                       err = -EIO;
+                                       err = BLK_STS_IOERR;
                                        goto cleanup;
                                }
                                /*
@@ -1882,7 +1882,7 @@ static void __raid_recover_end_io(struct btrfs_raid_bio *rbio)
                        if (rbio->bbio->raid_map[failb] == RAID6_Q_STRIPE) {
                                if (rbio->bbio->raid_map[faila] ==
                                    RAID5_P_STRIPE) {
-                                       err = -EIO;
+                                       err = BLK_STS_IOERR;
                                        goto cleanup;
                                }
                                /*
@@ -1954,13 +1954,13 @@ pstripe:
                }
        }
 
-       err = 0;
+       err = BLK_STS_OK;
 cleanup:
        kfree(pointers);
 
 cleanup_io:
        if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
-               if (err == 0)
+               if (err == BLK_STS_OK)
                        cache_rbio_pages(rbio);
                else
                        clear_bit(RBIO_CACHE_READY_BIT, &rbio->flags);
@@ -1968,7 +1968,7 @@ cleanup_io:
                rbio_orig_end_io(rbio, err);
        } else if (rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
                rbio_orig_end_io(rbio, err);
-       } else if (err == 0) {
+       } else if (err == BLK_STS_OK) {
                rbio->faila = -1;
                rbio->failb = -1;
 
@@ -2005,7 +2005,7 @@ static void raid_recover_end_io(struct bio *bio)
                return;
 
        if (atomic_read(&rbio->error) > rbio->bbio->max_errors)
-               rbio_orig_end_io(rbio, -EIO);
+               rbio_orig_end_io(rbio, BLK_STS_IOERR);
        else
                __raid_recover_end_io(rbio);
 }
@@ -2104,7 +2104,7 @@ out:
 cleanup:
        if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
            rbio->operation == BTRFS_RBIO_REBUILD_MISSING)
-               rbio_orig_end_io(rbio, -EIO);
+               rbio_orig_end_io(rbio, BLK_STS_IOERR);
        return -EIO;
 }
 
@@ -2431,7 +2431,7 @@ submit_write:
        nr_data = bio_list_size(&bio_list);
        if (!nr_data) {
                /* Every parity is right */
-               rbio_orig_end_io(rbio, 0);
+               rbio_orig_end_io(rbio, BLK_STS_OK);
                return;
        }
 
@@ -2451,7 +2451,7 @@ submit_write:
        return;
 
 cleanup:
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
 }
 
 static inline int is_data_stripe(struct btrfs_raid_bio *rbio, int stripe)
@@ -2519,7 +2519,7 @@ static void validate_rbio_for_parity_scrub(struct btrfs_raid_bio *rbio)
        return;
 
 cleanup:
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
 }
 
 /*
@@ -2633,7 +2633,7 @@ static void raid56_parity_scrub_stripe(struct btrfs_raid_bio *rbio)
        return;
 
 cleanup:
-       rbio_orig_end_io(rbio, -EIO);
+       rbio_orig_end_io(rbio, BLK_STS_IOERR);
        return;
 
 finish:
index e8b9a269fddec78fdf42adec32eabaffdf9c3636..bd679bc7a1a956cfc541fcb2a74ed4b49b815642 100644 (file)
@@ -6212,8 +6212,8 @@ static void bbio_error(struct btrfs_bio *bbio, struct bio *bio, u64 logical)
        }
 }
 
-int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
-                 int mirror_num, int async_submit)
+blk_status_t btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
+                          int mirror_num, int async_submit)
 {
        struct btrfs_device *dev;
        struct bio *first_bio = bio;
@@ -6233,7 +6233,7 @@ int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
                                &map_length, &bbio, mirror_num, 1);
        if (ret) {
                btrfs_bio_counter_dec(fs_info);
-               return ret;
+               return errno_to_blk_status(ret);
        }
 
        total_devs = bbio->num_stripes;
@@ -6256,7 +6256,7 @@ int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
                }
 
                btrfs_bio_counter_dec(fs_info);
-               return ret;
+               return errno_to_blk_status(ret);
        }
 
        if (map_length < length) {
@@ -6283,7 +6283,7 @@ int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
                                  dev_nr, async_submit);
        }
        btrfs_bio_counter_dec(fs_info);
-       return 0;
+       return BLK_STS_OK;
 }
 
 struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid,
index 6f45fd60d15a92d7aea7a5711bb365a201cb09f9..93277fc60930561a7ffb8c40d2e4ae61ce5908a0 100644 (file)
@@ -74,7 +74,7 @@ struct btrfs_device {
        int missing;
        int can_discard;
        int is_tgtdev_for_dev_replace;
-       int last_flush_error;
+       blk_status_t last_flush_error;
        int flush_bio_sent;
 
 #ifdef __BTRFS_NEED_DEVICE_DATA_ORDERED
@@ -416,8 +416,8 @@ int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
                      struct btrfs_fs_info *fs_info, u64 type);
 void btrfs_mapping_init(struct btrfs_mapping_tree *tree);
 void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree);
-int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
-                 int mirror_num, int async_submit);
+blk_status_t btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio,
+                          int mirror_num, int async_submit);
 int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
                       fmode_t flags, void *holder);
 int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
index 108df2e3602c2c63186b61b2fede52794cb46482..7eae33ffa3fcc1d709790b79225521472c6f8281 100644 (file)
@@ -133,6 +133,50 @@ static inline struct pts_fs_info *DEVPTS_SB(struct super_block *sb)
        return sb->s_fs_info;
 }
 
+static int devpts_ptmx_path(struct path *path)
+{
+       struct super_block *sb;
+       int err;
+
+       /* Has the devpts filesystem already been found? */
+       if (path->mnt->mnt_sb->s_magic == DEVPTS_SUPER_MAGIC)
+               return 0;
+
+       /* Is a devpts filesystem at "pts" in the same directory? */
+       err = path_pts(path);
+       if (err)
+               return err;
+
+       /* Is the path the root of a devpts filesystem? */
+       sb = path->mnt->mnt_sb;
+       if ((sb->s_magic != DEVPTS_SUPER_MAGIC) ||
+           (path->mnt->mnt_root != sb->s_root))
+               return -ENODEV;
+
+       return 0;
+}
+
+struct vfsmount *devpts_mntget(struct file *filp, struct pts_fs_info *fsi)
+{
+       struct path path;
+       int err;
+
+       path = filp->f_path;
+       path_get(&path);
+
+       err = devpts_ptmx_path(&path);
+       dput(path.dentry);
+       if (err) {
+               mntput(path.mnt);
+               path.mnt = ERR_PTR(err);
+       }
+       if (DEVPTS_SB(path.mnt->mnt_sb) != fsi) {
+               mntput(path.mnt);
+               path.mnt = ERR_PTR(-ENODEV);
+       }
+       return path.mnt;
+}
+
 struct pts_fs_info *devpts_acquire(struct file *filp)
 {
        struct pts_fs_info *result;
@@ -143,27 +187,16 @@ struct pts_fs_info *devpts_acquire(struct file *filp)
        path = filp->f_path;
        path_get(&path);
 
-       /* Has the devpts filesystem already been found? */
-       sb = path.mnt->mnt_sb;
-       if (sb->s_magic != DEVPTS_SUPER_MAGIC) {
-               /* Is a devpts filesystem at "pts" in the same directory? */
-               err = path_pts(&path);
-               if (err) {
-                       result = ERR_PTR(err);
-                       goto out;
-               }
-
-               /* Is the path the root of a devpts filesystem? */
-               result = ERR_PTR(-ENODEV);
-               sb = path.mnt->mnt_sb;
-               if ((sb->s_magic != DEVPTS_SUPER_MAGIC) ||
-                   (path.mnt->mnt_root != sb->s_root))
-                       goto out;
+       err = devpts_ptmx_path(&path);
+       if (err) {
+               result = ERR_PTR(err);
+               goto out;
        }
 
        /*
         * pty code needs to hold extra references in case of last /dev/tty close
         */
+       sb = path.mnt->mnt_sb;
        atomic_inc(&sb->s_active);
        result = DEVPTS_SB(sb);
 
index 5a1052627a81413685241f0c4669bbf755e2c5f9..701085620cd82cee3fda28ff6ca65921a9119fc1 100644 (file)
@@ -2300,7 +2300,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
                                             EXT4_MAX_BLOCK_LOG_SIZE);
        struct sg {
                struct ext4_group_info info;
-               ext4_grpblk_t counters[blocksize_bits + 2];
+               ext4_grpblk_t counters[EXT4_MAX_BLOCK_LOG_SIZE + 2];
        } sg;
 
        group--;
@@ -2309,6 +2309,9 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
                              " 2^0   2^1   2^2   2^3   2^4   2^5   2^6  "
                              " 2^7   2^8   2^9   2^10  2^11  2^12  2^13  ]\n");
 
+       i = (blocksize_bits + 2) * sizeof(sg.info.bb_counters[0]) +
+               sizeof(struct ext4_group_info);
+
        grinfo = ext4_get_group_info(sb, group);
        /* Load the group info in memory only if not already loaded. */
        if (unlikely(EXT4_MB_GRP_NEED_INIT(grinfo))) {
@@ -2320,7 +2323,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
                buddy_loaded = 1;
        }
 
-       memcpy(&sg, ext4_get_group_info(sb, group), sizeof(sg));
+       memcpy(&sg, ext4_get_group_info(sb, group), i);
 
        if (buddy_loaded)
                ext4_mb_unload_buddy(&e4b);
index 82a5af9f66685a6165a59a1f0c6ba603dd332e8b..3dd9701684488627d26294dd0489323b72afbba8 100644 (file)
@@ -1543,7 +1543,7 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
                        /* Clear padding bytes. */
                        memset(val + i->value_len, 0, new_size - i->value_len);
                }
-               return 0;
+               goto update_hash;
        }
 
        /* Compute min_offs and last. */
@@ -1707,6 +1707,7 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
                here->e_value_size = cpu_to_le32(i->value_len);
        }
 
+update_hash:
        if (i->value) {
                __le32 hash = 0;
 
@@ -1725,7 +1726,8 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i,
                                                     here->e_name_len,
                                                     &crc32c_hash, 1);
                } else if (is_block) {
-                       __le32 *value = s->base + min_offs - new_size;
+                       __le32 *value = s->base + le16_to_cpu(
+                                                       here->e_value_offs);
 
                        hash = ext4_xattr_hash_entry(here->e_name,
                                                     here->e_name_len, value,
index 3ee4fdc3da9ec359ad847afa36354240329a2da6..ab60051be6e533eb167a72e590494f0a46e3a488 100644 (file)
@@ -46,7 +46,7 @@ struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
 {
        struct fuse_file *ff;
 
-       ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
+       ff = kzalloc(sizeof(struct fuse_file), GFP_KERNEL);
        if (unlikely(!ff))
                return NULL;
 
@@ -609,7 +609,7 @@ static void fuse_aio_complete_req(struct fuse_conn *fc, struct fuse_req *req)
        struct fuse_io_priv *io = req->io;
        ssize_t pos = -1;
 
-       fuse_release_user_pages(req, !io->write);
+       fuse_release_user_pages(req, io->should_dirty);
 
        if (io->write) {
                if (req->misc.write.in.size != req->misc.write.out.size)
@@ -1316,7 +1316,6 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, struct iov_iter *iter,
                       loff_t *ppos, int flags)
 {
        int write = flags & FUSE_DIO_WRITE;
-       bool should_dirty = !write && iter_is_iovec(iter);
        int cuse = flags & FUSE_DIO_CUSE;
        struct file *file = io->file;
        struct inode *inode = file->f_mapping->host;
@@ -1346,6 +1345,7 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, struct iov_iter *iter,
                        inode_unlock(inode);
        }
 
+       io->should_dirty = !write && iter_is_iovec(iter);
        while (count) {
                size_t nres;
                fl_owner_t owner = current->files;
@@ -1360,7 +1360,7 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, struct iov_iter *iter,
                        nres = fuse_send_read(req, io, pos, nbytes, owner);
 
                if (!io->async)
-                       fuse_release_user_pages(req, should_dirty);
+                       fuse_release_user_pages(req, io->should_dirty);
                if (req->out.h.error) {
                        err = req->out.h.error;
                        break;
@@ -1669,6 +1669,7 @@ err_nofile:
 err_free:
        fuse_request_free(req);
 err:
+       mapping_set_error(page->mapping, error);
        end_page_writeback(page);
        return error;
 }
index 1bd7ffdad593977013c1ddd233b2a91093471471..bd4d2a3e1ec1b8cc0af29bcb8c26f13d82708804 100644 (file)
@@ -249,6 +249,7 @@ struct fuse_io_priv {
        size_t size;
        __u64 offset;
        bool write;
+       bool should_dirty;
        int err;
        struct kiocb *iocb;
        struct file *file;
index 039266128b7ff0b09ed2a55e621a9a0dc9e38720..59cc98ad7577b438a7928cdace1f2af3a7cabf75 100644 (file)
@@ -278,7 +278,7 @@ iomap_dirty_actor(struct inode *inode, loff_t pos, loff_t length, void *data,
                unsigned long bytes;    /* Bytes to write to page */
 
                offset = (pos & (PAGE_SIZE - 1));
-               bytes = min_t(unsigned long, PAGE_SIZE - offset, length);
+               bytes = min_t(loff_t, PAGE_SIZE - offset, length);
 
                rpage = __iomap_read_page(inode, pos);
                if (IS_ERR(rpage))
@@ -373,7 +373,7 @@ iomap_zero_range_actor(struct inode *inode, loff_t pos, loff_t count,
                unsigned offset, bytes;
 
                offset = pos & (PAGE_SIZE - 1); /* Within page */
-               bytes = min_t(unsigned, PAGE_SIZE - offset, count);
+               bytes = min_t(loff_t, PAGE_SIZE - offset, count);
 
                if (IS_DAX(inode))
                        status = iomap_dax_zero(pos, offset, bytes, iomap);
index 69d02cf8cf370678609175d6e5626d0dedf02c3a..5f93cfacb3d14b9befc6ba33a91d79ccde044d89 100644 (file)
@@ -121,6 +121,7 @@ config PNFS_FILE_LAYOUT
 config PNFS_BLOCK
        tristate
        depends on NFS_V4_1 && BLK_DEV_DM
+       depends on 64BIT || LBDAF
        default NFS_V4
 
 config PNFS_FLEXFILE_LAYOUT
index 6df7a0cf566015378aa3f76c480115675454297d..f32c58bbe55671cb75abdcb9934152d110e3537d 100644 (file)
@@ -32,6 +32,7 @@ void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds)
 {
        nfs4_print_deviceid(&mirror_ds->id_node.deviceid);
        nfs4_pnfs_ds_put(mirror_ds->ds);
+       kfree(mirror_ds->ds_versions);
        kfree_rcu(mirror_ds, id_node.rcu);
 }
 
index ffd2e712595d8ac875dc2780b52731543c687ca8..d901326423401c3e7d442f62bfa80d03d281ed02 100644 (file)
@@ -2553,9 +2553,8 @@ static int nfs41_check_open_stateid(struct nfs4_state *state)
                clear_bit(NFS_O_RDWR_STATE, &state->flags);
                clear_bit(NFS_OPEN_STATE, &state->flags);
                stateid->type = NFS4_INVALID_STATEID_TYPE;
-       }
-       if (status != NFS_OK)
                return status;
+       }
        if (nfs_open_stateid_recover_openmode(state))
                return -NFS4ERR_OPENMODE;
        return NFS_OK;
index 53a17496c5c536a9410cba73c16d70c5efdbba88..566e6ef99f077c76680cb005417ae995cb4e2878 100644 (file)
@@ -1124,6 +1124,10 @@ void dquot_free_reserved_space(struct dquot *dquot, qsize_t number)
                WARN_ON_ONCE(1);
                dquot->dq_dqb.dqb_rsvspace = 0;
        }
+       if (dquot->dq_dqb.dqb_curspace + dquot->dq_dqb.dqb_rsvspace <=
+           dquot->dq_dqb.dqb_bsoftlimit)
+               dquot->dq_dqb.dqb_btime = (time64_t) 0;
+       clear_bit(DQ_BLKS_B, &dquot->dq_flags);
 }
 
 static void dquot_decr_inodes(struct dquot *dquot, qsize_t number)
@@ -1145,7 +1149,8 @@ static void dquot_decr_space(struct dquot *dquot, qsize_t number)
                dquot->dq_dqb.dqb_curspace -= number;
        else
                dquot->dq_dqb.dqb_curspace = 0;
-       if (dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit)
+       if (dquot->dq_dqb.dqb_curspace + dquot->dq_dqb.dqb_rsvspace <=
+           dquot->dq_dqb.dqb_bsoftlimit)
                dquot->dq_dqb.dqb_btime = (time64_t) 0;
        clear_bit(DQ_BLKS_B, &dquot->dq_flags);
 }
@@ -1381,14 +1386,18 @@ static int info_idq_free(struct dquot *dquot, qsize_t inodes)
 
 static int info_bdq_free(struct dquot *dquot, qsize_t space)
 {
+       qsize_t tspace;
+
+       tspace = dquot->dq_dqb.dqb_curspace + dquot->dq_dqb.dqb_rsvspace;
+
        if (test_bit(DQ_FAKE_B, &dquot->dq_flags) ||
-           dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit)
+           tspace <= dquot->dq_dqb.dqb_bsoftlimit)
                return QUOTA_NL_NOWARN;
 
-       if (dquot->dq_dqb.dqb_curspace - space <= dquot->dq_dqb.dqb_bsoftlimit)
+       if (tspace - space <= dquot->dq_dqb.dqb_bsoftlimit)
                return QUOTA_NL_BSOFTBELOW;
-       if (dquot->dq_dqb.dqb_curspace >= dquot->dq_dqb.dqb_bhardlimit &&
-           dquot->dq_dqb.dqb_curspace - space < dquot->dq_dqb.dqb_bhardlimit)
+       if (tspace >= dquot->dq_dqb.dqb_bhardlimit &&
+           tspace - space < dquot->dq_dqb.dqb_bhardlimit)
                return QUOTA_NL_BHARDBELOW;
        return QUOTA_NL_NOWARN;
 }
@@ -2681,7 +2690,7 @@ static int do_set_dqblk(struct dquot *dquot, struct qc_dqblk *di)
 
        if (check_blim) {
                if (!dm->dqb_bsoftlimit ||
-                   dm->dqb_curspace < dm->dqb_bsoftlimit) {
+                   dm->dqb_curspace + dm->dqb_rsvspace < dm->dqb_bsoftlimit) {
                        dm->dqb_btime = 0;
                        clear_bit(DQ_BLKS_B, &dquot->dq_flags);
                } else if (!(di->d_fieldmask & QC_SPC_TIMER))
index ffd5a15d1bb6d0955633d4e1f61f6a84aa79f3f7..abf5beaae907d32b484128f182bee2d1d7a7cd9e 100644 (file)
@@ -1246,13 +1246,13 @@ xfs_dialloc_ag_inobt(
 
                        /* free inodes to the left? */
                        if (useleft && trec.ir_freecount) {
-                               rec = trec;
                                xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
                                cur = tcur;
 
                                pag->pagl_leftrec = trec.ir_startino;
                                pag->pagl_rightrec = rec.ir_startino;
                                pag->pagl_pagino = pagino;
+                               rec = trec;
                                goto alloc_inode;
                        }
 
index 0053bcf2b10a1de59c459de8c74a8cd1d0269d3b..4ebd0bafc914ce7b7f1e4d35134c9dedb09a23b5 100644 (file)
@@ -749,9 +749,20 @@ xfs_log_mount_finish(
                return 0;
        }
 
+       /*
+        * During the second phase of log recovery, we need iget and
+        * iput to behave like they do for an active filesystem.
+        * xfs_fs_drop_inode needs to be able to prevent the deletion
+        * of inodes before we're done replaying log items on those
+        * inodes.  Turn it off immediately after recovery finishes
+        * so that we don't leak the quota inodes if subsequent mount
+        * activities fail.
+        */
+       mp->m_super->s_flags |= MS_ACTIVE;
        error = xlog_recover_finish(mp->m_log);
        if (!error)
                xfs_log_work_queue(mp);
+       mp->m_super->s_flags &= ~MS_ACTIVE;
 
        return error;
 }
index 40d4e8b4e193b41a123903735736248568bf26b9..ea7d4b4e50d0ca3eedee85ffd840542e0bb105db 100644 (file)
@@ -944,15 +944,6 @@ xfs_mountfs(
                }
        }
 
-       /*
-        * During the second phase of log recovery, we need iget and
-        * iput to behave like they do for an active filesystem.
-        * xfs_fs_drop_inode needs to be able to prevent the deletion
-        * of inodes before we're done replaying log items on those
-        * inodes.
-        */
-       mp->m_super->s_flags |= MS_ACTIVE;
-
        /*
         * Finish recovering the file system.  This part needed to be delayed
         * until after the root and real-time bitmap inodes were consistently
@@ -1028,12 +1019,13 @@ xfs_mountfs(
  out_quota:
        xfs_qm_unmount_quotas(mp);
  out_rtunmount:
-       mp->m_super->s_flags &= ~MS_ACTIVE;
        xfs_rtunmount_inodes(mp);
  out_rele_rip:
        IRELE(rip);
        cancel_delayed_work_sync(&mp->m_reclaim_work);
        xfs_reclaim_inodes(mp, SYNC_WAIT);
+       /* Clean out dquots that might be in memory after quotacheck. */
+       xfs_qm_unmount(mp);
  out_log_dealloc:
        mp->m_flags |= XFS_MOUNT_UNMOUNTING;
        xfs_log_mount_cancel(mp);
index da0be9a8d1de4c239e53aec01a5e8b9cd87e3bdd..9623d78f84947e3d2fa51c7e7839932e5572441b 100644 (file)
 /* Align . to a 8 byte boundary equals to maximum function alignment. */
 #define ALIGN_FUNCTION()  . = ALIGN(8)
 
+/*
+ * LD_DEAD_CODE_DATA_ELIMINATION option enables -fdata-sections, which
+ * generates .data.identifier sections, which need to be pulled in with
+ * .data. We don't want to pull in .data..other sections, which Linux
+ * has defined. Same for text and bss.
+ */
+#ifdef CONFIG_LD_DEAD_CODE_DATA_ELIMINATION
+#define TEXT_MAIN .text .text.[0-9a-zA-Z_]*
+#define DATA_MAIN .data .data.[0-9a-zA-Z_]*
+#define BSS_MAIN .bss .bss.[0-9a-zA-Z_]*
+#else
+#define TEXT_MAIN .text
+#define DATA_MAIN .data
+#define BSS_MAIN .bss
+#endif
+
 /*
  * Align to a 32 byte boundary equal to the
  * alignment gcc 4.5 uses for a struct
 
 /*
  * .data section
- * LD_DEAD_CODE_DATA_ELIMINATION option enables -fdata-sections generates
- * .data.identifier which needs to be pulled in with .data, but don't want to
- * pull in .data..stuff which has its own requirements. Same for bss.
  */
 #define DATA_DATA                                                      \
-       *(.data .data.[0-9a-zA-Z_]*)                                    \
+       *(DATA_MAIN)                                                    \
        *(.ref.data)                                                    \
        *(.data..shared_aligned) /* percpu related */                   \
        MEM_KEEP(init.data)                                             \
                VMLINUX_SYMBOL(__security_initcall_end) = .;            \
        }
 
-/* .text section. Map to function alignment to avoid address changes
+/*
+ * .text section. Map to function alignment to avoid address changes
  * during second ld run in second ld pass when generating System.map
- * LD_DEAD_CODE_DATA_ELIMINATION option enables -ffunction-sections generates
- * .text.identifier which needs to be pulled in with .text , but some
- * architectures define .text.foo which is not intended to be pulled in here.
- * Those enabling LD_DEAD_CODE_DATA_ELIMINATION must ensure they don't have
- * conflicting section names, and must pull in .text.[0-9a-zA-Z_]* */
+ *
+ * TEXT_MAIN here will match .text.fixup and .text.unlikely if dead
+ * code elimination is enabled, so these sections should be converted
+ * to use ".." first.
+ */
 #define TEXT_TEXT                                                      \
                ALIGN_FUNCTION();                                       \
-               *(.text.hot .text .text.fixup .text.unlikely)           \
+               *(.text.hot TEXT_MAIN .text.fixup .text.unlikely)       \
                *(.ref.text)                                            \
        MEM_KEEP(init.text)                                             \
        MEM_KEEP(exit.text)                                             \
                BSS_FIRST_SECTIONS                                      \
                *(.bss..page_aligned)                                   \
                *(.dynbss)                                              \
-               *(.bss .bss.[0-9a-zA-Z_]*)                              \
+               *(BSS_MAIN)                                             \
                *(COMMON)                                               \
        }
 
index c749eef1daa1557910ec81c5f296dc907f5ccab2..27b4b66152637fe38b82711f0fa66d6ce946f7d7 100644 (file)
@@ -1209,6 +1209,7 @@ static inline bool acpi_has_watchdog(void) { return false; }
 #endif
 
 #ifdef CONFIG_ACPI_SPCR_TABLE
+extern bool qdf2400_e44_present;
 int parse_spcr(bool earlycon);
 #else
 static inline int parse_spcr(bool earlycon) { return 0; }
index 277ab9af9ac29a95773ce2201809f7c7613391a1..100cb43437630114127bfecc3d29ad4454c03446 100644 (file)
@@ -19,6 +19,7 @@
 
 struct pts_fs_info;
 
+struct vfsmount *devpts_mntget(struct file *, struct pts_fs_info *);
 struct pts_fs_info *devpts_acquire(struct file *);
 void devpts_release(struct pts_fs_info *);
 
@@ -32,6 +33,15 @@ void *devpts_get_priv(struct dentry *);
 /* unlink */
 void devpts_pty_kill(struct dentry *);
 
+/* in pty.c */
+int ptm_open_peer(struct file *master, struct tty_struct *tty, int flags);
+
+#else
+static inline int
+ptm_open_peer(struct file *master, struct tty_struct *tty, int flags)
+{
+       return -EIO;
+}
 #endif
 
 
index 497f2b3a5a62c8da6f87107de16519b176cc9f1f..97f1b465d04ff0b1ab33b0c1074a722ca41c1b0a 100644 (file)
@@ -105,6 +105,11 @@ struct st_sensor_fullscale {
        struct st_sensor_fullscale_avl fs_avl[ST_SENSORS_FULLSCALE_AVL_MAX];
 };
 
+struct st_sensor_sim {
+       u8 addr;
+       u8 value;
+};
+
 /**
  * struct st_sensor_bdu - ST sensor device block data update
  * @addr: address of the register.
@@ -197,6 +202,7 @@ struct st_sensor_transfer_function {
  * @bdu: Block data update register.
  * @das: Data Alignment Selection register.
  * @drdy_irq: Data ready register of the sensor.
+ * @sim: SPI serial interface mode register of the sensor.
  * @multi_read_bit: Use or not particular bit for [I2C/SPI] multi-read.
  * @bootime: samples to discard when sensor passing from power-down to power-up.
  */
@@ -213,6 +219,7 @@ struct st_sensor_settings {
        struct st_sensor_bdu bdu;
        struct st_sensor_das das;
        struct st_sensor_data_ready_irq drdy_irq;
+       struct st_sensor_sim sim;
        bool multi_read_bit;
        unsigned int bootime;
 };
index 77d427974f575699d181265a9eb4c8d1cdd4dad5..bae11c7e7bf31920fcf0b928b31dc29ad70f6f8b 100644 (file)
@@ -61,6 +61,7 @@ extern int memblock_debug;
 #ifdef CONFIG_ARCH_DISCARD_MEMBLOCK
 #define __init_memblock __meminit
 #define __initdata_memblock __meminitdata
+void memblock_discard(void);
 #else
 #define __init_memblock
 #define __initdata_memblock
@@ -74,8 +75,6 @@ phys_addr_t memblock_find_in_range_node(phys_addr_t size, phys_addr_t align,
                                        int nid, ulong flags);
 phys_addr_t memblock_find_in_range(phys_addr_t start, phys_addr_t end,
                                   phys_addr_t size, phys_addr_t align);
-phys_addr_t get_allocated_memblock_reserved_regions_info(phys_addr_t *addr);
-phys_addr_t get_allocated_memblock_memory_regions_info(phys_addr_t *addr);
 void memblock_allow_resize(void);
 int memblock_add_node(phys_addr_t base, phys_addr_t size, int nid);
 int memblock_add(phys_addr_t base, phys_addr_t size);
@@ -110,6 +109,9 @@ void __next_mem_range_rev(u64 *idx, int nid, ulong flags,
 void __next_reserved_mem_region(u64 *idx, phys_addr_t *out_start,
                                phys_addr_t *out_end);
 
+void __memblock_free_early(phys_addr_t base, phys_addr_t size);
+void __memblock_free_late(phys_addr_t base, phys_addr_t size);
+
 /**
  * for_each_mem_range - iterate through memblock areas from type_a and not
  * included in type_b. Or just type_a if type_b is NULL.
index 3914e3dd61680a9bc8814d9b9a386c59611eca81..9b15a4bcfa77dca9c51f88a195a753a018425e53 100644 (file)
@@ -484,7 +484,8 @@ bool mem_cgroup_oom_synchronize(bool wait);
 extern int do_swap_account;
 #endif
 
-void lock_page_memcg(struct page *page);
+struct mem_cgroup *lock_page_memcg(struct page *page);
+void __unlock_page_memcg(struct mem_cgroup *memcg);
 void unlock_page_memcg(struct page *page);
 
 static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
@@ -809,7 +810,12 @@ mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
 {
 }
 
-static inline void lock_page_memcg(struct page *page)
+static inline struct mem_cgroup *lock_page_memcg(struct page *page)
+{
+       return NULL;
+}
+
+static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
 {
 }
 
index dda2cc939a531dab67441c6ddf4b7869c6a06159..ebeb48c920054d705924453f6ed6c6aca0233af4 100644 (file)
@@ -37,7 +37,7 @@ struct net;
 
 /* Historically, SOCKWQ_ASYNC_NOSPACE & SOCKWQ_ASYNC_WAITDATA were located
  * in sock->flags, but moved into sk->sk_wq->flags to be RCU protected.
- * Eventually all flags will be in sk->sk_wq_flags.
+ * Eventually all flags will be in sk->sk_wq->flags.
  */
 #define SOCKWQ_ASYNC_NOSPACE   0
 #define SOCKWQ_ASYNC_WAITDATA  1
index 8aa01fd859fb84e5f64e9d7ffb0c20957afdfd2d..a36abe2da13e1a23edc8aa152205af143f658990 100644 (file)
@@ -168,6 +168,14 @@ extern int sysctl_hardlockup_all_cpu_backtrace;
 #define sysctl_softlockup_all_cpu_backtrace 0
 #define sysctl_hardlockup_all_cpu_backtrace 0
 #endif
+
+#if defined(CONFIG_HARDLOCKUP_CHECK_TIMESTAMP) && \
+    defined(CONFIG_HARDLOCKUP_DETECTOR)
+void watchdog_update_hrtimer_threshold(u64 period);
+#else
+static inline void watchdog_update_hrtimer_threshold(u64 period) { }
+#endif
+
 extern bool is_hardlockup(void);
 struct ctl_table;
 extern int proc_watchdog(struct ctl_table *, int ,
index 6c8c5d8041b72ec01097d1c0563b793ea7449f1f..2591878c1d4804d374d39491c2a3f64d1b43a214 100644 (file)
@@ -346,6 +346,11 @@ struct nvme_fc_remote_port {
  *       indicating an FC transport Aborted status.
  *       Entrypoint is Mandatory.
  *
+ * @defer_rcv:  Called by the transport to signal the LLLD that it has
+ *       begun processing of a previously received NVME CMD IU. The LLDD
+ *       is now free to re-use the rcv buffer associated with the
+ *       nvmefc_tgt_fcp_req.
+ *
  * @max_hw_queues:  indicates the maximum number of hw queues the LLDD
  *       supports for cpu affinitization.
  *       Value is Mandatory. Must be at least 1.
@@ -846,6 +851,8 @@ struct nvmet_fc_target_template {
                                struct nvmefc_tgt_fcp_req *fcpreq);
        void (*fcp_req_release)(struct nvmet_fc_target_port *tgtport,
                                struct nvmefc_tgt_fcp_req *fcpreq);
+       void (*defer_rcv)(struct nvmet_fc_target_port *tgtport,
+                               struct nvmefc_tgt_fcp_req *fcpreq);
 
        u32     max_hw_queues;
        u16     max_sgl_segments;
index 8a266e2be5a63a29a2de38ff76f0bdc2e4fac08e..76aac4ce39bcf59e2e5eeb38053cf05f0c9d05ae 100644 (file)
@@ -6,6 +6,8 @@
 #include <linux/types.h>
 #include <linux/nodemask.h>
 #include <uapi/linux/oom.h>
+#include <linux/sched/coredump.h> /* MMF_* */
+#include <linux/mm.h> /* VM_FAULT* */
 
 struct zonelist;
 struct notifier_block;
@@ -63,6 +65,26 @@ static inline bool tsk_is_oom_victim(struct task_struct * tsk)
        return tsk->signal->oom_mm;
 }
 
+/*
+ * Checks whether a page fault on the given mm is still reliable.
+ * This is no longer true if the oom reaper started to reap the
+ * address space which is reflected by MMF_UNSTABLE flag set in
+ * the mm. At that moment any !shared mapping would lose the content
+ * and could cause a memory corruption (zero pages instead of the
+ * original content).
+ *
+ * User should call this before establishing a page table entry for
+ * a !shared mapping and under the proper page table lock.
+ *
+ * Return 0 when the PF is safe VM_FAULT_SIGBUS otherwise.
+ */
+static inline int check_stable_address_space(struct mm_struct *mm)
+{
+       if (unlikely(test_bit(MMF_UNSTABLE, &mm->flags)))
+               return VM_FAULT_SIGBUS;
+       return 0;
+}
+
 extern unsigned long oom_badness(struct task_struct *p,
                struct mem_cgroup *memcg, const nodemask_t *nodemask,
                unsigned long totalpages);
index a75c136738529db410baf870f3baafc6e178a5a0..f958d0732af685761c8b996cc6c79ca87f99f43f 100644 (file)
@@ -188,6 +188,8 @@ enum pci_dev_flags {
         * the direct_complete optimization.
         */
        PCI_DEV_FLAGS_NEEDS_RESUME = (__force pci_dev_flags_t) (1 << 11),
+       /* Don't use Relaxed Ordering for TLPs directed at this device */
+       PCI_DEV_FLAGS_NO_RELAXED_ORDERING = (__force pci_dev_flags_t) (1 << 12),
 };
 
 enum pci_irq_reroute_variant {
@@ -1126,6 +1128,7 @@ bool pci_check_pme_status(struct pci_dev *dev);
 void pci_pme_wakeup_bus(struct pci_bus *bus);
 void pci_d3cold_enable(struct pci_dev *dev);
 void pci_d3cold_disable(struct pci_dev *dev);
+bool pcie_relaxed_ordering_enabled(struct pci_dev *dev);
 
 /* PCI Virtual Channel */
 int pci_save_vc_state(struct pci_dev *dev);
index a3b873fc59e41745c78bc31a70f0c71d7e4374c3..b14095bcf4bb40522b66716d4c1e7b59ebf72eb4 100644 (file)
@@ -310,8 +310,8 @@ struct pmu {
         * Notification that the event was mapped or unmapped.  Called
         * in the context of the mapping task.
         */
-       void (*event_mapped)            (struct perf_event *event); /*optional*/
-       void (*event_unmapped)          (struct perf_event *event); /*optional*/
+       void (*event_mapped)            (struct perf_event *event, struct mm_struct *mm); /* optional */
+       void (*event_unmapped)          (struct perf_event *event, struct mm_struct *mm); /* optional */
 
        /*
         * Flags for ->add()/->del()/ ->start()/->stop(). There are
index 4d179316e43108f3a14ef48659a1d3a94ac320fb..719582744a2e873a978dbf9a925e143ec29243da 100644 (file)
@@ -8,7 +8,9 @@ enum pid_type
        PIDTYPE_PID,
        PIDTYPE_PGID,
        PIDTYPE_SID,
-       PIDTYPE_MAX
+       PIDTYPE_MAX,
+       /* only valid to __task_pid_nr_ns() */
+       __PIDTYPE_TGID
 };
 
 /*
index 79b0e4cdb8141a10e73affb9a837d5bcc5f00a2a..f8274b0c68880ccbd02de1a85ece46a3cab2053f 100644 (file)
  *     Available only for accelerometer and pressure sensors.
  *     Accelerometer DRDY on LSM330 available only on pin 1 (see datasheet).
  * @open_drain: set the interrupt line to be open drain if possible.
+ * @spi_3wire: enable spi-3wire mode.
  */
 struct st_sensors_platform_data {
        u8 drdy_int_pin;
        bool open_drain;
+       bool spi_3wire;
 };
 
 #endif /* ST_SENSORS_PDATA_H */
index d8c97ec8a8e66d865f516f708a5b3f8fe567da29..37b4bb2545b32dc82633df6b76e629b9999f900b 100644 (file)
@@ -436,9 +436,9 @@ static inline int ptr_ring_consume_batched_bh(struct ptr_ring *r,
        __PTR_RING_PEEK_CALL_v; \
 })
 
-static inline void **__ptr_ring_init_queue_alloc(int size, gfp_t gfp)
+static inline void **__ptr_ring_init_queue_alloc(unsigned int size, gfp_t gfp)
 {
-       return kzalloc(ALIGN(size * sizeof(void *), SMP_CACHE_BYTES), gfp);
+       return kcalloc(size, sizeof(void *), gfp);
 }
 
 static inline void __ptr_ring_set_size(struct ptr_ring *r, int size)
@@ -582,7 +582,8 @@ static inline int ptr_ring_resize(struct ptr_ring *r, int size, gfp_t gfp,
  * In particular if you consume ring in interrupt or BH context, you must
  * disable interrupts/BH when doing so.
  */
-static inline int ptr_ring_resize_multiple(struct ptr_ring **rings, int nrings,
+static inline int ptr_ring_resize_multiple(struct ptr_ring **rings,
+                                          unsigned int nrings,
                                           int size,
                                           gfp_t gfp, void (*destroy)(void *))
 {
@@ -590,7 +591,7 @@ static inline int ptr_ring_resize_multiple(struct ptr_ring **rings, int nrings,
        void ***queues;
        int i;
 
-       queues = kmalloc(nrings * sizeof *queues, gfp);
+       queues = kmalloc_array(nrings, sizeof(*queues), gfp);
        if (!queues)
                goto noqueues;
 
index 772c5f643764b2de98a27980de9e093f2940198f..93be319e0cbf7eb83d4c61b3f67ade790b66ef03 100644 (file)
@@ -1173,13 +1173,6 @@ static inline pid_t task_tgid_nr(struct task_struct *tsk)
        return tsk->tgid;
 }
 
-extern pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns);
-
-static inline pid_t task_tgid_vnr(struct task_struct *tsk)
-{
-       return pid_vnr(task_tgid(tsk));
-}
-
 /**
  * pid_alive - check that a task structure is not stale
  * @p: Task structure to be checked.
@@ -1195,23 +1188,6 @@ static inline int pid_alive(const struct task_struct *p)
        return p->pids[PIDTYPE_PID].pid != NULL;
 }
 
-static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns)
-{
-       pid_t pid = 0;
-
-       rcu_read_lock();
-       if (pid_alive(tsk))
-               pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns);
-       rcu_read_unlock();
-
-       return pid;
-}
-
-static inline pid_t task_ppid_nr(const struct task_struct *tsk)
-{
-       return task_ppid_nr_ns(tsk, &init_pid_ns);
-}
-
 static inline pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns)
 {
        return __task_pid_nr_ns(tsk, PIDTYPE_PGID, ns);
@@ -1233,6 +1209,33 @@ static inline pid_t task_session_vnr(struct task_struct *tsk)
        return __task_pid_nr_ns(tsk, PIDTYPE_SID, NULL);
 }
 
+static inline pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns)
+{
+       return __task_pid_nr_ns(tsk, __PIDTYPE_TGID, ns);
+}
+
+static inline pid_t task_tgid_vnr(struct task_struct *tsk)
+{
+       return __task_pid_nr_ns(tsk, __PIDTYPE_TGID, NULL);
+}
+
+static inline pid_t task_ppid_nr_ns(const struct task_struct *tsk, struct pid_namespace *ns)
+{
+       pid_t pid = 0;
+
+       rcu_read_lock();
+       if (pid_alive(tsk))
+               pid = task_tgid_nr_ns(rcu_dereference(tsk->real_parent), ns);
+       rcu_read_unlock();
+
+       return pid;
+}
+
+static inline pid_t task_ppid_nr(const struct task_struct *tsk)
+{
+       return task_ppid_nr_ns(tsk, &init_pid_ns);
+}
+
 /* Obsolete, do not use: */
 static inline pid_t task_pgrp_nr(struct task_struct *tsk)
 {
index 35226cd4efb0f32afcf38cc4a97441e9524424f0..8621ffdeecbf0542e57c2d7fee12f53546af7774 100644 (file)
@@ -193,7 +193,8 @@ static inline int skb_array_resize(struct skb_array *a, int size, gfp_t gfp)
 }
 
 static inline int skb_array_resize_multiple(struct skb_array **rings,
-                                           int nrings, int size, gfp_t gfp)
+                                           int nrings, unsigned int size,
+                                           gfp_t gfp)
 {
        BUILD_BUG_ON(offsetof(struct skb_array, ring));
        return ptr_ring_resize_multiple((struct ptr_ring **)rings,
index 5b74e36c0ca896481acd04bbea92be002ad4ba64..dc19880c02f5eb9091d185c567b961754cfa46e2 100644 (file)
@@ -757,6 +757,43 @@ extern int do_wait_intr_irq(wait_queue_head_t *, wait_queue_entry_t *);
        __ret;                                                                  \
 })
 
+#define __wait_event_killable_timeout(wq_head, condition, timeout)             \
+       ___wait_event(wq_head, ___wait_cond_timeout(condition),                 \
+                     TASK_KILLABLE, 0, timeout,                                \
+                     __ret = schedule_timeout(__ret))
+
+/**
+ * wait_event_killable_timeout - sleep until a condition gets true or a timeout elapses
+ * @wq_head: the waitqueue to wait on
+ * @condition: a C expression for the event to wait for
+ * @timeout: timeout, in jiffies
+ *
+ * The process is put to sleep (TASK_KILLABLE) until the
+ * @condition evaluates to true or a kill signal is received.
+ * The @condition is checked each time the waitqueue @wq_head is woken up.
+ *
+ * wake_up() has to be called after changing any variable that could
+ * change the result of the wait condition.
+ *
+ * Returns:
+ * 0 if the @condition evaluated to %false after the @timeout elapsed,
+ * 1 if the @condition evaluated to %true after the @timeout elapsed,
+ * the remaining jiffies (at least 1) if the @condition evaluated
+ * to %true before the @timeout elapsed, or -%ERESTARTSYS if it was
+ * interrupted by a kill signal.
+ *
+ * Only kill signals interrupt this process.
+ */
+#define wait_event_killable_timeout(wq_head, condition, timeout)               \
+({                                                                             \
+       long __ret = timeout;                                                   \
+       might_sleep();                                                          \
+       if (!___wait_cond_timeout(condition))                                   \
+               __ret = __wait_event_killable_timeout(wq_head,                  \
+                                               condition, timeout);            \
+       __ret;                                                                  \
+})
+
 
 #define __wait_event_lock_irq(wq_head, condition, lock, cmd)                   \
        (void)___wait_event(wq_head, condition, TASK_UNINTERRUPTIBLE, 0, 0,     \
index 6df79e96a780b3bf0c141fee5b43eded5276775c..f44ff2476758e27ad9f9cc20d96d96f16e5ee98c 100644 (file)
@@ -336,6 +336,16 @@ static inline void in6_dev_put(struct inet6_dev *idev)
                in6_dev_finish_destroy(idev);
 }
 
+static inline void in6_dev_put_clear(struct inet6_dev **pidev)
+{
+       struct inet6_dev *idev = *pidev;
+
+       if (idev) {
+               in6_dev_put(idev);
+               *pidev = NULL;
+       }
+}
+
 static inline void __in6_dev_put(struct inet6_dev *idev)
 {
        refcount_dec(&idev->refcnt);
index b00508d22e0a6adc37ee4e69187377fea3bc102d..b2e68657a2162cd55b957c9230779f5d066fe6f3 100644 (file)
@@ -277,6 +277,11 @@ static inline bool bond_is_lb(const struct bonding *bond)
               BOND_MODE(bond) == BOND_MODE_ALB;
 }
 
+static inline bool bond_needs_speed_duplex(const struct bonding *bond)
+{
+       return BOND_MODE(bond) == BOND_MODE_8023AD || bond_is_lb(bond);
+}
+
 static inline bool bond_is_nondyn_tlb(const struct bonding *bond)
 {
        return (BOND_MODE(bond) == BOND_MODE_TLB)  &&
index 8ffd434676b7a270af73534f3dbcc26f370fe215..71c72a939bf8b21c5102994ea68796dc89c4f097 100644 (file)
 #include <linux/sched/signal.h>
 #include <net/ip.h>
 
-#ifdef CONFIG_NET_RX_BUSY_POLL
-
-struct napi_struct;
-extern unsigned int sysctl_net_busy_read __read_mostly;
-extern unsigned int sysctl_net_busy_poll __read_mostly;
-
 /*             0 - Reserved to indicate value not set
  *     1..NR_CPUS - Reserved for sender_cpu
  *  NR_CPUS+1..~0 - Region available for NAPI IDs
  */
 #define MIN_NAPI_ID ((unsigned int)(NR_CPUS + 1))
 
+#ifdef CONFIG_NET_RX_BUSY_POLL
+
+struct napi_struct;
+extern unsigned int sysctl_net_busy_read __read_mostly;
+extern unsigned int sysctl_net_busy_poll __read_mostly;
+
 static inline bool net_busy_loop_on(void)
 {
        return sysctl_net_busy_poll;
index 821cedcc8e73b68af72d1918242c25d757c63d24..0cf7f5a65fe6be2be30259aa5cd251d02de489d5 100644 (file)
@@ -352,7 +352,7 @@ static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
            !forwarding)
                return dst_mtu(dst);
 
-       return min(dst->dev->mtu, IP_MAX_MTU);
+       return min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU);
 }
 
 static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
@@ -364,7 +364,7 @@ static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
                return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
        }
 
-       return min(skb_dst(skb)->dev->mtu, IP_MAX_MTU);
+       return min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
 }
 
 u32 ip_idents_reserve(u32 hash, int segs);
index b2b5419467cc123b4a2b70c2d926b4576651248d..f8149ca192b430c181406fa9a3e4db8fc0a17252 100644 (file)
@@ -5499,6 +5499,21 @@ static inline void ieee80211_stop_rx_ba_session_offl(struct ieee80211_vif *vif,
        ieee80211_manage_rx_ba_offl(vif, addr, tid + IEEE80211_NUM_TIDS);
 }
 
+/**
+ * ieee80211_rx_ba_timer_expired - stop a Rx BA session due to timeout
+ *
+ * Some device drivers do not offload AddBa/DelBa negotiation, but handle rx
+ * buffer reording internally, and therefore also handle the session timer.
+ *
+ * Trigger the timeout flow, which sends a DelBa.
+ *
+ * @vif: &struct ieee80211_vif pointer from the add_interface callback
+ * @addr: station mac address
+ * @tid: the rx tid
+ */
+void ieee80211_rx_ba_timer_expired(struct ieee80211_vif *vif,
+                                  const u8 *addr, unsigned int tid);
+
 /* Rate control API */
 
 /**
index 1c123e2b2415797e7d32a0248df817cb9fbcb0c9..67f815e5d52517390226bc3531b1ea7b5f1020bc 100644 (file)
@@ -806,8 +806,11 @@ static inline struct Qdisc *qdisc_replace(struct Qdisc *sch, struct Qdisc *new,
        old = *pold;
        *pold = new;
        if (old != NULL) {
-               qdisc_tree_reduce_backlog(old, old->q.qlen, old->qstats.backlog);
+               unsigned int qlen = old->q.qlen;
+               unsigned int backlog = old->qstats.backlog;
+
                qdisc_reset(old);
+               qdisc_tree_reduce_backlog(old, qlen, backlog);
        }
        sch_tree_unlock(sch);
 
index 7c0632c7e87043ca18fe5d32d7d55792f75ca6e8..aeeec62992ca7dc5ff80f8d7164a1c143f606b03 100644 (file)
@@ -507,9 +507,7 @@ int sk_set_peek_off(struct sock *sk, int val);
 static inline int sk_peek_offset(struct sock *sk, int flags)
 {
        if (unlikely(flags & MSG_PEEK)) {
-               s32 off = READ_ONCE(sk->sk_peek_off);
-               if (off >= 0)
-                       return off;
+               return READ_ONCE(sk->sk_peek_off);
        }
 
        return 0;
index cc8036987dcb885012c6c5eda0fb2bed2e588841..586de4b811b5678c72a045a3ede9852164b9b3e2 100644 (file)
@@ -366,12 +366,13 @@ static inline bool udp_skb_is_linear(struct sk_buff *skb)
 static inline int copy_linear_skb(struct sk_buff *skb, int len, int off,
                                  struct iov_iter *to)
 {
-       int n, copy = len - off;
+       int n;
 
-       n = copy_to_iter(skb->data + off, copy, to);
-       if (n == copy)
+       n = copy_to_iter(skb->data + off, len, to);
+       if (n == len)
                return 0;
 
+       iov_iter_revert(to, n);
        return -EFAULT;
 }
 
index b5732432bb297dbf6067ae34b16c00073cb2949d..88c32aba32f71b15370e2d4b5cc724c31feed9d3 100644 (file)
@@ -1683,6 +1683,7 @@ struct ib_qp {
        enum ib_qp_type         qp_type;
        struct ib_rwq_ind_table *rwq_ind_tbl;
        struct ib_qp_security  *qp_sec;
+       u8                      port;
 };
 
 struct ib_mr {
index a1266d318c855dec12dbb41a6bb1ab9e4dc27f08..6af198d8120b098f31d674f15584ec7114832743 100644 (file)
@@ -57,6 +57,7 @@ struct scsi_pointer {
 /* for scmd->flags */
 #define SCMD_TAGGED            (1 << 0)
 #define SCMD_UNCHECKED_ISA_DMA (1 << 1)
+#define SCMD_ZONE_WRITE_LOCK   (1 << 2)
 
 struct scsi_cmnd {
        struct scsi_request req;
index 0ca1fb08805b254fa8ffda73f4decd4c48ac2211..fb87d32f5e513de3c2a1b7c6f402b5f48e401c45 100644 (file)
@@ -786,6 +786,7 @@ struct iscsi_np {
        int                     np_sock_type;
        enum np_thread_state_table np_thread_state;
        bool                    enabled;
+       atomic_t                np_reset_count;
        enum iscsi_timer_flags_table np_login_timer_flags;
        u32                     np_exports;
        enum np_flags_table     np_flags;
index 62d686d965813aeb4b8ef34877e886e08bd71912..9eb8b3511636e96e0607be2484119823b1034fb5 100644 (file)
@@ -66,7 +66,7 @@ static struct fsnotify_group *audit_watch_group;
 
 /* fsnotify events we care about. */
 #define AUDIT_FS_WATCH (FS_MOVE | FS_CREATE | FS_DELETE | FS_DELETE_SELF |\
-                       FS_MOVE_SELF | FS_EVENT_ON_CHILD)
+                       FS_MOVE_SELF | FS_EVENT_ON_CHILD | FS_UNMOUNT)
 
 static void audit_free_parent(struct audit_parent *parent)
 {
@@ -457,13 +457,15 @@ void audit_remove_watch_rule(struct audit_krule *krule)
        list_del(&krule->rlist);
 
        if (list_empty(&watch->rules)) {
+               /*
+                * audit_remove_watch() drops our reference to 'parent' which
+                * can get freed. Grab our own reference to be safe.
+                */
+               audit_get_parent(parent);
                audit_remove_watch(watch);
-
-               if (list_empty(&parent->watches)) {
-                       audit_get_parent(parent);
+               if (list_empty(&parent->watches))
                        fsnotify_destroy_mark(&parent->mark, audit_watch_group);
-                       audit_put_parent(parent);
-               }
+               audit_put_parent(parent);
        }
 }
 
index 426c2ffba16d4ce1474a798a4c43d78d5abedb1e..ee20d4c546b5ebc0248c084e11c483e0ef800c6f 100644 (file)
@@ -2217,6 +2217,33 @@ static int group_can_go_on(struct perf_event *event,
        return can_add_hw;
 }
 
+/*
+ * Complement to update_event_times(). This computes the tstamp_* values to
+ * continue 'enabled' state from @now, and effectively discards the time
+ * between the prior tstamp_stopped and now (as we were in the OFF state, or
+ * just switched (context) time base).
+ *
+ * This further assumes '@event->state == INACTIVE' (we just came from OFF) and
+ * cannot have been scheduled in yet. And going into INACTIVE state means
+ * '@event->tstamp_stopped = @now'.
+ *
+ * Thus given the rules of update_event_times():
+ *
+ *   total_time_enabled = tstamp_stopped - tstamp_enabled
+ *   total_time_running = tstamp_stopped - tstamp_running
+ *
+ * We can insert 'tstamp_stopped == now' and reverse them to compute new
+ * tstamp_* values.
+ */
+static void __perf_event_enable_time(struct perf_event *event, u64 now)
+{
+       WARN_ON_ONCE(event->state != PERF_EVENT_STATE_INACTIVE);
+
+       event->tstamp_stopped = now;
+       event->tstamp_enabled = now - event->total_time_enabled;
+       event->tstamp_running = now - event->total_time_running;
+}
+
 static void add_event_to_ctx(struct perf_event *event,
                               struct perf_event_context *ctx)
 {
@@ -2224,9 +2251,12 @@ static void add_event_to_ctx(struct perf_event *event,
 
        list_add_event(event, ctx);
        perf_group_attach(event);
-       event->tstamp_enabled = tstamp;
-       event->tstamp_running = tstamp;
-       event->tstamp_stopped = tstamp;
+       /*
+        * We can be called with event->state == STATE_OFF when we create with
+        * .disabled = 1. In that case the IOC_ENABLE will call this function.
+        */
+       if (event->state == PERF_EVENT_STATE_INACTIVE)
+               __perf_event_enable_time(event, tstamp);
 }
 
 static void ctx_sched_out(struct perf_event_context *ctx,
@@ -2471,10 +2501,11 @@ static void __perf_event_mark_enabled(struct perf_event *event)
        u64 tstamp = perf_event_time(event);
 
        event->state = PERF_EVENT_STATE_INACTIVE;
-       event->tstamp_enabled = tstamp - event->total_time_enabled;
+       __perf_event_enable_time(event, tstamp);
        list_for_each_entry(sub, &event->sibling_list, group_entry) {
+               /* XXX should not be > INACTIVE if event isn't */
                if (sub->state >= PERF_EVENT_STATE_INACTIVE)
-                       sub->tstamp_enabled = tstamp - sub->total_time_enabled;
+                       __perf_event_enable_time(sub, tstamp);
        }
 }
 
@@ -5090,7 +5121,7 @@ static void perf_mmap_open(struct vm_area_struct *vma)
                atomic_inc(&event->rb->aux_mmap_count);
 
        if (event->pmu->event_mapped)
-               event->pmu->event_mapped(event);
+               event->pmu->event_mapped(event, vma->vm_mm);
 }
 
 static void perf_pmu_output_stop(struct perf_event *event);
@@ -5113,7 +5144,7 @@ static void perf_mmap_close(struct vm_area_struct *vma)
        unsigned long size = perf_data_size(rb);
 
        if (event->pmu->event_unmapped)
-               event->pmu->event_unmapped(event);
+               event->pmu->event_unmapped(event, vma->vm_mm);
 
        /*
         * rb->aux_mmap_count will always drop before rb->mmap_count and
@@ -5411,7 +5442,7 @@ aux_unlock:
        vma->vm_ops = &perf_mmap_vmops;
 
        if (event->pmu->event_mapped)
-               event->pmu->event_mapped(event);
+               event->pmu->event_mapped(event, vma->vm_mm);
 
        return ret;
 }
index a3cc37c0c85e2267497f650611b656af90aaf5aa..3675c6004f2a68c9d60493c7c46e9c46a0baa73f 100644 (file)
@@ -1000,7 +1000,7 @@ EXPORT_SYMBOL_GPL(irq_set_chip_and_handler_name);
 
 void irq_modify_status(unsigned int irq, unsigned long clr, unsigned long set)
 {
-       unsigned long flags;
+       unsigned long flags, trigger, tmp;
        struct irq_desc *desc = irq_get_desc_lock(irq, &flags, 0);
 
        if (!desc)
@@ -1014,6 +1014,8 @@ void irq_modify_status(unsigned int irq, unsigned long clr, unsigned long set)
 
        irq_settings_clr_and_set(desc, clr, set);
 
+       trigger = irqd_get_trigger_type(&desc->irq_data);
+
        irqd_clear(&desc->irq_data, IRQD_NO_BALANCING | IRQD_PER_CPU |
                   IRQD_TRIGGER_MASK | IRQD_LEVEL | IRQD_MOVE_PCNTXT);
        if (irq_settings_has_no_balance_set(desc))
@@ -1025,7 +1027,11 @@ void irq_modify_status(unsigned int irq, unsigned long clr, unsigned long set)
        if (irq_settings_is_level(desc))
                irqd_set(&desc->irq_data, IRQD_LEVEL);
 
-       irqd_set(&desc->irq_data, irq_settings_get_trigger_mask(desc));
+       tmp = irq_settings_get_trigger_mask(desc);
+       if (tmp != IRQ_TYPE_NONE)
+               trigger = tmp;
+
+       irqd_set(&desc->irq_data, trigger);
 
        irq_put_desc_unlock(desc, flags);
 }
index 1a9abc1c8ea0046d6491d1786645f1c1e4e1eee2..259a22aa9934cf9d67eb779e5e3eae22db0a3412 100644 (file)
@@ -165,7 +165,7 @@ irq_hw_number_t ipi_get_hwirq(unsigned int irq, unsigned int cpu)
        struct irq_data *data = irq_get_irq_data(irq);
        struct cpumask *ipimask = data ? irq_data_get_affinity_mask(data) : NULL;
 
-       if (!data || !ipimask || cpu > nr_cpu_ids)
+       if (!data || !ipimask || cpu >= nr_cpu_ids)
                return INVALID_HWIRQ;
 
        if (!cpumask_test_cpu(cpu, ipimask))
@@ -195,7 +195,7 @@ static int ipi_send_verify(struct irq_chip *chip, struct irq_data *data,
        if (!chip->ipi_send_single && !chip->ipi_send_mask)
                return -EINVAL;
 
-       if (cpu > nr_cpu_ids)
+       if (cpu >= nr_cpu_ids)
                return -EINVAL;
 
        if (dest) {
index 6d016c5d97c8390f561e6a9456a4e0c1ca59bffe..2f37acde640b6c6e2ce73933a52dbad65f014cf3 100644 (file)
@@ -70,6 +70,18 @@ static DECLARE_RWSEM(umhelper_sem);
 static atomic_t kmod_concurrent_max = ATOMIC_INIT(MAX_KMOD_CONCURRENT);
 static DECLARE_WAIT_QUEUE_HEAD(kmod_wq);
 
+/*
+ * This is a restriction on having *all* MAX_KMOD_CONCURRENT threads
+ * running at the same time without returning. When this happens we
+ * believe you've somehow ended up with a recursive module dependency
+ * creating a loop.
+ *
+ * We have no option but to fail.
+ *
+ * Userspace should proactively try to detect and prevent these.
+ */
+#define MAX_KMOD_ALL_BUSY_TIMEOUT 5
+
 /*
        modprobe_path is set via /proc/sys.
 */
@@ -167,8 +179,17 @@ int __request_module(bool wait, const char *fmt, ...)
                pr_warn_ratelimited("request_module: kmod_concurrent_max (%u) close to 0 (max_modprobes: %u), for module %s, throttling...",
                                    atomic_read(&kmod_concurrent_max),
                                    MAX_KMOD_CONCURRENT, module_name);
-               wait_event_interruptible(kmod_wq,
-                                        atomic_dec_if_positive(&kmod_concurrent_max) >= 0);
+               ret = wait_event_killable_timeout(kmod_wq,
+                                                 atomic_dec_if_positive(&kmod_concurrent_max) >= 0,
+                                                 MAX_KMOD_ALL_BUSY_TIMEOUT * HZ);
+               if (!ret) {
+                       pr_warn_ratelimited("request_module: modprobe %s cannot be processed, kmod busy with %d threads for more than %d seconds now",
+                                           module_name, MAX_KMOD_CONCURRENT, MAX_KMOD_ALL_BUSY_TIMEOUT);
+                       return -ETIME;
+               } else if (ret == -ERESTARTSYS) {
+                       pr_warn_ratelimited("request_module: sigkill sent for modprobe %s, giving up", module_name);
+                       return ret;
+               }
        }
 
        trace_module_request(module_name, wait, _RET_IP_);
index c69c30d827e5a2fc2605c6449617eb3754d8479b..020dedbdf066bccbc370cba20be8dc7dbc914629 100644 (file)
@@ -527,8 +527,11 @@ pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
        if (!ns)
                ns = task_active_pid_ns(current);
        if (likely(pid_alive(task))) {
-               if (type != PIDTYPE_PID)
+               if (type != PIDTYPE_PID) {
+                       if (type == __PIDTYPE_TGID)
+                               type = PIDTYPE_PID;
                        task = task->group_leader;
+               }
                nr = pid_nr_ns(rcu_dereference(task->pids[type].pid), ns);
        }
        rcu_read_unlock();
@@ -537,12 +540,6 @@ pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type,
 }
 EXPORT_SYMBOL(__task_pid_nr_ns);
 
-pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns)
-{
-       return pid_nr_ns(task_tgid(tsk), ns);
-}
-EXPORT_SYMBOL(task_tgid_nr_ns);
-
 struct pid_namespace *task_active_pid_ns(struct task_struct *tsk)
 {
        return ns_of_pid(task_pid(tsk));
index 7e33f8c583e64c91d7cb6ac1fb330b01299b7be2..ed804a470dcd151c18915f956c2f325b6d22bb0f 100644 (file)
@@ -1194,7 +1194,11 @@ force_sig_info(int sig, struct siginfo *info, struct task_struct *t)
                        recalc_sigpending_and_wake(t);
                }
        }
-       if (action->sa.sa_handler == SIG_DFL)
+       /*
+        * Don't clear SIGNAL_UNKILLABLE for traced tasks, users won't expect
+        * debugging to leave init killable.
+        */
+       if (action->sa.sa_handler == SIG_DFL && !t->ptrace)
                t->signal->flags &= ~SIGNAL_UNKILLABLE;
        ret = specific_send_sig_info(sig, info, t);
        spin_unlock_irqrestore(&t->sighand->siglock, flags);
index 37385193a6084ed1b8fdd794eb0938c2021fcfa4..dc498b605d5dd36137eaba7bd0ee93da72a36c33 100644 (file)
@@ -204,10 +204,36 @@ BPF_CALL_5(bpf_trace_printk, char *, fmt, u32, fmt_size, u64, arg1,
                fmt_cnt++;
        }
 
-       return __trace_printk(1/* fake ip will not be printed */, fmt,
-                             mod[0] == 2 ? arg1 : mod[0] == 1 ? (long) arg1 : (u32) arg1,
-                             mod[1] == 2 ? arg2 : mod[1] == 1 ? (long) arg2 : (u32) arg2,
-                             mod[2] == 2 ? arg3 : mod[2] == 1 ? (long) arg3 : (u32) arg3);
+/* Horrid workaround for getting va_list handling working with different
+ * argument type combinations generically for 32 and 64 bit archs.
+ */
+#define __BPF_TP_EMIT()        __BPF_ARG3_TP()
+#define __BPF_TP(...)                                                  \
+       __trace_printk(1 /* Fake ip will not be printed. */,            \
+                      fmt, ##__VA_ARGS__)
+
+#define __BPF_ARG1_TP(...)                                             \
+       ((mod[0] == 2 || (mod[0] == 1 && __BITS_PER_LONG == 64))        \
+         ? __BPF_TP(arg1, ##__VA_ARGS__)                               \
+         : ((mod[0] == 1 || (mod[0] == 0 && __BITS_PER_LONG == 32))    \
+             ? __BPF_TP((long)arg1, ##__VA_ARGS__)                     \
+             : __BPF_TP((u32)arg1, ##__VA_ARGS__)))
+
+#define __BPF_ARG2_TP(...)                                             \
+       ((mod[1] == 2 || (mod[1] == 1 && __BITS_PER_LONG == 64))        \
+         ? __BPF_ARG1_TP(arg2, ##__VA_ARGS__)                          \
+         : ((mod[1] == 1 || (mod[1] == 0 && __BITS_PER_LONG == 32))    \
+             ? __BPF_ARG1_TP((long)arg2, ##__VA_ARGS__)                \
+             : __BPF_ARG1_TP((u32)arg2, ##__VA_ARGS__)))
+
+#define __BPF_ARG3_TP(...)                                             \
+       ((mod[2] == 2 || (mod[2] == 1 && __BITS_PER_LONG == 64))        \
+         ? __BPF_ARG2_TP(arg3, ##__VA_ARGS__)                          \
+         : ((mod[2] == 1 || (mod[2] == 0 && __BITS_PER_LONG == 32))    \
+             ? __BPF_ARG2_TP((long)arg3, ##__VA_ARGS__)                \
+             : __BPF_ARG2_TP((u32)arg3, ##__VA_ARGS__)))
+
+       return __BPF_TP_EMIT();
 }
 
 static const struct bpf_func_proto bpf_trace_printk_proto = {
index 02004ae918608f915b28341b6ed1b3b7f551c5a5..96cea88fa00fcdab41db0a2054ef6cfe50a778aa 100644 (file)
@@ -889,6 +889,10 @@ static int profile_graph_entry(struct ftrace_graph_ent *trace)
 
        function_profile_call(trace->func, 0, NULL, NULL);
 
+       /* If function graph is shutting down, ret_stack can be NULL */
+       if (!current->ret_stack)
+               return 0;
+
        if (index >= 0 && index < FTRACE_RETFUNC_DEPTH)
                current->ret_stack[index].subtime = 0;
 
index 529cc50d7243d6c1007e517bcdc83a5312f7ae6b..81279c6602ff1753d8c38c8def566c3576578962 100644 (file)
@@ -4386,15 +4386,19 @@ EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu);
  * the page that was allocated, with the read page of the buffer.
  *
  * Returns:
- *  The page allocated, or NULL on error.
+ *  The page allocated, or ERR_PTR
  */
 void *ring_buffer_alloc_read_page(struct ring_buffer *buffer, int cpu)
 {
-       struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
+       struct ring_buffer_per_cpu *cpu_buffer;
        struct buffer_data_page *bpage = NULL;
        unsigned long flags;
        struct page *page;
 
+       if (!cpumask_test_cpu(cpu, buffer->cpumask))
+               return ERR_PTR(-ENODEV);
+
+       cpu_buffer = buffer->buffers[cpu];
        local_irq_save(flags);
        arch_spin_lock(&cpu_buffer->lock);
 
@@ -4412,7 +4416,7 @@ void *ring_buffer_alloc_read_page(struct ring_buffer *buffer, int cpu)
        page = alloc_pages_node(cpu_to_node(cpu),
                                GFP_KERNEL | __GFP_NORETRY, 0);
        if (!page)
-               return NULL;
+               return ERR_PTR(-ENOMEM);
 
        bpage = page_address(page);
 
@@ -4467,8 +4471,8 @@ EXPORT_SYMBOL_GPL(ring_buffer_free_read_page);
  *
  * for example:
  *     rpage = ring_buffer_alloc_read_page(buffer, cpu);
- *     if (!rpage)
- *             return error;
+ *     if (IS_ERR(rpage))
+ *             return PTR_ERR(rpage);
  *     ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0);
  *     if (ret >= 0)
  *             process_page(rpage, ret);
index 9fbcaf56788626335bfd5412dffaf820a9b308bc..68ee79afe31c223218e8fab1ffd829b7be1226d3 100644 (file)
@@ -113,7 +113,7 @@ static enum event_status read_page(int cpu)
        int i;
 
        bpage = ring_buffer_alloc_read_page(buffer, cpu);
-       if (!bpage)
+       if (IS_ERR(bpage))
                return EVENT_DROPPED;
 
        ret = ring_buffer_read_page(buffer, &bpage, PAGE_SIZE, cpu, 1);
index 42b9355033d45d5a7c27bf0606799772e70d7421..44004d8aa3b33f259ed90dd13ce4c102034da23a 100644 (file)
@@ -6598,7 +6598,7 @@ tracing_buffers_read(struct file *filp, char __user *ubuf,
 {
        struct ftrace_buffer_info *info = filp->private_data;
        struct trace_iterator *iter = &info->iter;
-       ssize_t ret;
+       ssize_t ret = 0;
        ssize_t size;
 
        if (!count)
@@ -6612,10 +6612,15 @@ tracing_buffers_read(struct file *filp, char __user *ubuf,
        if (!info->spare) {
                info->spare = ring_buffer_alloc_read_page(iter->trace_buffer->buffer,
                                                          iter->cpu_file);
-               info->spare_cpu = iter->cpu_file;
+               if (IS_ERR(info->spare)) {
+                       ret = PTR_ERR(info->spare);
+                       info->spare = NULL;
+               } else {
+                       info->spare_cpu = iter->cpu_file;
+               }
        }
        if (!info->spare)
-               return -ENOMEM;
+               return ret;
 
        /* Do we have previous read data to read? */
        if (info->read < PAGE_SIZE)
@@ -6790,8 +6795,9 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos,
                ref->ref = 1;
                ref->buffer = iter->trace_buffer->buffer;
                ref->page = ring_buffer_alloc_read_page(ref->buffer, iter->cpu_file);
-               if (!ref->page) {
-                       ret = -ENOMEM;
+               if (IS_ERR(ref->page)) {
+                       ret = PTR_ERR(ref->page);
+                       ref->page = NULL;
                        kfree(ref);
                        break;
                }
@@ -8293,6 +8299,7 @@ __init static int tracer_alloc_buffers(void)
        if (ret < 0)
                goto out_free_cpumask;
        /* Used for event triggers */
+       ret = -ENOMEM;
        temp_buffer = ring_buffer_alloc(PAGE_SIZE, RB_FL_OVERWRITE);
        if (!temp_buffer)
                goto out_rm_hp_state;
@@ -8407,4 +8414,4 @@ __init static int clear_boot_tracer(void)
 }
 
 fs_initcall(tracer_init_tracefs);
-late_initcall(clear_boot_tracer);
+late_initcall_sync(clear_boot_tracer);
index 59a411ff60c709834121fc70363c74560dfe1ff5..181e139a8057bed695ad6d1d7460036e378f5ca0 100644 (file)
@@ -1959,6 +1959,10 @@ static int create_filter(struct trace_event_call *call,
                if (err && set_str)
                        append_filter_err(ps, filter);
        }
+       if (err && !set_str) {
+               free_event_filter(filter);
+               filter = NULL;
+       }
        create_filter_finish(ps);
 
        *filterp = filter;
index 0a689bbb78ef4fe48d9e55cb867e836aea87a726..305039b122fafba242f73b5982289ce4a12a6e20 100644 (file)
@@ -221,16 +221,19 @@ void tracing_map_array_free(struct tracing_map_array *a)
        if (!a)
                return;
 
-       if (!a->pages) {
-               kfree(a);
-               return;
-       }
+       if (!a->pages)
+               goto free;
 
        for (i = 0; i < a->n_pages; i++) {
                if (!a->pages[i])
                        break;
                free_page((unsigned long)a->pages[i]);
        }
+
+       kfree(a->pages);
+
+ free:
+       kfree(a);
 }
 
 struct tracing_map_array *tracing_map_array_alloc(unsigned int n_elts,
index 06d3389bca0df01c8a730f7b270767f6ed35a8a9..f5d52024f6b72a9d1354b1a44c12c3e3b6af06a9 100644 (file)
@@ -240,6 +240,7 @@ static void set_sample_period(void)
         * hardlockup detector generates a warning
         */
        sample_period = get_softlockup_thresh() * ((u64)NSEC_PER_SEC / 5);
+       watchdog_update_hrtimer_threshold(sample_period);
 }
 
 /* Commands for resetting the watchdog */
index 295a0d84934cb1d3a9a87abd4a8ac0f1d38656a5..3a09ea1b1d3d5e6e284d058052403ac1396804ca 100644 (file)
@@ -37,6 +37,62 @@ void arch_touch_nmi_watchdog(void)
 }
 EXPORT_SYMBOL(arch_touch_nmi_watchdog);
 
+#ifdef CONFIG_HARDLOCKUP_CHECK_TIMESTAMP
+static DEFINE_PER_CPU(ktime_t, last_timestamp);
+static DEFINE_PER_CPU(unsigned int, nmi_rearmed);
+static ktime_t watchdog_hrtimer_sample_threshold __read_mostly;
+
+void watchdog_update_hrtimer_threshold(u64 period)
+{
+       /*
+        * The hrtimer runs with a period of (watchdog_threshold * 2) / 5
+        *
+        * So it runs effectively with 2.5 times the rate of the NMI
+        * watchdog. That means the hrtimer should fire 2-3 times before
+        * the NMI watchdog expires. The NMI watchdog on x86 is based on
+        * unhalted CPU cycles, so if Turbo-Mode is enabled the CPU cycles
+        * might run way faster than expected and the NMI fires in a
+        * smaller period than the one deduced from the nominal CPU
+        * frequency. Depending on the Turbo-Mode factor this might be fast
+        * enough to get the NMI period smaller than the hrtimer watchdog
+        * period and trigger false positives.
+        *
+        * The sample threshold is used to check in the NMI handler whether
+        * the minimum time between two NMI samples has elapsed. That
+        * prevents false positives.
+        *
+        * Set this to 4/5 of the actual watchdog threshold period so the
+        * hrtimer is guaranteed to fire at least once within the real
+        * watchdog threshold.
+        */
+       watchdog_hrtimer_sample_threshold = period * 2;
+}
+
+static bool watchdog_check_timestamp(void)
+{
+       ktime_t delta, now = ktime_get_mono_fast_ns();
+
+       delta = now - __this_cpu_read(last_timestamp);
+       if (delta < watchdog_hrtimer_sample_threshold) {
+               /*
+                * If ktime is jiffies based, a stalled timer would prevent
+                * jiffies from being incremented and the filter would look
+                * at a stale timestamp and never trigger.
+                */
+               if (__this_cpu_inc_return(nmi_rearmed) < 10)
+                       return false;
+       }
+       __this_cpu_write(nmi_rearmed, 0);
+       __this_cpu_write(last_timestamp, now);
+       return true;
+}
+#else
+static inline bool watchdog_check_timestamp(void)
+{
+       return true;
+}
+#endif
+
 static struct perf_event_attr wd_hw_attr = {
        .type           = PERF_TYPE_HARDWARE,
        .config         = PERF_COUNT_HW_CPU_CYCLES,
@@ -61,6 +117,9 @@ static void watchdog_overflow_callback(struct perf_event *event,
                return;
        }
 
+       if (!watchdog_check_timestamp())
+               return;
+
        /* check for a hardlockup
         * This is done by making sure our timer interrupt
         * is incrementing.  The timer interrupt should have
index eb0f316b731abf57a94e0662d87fc179b0d57a88..e718df3cbd464b9b1f873c4224f149d5a6d68cb6 100644 (file)
@@ -797,6 +797,13 @@ config HARDLOCKUP_DETECTOR_PERF
        bool
        select SOFTLOCKUP_DETECTOR
 
+#
+# Enables a timestamp based low pass filter to compensate for perf based
+# hard lockup detection which runs too fast due to turbo modes.
+#
+config HARDLOCKUP_CHECK_TIMESTAMP
+       bool
+
 #
 # arch/ can define HAVE_HARDLOCKUP_DETECTOR_ARCH to provide their own hard
 # lockup detector rather than the perf based detector.
index 595b757bef72722fb2715afd9ab0cd1536181c9e..c03ccbc405a066038619361cf995f675dc2efaa1 100644 (file)
@@ -167,7 +167,7 @@ static void cma_debugfs_add_one(struct cma *cma, int idx)
        char name[16];
        int u32s;
 
-       sprintf(name, "cma-%s", cma->name);
+       scnprintf(name, sizeof(name), "cma-%s", cma->name);
 
        tmp = debugfs_create_dir(name, cma_debugfs_root);
 
index 08f6c19938328b7a1e946b1fa247a89570e70146..3644ff91843490a282f25f44fab8d3ed1d632433 100644 (file)
@@ -32,6 +32,7 @@
 #include <linux/userfaultfd_k.h>
 #include <linux/page_idle.h>
 #include <linux/shmem_fs.h>
+#include <linux/oom.h>
 
 #include <asm/tlb.h>
 #include <asm/pgalloc.h>
@@ -550,6 +551,7 @@ static int __do_huge_pmd_anonymous_page(struct vm_fault *vmf, struct page *page,
        struct mem_cgroup *memcg;
        pgtable_t pgtable;
        unsigned long haddr = vmf->address & HPAGE_PMD_MASK;
+       int ret = 0;
 
        VM_BUG_ON_PAGE(!PageCompound(page), page);
 
@@ -561,9 +563,8 @@ static int __do_huge_pmd_anonymous_page(struct vm_fault *vmf, struct page *page,
 
        pgtable = pte_alloc_one(vma->vm_mm, haddr);
        if (unlikely(!pgtable)) {
-               mem_cgroup_cancel_charge(page, memcg, true);
-               put_page(page);
-               return VM_FAULT_OOM;
+               ret = VM_FAULT_OOM;
+               goto release;
        }
 
        clear_huge_page(page, haddr, HPAGE_PMD_NR);
@@ -576,13 +577,14 @@ static int __do_huge_pmd_anonymous_page(struct vm_fault *vmf, struct page *page,
 
        vmf->ptl = pmd_lock(vma->vm_mm, vmf->pmd);
        if (unlikely(!pmd_none(*vmf->pmd))) {
-               spin_unlock(vmf->ptl);
-               mem_cgroup_cancel_charge(page, memcg, true);
-               put_page(page);
-               pte_free(vma->vm_mm, pgtable);
+               goto unlock_release;
        } else {
                pmd_t entry;
 
+               ret = check_stable_address_space(vma->vm_mm);
+               if (ret)
+                       goto unlock_release;
+
                /* Deliver the page fault to userland */
                if (userfaultfd_missing(vma)) {
                        int ret;
@@ -610,6 +612,15 @@ static int __do_huge_pmd_anonymous_page(struct vm_fault *vmf, struct page *page,
        }
 
        return 0;
+unlock_release:
+       spin_unlock(vmf->ptl);
+release:
+       if (pgtable)
+               pte_free(vma->vm_mm, pgtable);
+       mem_cgroup_cancel_charge(page, memcg, true);
+       put_page(page);
+       return ret;
+
 }
 
 /*
@@ -688,7 +699,10 @@ int do_huge_pmd_anonymous_page(struct vm_fault *vmf)
                ret = 0;
                set = false;
                if (pmd_none(*vmf->pmd)) {
-                       if (userfaultfd_missing(vma)) {
+                       ret = check_stable_address_space(vma->vm_mm);
+                       if (ret) {
+                               spin_unlock(vmf->ptl);
+                       } else if (userfaultfd_missing(vma)) {
                                spin_unlock(vmf->ptl);
                                ret = handle_userfault(vmf, VM_UFFD_MISSING);
                                VM_BUG_ON(ret & VM_FAULT_FALLBACK);
index 2cb25fe4452c279c5ff6ff74cbbfee64128d820e..bf14aea6ab709dc61666c1994718d9d244291e22 100644 (file)
@@ -285,31 +285,27 @@ static void __init_memblock memblock_remove_region(struct memblock_type *type, u
 }
 
 #ifdef CONFIG_ARCH_DISCARD_MEMBLOCK
-
-phys_addr_t __init_memblock get_allocated_memblock_reserved_regions_info(
-                                       phys_addr_t *addr)
-{
-       if (memblock.reserved.regions == memblock_reserved_init_regions)
-               return 0;
-
-       *addr = __pa(memblock.reserved.regions);
-
-       return PAGE_ALIGN(sizeof(struct memblock_region) *
-                         memblock.reserved.max);
-}
-
-phys_addr_t __init_memblock get_allocated_memblock_memory_regions_info(
-                                       phys_addr_t *addr)
+/**
+ * Discard memory and reserved arrays if they were allocated
+ */
+void __init memblock_discard(void)
 {
-       if (memblock.memory.regions == memblock_memory_init_regions)
-               return 0;
+       phys_addr_t addr, size;
 
-       *addr = __pa(memblock.memory.regions);
+       if (memblock.reserved.regions != memblock_reserved_init_regions) {
+               addr = __pa(memblock.reserved.regions);
+               size = PAGE_ALIGN(sizeof(struct memblock_region) *
+                                 memblock.reserved.max);
+               __memblock_free_late(addr, size);
+       }
 
-       return PAGE_ALIGN(sizeof(struct memblock_region) *
-                         memblock.memory.max);
+       if (memblock.memory.regions == memblock_memory_init_regions) {
+               addr = __pa(memblock.memory.regions);
+               size = PAGE_ALIGN(sizeof(struct memblock_region) *
+                                 memblock.memory.max);
+               __memblock_free_late(addr, size);
+       }
 }
-
 #endif
 
 /**
index 3df3c04d73ab08e3bbb663f25b2e195b396d2149..e09741af816f8a6d5343546ebd23eb7f25f8ab54 100644 (file)
@@ -1611,9 +1611,13 @@ cleanup:
  * @page: the page
  *
  * This function protects unlocked LRU pages from being moved to
- * another cgroup and stabilizes their page->mem_cgroup binding.
+ * another cgroup.
+ *
+ * It ensures lifetime of the returned memcg. Caller is responsible
+ * for the lifetime of the page; __unlock_page_memcg() is available
+ * when @page might get freed inside the locked section.
  */
-void lock_page_memcg(struct page *page)
+struct mem_cgroup *lock_page_memcg(struct page *page)
 {
        struct mem_cgroup *memcg;
        unsigned long flags;
@@ -1622,18 +1626,24 @@ void lock_page_memcg(struct page *page)
         * The RCU lock is held throughout the transaction.  The fast
         * path can get away without acquiring the memcg->move_lock
         * because page moving starts with an RCU grace period.
-        */
+        *
+        * The RCU lock also protects the memcg from being freed when
+        * the page state that is going to change is the only thing
+        * preventing the page itself from being freed. E.g. writeback
+        * doesn't hold a page reference and relies on PG_writeback to
+        * keep off truncation, migration and so forth.
+         */
        rcu_read_lock();
 
        if (mem_cgroup_disabled())
-               return;
+               return NULL;
 again:
        memcg = page->mem_cgroup;
        if (unlikely(!memcg))
-               return;
+               return NULL;
 
        if (atomic_read(&memcg->moving_account) <= 0)
-               return;
+               return memcg;
 
        spin_lock_irqsave(&memcg->move_lock, flags);
        if (memcg != page->mem_cgroup) {
@@ -1649,18 +1659,18 @@ again:
        memcg->move_lock_task = current;
        memcg->move_lock_flags = flags;
 
-       return;
+       return memcg;
 }
 EXPORT_SYMBOL(lock_page_memcg);
 
 /**
- * unlock_page_memcg - unlock a page->mem_cgroup binding
- * @page: the page
+ * __unlock_page_memcg - unlock and unpin a memcg
+ * @memcg: the memcg
+ *
+ * Unlock and unpin a memcg returned by lock_page_memcg().
  */
-void unlock_page_memcg(struct page *page)
+void __unlock_page_memcg(struct mem_cgroup *memcg)
 {
-       struct mem_cgroup *memcg = page->mem_cgroup;
-
        if (memcg && memcg->move_lock_task == current) {
                unsigned long flags = memcg->move_lock_flags;
 
@@ -1672,6 +1682,15 @@ void unlock_page_memcg(struct page *page)
 
        rcu_read_unlock();
 }
+
+/**
+ * unlock_page_memcg - unlock a page->mem_cgroup binding
+ * @page: the page
+ */
+void unlock_page_memcg(struct page *page)
+{
+       __unlock_page_memcg(page->mem_cgroup);
+}
 EXPORT_SYMBOL(unlock_page_memcg);
 
 /*
index e158f7ac67300b10b8827fe6825667506095f550..fe2fba27ded2fab229d0ef7a4908551343d31b89 100644 (file)
@@ -68,6 +68,7 @@
 #include <linux/debugfs.h>
 #include <linux/userfaultfd_k.h>
 #include <linux/dax.h>
+#include <linux/oom.h>
 
 #include <asm/io.h>
 #include <asm/mmu_context.h>
@@ -2893,6 +2894,7 @@ static int do_anonymous_page(struct vm_fault *vmf)
        struct vm_area_struct *vma = vmf->vma;
        struct mem_cgroup *memcg;
        struct page *page;
+       int ret = 0;
        pte_t entry;
 
        /* File mapping without ->vm_ops ? */
@@ -2925,6 +2927,9 @@ static int do_anonymous_page(struct vm_fault *vmf)
                                vmf->address, &vmf->ptl);
                if (!pte_none(*vmf->pte))
                        goto unlock;
+               ret = check_stable_address_space(vma->vm_mm);
+               if (ret)
+                       goto unlock;
                /* Deliver the page fault to userland, check inside PT lock */
                if (userfaultfd_missing(vma)) {
                        pte_unmap_unlock(vmf->pte, vmf->ptl);
@@ -2959,6 +2964,10 @@ static int do_anonymous_page(struct vm_fault *vmf)
        if (!pte_none(*vmf->pte))
                goto release;
 
+       ret = check_stable_address_space(vma->vm_mm);
+       if (ret)
+               goto release;
+
        /* Deliver the page fault to userland, check inside PT lock */
        if (userfaultfd_missing(vma)) {
                pte_unmap_unlock(vmf->pte, vmf->ptl);
@@ -2978,7 +2987,7 @@ setpte:
        update_mmu_cache(vma, vmf->address, vmf->pte);
 unlock:
        pte_unmap_unlock(vmf->pte, vmf->ptl);
-       return 0;
+       return ret;
 release:
        mem_cgroup_cancel_charge(page, memcg, false);
        put_page(page);
@@ -3252,7 +3261,7 @@ int alloc_set_pte(struct vm_fault *vmf, struct mem_cgroup *memcg,
 int finish_fault(struct vm_fault *vmf)
 {
        struct page *page;
-       int ret;
+       int ret = 0;
 
        /* Did we COW the page? */
        if ((vmf->flags & FAULT_FLAG_WRITE) &&
@@ -3260,7 +3269,15 @@ int finish_fault(struct vm_fault *vmf)
                page = vmf->cow_page;
        else
                page = vmf->page;
-       ret = alloc_set_pte(vmf, vmf->memcg, page);
+
+       /*
+        * check even for read faults because we might have lost our CoWed
+        * page
+        */
+       if (!(vmf->vma->vm_flags & VM_SHARED))
+               ret = check_stable_address_space(vmf->vma->vm_mm);
+       if (!ret)
+               ret = alloc_set_pte(vmf, vmf->memcg, page);
        if (vmf->pte)
                pte_unmap_unlock(vmf->pte, vmf->ptl);
        return ret;
@@ -3900,19 +3917,6 @@ int handle_mm_fault(struct vm_area_struct *vma, unsigned long address,
                        mem_cgroup_oom_synchronize(false);
        }
 
-       /*
-        * This mm has been already reaped by the oom reaper and so the
-        * refault cannot be trusted in general. Anonymous refaults would
-        * lose data and give a zero page instead e.g. This is especially
-        * problem for use_mm() because regular tasks will just die and
-        * the corrupted data will not be visible anywhere while kthread
-        * will outlive the oom victim and potentially propagate the date
-        * further.
-        */
-       if (unlikely((current->flags & PF_KTHREAD) && !(ret & VM_FAULT_ERROR)
-                               && test_bit(MMF_UNSTABLE, &vma->vm_mm->flags)))
-               ret = VM_FAULT_SIGBUS;
-
        return ret;
 }
 EXPORT_SYMBOL_GPL(handle_mm_fault);
index d911fa5cb2a73fe464042a59e1c2676c1129239d..618ab125228baec0810146a0638ce80ce4d50284 100644 (file)
@@ -861,11 +861,6 @@ static long do_get_mempolicy(int *policy, nodemask_t *nmask,
                *policy |= (pol->flags & MPOL_MODE_FLAGS);
        }
 
-       if (vma) {
-               up_read(&current->mm->mmap_sem);
-               vma = NULL;
-       }
-
        err = 0;
        if (nmask) {
                if (mpol_store_user_nodemask(pol)) {
index d68a41da6abb0743d6b09cc49c5c9524463715c3..e84eeb4e43566c7b1ee85e3759b1b60b72c1c532 100644 (file)
@@ -41,6 +41,7 @@
 #include <linux/page_idle.h>
 #include <linux/page_owner.h>
 #include <linux/sched/mm.h>
+#include <linux/ptrace.h>
 
 #include <asm/tlbflush.h>
 
@@ -1652,7 +1653,6 @@ SYSCALL_DEFINE6(move_pages, pid_t, pid, unsigned long, nr_pages,
                const int __user *, nodes,
                int __user *, status, int, flags)
 {
-       const struct cred *cred = current_cred(), *tcred;
        struct task_struct *task;
        struct mm_struct *mm;
        int err;
@@ -1676,14 +1676,9 @@ SYSCALL_DEFINE6(move_pages, pid_t, pid, unsigned long, nr_pages,
 
        /*
         * Check if this process has the right to modify the specified
-        * process. The right exists if the process has administrative
-        * capabilities, superuser privileges or the same
-        * userid as the target process.
+        * process. Use the regular "ptrace_may_access()" checks.
         */
-       tcred = __task_cred(task);
-       if (!uid_eq(cred->euid, tcred->suid) && !uid_eq(cred->euid, tcred->uid) &&
-           !uid_eq(cred->uid,  tcred->suid) && !uid_eq(cred->uid,  tcred->uid) &&
-           !capable(CAP_SYS_NICE)) {
+       if (!ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS)) {
                rcu_read_unlock();
                err = -EPERM;
                goto out;
index 36454d0f96ee6b91383554c83015a2b47e66f038..3637809a18d04f9c20d1b00d70687ae1eb00b282 100644 (file)
@@ -146,22 +146,6 @@ static unsigned long __init free_low_memory_core_early(void)
                                NULL)
                count += __free_memory_core(start, end);
 
-#ifdef CONFIG_ARCH_DISCARD_MEMBLOCK
-       {
-               phys_addr_t size;
-
-               /* Free memblock.reserved array if it was allocated */
-               size = get_allocated_memblock_reserved_regions_info(&start);
-               if (size)
-                       count += __free_memory_core(start, start + size);
-
-               /* Free memblock.memory array if it was allocated */
-               size = get_allocated_memblock_memory_regions_info(&start);
-               if (size)
-                       count += __free_memory_core(start, start + size);
-       }
-#endif
-
        return count;
 }
 
index 96e93b214d317baf4fb4ffbb5fcbe726e110980d..bf050ab025b76a268cd09a37173eac4f86febcb8 100644 (file)
@@ -2724,9 +2724,12 @@ EXPORT_SYMBOL(clear_page_dirty_for_io);
 int test_clear_page_writeback(struct page *page)
 {
        struct address_space *mapping = page_mapping(page);
+       struct mem_cgroup *memcg;
+       struct lruvec *lruvec;
        int ret;
 
-       lock_page_memcg(page);
+       memcg = lock_page_memcg(page);
+       lruvec = mem_cgroup_page_lruvec(page, page_pgdat(page));
        if (mapping && mapping_use_writeback_tags(mapping)) {
                struct inode *inode = mapping->host;
                struct backing_dev_info *bdi = inode_to_bdi(inode);
@@ -2754,12 +2757,18 @@ int test_clear_page_writeback(struct page *page)
        } else {
                ret = TestClearPageWriteback(page);
        }
+       /*
+        * NOTE: Page might be free now! Writeback doesn't hold a page
+        * reference on its own, it relies on truncation to wait for
+        * the clearing of PG_writeback. The below can only access
+        * page state that is static across allocation cycles.
+        */
        if (ret) {
-               dec_lruvec_page_state(page, NR_WRITEBACK);
+               dec_lruvec_state(lruvec, NR_WRITEBACK);
                dec_zone_page_state(page, NR_ZONE_WRITE_PENDING);
                inc_node_page_state(page, NR_WRITTEN);
        }
-       unlock_page_memcg(page);
+       __unlock_page_memcg(memcg);
        return ret;
 }
 
index 6acf612fdd990005e4a2dd3f05f0105ddf4f1513..471b0526b876db35a25901d89a1aa300b7f47bb3 100644 (file)
@@ -1585,6 +1585,10 @@ void __init page_alloc_init_late(void)
        /* Reinit limits that are based on free pages after the kernel is up */
        files_maxfiles_init();
 #endif
+#ifdef CONFIG_ARCH_DISCARD_MEMBLOCK
+       /* Discard memblock private memory */
+       memblock_discard();
+#endif
 
        for_each_populated_zone(zone)
                set_zone_contiguous(zone);
index 1d3f9835f4eabe91494f48d8ace08a5e42a895f9..e8b4e31162cae8c4d8e473ae2d78769aaa55089e 100644 (file)
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -5642,13 +5642,14 @@ static void sysfs_slab_remove_workfn(struct work_struct *work)
                 * A cache is never shut down before deactivation is
                 * complete, so no need to worry about synchronization.
                 */
-               return;
+               goto out;
 
 #ifdef CONFIG_MEMCG
        kset_unregister(s->memcg_kset);
 #endif
        kobject_uevent(&s->kobj, KOBJ_REMOVE);
        kobject_del(&s->kobj);
+out:
        kobject_put(&s->kobj);
 }
 
index 8698c1c86c4dbed685269eae1ecded2f5e714368..a47e3894c775646cd636c3f48bcb75fd4b8771d6 100644 (file)
@@ -1671,7 +1671,10 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
        struct page **pages;
        unsigned int nr_pages, array_size, i;
        const gfp_t nested_gfp = (gfp_mask & GFP_RECLAIM_MASK) | __GFP_ZERO;
-       const gfp_t alloc_mask = gfp_mask | __GFP_HIGHMEM | __GFP_NOWARN;
+       const gfp_t alloc_mask = gfp_mask | __GFP_NOWARN;
+       const gfp_t highmem_mask = (gfp_mask & (GFP_DMA | GFP_DMA32)) ?
+                                       0 :
+                                       __GFP_HIGHMEM;
 
        nr_pages = get_vm_area_size(area) >> PAGE_SHIFT;
        array_size = (nr_pages * sizeof(struct page *));
@@ -1679,7 +1682,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
        area->nr_pages = nr_pages;
        /* Please note that the recursion is strictly bounded. */
        if (array_size > PAGE_SIZE) {
-               pages = __vmalloc_node(array_size, 1, nested_gfp|__GFP_HIGHMEM,
+               pages = __vmalloc_node(array_size, 1, nested_gfp|highmem_mask,
                                PAGE_KERNEL, node, area->caller);
        } else {
                pages = kmalloc_node(array_size, nested_gfp, node);
@@ -1700,9 +1703,9 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
                }
 
                if (node == NUMA_NO_NODE)
-                       page = alloc_page(alloc_mask);
+                       page = alloc_page(alloc_mask|highmem_mask);
                else
-                       page = alloc_pages_node(node, alloc_mask, 0);
+                       page = alloc_pages_node(node, alloc_mask|highmem_mask, 0);
 
                if (unlikely(!page)) {
                        /* Successfully allocated i pages, free them in __vunmap() */
@@ -1710,7 +1713,7 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
                        goto fail;
                }
                area->pages[i] = page;
-               if (gfpflags_allow_blocking(gfp_mask))
+               if (gfpflags_allow_blocking(gfp_mask|highmem_mask))
                        cond_resched();
        }
 
index ee5647bd91b3f3a864dccdc46098b30e8561377d..a21ca8dee5eadca0d9ab7c78a939ac90bb3963b0 100644 (file)
@@ -169,14 +169,20 @@ struct sk_buff *__skb_try_recv_from_queue(struct sock *sk,
                                          int *peeked, int *off, int *err,
                                          struct sk_buff **last)
 {
+       bool peek_at_off = false;
        struct sk_buff *skb;
-       int _off = *off;
+       int _off = 0;
+
+       if (unlikely(flags & MSG_PEEK && *off >= 0)) {
+               peek_at_off = true;
+               _off = *off;
+       }
 
        *last = queue->prev;
        skb_queue_walk(queue, skb) {
                if (flags & MSG_PEEK) {
-                       if (_off >= skb->len && (skb->len || _off ||
-                                                skb->peeked)) {
+                       if (peek_at_off && _off >= skb->len &&
+                           (_off || skb->peeked)) {
                                _off -= skb->len;
                                continue;
                        }
index f44fc22fd45aca4941c618abf97eb48494a67c16..6280a602604c2e0e05c57d0c2104b5f4791f6928 100644 (file)
@@ -3505,6 +3505,7 @@ static u32 bpf_convert_ctx_access(enum bpf_access_type type,
                                              bpf_target_off(struct sk_buff, tc_index, 2,
                                                             target_size));
 #else
+               *target_size = 2;
                if (type == BPF_WRITE)
                        *insn++ = BPF_MOV64_REG(si->dst_reg, si->dst_reg);
                else
@@ -3520,6 +3521,7 @@ static u32 bpf_convert_ctx_access(enum bpf_access_type type,
                *insn++ = BPF_JMP_IMM(BPF_JGE, si->dst_reg, MIN_NAPI_ID, 1);
                *insn++ = BPF_MOV64_IMM(si->dst_reg, 0);
 #else
+               *target_size = 4;
                *insn++ = BPF_MOV64_IMM(si->dst_reg, 0);
 #endif
                break;
index 9fe25bf6329691ecf0acdc35df7278b074d446c1..b68168fcc06aa1981258eca4857511329af62f9a 100644 (file)
@@ -24,6 +24,7 @@
 #include <net/checksum.h>
 
 #include <net/inet_sock.h>
+#include <net/inet_common.h>
 #include <net/sock.h>
 #include <net/xfrm.h>
 
@@ -170,6 +171,15 @@ const char *dccp_packet_name(const int type)
 
 EXPORT_SYMBOL_GPL(dccp_packet_name);
 
+static void dccp_sk_destruct(struct sock *sk)
+{
+       struct dccp_sock *dp = dccp_sk(sk);
+
+       ccid_hc_tx_delete(dp->dccps_hc_tx_ccid, sk);
+       dp->dccps_hc_tx_ccid = NULL;
+       inet_sock_destruct(sk);
+}
+
 int dccp_init_sock(struct sock *sk, const __u8 ctl_sock_initialized)
 {
        struct dccp_sock *dp = dccp_sk(sk);
@@ -179,6 +189,7 @@ int dccp_init_sock(struct sock *sk, const __u8 ctl_sock_initialized)
        icsk->icsk_syn_retries  = sysctl_dccp_request_retries;
        sk->sk_state            = DCCP_CLOSED;
        sk->sk_write_space      = dccp_write_space;
+       sk->sk_destruct         = dccp_sk_destruct;
        icsk->icsk_sync_mss     = dccp_sync_mss;
        dp->dccps_mss_cache     = 536;
        dp->dccps_rate_last     = jiffies;
@@ -201,10 +212,7 @@ void dccp_destroy_sock(struct sock *sk)
 {
        struct dccp_sock *dp = dccp_sk(sk);
 
-       /*
-        * DCCP doesn't use sk_write_queue, just sk_send_head
-        * for retransmissions
-        */
+       __skb_queue_purge(&sk->sk_write_queue);
        if (sk->sk_send_head != NULL) {
                kfree_skb(sk->sk_send_head);
                sk->sk_send_head = NULL;
@@ -222,8 +230,7 @@ void dccp_destroy_sock(struct sock *sk)
                dp->dccps_hc_rx_ackvec = NULL;
        }
        ccid_hc_rx_delete(dp->dccps_hc_rx_ccid, sk);
-       ccid_hc_tx_delete(dp->dccps_hc_tx_ccid, sk);
-       dp->dccps_hc_rx_ccid = dp->dccps_hc_tx_ccid = NULL;
+       dp->dccps_hc_rx_ccid = NULL;
 
        /* clean up feature negotiation state */
        dccp_feat_list_purge(&dp->dccps_featneg);
index fab41de8e9837b512709b3c853ffb4831fe08317..de66ca8e620177693e55b54492404d8d3ee52197 100644 (file)
@@ -42,6 +42,9 @@ static struct sk_buff *ksz_xmit(struct sk_buff *skb, struct net_device *dev)
        padlen = (skb->len >= ETH_ZLEN) ? 0 : ETH_ZLEN - skb->len;
 
        if (skb_tailroom(skb) >= padlen + KSZ_INGRESS_TAG_LEN) {
+               if (skb_put_padto(skb, skb->len + padlen))
+                       return NULL;
+
                nskb = skb;
        } else {
                nskb = alloc_skb(NET_IP_ALIGN + skb->len +
@@ -56,13 +59,15 @@ static struct sk_buff *ksz_xmit(struct sk_buff *skb, struct net_device *dev)
                skb_set_transport_header(nskb,
                                         skb_transport_header(skb) - skb->head);
                skb_copy_and_csum_dev(skb, skb_put(nskb, skb->len));
+
+               if (skb_put_padto(nskb, nskb->len + padlen)) {
+                       kfree_skb(nskb);
+                       return NULL;
+               }
+
                kfree_skb(skb);
        }
 
-       /* skb is freed when it fails */
-       if (skb_put_padto(nskb, nskb->len + padlen))
-               return NULL;
-
        tag = skb_put(nskb, KSZ_INGRESS_TAG_LEN);
        tag[0] = 0;
        tag[1] = 1 << p->dp->index; /* destination port */
index b8d18171cca33ab5dab67408c3cd11ad57f25b83..ec3a9ce281a6ffb86b62e21f7284fd7c801668f0 100644 (file)
@@ -1083,15 +1083,17 @@ struct fib_info *fib_create_info(struct fib_config *cfg,
        fi = kzalloc(sizeof(*fi)+nhs*sizeof(struct fib_nh), GFP_KERNEL);
        if (!fi)
                goto failure;
-       fib_info_cnt++;
        if (cfg->fc_mx) {
                fi->fib_metrics = kzalloc(sizeof(*fi->fib_metrics), GFP_KERNEL);
-               if (!fi->fib_metrics)
-                       goto failure;
+               if (unlikely(!fi->fib_metrics)) {
+                       kfree(fi);
+                       return ERR_PTR(err);
+               }
                atomic_set(&fi->fib_metrics->refcnt, 1);
-       } else
+       } else {
                fi->fib_metrics = (struct dst_metrics *)&dst_default_metrics;
-
+       }
+       fib_info_cnt++;
        fi->fib_net = net;
        fi->fib_protocol = cfg->fc_protocol;
        fi->fib_scope = cfg->fc_scope;
index 498706b072fb70e1ffe6b5dba817816db5a4cfa7..caf2f1101d027b7b6e8d9683887e16c7bd4a8438 100644 (file)
@@ -1007,10 +1007,18 @@ int igmp_rcv(struct sk_buff *skb)
 {
        /* This basically follows the spec line by line -- see RFC1112 */
        struct igmphdr *ih;
-       struct in_device *in_dev = __in_dev_get_rcu(skb->dev);
+       struct net_device *dev = skb->dev;
+       struct in_device *in_dev;
        int len = skb->len;
        bool dropped = true;
 
+       if (netif_is_l3_master(dev)) {
+               dev = dev_get_by_index_rcu(dev_net(dev), IPCB(skb)->iif);
+               if (!dev)
+                       goto drop;
+       }
+
+       in_dev = __in_dev_get_rcu(dev);
        if (!in_dev)
                goto drop;
 
index 0383e66f59bcef3bd6b8627edae9aa2d34139f5d..2331de20ca505d7f25fe9d93d5320e9e39af6c39 100644 (file)
@@ -1267,7 +1267,7 @@ static unsigned int ipv4_mtu(const struct dst_entry *dst)
        if (mtu)
                return mtu;
 
-       mtu = dst->dev->mtu;
+       mtu = READ_ONCE(dst->dev->mtu);
 
        if (unlikely(dst_metric_locked(dst, RTAX_MTU))) {
                if (rt->rt_uses_gateway && mtu > 576)
@@ -2750,26 +2750,34 @@ static int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh,
                err = 0;
                if (IS_ERR(rt))
                        err = PTR_ERR(rt);
+               else
+                       skb_dst_set(skb, &rt->dst);
        }
 
        if (err)
                goto errout_free;
 
-       skb_dst_set(skb, &rt->dst);
        if (rtm->rtm_flags & RTM_F_NOTIFY)
                rt->rt_flags |= RTCF_NOTIFY;
 
        if (rtm->rtm_flags & RTM_F_LOOKUP_TABLE)
                table_id = rt->rt_table_id;
 
-       if (rtm->rtm_flags & RTM_F_FIB_MATCH)
+       if (rtm->rtm_flags & RTM_F_FIB_MATCH) {
+               if (!res.fi) {
+                       err = fib_props[res.type].error;
+                       if (!err)
+                               err = -EHOSTUNREACH;
+                       goto errout_free;
+               }
                err = fib_dump_info(skb, NETLINK_CB(in_skb).portid,
                                    nlh->nlmsg_seq, RTM_NEWROUTE, table_id,
                                    rt->rt_type, res.prefix, res.prefixlen,
                                    fl4.flowi4_tos, res.fi, 0);
-       else
+       } else {
                err = rt_fill_info(net, dst, src, table_id, &fl4, skb,
                                   NETLINK_CB(in_skb).portid, nlh->nlmsg_seq);
+       }
        if (err < 0)
                goto errout_free;
 
index 53de1424c13cda5d1fec826b97cacf4f95adc99a..bab7f0493098c6521f445923d721d296f326f7e1 100644 (file)
@@ -3009,8 +3009,7 @@ void tcp_rearm_rto(struct sock *sk)
                        /* delta_us may not be positive if the socket is locked
                         * when the retrans timer fires and is rescheduled.
                         */
-                       if (delta_us > 0)
-                               rto = usecs_to_jiffies(delta_us);
+                       rto = usecs_to_jiffies(max_t(int, delta_us, 1));
                }
                inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, rto,
                                          TCP_RTO_MAX);
index a20e7f03d5f7d81ccf7e92de9bfbbcc3e2df6718..e9252c7df8091a8e0d2fc9d7e5722e9fd605a857 100644 (file)
@@ -1722,6 +1722,8 @@ process:
                 */
                sock_hold(sk);
                refcounted = true;
+               if (tcp_filter(sk, skb))
+                       goto discard_and_relse;
                nsk = tcp_check_req(sk, skb, req, false);
                if (!nsk) {
                        reqsk_put(req);
@@ -1729,8 +1731,6 @@ process:
                }
                if (nsk == sk) {
                        reqsk_put(req);
-               } else if (tcp_filter(sk, skb)) {
-                       goto discard_and_relse;
                } else if (tcp_child_process(sk, nsk, skb)) {
                        tcp_v4_send_reset(nsk, skb);
                        goto discard_and_relse;
index 2417f55374c593c89b2aeb1ec4f6e6e74bb1395f..6bb9e14c710a7e2bfa58ee63ff6e02461a22cbec 100644 (file)
@@ -122,14 +122,14 @@ int tcp_set_ulp(struct sock *sk, const char *name)
 
        ulp_ops = __tcp_ulp_find_autoload(name);
        if (!ulp_ops)
-               err = -ENOENT;
-       else
-               err = ulp_ops->init(sk);
+               return -ENOENT;
 
-       if (err)
-               goto out;
+       err = ulp_ops->init(sk);
+       if (err) {
+               module_put(ulp_ops->owner);
+               return err;
+       }
 
        icsk->icsk_ulp_ops = ulp_ops;
- out:
-       return err;
+       return 0;
 }
index b34f09b20fef1f69fd2d95d776e572eed6485a04..ebe46ed997cbbb4901191edda46578a44c0cfe48 100644 (file)
@@ -1574,7 +1574,8 @@ int udp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int noblock,
                return ip_recv_error(sk, msg, len, addr_len);
 
 try_again:
-       peeking = off = sk_peek_offset(sk, flags);
+       peeking = flags & MSG_PEEK;
+       off = sk_peek_offset(sk, flags);
        skb = __skb_recv_udp(sk, flags, noblock, &peeked, &off, &err);
        if (!skb)
                return err;
index ebb299cf72b7e81b77999d43d2e7f83212386f73..5cc0ea0381981b0539d5d6e67401d962a6f6a230 100644 (file)
@@ -914,6 +914,8 @@ add:
                }
                nsiblings = iter->rt6i_nsiblings;
                fib6_purge_rt(iter, fn, info->nl_net);
+               if (fn->rr_ptr == iter)
+                       fn->rr_ptr = NULL;
                rt6_release(iter);
 
                if (nsiblings) {
@@ -926,6 +928,8 @@ add:
                                if (rt6_qualify_for_ecmp(iter)) {
                                        *ins = iter->dst.rt6_next;
                                        fib6_purge_rt(iter, fn, info->nl_net);
+                                       if (fn->rr_ptr == iter)
+                                               fn->rr_ptr = NULL;
                                        rt6_release(iter);
                                        nsiblings--;
                                } else {
@@ -1014,7 +1018,7 @@ int fib6_add(struct fib6_node *root, struct rt6_info *rt,
                        /* Create subtree root node */
                        sfn = node_alloc();
                        if (!sfn)
-                               goto st_failure;
+                               goto failure;
 
                        sfn->leaf = info->nl_net->ipv6.ip6_null_entry;
                        atomic_inc(&info->nl_net->ipv6.ip6_null_entry->rt6i_ref);
@@ -1031,12 +1035,12 @@ int fib6_add(struct fib6_node *root, struct rt6_info *rt,
 
                        if (IS_ERR(sn)) {
                                /* If it is failed, discard just allocated
-                                  root, and then (in st_failure) stale node
+                                  root, and then (in failure) stale node
                                   in main tree.
                                 */
                                node_free(sfn);
                                err = PTR_ERR(sn);
-                               goto st_failure;
+                               goto failure;
                        }
 
                        /* Now link new subtree to main tree */
@@ -1051,7 +1055,7 @@ int fib6_add(struct fib6_node *root, struct rt6_info *rt,
 
                        if (IS_ERR(sn)) {
                                err = PTR_ERR(sn);
-                               goto st_failure;
+                               goto failure;
                        }
                }
 
@@ -1092,18 +1096,17 @@ out:
                        atomic_inc(&pn->leaf->rt6i_ref);
                }
 #endif
-               /* Always release dst as dst->__refcnt is guaranteed
-                * to be taken before entering this function
-                */
-               dst_release_immediate(&rt->dst);
+               goto failure;
        }
        return err;
 
-#ifdef CONFIG_IPV6_SUBTREES
-       /* Subtree creation failed, probably main tree node
-          is orphan. If it is, shoot it.
+failure:
+       /* fn->leaf could be NULL if fn is an intermediate node and we
+        * failed to add the new route to it in both subtree creation
+        * failure and fib6_add_rt2node() failure case.
+        * In both cases, fib6_repair_tree() should be called to fix
+        * fn->leaf.
         */
-st_failure:
        if (fn && !(fn->fn_flags & (RTN_RTINFO|RTN_ROOT)))
                fib6_repair_tree(info->nl_net, fn);
        /* Always release dst as dst->__refcnt is guaranteed
@@ -1111,7 +1114,6 @@ st_failure:
         */
        dst_release_immediate(&rt->dst);
        return err;
-#endif
 }
 
 /*
index a640fbcba15dbf246e419d3e03da8eca0fa6901a..94d6a13d47f0e9ec5ff4cbc50b90d4cd9ca3f38a 100644 (file)
@@ -417,14 +417,11 @@ static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
        struct net_device *loopback_dev =
                dev_net(dev)->loopback_dev;
 
-       if (dev != loopback_dev) {
-               if (idev && idev->dev == dev) {
-                       struct inet6_dev *loopback_idev =
-                               in6_dev_get(loopback_dev);
-                       if (loopback_idev) {
-                               rt->rt6i_idev = loopback_idev;
-                               in6_dev_put(idev);
-                       }
+       if (idev && idev->dev != loopback_dev) {
+               struct inet6_dev *loopback_idev = in6_dev_get(loopback_dev);
+               if (loopback_idev) {
+                       rt->rt6i_idev = loopback_idev;
+                       in6_dev_put(idev);
                }
        }
 }
@@ -3724,10 +3721,10 @@ static int ip6_route_dev_notify(struct notifier_block *this,
                /* NETDEV_UNREGISTER could be fired for multiple times by
                 * netdev_wait_allrefs(). Make sure we only call this once.
                 */
-               in6_dev_put(net->ipv6.ip6_null_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_null_entry->rt6i_idev);
 #ifdef CONFIG_IPV6_MULTIPLE_TABLES
-               in6_dev_put(net->ipv6.ip6_prohibit_entry->rt6i_idev);
-               in6_dev_put(net->ipv6.ip6_blk_hole_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_prohibit_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_blk_hole_entry->rt6i_idev);
 #endif
        }
 
index 2521690d62d6e591af594c3629f71f004240ed68..206210125fd71d129a9ed2ead51a536749ab62b8 100644 (file)
@@ -1456,6 +1456,8 @@ process:
                }
                sock_hold(sk);
                refcounted = true;
+               if (tcp_filter(sk, skb))
+                       goto discard_and_relse;
                nsk = tcp_check_req(sk, skb, req, false);
                if (!nsk) {
                        reqsk_put(req);
@@ -1464,8 +1466,6 @@ process:
                if (nsk == sk) {
                        reqsk_put(req);
                        tcp_v6_restore_cb(skb);
-               } else if (tcp_filter(sk, skb)) {
-                       goto discard_and_relse;
                } else if (tcp_child_process(sk, nsk, skb)) {
                        tcp_v6_send_reset(nsk, skb);
                        goto discard_and_relse;
index 96d240798c38bfaa32e94c0adcf823626dcc9797..8cd9b628cdc769c09c98c71b71bbb7e4829438e8 100644 (file)
@@ -362,7 +362,8 @@ int udpv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
                return ipv6_recv_rxpmtu(sk, msg, len, addr_len);
 
 try_again:
-       peeking = off = sk_peek_offset(sk, flags);
+       peeking = flags & MSG_PEEK;
+       off = sk_peek_offset(sk, flags);
        skb = __skb_recv_udp(sk, flags, noblock, &peeked, &off, &err);
        if (!skb)
                return err;
index 2e6990f8b80b6b4cc1d59665b0154a81790e1788..23fa7c8b09a5861e2acc042775830cc2dac4e0f4 100644 (file)
@@ -2213,7 +2213,7 @@ static int irda_getsockopt(struct socket *sock, int level, int optname,
 {
        struct sock *sk = sock->sk;
        struct irda_sock *self = irda_sk(sk);
-       struct irda_device_list list;
+       struct irda_device_list list = { 0 };
        struct irda_device_info *discoveries;
        struct irda_ias_set *   ias_opt;        /* IAS get/query params */
        struct ias_object *     ias_obj;        /* Object in IAS */
index ca9d3ae665e76ea847a4ce03b4d275f80d7705bc..98f4d8211b9a9d9bc26e7d9979c5c2bf27c1b344 100644 (file)
@@ -228,7 +228,7 @@ static int pfkey_broadcast_one(struct sk_buff *skb, struct sk_buff **skb2,
 #define BROADCAST_ONE          1
 #define BROADCAST_REGISTERED   2
 #define BROADCAST_PROMISC_ONLY 4
-static int pfkey_broadcast(struct sk_buff *skb,
+static int pfkey_broadcast(struct sk_buff *skb, gfp_t allocation,
                           int broadcast_flags, struct sock *one_sk,
                           struct net *net)
 {
@@ -278,7 +278,7 @@ static int pfkey_broadcast(struct sk_buff *skb,
        rcu_read_unlock();
 
        if (one_sk != NULL)
-               err = pfkey_broadcast_one(skb, &skb2, GFP_KERNEL, one_sk);
+               err = pfkey_broadcast_one(skb, &skb2, allocation, one_sk);
 
        kfree_skb(skb2);
        kfree_skb(skb);
@@ -311,7 +311,7 @@ static int pfkey_do_dump(struct pfkey_sock *pfk)
                hdr = (struct sadb_msg *) pfk->dump.skb->data;
                hdr->sadb_msg_seq = 0;
                hdr->sadb_msg_errno = rc;
-               pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
+               pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
                                &pfk->sk, sock_net(&pfk->sk));
                pfk->dump.skb = NULL;
        }
@@ -355,7 +355,7 @@ static int pfkey_error(const struct sadb_msg *orig, int err, struct sock *sk)
        hdr->sadb_msg_len = (sizeof(struct sadb_msg) /
                             sizeof(uint64_t));
 
-       pfkey_broadcast(skb, BROADCAST_ONE, sk, sock_net(sk));
+       pfkey_broadcast(skb, GFP_KERNEL, BROADCAST_ONE, sk, sock_net(sk));
 
        return 0;
 }
@@ -1389,7 +1389,7 @@ static int pfkey_getspi(struct sock *sk, struct sk_buff *skb, const struct sadb_
 
        xfrm_state_put(x);
 
-       pfkey_broadcast(resp_skb, BROADCAST_ONE, sk, net);
+       pfkey_broadcast(resp_skb, GFP_KERNEL, BROADCAST_ONE, sk, net);
 
        return 0;
 }
@@ -1476,7 +1476,7 @@ static int key_notify_sa(struct xfrm_state *x, const struct km_event *c)
        hdr->sadb_msg_seq = c->seq;
        hdr->sadb_msg_pid = c->portid;
 
-       pfkey_broadcast(skb, BROADCAST_ALL, NULL, xs_net(x));
+       pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, xs_net(x));
 
        return 0;
 }
@@ -1589,7 +1589,7 @@ static int pfkey_get(struct sock *sk, struct sk_buff *skb, const struct sadb_msg
        out_hdr->sadb_msg_reserved = 0;
        out_hdr->sadb_msg_seq = hdr->sadb_msg_seq;
        out_hdr->sadb_msg_pid = hdr->sadb_msg_pid;
-       pfkey_broadcast(out_skb, BROADCAST_ONE, sk, sock_net(sk));
+       pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, sk, sock_net(sk));
 
        return 0;
 }
@@ -1694,8 +1694,8 @@ static int pfkey_register(struct sock *sk, struct sk_buff *skb, const struct sad
                return -ENOBUFS;
        }
 
-       pfkey_broadcast(supp_skb, BROADCAST_REGISTERED, sk, sock_net(sk));
-
+       pfkey_broadcast(supp_skb, GFP_KERNEL, BROADCAST_REGISTERED, sk,
+                       sock_net(sk));
        return 0;
 }
 
@@ -1712,7 +1712,8 @@ static int unicast_flush_resp(struct sock *sk, const struct sadb_msg *ihdr)
        hdr->sadb_msg_errno = (uint8_t) 0;
        hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));
 
-       return pfkey_broadcast(skb, BROADCAST_ONE, sk, sock_net(sk));
+       return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ONE, sk,
+                              sock_net(sk));
 }
 
 static int key_notify_sa_flush(const struct km_event *c)
@@ -1733,7 +1734,7 @@ static int key_notify_sa_flush(const struct km_event *c)
        hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));
        hdr->sadb_msg_reserved = 0;
 
-       pfkey_broadcast(skb, BROADCAST_ALL, NULL, c->net);
+       pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, c->net);
 
        return 0;
 }
@@ -1790,7 +1791,7 @@ static int dump_sa(struct xfrm_state *x, int count, void *ptr)
        out_hdr->sadb_msg_pid = pfk->dump.msg_portid;
 
        if (pfk->dump.skb)
-               pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
+               pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
                                &pfk->sk, sock_net(&pfk->sk));
        pfk->dump.skb = out_skb;
 
@@ -1878,7 +1879,7 @@ static int pfkey_promisc(struct sock *sk, struct sk_buff *skb, const struct sadb
                new_hdr->sadb_msg_errno = 0;
        }
 
-       pfkey_broadcast(skb, BROADCAST_ALL, NULL, sock_net(sk));
+       pfkey_broadcast(skb, GFP_KERNEL, BROADCAST_ALL, NULL, sock_net(sk));
        return 0;
 }
 
@@ -2206,7 +2207,7 @@ static int key_notify_policy(struct xfrm_policy *xp, int dir, const struct km_ev
        out_hdr->sadb_msg_errno = 0;
        out_hdr->sadb_msg_seq = c->seq;
        out_hdr->sadb_msg_pid = c->portid;
-       pfkey_broadcast(out_skb, BROADCAST_ALL, NULL, xp_net(xp));
+       pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ALL, NULL, xp_net(xp));
        return 0;
 
 }
@@ -2426,7 +2427,7 @@ static int key_pol_get_resp(struct sock *sk, struct xfrm_policy *xp, const struc
        out_hdr->sadb_msg_errno = 0;
        out_hdr->sadb_msg_seq = hdr->sadb_msg_seq;
        out_hdr->sadb_msg_pid = hdr->sadb_msg_pid;
-       pfkey_broadcast(out_skb, BROADCAST_ONE, sk, xp_net(xp));
+       pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, sk, xp_net(xp));
        err = 0;
 
 out:
@@ -2682,7 +2683,7 @@ static int dump_sp(struct xfrm_policy *xp, int dir, int count, void *ptr)
        out_hdr->sadb_msg_pid = pfk->dump.msg_portid;
 
        if (pfk->dump.skb)
-               pfkey_broadcast(pfk->dump.skb, BROADCAST_ONE,
+               pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
                                &pfk->sk, sock_net(&pfk->sk));
        pfk->dump.skb = out_skb;
 
@@ -2739,7 +2740,7 @@ static int key_notify_policy_flush(const struct km_event *c)
        hdr->sadb_msg_satype = SADB_SATYPE_UNSPEC;
        hdr->sadb_msg_len = (sizeof(struct sadb_msg) / sizeof(uint64_t));
        hdr->sadb_msg_reserved = 0;
-       pfkey_broadcast(skb_out, BROADCAST_ALL, NULL, c->net);
+       pfkey_broadcast(skb_out, GFP_ATOMIC, BROADCAST_ALL, NULL, c->net);
        return 0;
 
 }
@@ -2803,7 +2804,7 @@ static int pfkey_process(struct sock *sk, struct sk_buff *skb, const struct sadb
        void *ext_hdrs[SADB_EXT_MAX];
        int err;
 
-       pfkey_broadcast(skb_clone(skb, GFP_KERNEL),
+       pfkey_broadcast(skb_clone(skb, GFP_KERNEL), GFP_KERNEL,
                        BROADCAST_PROMISC_ONLY, NULL, sock_net(sk));
 
        memset(ext_hdrs, 0, sizeof(ext_hdrs));
@@ -3024,7 +3025,8 @@ static int key_notify_sa_expire(struct xfrm_state *x, const struct km_event *c)
        out_hdr->sadb_msg_seq = 0;
        out_hdr->sadb_msg_pid = 0;
 
-       pfkey_broadcast(out_skb, BROADCAST_REGISTERED, NULL, xs_net(x));
+       pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL,
+                       xs_net(x));
        return 0;
 }
 
@@ -3212,7 +3214,8 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct
                       xfrm_ctx->ctx_len);
        }
 
-       return pfkey_broadcast(skb, BROADCAST_REGISTERED, NULL, xs_net(x));
+       return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL,
+                              xs_net(x));
 }
 
 static struct xfrm_policy *pfkey_compile_policy(struct sock *sk, int opt,
@@ -3408,7 +3411,8 @@ static int pfkey_send_new_mapping(struct xfrm_state *x, xfrm_address_t *ipaddr,
        n_port->sadb_x_nat_t_port_port = sport;
        n_port->sadb_x_nat_t_port_reserved = 0;
 
-       return pfkey_broadcast(skb, BROADCAST_REGISTERED, NULL, xs_net(x));
+       return pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_REGISTERED, NULL,
+                              xs_net(x));
 }
 
 #ifdef CONFIG_NET_KEY_MIGRATE
@@ -3599,7 +3603,7 @@ static int pfkey_send_migrate(const struct xfrm_selector *sel, u8 dir, u8 type,
        }
 
        /* broadcast migrate message to sockets */
-       pfkey_broadcast(skb, BROADCAST_ALL, NULL, &init_net);
+       pfkey_broadcast(skb, GFP_ATOMIC, BROADCAST_ALL, NULL, &init_net);
 
        return 0;
 
index 8708cbe8af5bbcabc5c7ea295c6a0abd15a8eb84..2b36eff5d97ea7c1cef117033c66dd35058ac4e7 100644 (file)
@@ -7,7 +7,7 @@
  * Copyright 2006-2007 Jiri Benc <jbenc@suse.cz>
  * Copyright 2007, Michael Wu <flamingice@sourmilk.net>
  * Copyright 2007-2010, Intel Corporation
- * Copyright(c) 2015 Intel Deutschland GmbH
+ * Copyright(c) 2015-2017 Intel Deutschland GmbH
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
@@ -466,3 +466,23 @@ void ieee80211_manage_rx_ba_offl(struct ieee80211_vif *vif,
        rcu_read_unlock();
 }
 EXPORT_SYMBOL(ieee80211_manage_rx_ba_offl);
+
+void ieee80211_rx_ba_timer_expired(struct ieee80211_vif *vif,
+                                  const u8 *addr, unsigned int tid)
+{
+       struct ieee80211_sub_if_data *sdata = vif_to_sdata(vif);
+       struct ieee80211_local *local = sdata->local;
+       struct sta_info *sta;
+
+       rcu_read_lock();
+       sta = sta_info_get_bss(sdata, addr);
+       if (!sta)
+               goto unlock;
+
+       set_bit(tid, sta->ampdu_mlme.tid_rx_timer_expired);
+       ieee80211_queue_work(&local->hw, &sta->ampdu_mlme.work);
+
+ unlock:
+       rcu_read_unlock();
+}
+EXPORT_SYMBOL(ieee80211_rx_ba_timer_expired);
index e4610676299bcdac626db1a30cd4da44ccc62c0b..a54a556fcdb57d95b4a4a6606016ead527b93d64 100644 (file)
@@ -1337,6 +1337,7 @@ int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb,
                goto out;
        }
 
+       OVS_CB(skb)->acts_origlen = acts->orig_len;
        err = do_execute_actions(dp, skb, key,
                                 acts->actions, acts->actions_len);
 
index 45fe8c8a884df36100bc0cf26d279cb92e51bdf8..6b44fe4052825a87b373bafb58ca014e3ec99015 100644 (file)
@@ -381,7 +381,7 @@ static int queue_gso_packets(struct datapath *dp, struct sk_buff *skb,
 }
 
 static size_t upcall_msg_size(const struct dp_upcall_info *upcall_info,
-                             unsigned int hdrlen)
+                             unsigned int hdrlen, int actions_attrlen)
 {
        size_t size = NLMSG_ALIGN(sizeof(struct ovs_header))
                + nla_total_size(hdrlen) /* OVS_PACKET_ATTR_PACKET */
@@ -398,7 +398,7 @@ static size_t upcall_msg_size(const struct dp_upcall_info *upcall_info,
 
        /* OVS_PACKET_ATTR_ACTIONS */
        if (upcall_info->actions_len)
-               size += nla_total_size(upcall_info->actions_len);
+               size += nla_total_size(actions_attrlen);
 
        /* OVS_PACKET_ATTR_MRU */
        if (upcall_info->mru)
@@ -465,7 +465,8 @@ static int queue_userspace_packet(struct datapath *dp, struct sk_buff *skb,
        else
                hlen = skb->len;
 
-       len = upcall_msg_size(upcall_info, hlen - cutlen);
+       len = upcall_msg_size(upcall_info, hlen - cutlen,
+                             OVS_CB(skb)->acts_origlen);
        user_skb = genlmsg_new(len, GFP_ATOMIC);
        if (!user_skb) {
                err = -ENOMEM;
index 5d8dcd88815f0622a48a54f14eaf5b5b9c53b18c..480600649d0b03a185f73b587181a3136d75e9b2 100644 (file)
@@ -99,11 +99,13 @@ struct datapath {
  * when a packet is received by OVS.
  * @mru: The maximum received fragement size; 0 if the packet is not
  * fragmented.
+ * @acts_origlen: The netlink size of the flow actions applied to this skb.
  * @cutlen: The number of bytes from the packet end to be removed.
  */
 struct ovs_skb_cb {
        struct vport            *input_vport;
        u16                     mru;
+       u16                     acts_origlen;
        u32                     cutlen;
 };
 #define OVS_CB(skb) ((struct ovs_skb_cb *)(skb)->cb)
index dd30d74824b0de42d1866f4449300f89c83d6724..ec3383f97d4c3e863c7fe77af8d3ed3ba5ea039f 100644 (file)
@@ -223,6 +223,7 @@ void rxrpc_discard_prealloc(struct rxrpc_sock *rx)
        tail = b->call_backlog_tail;
        while (CIRC_CNT(head, tail, size) > 0) {
                struct rxrpc_call *call = b->call_backlog[tail];
+               call->socket = rx;
                if (rx->discard_new_call) {
                        _debug("discard %lx", call->user_call_ID);
                        rx->discard_new_call(call, call->user_call_ID);
index d516ba8178b8099f5e8e180f2e60e7a61de37811..541707802a2380e83a96821dd20eb5946824cfe5 100644 (file)
@@ -41,6 +41,7 @@ static int ipt_init_target(struct net *net, struct xt_entry_target *t,
 {
        struct xt_tgchk_param par;
        struct xt_target *target;
+       struct ipt_entry e = {};
        int ret = 0;
 
        target = xt_request_find_target(AF_INET, t->u.user.name,
@@ -52,6 +53,7 @@ static int ipt_init_target(struct net *net, struct xt_entry_target *t,
        memset(&par, 0, sizeof(par));
        par.net       = net;
        par.table     = table;
+       par.entryinfo = &e;
        par.target    = target;
        par.targinfo  = t->data;
        par.hook_mask = hook;
index 39da0c5801c908e28c4e258319adefde836c6b0c..9fd44c22134783edf3db4f62ff1e8184c455cbd7 100644 (file)
@@ -205,7 +205,7 @@ static void tcf_chain_flush(struct tcf_chain *chain)
 {
        struct tcf_proto *tp;
 
-       if (*chain->p_filter_chain)
+       if (chain->p_filter_chain)
                RCU_INIT_POINTER(*chain->p_filter_chain, NULL);
        while ((tp = rtnl_dereference(chain->filter_chain)) != NULL) {
                RCU_INIT_POINTER(chain->filter_chain, tp->next);
index bd24a550e0f9f114598f4a398d6efd4c72672f50..a3fa144b864871088209386fd573bded1886432f 100644 (file)
@@ -286,9 +286,6 @@ static struct Qdisc *qdisc_match_from_root(struct Qdisc *root, u32 handle)
 void qdisc_hash_add(struct Qdisc *q, bool invisible)
 {
        if ((q->parent != TC_H_ROOT) && !(q->flags & TCQ_F_INGRESS)) {
-               struct Qdisc *root = qdisc_dev(q)->qdisc;
-
-               WARN_ON_ONCE(root == &noop_qdisc);
                ASSERT_RTNL();
                hash_add_rcu(qdisc_dev(q)->qdisc_hash, &q->hash, q->handle);
                if (invisible)
index 572fe2584e48c81dbf58d90ce9d6a4ae68d2a385..c403c87aff7a44bccfdd5f07e2e00ea0698a5c90 100644 (file)
@@ -572,8 +572,10 @@ static void atm_tc_destroy(struct Qdisc *sch)
        struct atm_flow_data *flow, *tmp;
 
        pr_debug("atm_tc_destroy(sch %p,[qdisc %p])\n", sch, p);
-       list_for_each_entry(flow, &p->flows, list)
+       list_for_each_entry(flow, &p->flows, list) {
                tcf_block_put(flow->block);
+               flow->block = NULL;
+       }
 
        list_for_each_entry_safe(flow, tmp, &p->flows, list) {
                if (flow->ref > 1)
index 481036f6b54e4730ee27fae6236277c64d3eaa1a..780db43300b16284192b24006b0ae8677adbe505 100644 (file)
@@ -1431,8 +1431,10 @@ static void cbq_destroy(struct Qdisc *sch)
         * be bound to classes which have been destroyed already. --TGR '04
         */
        for (h = 0; h < q->clhash.hashsize; h++) {
-               hlist_for_each_entry(cl, &q->clhash.hash[h], common.hnode)
+               hlist_for_each_entry(cl, &q->clhash.hash[h], common.hnode) {
                        tcf_block_put(cl->block);
+                       cl->block = NULL;
+               }
        }
        for (h = 0; h < q->clhash.hashsize; h++) {
                hlist_for_each_entry_safe(cl, next, &q->clhash.hash[h],
index b52f74610dc7539fd7804403a6ff41b690243df7..fd15200f86273add7d6c8c4a18aaef912aba7411 100644 (file)
@@ -1428,6 +1428,10 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt)
                return err;
        q->eligible = RB_ROOT;
 
+       err = tcf_block_get(&q->root.block, &q->root.filter_list);
+       if (err)
+               goto err_tcf;
+
        q->root.cl_common.classid = sch->handle;
        q->root.refcnt  = 1;
        q->root.sched   = q;
@@ -1447,6 +1451,10 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt)
        qdisc_watchdog_init(&q->watchdog, sch);
 
        return 0;
+
+err_tcf:
+       qdisc_class_hash_destroy(&q->clhash);
+       return err;
 }
 
 static int
@@ -1522,8 +1530,10 @@ hfsc_destroy_qdisc(struct Qdisc *sch)
        unsigned int i;
 
        for (i = 0; i < q->clhash.hashsize; i++) {
-               hlist_for_each_entry(cl, &q->clhash.hash[i], cl_common.hnode)
+               hlist_for_each_entry(cl, &q->clhash.hash[i], cl_common.hnode) {
                        tcf_block_put(cl->block);
+                       cl->block = NULL;
+               }
        }
        for (i = 0; i < q->clhash.hashsize; i++) {
                hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i],
index 203286ab442799a07808bd8f73af9f07b0482cbd..5d65ec5207e91202d501a83c983793f2e923f075 100644 (file)
@@ -1258,8 +1258,10 @@ static void htb_destroy(struct Qdisc *sch)
        tcf_block_put(q->block);
 
        for (i = 0; i < q->clhash.hashsize; i++) {
-               hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode)
+               hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) {
                        tcf_block_put(cl->block);
+                       cl->block = NULL;
+               }
        }
        for (i = 0; i < q->clhash.hashsize; i++) {
                hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i],
index f80ea2cc5f1f4bdee2452c930e3650f79397f3a1..82469ef9655eb65431477215bb2df0fb2737f905 100644 (file)
@@ -437,6 +437,7 @@ congestion_drop:
                qdisc_drop(head, sch, to_free);
 
                slot_queue_add(slot, skb);
+               qdisc_tree_reduce_backlog(sch, 0, delta);
                return NET_XMIT_CN;
        }
 
@@ -468,8 +469,10 @@ enqueue:
        /* Return Congestion Notification only if we dropped a packet
         * from this flow.
         */
-       if (qlen != slot->qlen)
+       if (qlen != slot->qlen) {
+               qdisc_tree_reduce_backlog(sch, 0, dropped - qdisc_pkt_len(skb));
                return NET_XMIT_CN;
+       }
 
        /* As we dropped a packet, better let upper stack know this */
        qdisc_tree_reduce_backlog(sch, 1, dropped);
index 2a186b201ad2c20594169633b241dacd0cc971c0..a4b6ffb6149541b78e39aceae859224d13487106 100644 (file)
@@ -512,7 +512,9 @@ static void sctp_v6_to_addr(union sctp_addr *addr, struct in6_addr *saddr,
 {
        addr->sa.sa_family = AF_INET6;
        addr->v6.sin6_port = port;
+       addr->v6.sin6_flowinfo = 0;
        addr->v6.sin6_addr = *saddr;
+       addr->v6.sin6_scope_id = 0;
 }
 
 /* Compare addresses exactly.
index d174ee3254eecb523dbc86061d80a1e812dcc305..767e0537dde5a8d0cfac97f6b11ee1d568ef87ed 100644 (file)
@@ -596,7 +596,7 @@ static int tipc_l2_rcv_msg(struct sk_buff *skb, struct net_device *dev,
        rcu_read_lock();
        b = rcu_dereference_rtnl(dev->tipc_ptr);
        if (likely(b && test_bit(0, &b->up) &&
-                  (skb->pkt_type <= PACKET_BROADCAST))) {
+                  (skb->pkt_type <= PACKET_MULTICAST))) {
                skb->next = NULL;
                tipc_rcv(dev_net(dev), skb, b);
                rcu_read_unlock();
index ab3087687a32446ffa3bbfaccf206028886e6945..dcd90e6fa7c39c962eb6f6684cc31dd2b2443e4a 100644 (file)
@@ -513,6 +513,7 @@ bool tipc_msg_reverse(u32 own_node,  struct sk_buff **skb, int err)
 
        /* Now reverse the concerned fields */
        msg_set_errcode(hdr, err);
+       msg_set_non_seq(hdr, 0);
        msg_set_origport(hdr, msg_destport(&ohdr));
        msg_set_destport(hdr, msg_origport(&ohdr));
        msg_set_destnode(hdr, msg_prevnode(&ohdr));
index 9bfe886ab33080f653ec1c39e0255b15709d0b76..750949dfc1d7d9bc81c55de55f0e7dafb4023a2c 100644 (file)
@@ -258,13 +258,15 @@ static int tipc_nl_compat_dumpit(struct tipc_nl_compat_cmd_dump *cmd,
        arg = nlmsg_new(0, GFP_KERNEL);
        if (!arg) {
                kfree_skb(msg->rep);
+               msg->rep = NULL;
                return -ENOMEM;
        }
 
        err = __tipc_nl_compat_dumpit(cmd, msg, arg);
-       if (err)
+       if (err) {
                kfree_skb(msg->rep);
-
+               msg->rep = NULL;
+       }
        kfree_skb(arg);
 
        return err;
index 7b52a380d710d238c440293b6d3b8973fc30c90c..be8982b4f8c00be8bb95748c1c33a76e13079dff 100644 (file)
@@ -2304,10 +2304,7 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
         */
        mutex_lock(&u->iolock);
 
-       if (flags & MSG_PEEK)
-               skip = sk_peek_offset(sk, flags);
-       else
-               skip = 0;
+       skip = max(sk_peek_offset(sk, flags), 0);
 
        do {
                int chunk;
index dd8e2dde0b34b3759fe6061eb5c2887f4ecd21fd..9ffd3dda3889c56a7a72229bed21ff5c49d62856 100644 (file)
@@ -85,8 +85,8 @@ TMPOUT := $(if $(KBUILD_EXTMOD),$(firstword $(KBUILD_EXTMOD))/)
 
 # try-run
 # Usage: option = $(call try-run, $(CC)...-o "$$TMP",option-ok,otherwise)
-# Exit code chooses option. "$$TMP" is can be used as temporary file and
-# is automatically cleaned up.
+# Exit code chooses option. "$$TMP" serves as a temporary file and is
+# automatically cleaned up.
 try-run = $(shell set -e;              \
        TMP="$(TMPOUT).$$$$.tmp";       \
        TMPO="$(TMPOUT).$$$$.o";        \
@@ -261,7 +261,6 @@ make-cmd = $(call escsq,$(subst \#,\\\#,$(subst $$,$$$$,$(cmd_$(1)))))
 any-prereq = $(filter-out $(PHONY),$?) $(filter-out $(PHONY) $(wildcard $^),$^)
 
 # Execute command if command has changed or prerequisite(s) are updated.
-#
 if_changed = $(if $(strip $(any-prereq) $(arg-check)),                       \
        @set -e;                                                             \
        $(echo-cmd) $(cmd_$(1));                                             \
@@ -315,7 +314,7 @@ if_changed_rule = $(if $(strip $(any-prereq) $(arg-check) ),                 \
        $(rule_$(1)), @:)
 
 ###
-# why - tell why a a target got build
+# why - tell why a target got built
 #       enabled by make V=2
 #       Output (listed in the order they are checked):
 #          (1) - due to target is PHONY
index 95f7d80901524a4c966505397af6931f3e3316f3..a6c8c17808551aa146102ccb4132f8f5605ea74d 100644 (file)
@@ -1,9 +1,9 @@
 # include/asm-generic contains a lot of files that are used
 # verbatim by several architectures.
 #
-# This Makefile reads the file arch/$(SRCARCH)/include/asm/Kbuild
+# This Makefile reads the file arch/$(SRCARCH)/include/$(src)/Kbuild
 # and for each file listed in this file with generic-y creates
-# a small wrapper file in $(obj) (arch/$(SRCARCH)/include/generated/asm)
+# a small wrapper file in $(obj) (arch/$(SRCARCH)/include/generated/$(src))
 
 kbuild-file := $(srctree)/arch/$(SRCARCH)/include/$(src)/Kbuild
 -include $(kbuild-file)
index 854608d42e858ac95c37741081eb5e97b7abc871..a18cb4496e1eedca52803c783e5d2879a79af40d 100644 (file)
@@ -229,8 +229,8 @@ ifeq ("$(origin RECORDMCOUNT_WARN)", "command line")
 endif
 # Due to recursion, we must skip empty.o.
 # The empty.o file is created in the make process in order to determine
-#  the target endianness and word size. It is made before all other C
-#  files, including recordmcount.
+# the target endianness and word size. It is made before all other C
+# files, including recordmcount.
 sub_cmd_record_mcount =                                        \
        if [ $(@) != "scripts/mod/empty.o" ]; then      \
                $(objtree)/scripts/recordmcount $(RECORDMCOUNT_FLAGS) "$(@)";   \
@@ -245,13 +245,13 @@ sub_cmd_record_mcount = set -e ; perl $(srctree)/scripts/recordmcount.pl "$(ARCH
        "$(LD)" "$(NM)" "$(RM)" "$(MV)" \
        "$(if $(part-of-module),1,0)" "$(@)";
 recordmcount_source := $(srctree)/scripts/recordmcount.pl
-endif
+endif # BUILD_C_RECORDMCOUNT
 cmd_record_mcount =                                            \
        if [ "$(findstring $(CC_FLAGS_FTRACE),$(_c_flags))" =   \
             "$(CC_FLAGS_FTRACE)" ]; then                       \
                $(sub_cmd_record_mcount)                        \
        fi;
-endif
+endif # CONFIG_FTRACE_MCOUNT_RECORD
 
 ifdef CONFIG_STACK_VALIDATION
 ifneq ($(SKIP_STACK_VALIDATION),1)
index 34614a48b717eafa2c15f418e6e6769905e32ec1..993fb85982df2df2a7f8e20780d0cfbe3cb732e8 100644 (file)
@@ -14,7 +14,7 @@ src := $(obj)
 PHONY := __dtbs_install
 __dtbs_install:
 
-export dtbinst-root ?= $(obj)
+export dtbinst_root ?= $(obj)
 
 include include/config/auto.conf
 include scripts/Kbuild.include
@@ -27,7 +27,7 @@ dtbinst-dirs  := $(dts-dirs)
 quiet_cmd_dtb_install =        INSTALL $<
       cmd_dtb_install =        mkdir -p $(2); cp $< $(2)
 
-install-dir = $(patsubst $(dtbinst-root)%,$(INSTALL_DTBS_PATH)%,$(obj))
+install-dir = $(patsubst $(dtbinst_root)%,$(INSTALL_DTBS_PATH)%,$(obj))
 
 $(dtbinst-files): %.dtb: $(obj)/%.dtb
        $(call cmd,dtb_install,$(install-dir))
index ec10d9345bc2d9b978d8cae4d85bf2fd89ab121c..0372b33febe52f74978420c2c559adef5e815374 100644 (file)
@@ -1,5 +1,5 @@
 ###
-# Makefile.basic lists the most basic programs used during the build process.
+# This Makefile lists the most basic programs used during the build process.
 # The programs listed herein are what are needed to do the basic stuff,
 # such as fix file dependencies.
 # This initial step is needed to avoid files to be recompiled
index fff818b92acb7e3e469ce06cc722f3ab0f275171..bbf62cb1f8190917e930d7003eddbf20f0fececc 100644 (file)
@@ -25,7 +25,7 @@
  *
  * So we play the same trick that "mkdep" played before. We replace
  * the dependency on autoconf.h by a dependency on every config
- * option which is mentioned in any of the listed prequisites.
+ * option which is mentioned in any of the listed prerequisites.
  *
  * kconfig populates a tree in include/config/ with an empty file
  * for each config symbol and when the configuration is updated
@@ -34,7 +34,7 @@
  * the config symbols are rebuilt.
  *
  * So if the user changes his CONFIG_HIS_DRIVER option, only the objects
- * which depend on "include/linux/config/his/driver.h" will be rebuilt,
+ * which depend on "include/config/his/driver.h" will be rebuilt,
  * so most likely only his driver ;-)
  *
  * The idea above dates, by the way, back to Michael E Chastain, AFAIK.
@@ -75,7 +75,7 @@
  * and then basically copies the .<target>.d file to stdout, in the
  * process filtering out the dependency on autoconf.h and adding
  * dependencies on include/config/my/option.h for every
- * CONFIG_MY_OPTION encountered in any of the prequisites.
+ * CONFIG_MY_OPTION encountered in any of the prerequisites.
  *
  * It will also filter out all the dependencies on *.ver. We need
  * to make sure that the generated version checksum are globally up
index a536760a94c26521a04cbe4233b1d5fa4b69e625..45c1336c6597f1b49a343b8629613fb979b0ef8b 100644 (file)
@@ -47,10 +47,10 @@ config SND_SEQ_HRTIMER_DEFAULT
          timer.
 
 config SND_SEQ_MIDI_EVENT
-       def_tristate SND_RAWMIDI
+       tristate
 
 config SND_SEQ_MIDI
-       tristate
+       def_tristate SND_RAWMIDI
        select SND_SEQ_MIDI_EVENT
 
 config SND_SEQ_MIDI_EMUL
index 272c55fe17c88aec700a7552da4473349c4ce359..ea2d0ae85bd367d5ea70068ee74d925a349789c3 100644 (file)
@@ -1502,16 +1502,11 @@ static int snd_seq_ioctl_unsubscribe_port(struct snd_seq_client *client,
 static int snd_seq_ioctl_create_queue(struct snd_seq_client *client, void *arg)
 {
        struct snd_seq_queue_info *info = arg;
-       int result;
        struct snd_seq_queue *q;
 
-       result = snd_seq_queue_alloc(client->number, info->locked, info->flags);
-       if (result < 0)
-               return result;
-
-       q = queueptr(result);
-       if (q == NULL)
-               return -EINVAL;
+       q = snd_seq_queue_alloc(client->number, info->locked, info->flags);
+       if (IS_ERR(q))
+               return PTR_ERR(q);
 
        info->queue = q->queue;
        info->locked = q->locked;
@@ -1521,7 +1516,7 @@ static int snd_seq_ioctl_create_queue(struct snd_seq_client *client, void *arg)
        if (!info->name[0])
                snprintf(info->name, sizeof(info->name), "Queue-%d", q->queue);
        strlcpy(q->name, info->name, sizeof(q->name));
-       queuefree(q);
+       snd_use_lock_free(&q->use_lock);
 
        return 0;
 }
index 450c5187eecb6bb083736d2d2a1aad43b98c7c3f..79e0c5604ef806d62eca084293e66222b8fe6828 100644 (file)
@@ -184,22 +184,26 @@ void __exit snd_seq_queues_delete(void)
 static void queue_use(struct snd_seq_queue *queue, int client, int use);
 
 /* allocate a new queue -
- * return queue index value or negative value for error
+ * return pointer to new queue or ERR_PTR(-errno) for error
+ * The new queue's use_lock is set to 1. It is the caller's responsibility to
+ * call snd_use_lock_free(&q->use_lock).
  */
-int snd_seq_queue_alloc(int client, int locked, unsigned int info_flags)
+struct snd_seq_queue *snd_seq_queue_alloc(int client, int locked, unsigned int info_flags)
 {
        struct snd_seq_queue *q;
 
        q = queue_new(client, locked);
        if (q == NULL)
-               return -ENOMEM;
+               return ERR_PTR(-ENOMEM);
        q->info_flags = info_flags;
        queue_use(q, client, 1);
+       snd_use_lock_use(&q->use_lock);
        if (queue_list_add(q) < 0) {
+               snd_use_lock_free(&q->use_lock);
                queue_delete(q);
-               return -ENOMEM;
+               return ERR_PTR(-ENOMEM);
        }
-       return q->queue;
+       return q;
 }
 
 /* delete a queue - queue must be owned by the client */
index 30c8111477f61ed26987dc03abbe9670b36db221..719093489a2c4eec57fed70d4ba2b862cb64a9cf 100644 (file)
@@ -71,7 +71,7 @@ void snd_seq_queues_delete(void);
 
 
 /* create new queue (constructor) */
-int snd_seq_queue_alloc(int client, int locked, unsigned int flags);
+struct snd_seq_queue *snd_seq_queue_alloc(int client, int locked, unsigned int flags);
 
 /* delete queue (destructor) */
 int snd_seq_queue_delete(int client, int queueid);
index dc585959ca32c8cfbfa3aff0557a4e20c7ec1430..a2b56b188be4d90d9c51547e5fd8a7013700b14f 100644 (file)
@@ -698,10 +698,18 @@ static int copy_gctl(struct snd_emu10k1 *emu,
 {
        struct snd_emu10k1_fx8010_control_old_gpr __user *octl;
 
-       if (emu->support_tlv)
-               return copy_from_user(gctl, &_gctl[idx], sizeof(*gctl));
+       if (emu->support_tlv) {
+               if (in_kernel)
+                       memcpy(gctl, (void *)&_gctl[idx], sizeof(*gctl));
+               else if (copy_from_user(gctl, &_gctl[idx], sizeof(*gctl)))
+                       return -EFAULT;
+               return 0;
+       }
+
        octl = (struct snd_emu10k1_fx8010_control_old_gpr __user *)_gctl;
-       if (copy_from_user(gctl, &octl[idx], sizeof(*octl)))
+       if (in_kernel)
+               memcpy(gctl, (void *)&octl[idx], sizeof(*octl));
+       else if (copy_from_user(gctl, &octl[idx], sizeof(*octl)))
                return -EFAULT;
        gctl->tlv = NULL;
        return 0;
index a91a9ef00c40611db8f19ffcd14d414cd7d42d71..217bb582aff16a6ec428311fddc53370e0355288 100644 (file)
@@ -6647,7 +6647,6 @@ static const struct snd_hda_pin_quirk alc269_pin_fixup_tbl[] = {
        SND_HDA_PIN_QUIRK(0x10ec0299, 0x1028, "Dell", ALC269_FIXUP_DELL4_MIC_NO_PRESENCE,
                ALC225_STANDARD_PINS,
                {0x12, 0xb7a60130},
-               {0x13, 0xb8a61140},
                {0x17, 0x90170110}),
        {}
 };
index 082736c539bc14eec73a1afad11e2a3cd5577d89..e630813c5008627d50e41637e03eeaf8ef2bf0c3 100644 (file)
@@ -542,6 +542,8 @@ int snd_usb_mixer_vol_tlv(struct snd_kcontrol *kcontrol, int op_flag,
 
        if (size < sizeof(scale))
                return -ENOMEM;
+       if (cval->min_mute)
+               scale[0] = SNDRV_CTL_TLVT_DB_MINMAX_MUTE;
        scale[2] = cval->dBmin;
        scale[3] = cval->dBmax;
        if (copy_to_user(_tlv, scale, sizeof(scale)))
index 3417ef347e40432482b84de271a3bb98c8724297..2b4b067646ab099653fe7ea79d9af1570e2971f6 100644 (file)
@@ -64,6 +64,7 @@ struct usb_mixer_elem_info {
        int cached;
        int cache_val[MAX_CHANNELS];
        u8 initialized;
+       u8 min_mute;
        void *private_data;
 };
 
index e3d1dec48ee49f21d4efe4627fc3c264d76fcfe2..e1e7ce9ab217f6f716fc95da5d1544279c0d1b75 100644 (file)
@@ -1878,6 +1878,12 @@ void snd_usb_mixer_fu_apply_quirk(struct usb_mixer_interface *mixer,
                if (unitid == 7 && cval->control == UAC_FU_VOLUME)
                        snd_dragonfly_quirk_db_scale(mixer, cval, kctl);
                break;
+       /* lowest playback value is muted on C-Media devices */
+       case USB_ID(0x0d8c, 0x000c):
+       case USB_ID(0x0d8c, 0x0014):
+               if (strstr(kctl->id.name, "Playback"))
+                       cval->min_mute = 1;
+               break;
        }
 }
 
index d7b0b0a3a2db55617a908e2fe4a8a2af90082e02..6a03f9697039cc3d157bba92d5253b328e6ad45d 100644 (file)
@@ -1142,6 +1142,7 @@ bool snd_usb_get_sample_rate_quirk(struct snd_usb_audio *chip)
        case USB_ID(0x0556, 0x0014): /* Phoenix Audio TMX320VC */
        case USB_ID(0x05A3, 0x9420): /* ELP HD USB Camera */
        case USB_ID(0x074D, 0x3553): /* Outlaw RR2150 (Micronas UAC3553B) */
+       case USB_ID(0x1395, 0x740a): /* Sennheiser DECT */
        case USB_ID(0x1901, 0x0191): /* GE B850V3 CP2114 audio interface */
        case USB_ID(0x1de7, 0x0013): /* Phoenix Audio MT202exe */
        case USB_ID(0x1de7, 0x0014): /* Phoenix Audio TMX320 */
@@ -1374,6 +1375,10 @@ u64 snd_usb_interface_dsd_format_quirks(struct snd_usb_audio *chip,
                        }
                }
                break;
+       case USB_ID(0x16d0, 0x0a23):
+               if (fp->altsetting == 2)
+                       return SNDRV_PCM_FMTBIT_DSD_U32_BE;
+               break;
 
        default:
                break;
index 1a2c07eb7795bb4fb43e4a97bdc721d7cbc7f3b8..8c67a90dbd8229062fe1c0a9ae317fd5f3c1689b 100644 (file)
@@ -879,7 +879,8 @@ bpf_object__create_maps(struct bpf_object *obj)
                        size_t j;
                        int err = *pfd;
 
-                       pr_warning("failed to create map: %s\n",
+                       pr_warning("failed to create map (name: '%s'): %s\n",
+                                  obj->maps[i].name,
                                   strerror(errno));
                        for (j = 0; j < i; j++)
                                zclose(obj->maps[j].fd);
index e2fbb890aef9549281318a4241a2e5a8ad68b9d1..7c647f619d63faf3eff4c135c1b302565b7f78d5 100644 (file)
@@ -14,7 +14,7 @@ all:
        done
 
 override define RUN_TESTS
-       @if [ `dirname $(OUTPUT)` = $(PWD) ]; then ./run.sh; fi
+       $(OUTPUT)/run.sh
 endef
 
 override define INSTALL_RULE
old mode 100644 (file)
new mode 100755 (executable)
index 8cecae9..7956ea3
@@ -473,8 +473,8 @@ usage()
        echo "    all     Runs all tests (default)"
        echo "    -t      Run test ID the number amount of times is recommended"
        echo "    -w      Watch test ID run until it runs into an error"
-       echo "    -c      Run test ID once"
-       echo "    -s      Run test ID x test-count number of times"
+       echo "    -s      Run test ID once"
+       echo "    -c      Run test ID x test-count number of times"
        echo "    -l      List all test ID list"
        echo " -h|--help  Help"
        echo
old mode 100644 (file)
new mode 100755 (executable)
index e8c61830825a4ba1fedc9b4949561f5c61b185a5..22312eb4c9419a5fc3ee7054b4a0b604f7cdafce 100644 (file)
@@ -229,10 +229,9 @@ static void init_test(void)
        printf("CLOCK_MONOTONIC_RAW+CLOCK_MONOTONIC precision: %.0f ns\t\t",
               1e9 * precision);
 
-       if (precision > MAX_PRECISION) {
-               printf("[SKIP]\n");
-               ksft_exit_skip();
-       }
+       if (precision > MAX_PRECISION)
+               ksft_exit_skip("precision: %.0f ns > MAX_PRECISION: %.0f ns\n",
+                               1e9 * precision, 1e9 * MAX_PRECISION);
 
        printf("[OK]\n");
        srand(ts.tv_sec ^ ts.tv_nsec);