Back-to-back LPCG writes can be ignored by the LPCG register due to
a HW bug. The writes need to be separated by at least 4 cycles of
the gated clock. See https://www.nxp.com.cn/docs/en/errata/IMX8_1N94W.pdf
The workaround is implemented as follows:
1. For clocks running greater than or equal to 24MHz, a read
followed by the write will provide sufficient delay.
2. For clocks running below 24MHz, add a delay of 4 clock cylces
after the write to the LPCG register.
Fixes: 2f77296d3df9 ("clk: imx: add lpcg clock support") Signed-off-by: Peng Fan <peng.fan@nxp.com> Reviewed-by: Abel Vesa <abel.vesa@linaro.org> Link: https://lore.kernel.org/r/20241027-imx-clk-v1-v3-1-89152574d1d7@nxp.com Signed-off-by: Abel Vesa <abel.vesa@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In kvm_riscv_vcpu_sbi_init() the entry->ext_idx can contain an
out-of-bound index. This is used as a special marker for the base
extensions, that cannot be disabled. However, when traversing the
extensions, that special marker is not checked prior indexing the
array.
Add an out-of-bounds check to the function.
Fixes: 56d8a385b605 ("RISC-V: KVM: Allow some SBI extensions to be disabled by default") Signed-off-by: Björn Töpel <bjorn@rivosinc.com> Reviewed-by: Anup Patel <anup@brainfault.org> Link: https://lore.kernel.org/r/20241104191503.74725-1-bjorn@kernel.org Signed-off-by: Anup Patel <anup@brainfault.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the section "4.7 Precise effects on interrupt-pending bits"
of the RISC-V AIA specification defines that:
"If the source mode is Level1 or Level0 and the interrupt domain
is configured in MSI delivery mode (domaincfg.DM = 1):
The pending bit is cleared whenever the rectified input value is
low, when the interrupt is forwarded by MSI, or by a relevant
write to an in_clrip register or to clripnum."
Update the aplic_write_pending() to match the spec.
Same with commit e375b9c92985 ("RDMA/cxgb4: Set queue pair state when
being queried"). The API for ib_query_qp requires the driver to set
cur_qp_state on return, add the missing set.
Fixes: 8700e3e7c485 ("Soft RoCE driver") Signed-off-by: Liu Jian <liujian56@huawei.com> Link: https://patch.msgid.link/20241031092019.2138467-1-liujian56@huawei.com Reviewed-by: Zhu Yanjun <yanjun.zhu@linux.dev> Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
While computing foutpostdiv_rate, the value of params->pl5_fracin
is discarded, which results in the wrong refresh rate. Fix the formula
for computing foutpostdiv_rate.
To work around a limitation in our clock modelling, we try to force two
bits in the AUDIO0 PLL to 0, in the CCU probe routine.
However the ~ operator only applies to the first expression, and does
not cover the second bit, so we end up clearing only bit 1.
Group the bit-ORing with parentheses, to make it both clearer to read
and actually correct.
Fixes: 35b97bb94111 ("clk: sunxi-ng: Add support for the D1 SoC clocks") Signed-off-by: Andre Przywara <andre.przywara@arm.com> Link: https://patch.msgid.link/20241001105016.1068558-1-andre.przywara@arm.com Signed-off-by: Chen-Yu Tsai <wens@csie.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Invalidate rkey is cpu endian and immediate data is in big endian format.
Both immediate data and invalidate the remote key returned by
HW is in little endian format.
While handling the commit in fixes tag, the difference between
immediate data and invalidate rkey endianness was not considered.
Without changes of this patch, Kernel ULP was failing while processing
inv_rkey.
dmesg log snippet -
nvme nvme0: Bogus remote invalidation for rkey 0x2000019Fix in this patch
Do endianness conversion based on completion queue entry flag.
Also, the HW completions are already converted to host endianness in
bnxt_qplib_cq_process_res_rc and bnxt_qplib_cq_process_res_ud and there
is no need to convert it again in bnxt_re_poll_cq. Modified the union to
hold the correct data type.
During reset, cmd to destroy resources such as qp, cq, and mr may fail,
and error logs will be printed. When a large number of resources are
destroyed, there will be lots of printings, and it may lead to a cpu
stuck.
Delete some unnecessary printings and replace other printing functions
in these paths with the ratelimited version.
Fixes: 9a4435375cd1 ("IB/hns: Add driver files for hns RoCE driver") Fixes: c7bcb13442e1 ("RDMA/hns: Add SRQ support for hip08 kernel mode") Fixes: 70f92521584f ("RDMA/hns: Use the reserved loopback QPs to free MR before destroying MPT") Fixes: 926a01dc000d ("RDMA/hns: Add QP operations support for hip08 SoC") Signed-off-by: wenglianfa <wenglianfa@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20241024124000.2931869-6-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The sub-directory of hns_roce debugfs is named after the device's
kernel name currently, but it will be inconvenient to use when
the device is renamed.
Modify the name to pci name as users can always easily find the
correspondence between an RDMA device and its pci name.
Fixes: eb7854d63db5 ("RDMA/hns: Support SW stats with debugfs") Signed-off-by: Yuyu Li <liyuyu6@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20241024124000.2931869-4-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
QP needs to be modified to IB_QPS_ERROR to trigger HW flush cqe. But
when this process races with destroy qp, the destroy-qp process may
modify the QP to IB_QPS_RESET first. In this case flush cqe will fail
since it is invalid to modify qp from IB_QPS_RESET to IB_QPS_ERROR.
Add lock and bit flag to make sure pending flush cqe work is completed
first and no more new works will be added.
eq_db_ci is updated only after all AEQEs are processed in the AEQ
interrupt handler, which is not timely enough and may result in
AEQ overflow. Two optimization methods are proposed:
1. Set an upper limit for AEQE processing.
2. Move time-consuming operations such as printings to the bottom
half of the interrupt.
cmd events and flush_cqe events are still fully processed in the top half
to ensure timely handling.
Fixes: a5073d6054f7 ("RDMA/hns: Add eq support of hip08") Signed-off-by: wenglianfa <wenglianfa@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20241024124000.2931869-2-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit c7fc12354be0 ("iommu/amd/pgtbl_v2: Invalidate updated page ranges
only") missed to take domain lock before calling
amd_iommu_domain_flush_pages(). Fix this by taking protection domain
lock before calling TLB invalidation function.
The AMD io_pgtable stuff doesn't implement the tlb ops callbacks, instead
it invokes the invalidation ops directly on the struct protection_domain.
Narrow the use of struct protection_domain to only those few code paths.
Make everything else properly use struct amd_io_pgtable through the call
chains, which is the correct modular type for an io-pgtable module.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Vasant Hegde <vasant.hegde@amd.com> Link: https://lore.kernel.org/r/9-v2-831cdc4d00f3+1a315-amd_iopgtbl_jgg@nvidia.com Signed-off-by: Joerg Roedel <jroedel@suse.de>
Stable-dep-of: 016991606aa0 ("iommu/amd/pgtbl_v2: Take protection domain lock before invalidating TLB") Signed-off-by: Sasha Levin <sashal@kernel.org>
We already have memory in the union here that is being wasted in AMD's
case, use it to store the nid.
Putting the nid here further isolates the io_pgtable code from the struct
protection_domain.
Fixup protection_domain_alloc so that the NID from the device is provided,
at this point dev is never NULL for AMD so this will now allocate the
first table pointer on the correct NUMA node.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Vasant Hegde <vasant.hegde@amd.com> Link: https://lore.kernel.org/r/8-v2-831cdc4d00f3+1a315-amd_iopgtbl_jgg@nvidia.com Signed-off-by: Joerg Roedel <jroedel@suse.de>
Stable-dep-of: 016991606aa0 ("iommu/amd/pgtbl_v2: Take protection domain lock before invalidating TLB") Signed-off-by: Sasha Levin <sashal@kernel.org>
There is struct protection_domain iopt and struct amd_io_pgtable iopt.
Next patches are going to want to write domain.iopt.iopt.xx which is quite
unnatural to read.
Give one of them a different name, amd_io_pgtable has fewer references so
call it pgtbl, to match pgtbl_cfg, instead.
It is a serious bug if the domain is still mapped to any DTEs when it is
freed as we immediately start freeing page table memory, so any remaining
HW touch will UAF.
If it is not mapped then dev_list is empty and amd_iommu_domain_update()
does nothing.
Remove it and add a WARN_ON() to catch this class of bug.
Reviewed-by: Vasant Hegde <vasant.hegde@amd.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Link: https://lore.kernel.org/r/4-v2-831cdc4d00f3+1a315-amd_iopgtbl_jgg@nvidia.com Signed-off-by: Joerg Roedel <jroedel@suse.de>
Stable-dep-of: 016991606aa0 ("iommu/amd/pgtbl_v2: Take protection domain lock before invalidating TLB") Signed-off-by: Sasha Levin <sashal@kernel.org>
cpufreq_cpu_get_raw() may return NULL if the cpu is not in
policy->cpus cpu mask and it will cause null pointer dereference,
so check NULL for cppc_get_cpu_cost().
Fixes: 740fcdc2c20e ("cpufreq: CPPC: Register EM based on efficiency class information") Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In octal DTR mode, RD_ANY_REG_OP needs to use 4-byte address regardless
of flash's internal address mode. Use nor->addr_nbytes which is set to 4
during setup.
This was found by a static analyzer.
There may be a potential integer overflow issue in
sg2042_pll_recalc_rate(). numerator is defined as u64 while
parent_rate is defined as unsigned long and ctrl_table.fbdiv
is defined as unsigned int. On 32-bit machine, the result of
the calculation will be limited to "u32" without correct casting.
Integer overflow may occur on high-performance systems.
Fixes: 48cf7e01386e ("clk: sophgo: Add SG2042 clock driver") Signed-off-by: Zichen Xie <zichenxie0106@gmail.com> Reviewed-by: Chen Wang <unicorn_wang@outlook.com> Link: https://lore.kernel.org/r/20241023145146.13130-1-zichenxie0106@gmail.com Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org> Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
copy_from_kernel_nofault() can be called when doing read of /proc/kcore.
/proc/kcore can have some unmapped kfence objects which when read via
copy_from_kernel_nofault() can cause page faults. Since *_nofault()
functions define their own fixup table for handling fault, use that
instead of asking kfence to handle such faults.
Hence we search the exception tables for the nip which generated the
fault. If there is an entry then we let the fixup table handler handle the
page fault by returning an error from within ___do_page_fault().
This can be easily triggered if someone tries to do dd from /proc/kcore.
eg. dd if=/proc/kcore of=/dev/null bs=1M
Some example false negatives:
===============================
BUG: KFENCE: invalid read in copy_from_kernel_nofault+0x9c/0x1a0
Invalid read at 0xc0000000fdff0000:
copy_from_kernel_nofault+0x9c/0x1a0
0xc00000000665f950
read_kcore_iter+0x57c/0xa04
proc_reg_read_iter+0xe4/0x16c
vfs_read+0x320/0x3ec
ksys_read+0x90/0x154
system_call_exception+0x120/0x310
system_call_vectored_common+0x15c/0x2ec
BUG: KFENCE: use-after-free read in copy_from_kernel_nofault+0x9c/0x1a0
Use-after-free read at 0xc0000000fe050000 (in kfence-#2):
copy_from_kernel_nofault+0x9c/0x1a0
0xc00000000665f950
read_kcore_iter+0x57c/0xa04
proc_reg_read_iter+0xe4/0x16c
vfs_read+0x320/0x3ec
ksys_read+0x90/0x154
system_call_exception+0x120/0x310
system_call_vectored_common+0x15c/0x2ec
The pmecc "user" structure is allocated in atmel_pmecc_create_user() and
was supposed to be freed with atmel_pmecc_destroy_user(), but this other
helper is never called. One solution would be to find the proper
location to call the destructor, but the trend today is to switch to
device managed allocations, which in this case fits pretty well.
Replace kzalloc() by devm_kzalloc() and drop the destructor entirely.
Reported-by: "Dr. David Alan Gilbert" <linux@treblig.org> Closes: https://lore.kernel.org/all/ZvmIvRJCf6VhHvpo@gallifrey/ Fixes: f88fc122cc34 ("mtd: nand: Cleanup/rework the atmel_nand driver") Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/linux-mtd/20241001203149.387655-1-miquel.raynal@bootlin.com Signed-off-by: Sasha Levin <sashal@kernel.org>
During early init CMA_MIN_ALIGNMENT_BYTES can be PAGE_SIZE,
since pageblock_order is still zero and it gets initialized
later during initmem_init() e.g.
setup_arch() -> initmem_init() -> sparse_init() -> set_pageblock_order()
One such use case where this causes issue is -
early_setup() -> early_init_devtree() -> fadump_reserve_mem() -> fadump_cma_init()
This causes CMA memory alignment check to be bypassed in
cma_init_reserved_mem(). Then later cma_activate_area() can hit
a VM_BUG_ON_PAGE(pfn & ((1 << order) - 1)) if the reserved memory
area was not pageblock_order aligned.
Fix it by moving the fadump_cma_init() after initmem_init(),
where other such cma reservations also gets called.
We anyway don't use any return values from fadump_cma_init(). Since
fadump_reserve_mem() from where fadump_cma_init() gets called today,
already has the required checks.
This patch makes this function return type as void. Let's also handle
extra cases like return if fadump_supported is false or dump_active, so
that in later patches we can call fadump_cma_init() separately from
setup_arch().
Acked-by: Hari Bathini <hbathini@linux.ibm.com> Reviewed-by: Madhavan Srinivasan <maddy@linux.ibm.com> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://patch.msgid.link/a2afc3d6481a87a305e89cfc4a3f3d2a0b8ceab3.1729146153.git.ritesh.list@gmail.com
Stable-dep-of: 05b94cae1c47 ("powerpc/fadump: Move fadump_cma_init to setup_arch() after initmem_init()") Signed-off-by: Sasha Levin <sashal@kernel.org>
When cpufreq_register_driver() returns error, the cpufreq_init() returns
without unregister platform_driver, fix by add missing
platform_driver_unregister() when cpufreq_register_driver() failed.
Fixes: f8ede0f700f5 ("MIPS: Loongson 2F: Add CPU frequency scaling support") Signed-off-by: Yuan Can <yuancan@huawei.com> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When multiple IRQ domains are created from the same device-tree node they
will get the same name based on the device-tree path. This will cause a
naming collision in debugFS when IRQ domain specific entries are created.
The regmap-IRQ creates per instance IRQ domains. This will lead to a
domain name conflict when a device which provides more than one
interrupt line uses the regmap-IRQ.
Add support for specifying an IRQ domain name suffix when creating a
regmap-IRQ controller.
Devices can provide multiple interrupt lines. One reason for this is that
a device has multiple subfunctions, each providing its own interrupt line.
Another reason is that a device can be designed to be used (also) on a
system where some of the interrupts can be routed to another processor.
A line often further acts as a demultiplex for specific interrupts
and has it's respective set of interrupt (status, mask, ack, ...)
registers.
Regmap supports the handling of these registers and demultiplexing
interrupts, but the interrupt domain code ends up assigning the same name
for the per interrupt line domains. This causes a naming collision in the
debugFS code and leads to confusion, as /proc/interrupts shows two separate
interrupts with the same domain name and hardware interrupt number.
Instead of adding a workaround in regmap or driver code, allow giving a
name suffix for the domain name when the domain is created.
Add a name_suffix field in the irq_domain_info structure and make
irq_domain_instantiate() use this suffix if it is given when a domain is
created.
[ tglx: Adopt it to the cleanup patch and fixup the invalid NULL return ]
irq_domain_create_simple() and irq_domain_create_legacy() use
__irq_domain_instantiate(), but have extra handling of allocating interrupt
descriptors and associating interrupts in them. Some of that is duplicated.
There are also call sites which have conditonals to invoke different
interrupt domain creator functions, where one of them is usually
irq_domain_create_legacy(). Alternatively they associate the interrupts for
the legacy case after creating the domain.
Moving the extra logic of irq_domain_create_simple()/legacy() into
__irq_domain_instantiate() allows to consolidate that.
Introduce hwirq_base and virq_base members in the irq_domain_info
structure, which allows to transport the required information and add the
conditional interrupt descriptor allocation and interrupt association into
__irq_domain_instantiate().
This reduces irq_domain_create_legacy() and irq_domain_create_simple() to
trivial wrappers which fill in the info structure and allows call sites
which must support the legacy case along with more modern mechanism to
select the domain type via the parameters of the info struct.
While design wise the idea of converting the driver to use
the hierarchy of the IRQ chips is correct, the implementation
has (inherited) flaws. This was unveiled when platform_get_irq()
had started WARN() on IRQ 0 that is supposed to be a Linux
IRQ number (also known as vIRQ).
Rework the driver to respect IRQ domain when creating each MFD
device separately, as the domain is not the same for all of them.
Fixes: 57129044f504 ("mfd: intel_soc_pmic_bxtwc: Use chained IRQs for second level IRQ chips") Tested-by: Zhang Ning <zhangn1985@outlook.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Link: https://lore.kernel.org/r/20241005193029.1929139-4-andriy.shevchenko@linux.intel.com Signed-off-by: Lee Jones <lee@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
While design wise the idea of converting the driver to use
the hierarchy of the IRQ chips is correct, the implementation
has (inherited) flaws. This was unveiled when platform_get_irq()
had started WARN() on IRQ 0 that is supposed to be a Linux
IRQ number (also known as vIRQ).
Rework the driver to respect IRQ domain when creating each MFD
device separately, as the domain is not the same for all of them.
Fixes: 957ae5098185 ("platform/x86: Add Whiskey Cove PMIC TMU support") Fixes: 57129044f504 ("mfd: intel_soc_pmic_bxtwc: Use chained IRQs for second level IRQ chips") Reported-by: Zhang Ning <zhangn1985@outlook.com> Closes: https://lore.kernel.org/r/TY2PR01MB3322FEDCDC048B7D3794F922CDBA2@TY2PR01MB3322.jpnprd01.prod.outlook.com Tested-by: Zhang Ning <zhangn1985@outlook.com> Acked-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Link: https://lore.kernel.org/r/20241005193029.1929139-3-andriy.shevchenko@linux.intel.com Signed-off-by: Lee Jones <lee@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
While design wise the idea of converting the driver to use
the hierarchy of the IRQ chips is correct, the implementation
has (inherited) flaws. This was unveiled when platform_get_irq()
had started WARN() on IRQ 0 that is supposed to be a Linux
IRQ number (also known as vIRQ).
Rework the driver to respect IRQ domain when creating each MFD
device separately, as the domain is not the same for all of them.
Fixes: 9c6235c86332 ("mfd: intel_soc_pmic_bxtwc: Add bxt_wcove_usbc device") Fixes: d2061f9cc32d ("usb: typec: add driver for Intel Whiskey Cove PMIC USB Type-C PHY") Fixes: 57129044f504 ("mfd: intel_soc_pmic_bxtwc: Use chained IRQs for second level IRQ chips") Reported-by: Zhang Ning <zhangn1985@outlook.com> Closes: https://lore.kernel.org/r/TY2PR01MB3322FEDCDC048B7D3794F922CDBA2@TY2PR01MB3322.jpnprd01.prod.outlook.com Tested-by: Zhang Ning <zhangn1985@outlook.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20241005193029.1929139-2-andriy.shevchenko@linux.intel.com Signed-off-by: Lee Jones <lee@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
As the comment said, disable_irq() after request_irq() still has a
time gap in which interrupts can come. request_irq() with IRQF_NO_AUTOEN
flag will disable IRQ auto-enable when request IRQ.
Symbol table '.dynsym' contains 12 entries:
Num: Value Size Type Bind Vis Ndx Name
...
1: 0000000000000524 84 NOTYPE GLOBAL DEFAULT 8 __[...]@@LINUX_2.6.15
...
4: 0000000000000000 0 OBJECT GLOBAL DEFAULT ABS LINUX_2.6.15
5: 00000000000006c0 48 NOTYPE GLOBAL DEFAULT 8 __[...]@@LINUX_2.6.15
Symbol table '.symtab' contains 56 entries:
Num: Value Size Type Bind Vis Ndx Name
...
45: 0000000000000000 0 OBJECT GLOBAL DEFAULT ABS LINUX_2.6.15
46: 00000000000006c0 48 NOTYPE GLOBAL DEFAULT 8 __kernel_getcpu
47: 0000000000000524 84 NOTYPE GLOBAL DEFAULT 8 __kernel_clock_getres
To overcome that, commit ba83b3239e65 ("selftests: vDSO: fix vDSO
symbols lookup for powerpc64") was applied to have selftests also
look for NOTYPE symbols, but the correct fix should be to flag VDSO
entry points as functions.
The original commit that brought VDSO support into powerpc/64 has the
following explanation:
Note that the symbols exposed by the vDSO aren't "normal" function symbols, apps
can't be expected to link against them directly, the vDSO's are both seen
as if they were linked at 0 and the symbols just contain offsets to the
various functions. This is done on purpose to avoid a relocation step
(ppc64 functions normally have descriptors with abs addresses in them).
When glibc uses those functions, it's expected to use it's own trampolines
that know how to reach them.
The descriptors it's talking about are the OPD function descriptors
used on ABI v1 (big endian). But it would be more correct for a text
symbol to have type function, even if there's no function descriptor
for it.
glibc has a special case already for handling the VDSO symbols which
creates a fake opd pointing at the kernel symbol. So changing the VDSO
symbol type to function shouldn't affect that.
For ABI v2, there is no function descriptors and VDSO functions can
safely have function type.
So lets flag VDSO entry points as functions and revert the
selftest change.
For the controller reset operation(such as FLR or clear nexus ha in SCSI
EH), we will disable all PHYs and then enable PHY based on the
hisi_hba->phy_state obtained in hisi_sas_controller_reset_prepare(). If
the device is removed before controller reset or the PHY is not attached
to any device in directly attached scenario, the corresponding bit of
phy_state is not set. After controller reset done, the PHY is disabled.
The device cannot be identified even if user reconnect the disk.
Therefore, for PHYs that are not disabled by user, hisi_sas_phy_enable()
needs to be executed even if the corresponding bit of phy_state is not
set.
Fixes: 89954f024c3a ("scsi: hisi_sas: Ensure all enabled PHYs up during controller reset") Signed-off-by: Yihang Li <liyihang9@huawei.com> Link: https://lore.kernel.org/r/20241008021822.2617339-5-liyihang9@huawei.com Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This fixes a crash when surprise hot-unplugging a PCI device. This crash
happens because during hot-unplug __iommu_group_set_domain_nofail()
attaching the default domain fails when the platform no longer
recognizes the device as it has already been removed and we end up with
a NULL domain pointer and UAF. This is exactly the case referred to in
the second comment in __iommu_device_set_domain() and just as stated
there if we can instead attach the blocking domain the UAF is prevented
as this can handle the already removed device. Implement the blocking
domain to use this handling. With this change, the crash is fixed but
we still hit a warning attempting to change DMA ownership on a blocked
device.
When a tracepoint event is created with attr.freq = 1,
'hwc->period_left' is not initialized correctly. As a result,
in the perf_swevent_overflow() function, when the first time the event occurs,
it calculates the event overflow and the perf_swevent_set_period() returns 3,
this leads to the event are recorded for three duplicate times.
Commit 0f471d31e5e8 ("clk: mediatek: Split MT8195 clock drivers and allow
module build") adds a number of new COMMON_CLK_MT8195_* config options.
Among those, the config options COMMON_CLK_MT8195_AUDSYS and
COMMON_CLK_MT8195_MSDC have no reference in the source tree and are not
used in the Makefile to include a specific file.
Drop the dead config options COMMON_CLK_MT8195_AUDSYS and
COMMON_CLK_MT8195_MSDC.
Fixes: 0f471d31e5e8 ("clk: mediatek: Split MT8195 clock drivers and allow module build") Signed-off-by: Lukas Bulwahn <lukas.bulwahn@redhat.com> Link: https://lore.kernel.org/r/20240927092232.386511-1-lukas.bulwahn@redhat.com Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When HW is being reset, userspace should not ring doorbell otherwise
it may lead to abnormal consequence such as RAS.
Disassociate mmap pages for all uctx to prevent userspace from ringing
doorbell to HW. Since all resources will be destroyed during HW reset,
no new mmap is allowed after HW reset is completed.
Fixes: 9a4435375cd1 ("IB/hns: Add driver files for hns RoCE driver") Signed-off-by: Chengchang Tang <tangchengchang@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20240927103323.1897094-3-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Provide a new api rdma_user_mmap_disassociate() for drivers to
disassociate mmap pages for a device.
Since drivers can now disassociate mmaps by calling this api,
introduce a new disassociation_lock to specifically prevent
races between this disassociation process and new mmaps. And
thus the old hw_destroy_rwsem is not needed in this api.
Signed-off-by: Chengchang Tang <tangchengchang@huawei.com> Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com> Link: https://patch.msgid.link/20240927103323.1897094-2-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky <leon@kernel.org>
Stable-dep-of: 615b94746a54 ("RDMA/hns: Disassociate mmap pages for all uctx when HW is being reset") Signed-off-by: Sasha Levin <sashal@kernel.org>
The CPPC performance feedback counters could be 0 or unchanged when the
target cpu is in a low-power idle state, e.g. power-gated or clock-gated.
When the counters are 0, cppc_cpufreq_get_rate() returns 0 KHz, which makes
cpufreq_online() get a false error and fail to generate a cpufreq policy.
When the counters are unchanged, the existing cppc_perf_from_fbctrs()
returns a cached desired perf, but some platforms may update the real
frequency back to the desired perf reg.
For the above cases in cppc_cpufreq_get_rate(), get the latest desired perf
from the CPPC reg to reflect the frequency because some platforms may
update the actual frequency back there; if failed, use the cached desired
perf.
Fixes: 6a4fec4f6d30 ("cpufreq: cppc: cppc_cpufreq_get_rate() returns zero in all error cases.") Signed-off-by: Jie Zhan <zhanjie9@hisilicon.com> Reviewed-by: Zeng Heng <zengheng4@huawei.com> Reviewed-by: Ionela Voinescu <ionela.voinescu@arm.com> Reviewed-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
utf8_load() requests the symbol "utf8_data_table" and then checks if the
requested UTF-8 version is supported. If it's unsupported, it tries to
put the data table using symbol_put(). If an unsupported version is
requested, symbol_put() fails like this:
That happens because symbol_put() expects the unique string that
identify the symbol, instead of a pointer to the loaded symbol. Fix that
by using such string.
When the stream_verdict program returns SK_PASS, it places the received skb
into its own receive queue, but a recursive lock eventually occurs, leading
to an operating system deadlock. This issue has been present since v6.9.
Some distros may not load nf_conntrack by default, which will cause
subsequent nf_conntrack sets to fail. Load this module if it is not
already loaded.
Fixes: e7096c131e51 ("net: WireGuard secure network tunnel") Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Reviewed-by: Simon Horman <horms@kernel.org>
[ Jason: add [[ -e ... ]] check so this works in the qemu harness. ] Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Link: https://patch.msgid.link/20241117212030.629159-4-Jason@zx2c4.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The ndev->npinfo pointer in netpoll_poll_lock() is RCU-protected but is
being accessed directly for a NULL check. While no RCU read lock is held
in this context, we should still use proper RCU primitives for
consistency and correctness.
Replace the direct NULL check with rcu_access_pointer(), which is the
appropriate primitive when only checking for NULL without dereferencing
the pointer. This function provides the necessary ordering guarantees
without requiring RCU read-side protection.
Fixes: bea3348eef27 ("[NET]: Make NAPI polling independent of struct net_device objects.") Signed-off-by: Breno Leitao <leitao@debian.org> Reviewed-by: Michal Kubiak <michal.kubiak@intel.com> Link: https://patch.msgid.link/20241118-netpoll_rcu-v1-2-a1888dcb4a02@debian.org Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since the GPIO interrupt controller is always not working properly, we need
to constantly add workaround to cope with hardware deficiencies. So just
remove GPIO interrupt controller, and let the SFP driver poll the GPIO
status.
If dlm_recover_members() fails we don't drop the references of the
previous created root_list that holds and keep all rsbs alive during the
recovery. It might be not an unlikely event because ping_members() could
run into an -EINTR if another recovery progress was triggered again.
Fixes: 3a747f4a2ee8 ("dlm: move rsb root_list to ls_recover() stack") Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
syzbot reported a WARNING in iomap_iter_done:
iomap_fiemap+0x73b/0x9b0 fs/iomap/fiemap.c:80
ioctl_fiemap fs/ioctl.c:220 [inline]
Generally, NONHEAD lclusters won't have delta[1]==0, except for crafted
images and filesystems created by pre-1.0 mkfs versions.
Previously, it would immediately bail out if delta[1]==0, which led to
inadequate decompressed lengths (thus FIEMAP is impacted). Treat it as
delta[1]=1 to work around these legacy mkfs versions.
`lclusterbits > 14` is illegal for compact indexes, error out too.
Reported-by: syzbot+6c0b301317aa0156f9eb@syzkaller.appspotmail.com Closes: https://lore.kernel.org/r/67373c0c.050a0220.2a2fcc.0079.GAE@google.com Tested-by: syzbot+6c0b301317aa0156f9eb@syzkaller.appspotmail.com Fixes: d95ae5e25326 ("erofs: add support for the full decompressed length") Fixes: 001b8ccd0650 ("erofs: fix compact 4B support for 16k block size") Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com> Link: https://lore.kernel.org/r/20241115173651.3339514-1-hsiangkao@linux.alibaba.com Signed-off-by: Sasha Levin <sashal@kernel.org>
When a new skb is allocated for transmitting an xsk descriptor, i.e., for
every non-multibuf descriptor or the first frag of a multibuf descriptor,
but the descriptor is later found to have invalid options set for the TX
metadata, the new skb is never freed. This can leak skbs until the send
buffer is full which makes sending more packets impossible.
Fix this by freeing the skb in the error path if we are currently dealing
with the first frag, i.e., an skb allocated in this iteration of
xsk_build_skb.
Fixes: 48eb03dd2630 ("xsk: Add TX timestamp and TX checksum offload support") Reported-by: Michal Schmidt <mschmidt@redhat.com> Signed-off-by: Felix Maurer <fmaurer@redhat.com> Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com> Acked-by: Stanislav Fomichev <sdf@fomichev.me> Acked-by: Martin KaFai Lau <martin.lau@kernel.org> Link: https://patch.msgid.link/edb9b00fb19e680dff5a3350cd7581c5927975a8.1731581697.git.fmaurer@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In 'hci_conn_del_sysfs()', 'device_unregister()' may be called when
an underlying (kobject) reference counter is greater than 1. This
means that reparenting (happened when the device is actually freed)
is delayed and, during that delay, parent controller device (hciX)
may be deleted. Since the latter may create a dangling pointer to
freed parent, avoid that scenario by reparenting to NULL explicitly.
Reported-by: syzbot+6cf5652d3df49fae2e3f@syzkaller.appspotmail.com Tested-by: syzbot+6cf5652d3df49fae2e3f@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=6cf5652d3df49fae2e3f Fixes: a85fb91e3d72 ("Bluetooth: Fix double free in hci_conn_cleanup") Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Before issuing the LE BIG Create Sync command, an available BIG handle
is chosen by iterating through the conn_hash list and finding the first
unused value.
If a BIG is terminated, the associated hcons are removed from the list
and the LE BIG Terminate Sync command is sent via hci_sync queue.
However, a new LE BIG Create sync command might be issued via
hci_send_cmd, before the previous BIG sync was terminated. This
can cause the same BIG handle to be reused and the LE BIG Create Sync
to fail with Command Disallowed.
< HCI Command: LE Broadcast Isochronous Group Create Sync (0x08|0x006b)
BIG Handle: 0x00
BIG Sync Handle: 0x0002
Encryption: Unencrypted (0x00)
Broadcast Code[16]: 00000000000000000000000000000000
Maximum Number Subevents: 0x00
Timeout: 20000 ms (0x07d0)
Number of BIS: 1
BIS ID: 0x01
> HCI Event: Command Status (0x0f) plen 4
LE Broadcast Isochronous Group Create Sync (0x08|0x006b) ncmd 1
Status: Command Disallowed (0x0c)
< HCI Command: LE Broadcast Isochronous Group Terminate Sync (0x08|0x006c)
BIG Handle: 0x00
This commit fixes the ordering of the LE BIG Create Sync/LE BIG Terminate
Sync commands, to make sure that either the previous BIG sync is
terminated before reusing the handle, or that a new handle is chosen
for a new sync.
Fixes: eca0ae4aea66 ("Bluetooth: Add initial implementation of BIS connections") Signed-off-by: Iulia Tanasescu <iulia.tanasescu@nxp.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The Bluetooth Core spec does not allow a LE BIG Create sync command to be
sent to Controller if another one is pending (Vol 4, Part E, page 2586).
In order to avoid this issue, the HCI_CONN_CREATE_BIG_SYNC was added
to mark that the LE BIG Create Sync command has been sent for a hcon.
Once the BIG Sync Established event is received, the hcon flag is
erased and the next pending hcon is handled.
Signed-off-by: Iulia Tanasescu <iulia.tanasescu@nxp.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Stable-dep-of: 07a9342b94a9 ("Bluetooth: ISO: Send BIG Create Sync via hci_sync") Signed-off-by: Sasha Levin <sashal@kernel.org>
The Bluetooth Core spec does not allow a LE PA Create sync command to be
sent to Controller if another one is pending (Vol 4, Part E, page 2493).
In order to avoid this issue, the HCI_CONN_CREATE_PA_SYNC was added
to mark that the LE PA Create Sync command has been sent for a hcon.
Once the PA Sync Established event is received, the hcon flag is
erased and the next pending hcon is handled.
Signed-off-by: Iulia Tanasescu <iulia.tanasescu@nxp.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Stable-dep-of: 07a9342b94a9 ("Bluetooth: ISO: Send BIG Create Sync via hci_sync") Signed-off-by: Sasha Levin <sashal@kernel.org>
This make use of kref to keep track of reference of iso_conn which
allows better tracking of its lifetime with usage of things like
kref_get_unless_zero in a similar way as used in l2cap_chan.
In addition to it remove call to iso_sock_set_timer on iso_sock_disconn
since at that point it is useless to set a timer as the sk will be freed
there is nothing to be done in iso_sock_timeout.
Fixes: ccf74f2390d6 ("Bluetooth: Add BTPROTO_ISO socket type") Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
of_find_node_by_path() returns a pointer to a device_node with its
refcount incremented, and a call to of_node_put() is required to
decrement the refcount again and avoid leaking the resource.
If 'of_property_read_string_index(root, "compatible", 0, &tmp)' fails,
the function returns without calling of_node_put(root) before doing so.
The automatic cleanup attribute can be used by means of the __free()
macro to automatically call of_node_put() when the variable goes out of
scope, fixing the issue and also accounting for new error paths.
Fixes: 63fac3343b99 ("Bluetooth: btbcm: Support per-board firmware variants") Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
MediaTek iso data anchor init should be moved to where MediaTek
claims iso data interface.
If there is an unexpected BT usb disconnect during setup flow,
it will cause a NULL pointer crash issue when releasing iso
anchor since the anchor wasn't been init yet. Adjust the position
to do iso data anchor init.
Fixes: ceac1cb0259d ("Bluetooth: btusb: mediatek: add ISO data transmission functions") Signed-off-by: Chris Lu <chris.lu@mediatek.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
During firmware download, vendor specific events like boot up and
secure send result are generated. These events can be safely processed at
the driver level. Passing on these events to stack prints unnecessary
log as below.
The following handshake mechanism needs be followed after firmware
download is completed to bring the firmware to running state.
After firmware fragments of Operational image are downloaded and
secure sends result of the image succeeds,
1. Driver sends HCI Intel reset with boot option #1 to switch FW image.
2. FW sends Alive GP[0] MSIx
3. Driver enables data path (doorbell 0x460 for RBDs, etc...)
4. Driver gets Bootup event from firmware
5. Driver performs D0 entry to device (WRITE to IPC_Sleep_Control =0x0)
6. FW sends Alive GP[0] MSIx
7. Device host interface is fully set for BT protocol stack operation.
8. Driver may optionally get debug event with ID 0x97 which can be dropped
For Intermediate loadger image, all the above steps are applicable
expcept #5 and #6.
On HCI_OP_RESET, firmware raises alive interrupt. Driver needs to wait
for it before passing control over to bluetooth stack.
Co-developed-by: Devegowda Chandrashekar <chandrashekar.devegowda@intel.com> Signed-off-by: Devegowda Chandrashekar <chandrashekar.devegowda@intel.com> Signed-off-by: Kiran K <kiran.k@intel.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Stable-dep-of: 510e8380b038 ("Bluetooth: btintel: Do no pass vendor events to stack") Signed-off-by: Sasha Levin <sashal@kernel.org>
Early return in i2cdev_ioctl_rdwr() failed to free the memory allocated
by the caller. Move freeing the memory to the function where it has been
allocated to prevent similar leaks in the future.
Fixes: 97ca843f6ad3 ("i2c: dev: Check for I2C_FUNC_I2C before calling i2c_transfer") Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
[wsa: replaced '== NULL' with '!'] Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The current 6fire code tries to release the resources right after the
call of usb6fire_chip_abort(). But at this moment, the card object
might be still in use (as we're calling snd_card_free_when_closed()).
For avoid potential UAFs, move the release of resources to the card's
private_free instead of the manual call of usb6fire_chip_destroy() at
the USB disconnect callback.
The USB disconnect callback is supposed to be short and not too-long
waiting. OTOH, the current code uses snd_card_free() at
disconnection, but this waits for the close of all used fds, hence it
can take long. It eventually blocks the upper layer USB ioctls, which
may trigger a soft lockup.
An easy workaround is to replace snd_card_free() with
snd_card_free_when_closed(). This variant returns immediately while
the release of resources is done asynchronously by the card device
release at the last close.
This patch also splits the code to the disconnect and the free phases;
the former is called immediately at the USB disconnect callback while
the latter is called from the card destructor.
The USB disconnect callback is supposed to be short and not too-long
waiting. OTOH, the current code uses snd_card_free() at
disconnection, but this waits for the close of all used fds, hence it
can take long. It eventually blocks the upper layer USB ioctls, which
may trigger a soft lockup.
An easy workaround is to replace snd_card_free() with
snd_card_free_when_closed(). This variant returns immediately while
the release of resources is done asynchronously by the card device
release at the last close.
The loop of us122l->mmap_count check is dropped as well. The check is
useless for the asynchronous operation with *_when_closed().
The USB disconnect callback is supposed to be short and not too-long
waiting. OTOH, the current code uses snd_card_free() at
disconnection, but this waits for the close of all used fds, hence it
can take long. It eventually blocks the upper layer USB ioctls, which
may trigger a soft lockup.
An easy workaround is to replace snd_card_free() with
snd_card_free_when_closed(). This variant returns immediately while
the release of resources is done asynchronously by the card device
release at the last close.
Fixes: 230cd5e24853 ("[ALSA] prevent oops & dead keyboard on usb unplugging while the device is be ing used") Reported-by: syzbot+73582d08864d8268b6fd@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=73582d08864d8268b6fd Signed-off-by: Takashi Iwai <tiwai@suse.de> Link: https://patch.msgid.link/20241113111042.15058-2-tiwai@suse.de Signed-off-by: Sasha Levin <sashal@kernel.org>
Without kernel symbols for struct_ops trampoline, the unwinder may
produce unexpected stacktraces.
For example, the x86 ORC and FP unwinders check if an IP is in kernel
text by verifying the presence of the IP's kernel symbol. When a
struct_ops trampoline address is encountered, the unwinder stops due
to the absence of symbol, resulting in an incomplete stacktrace that
consists only of direct and indirect child functions called from the
trampoline.
The arm64 unwinder is another example. While the arm64 unwinder can
proceed across a struct_ops trampoline address, the corresponding
symbol name is displayed as "unknown", which is confusing.
Thus, add kernel symbol for struct_ops trampoline. The name is
bpf__<struct_ops_name>_<member_name>, where <struct_ops_name> is the
type name of the struct_ops, and <member_name> is the name of
the member that the trampoline is linked to.
Below is a comparison of stacktraces captured on x86 by perf record,
before and after this patch.
Only function pointers in a struct_ops structure can be linked to bpf
progs, so set the links count to the function pointers count, instead
of the total members count in the structure.
Suggested-by: Martin KaFai Lau <martin.lau@linux.dev> Signed-off-by: Xu Kuohai <xukuohai@huawei.com> Link: https://lore.kernel.org/r/20241112145849.3436772-3-xukuohai@huaweicloud.com Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Stable-dep-of: 7c8ce4ffb684 ("bpf: Add kernel symbol for struct_ops trampoline") Signed-off-by: Sasha Levin <sashal@kernel.org>
Alf reports that this commit causes the connection to eventually die on
iwl4965. The reason is that rx_status.flag is zeroed after
RX_FLAG_FAILED_FCS_CRC is set and mac80211 doesn't know the received frame is
corrupted.
Fixes: 02b682d54598 ("wifi: iwlegacy: do not skip frames with bad FCS") Reported-by: Alf Marius <post@alfmarius.net> Closes: https://lore.kernel.org/r/60f752e8-787e-44a8-92ae-48bdfc9b43e7@app.fastmail.com/ Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20241112142419.1023743-1-kvalo@kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Soft lockups have been observed on a cluster of Linux-based edge routers
located in a highly dynamic environment. Using the `bird` service, these
routers continuously update BGP-advertised routes due to frequently
changing nexthop destinations, while also managing significant IPv6
traffic. The lockups occur during the traversal of the multipath
circular linked-list in the `fib6_select_path` function, particularly
while iterating through the siblings in the list. The issue typically
arises when the nodes of the linked list are unexpectedly deleted
concurrently on a different core—indicated by their 'next' and
'previous' elements pointing back to the node itself and their reference
count dropping to zero. This results in an infinite loop, leading to a
soft lockup that triggers a system panic via the watchdog timer.
Apply RCU primitives in the problematic code sections to resolve the
issue. Where necessary, update the references to fib6_siblings to
annotate or use the RCU APIs.
Include a test script that reproduces the issue. The script
periodically updates the routing table while generating a heavy load
of outgoing IPv6 traffic through multiple iperf3 clients. It
consistently induces infinite soft lockups within a couple of minutes.
In kfd_procfs_show(), the sdma_activity_work_handler is a local variable
and the sdma_activity_work_handler.sdma_activity_work should initialize
with INIT_WORK_ONSTACK() instead of INIT_WORK().
Fixes: 32cb59f31362 ("drm/amdkfd: Track SDMA utilization per process") Signed-off-by: Yuan Can <yuancan@huawei.com> Signed-off-by: Felix Kuehling <felix.kuehling@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In current logic, it calls ring_alloc followed by a ring_test. ring_test
in turn will call another ring_alloc. This is illegal usage as a
ring_alloc is expected to be closed properly with a ring_commit. Change
to commit the map/unmap queue packet first followed by a ring_test. Add a
comment about the usage of ring_test.
Also, reorder the current pre-condition checks of job hang or kiq ring
scheduler not ready. Without them being met, it is not useful to attempt
ring or memory allocations.
Fixes tag refers to the original patch which introduced this issue which
then got carried over into newer code.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Le Ma <le.ma@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Fixes: 6c10b5cc4eaa ("drm/amdgpu: Remove duplicate code in gfx_v8_0.c") Signed-off-by: Sasha Levin <sashal@kernel.org>
Since we no longer shut down the device in suspend, we also no longer
call iwl_mvm_mei_device_state() and this is a problem because iwlmei
expects this to be called when it runs its own suspend sequence. It
checks mei->device_down in iwl_mei_remove() which is called upon
suspend.
Fix this by telling iwlmei when we're done accessing the device.
When we'll wake up, the device should be untouched if CSME didn't use it
during the suspend time. If CSME used it, we'll notice it through the
CSR_FUNC_SCRATCH register.
This feature can be used on ax200 as well. It'll avoid to restart the
firmware upon suspend / resume flow. Doing so also avoids releasing and
re-allocating all the device related memory which makes the memory's
subsystem task easier.
Currently, when the driver attempts to connect to an AP MLD with multiple
APs, the cfg80211_mlme_check_mlo_compat() function requires the Medium
Synchronization Delay values from different APs of the same AP MLD to be
equal, which may result in connection failures.
This is because when the driver receives a multi-link probe response from
an AP MLD with multiple APs, cfg80211 updates the Elements for each AP
based on the multi-link probe response. If the Medium Synchronization Delay
is set in the multi-link probe response, the Elements for each AP belonging
to the same AP MLD will have the Medium Synchronization Delay set
simultaneously. If non-multi-link probe responses are received from
different APs of the same MLD AP, cfg80211 will still update the Elements
based on the non-multi-link probe response. Since the non-multi-link probe
response does not set the Medium Synchronization Delay
(IEEE 802.11be-2024-35.3.4.4), if the Elements from a non-multi-link probe
response overwrite those from a multi-link probe response that has set the
Medium Synchronization Delay, the Medium Synchronization Delay values for
APs belonging to the same AP MLD will not be equal. This discrepancy causes
the cfg80211_mlme_check_mlo_compat() function to fail, leading to
connection failures. Commit ccb964b4ab16
("wifi: cfg80211: validate MLO connections better") did not take this into
account.
To address this issue, remove this validity check.
The CI is hitting some aperiodic hangup at device removal time in the
pmtu.sh self-test:
unregister_netdevice: waiting for veth_A-R1 to become free. Usage count = 6
ref_tracker: veth_A-R1@ffff888013df15d8 has 1/5 users at
dst_init+0x84/0x4a0
dst_alloc+0x97/0x150
ip6_dst_alloc+0x23/0x90
ip6_rt_pcpu_alloc+0x1e6/0x520
ip6_pol_route+0x56f/0x840
fib6_rule_lookup+0x334/0x630
ip6_route_output_flags+0x259/0x480
ip6_dst_lookup_tail.constprop.0+0x5c2/0x940
ip6_dst_lookup_flow+0x88/0x190
udp_tunnel6_dst_lookup+0x2a7/0x4c0
vxlan_xmit_one+0xbde/0x4a50 [vxlan]
vxlan_xmit+0x9ad/0xf20 [vxlan]
dev_hard_start_xmit+0x10e/0x360
__dev_queue_xmit+0xf95/0x18c0
arp_solicit+0x4a2/0xe00
neigh_probe+0xaa/0xf0
While the first suspect is the dst_cache, explicitly tracking the dst
owing the last device reference via probes proved such dst is held by
the nexthop in the originating fib6_info.
Similar to commit f5b51fe804ec ("ipv6: route: purge exception on
removal"), we need to explicitly release the originating fib info when
disconnecting a to-be-removed device from a live ipv6 dst: move the
fib6_info cleanup into ip6_dst_ifdown().
Tested running:
./pmtu.sh cleanup_ipv6_exception
in a tight loop for more than 400 iterations with no spat, running an
unpatched kernel I observed a splat every ~10 iterations.
Fixes: f88d8ea67fbd ("ipv6: Plumb support for nexthop object in a fib6_info") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: David Ahern <dsahern@kernel.org> Link: https://patch.msgid.link/604c45c188c609b732286b47ac2a451a40f6cf6d.1730828007.git.pabeni@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Found in the test_txmsg_pull in test_sockmap,
```
txmsg_cork = 512; // corking is importrant here
opt->iov_length = 3;
opt->iov_count = 1;
opt->rate = 512; // sendmsg will be invoked 512 times
```
The first sendmsg will send an sk_msg with size 3, and bpf_msg_pull_data
will be invoked the first time. sk_msg_reset_curr will reset the copybreak
from 3 to 0. In the second sendmsg, since we are in the stage of corking,
psock->cork will be reused in func sk_msg_alloc. msg->sg.copybreak is 0
now, the second msg will overwrite the first msg. As a result, we could
not pass the data integrity test.
The same problem happens in push and pop test. Thus, fix sk_msg_reset_curr
to restore the correct copybreak.
Fixes: bb9aefde5bba ("bpf: sockmap, updating the sg structure should also update curr") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Link: https://lore.kernel.org/r/20241106222520.527076-9-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Several fixes to bpf_msg_pop_data,
1. In sk_msg_shift_left, we should put_page
2. if (len == 0), return early is better
3. pop the entire sk_msg (last == msg->sg.size) should be supported
4. Fix for the value of variable "a"
5. In sk_msg_shift_left, after shifting, i has already pointed to the next
element. Addtional sk_msg_iter_var_next may result in BUG.
Fixes: 7246d8ed4dcc ("bpf: helper to pop data from messages") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20241106222520.527076-8-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Several fixes to bpf_msg_push_data,
1. test_sockmap has tests where bpf_msg_push_data is invoked to push some
data at the end of a message, but -EINVAL is returned. In this case, in
bpf_msg_push_data, after the first loop, i will be set to msg->sg.end, add
the logic to handle it.
2. In the code block of "if (start - offset)", it's possible that "i"
points to the last of sk_msg_elem. In this case, "sk_msg_iter_next(msg,
end)" might still be called twice, another invoking is in "if (!copy)"
code block, but actually only one is needed. Add the logic to handle it,
and reconstruct the code to make the logic more clear.
Fixes: 6fff607e2f14 ("bpf: sk_msg program helper bpf_msg_push_data") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Link: https://lore.kernel.org/r/20241106222520.527076-7-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add push/pop checking for msg_verify_data in test_sockmap, except for
pop/push with cork tests, in these tests the logic will be different.
1. With corking, pop/push might not be invoked in each sendmsg, it makes
the layout of the received data difficult
2. It makes it hard to calculate the total_bytes in the recvmsg
Temporarily skip the data integrity test for these cases now, added a TODO
Fixes: ee9b352ce465 ("selftests/bpf: Fix msg_verify_data in test_sockmap") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20241106222520.527076-5-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
total_bytes in msg_loop_rx should also take push into account, otherwise
total_bytes will be a smaller value, which makes the msg_loop_rx end early.
Besides, total_bytes has already taken pop into account, so we don't need
to subtract some bytes from iov_buf in sendmsg_test. The additional
subtraction may make total_bytes a negative number, and msg_loop_rx will
just end without checking anything.
Fixes: 18d4e900a450 ("bpf: Selftests, improve test_sockmap total bytes counter") Fixes: d69672147faa ("selftests, bpf: Add one test for sockmap with strparser") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20241106222520.527076-4-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In the SENDPAGE test, "opt->iov_length * cnt" size of data will be sent
cnt times by sendfile.
1. In push/pop tests, they will be invoked cnt times, for the simplicity of
msg_verify_data, change chunk_sz to iov_length
2. Change iov_length in test_send_large from 1024 to 8192. We have pop test
where txmsg_start_pop is 4096. 4096 > 1024, an error will be returned.
Fixes: 328aa08a081b ("bpf: Selftests, break down test_sockmap into subtests") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20241106222520.527076-3-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Add txmsg_pass to test_txmsg_pull/push/pop. If txmsg_pass is missing,
tx_prog will be NULL, and no program will be attached to the sockmap.
As a result, pull/push/pop are never invoked.
Fixes: 328aa08a081b ("bpf: Selftests, break down test_sockmap into subtests") Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com> Reviewed-by: John Fastabend <john.fastabend@gmail.com> Link: https://lore.kernel.org/r/20241106222520.527076-2-zijianzhang@bytedance.com Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The opt->iocharset is freed inside the isofs_fill_super function,
But there may be situations where it's not possible to
enter this function.
For example, in the get_tree_bdev_flags function,when
encountering the situation where "Can't mount, would change RO state,"
In such a case, isofs_fill_super will not have the opportunity
to be called,which means that opt->iocharset will not have the chance
to be freed,ultimately leading to a memory leak.
Let's move the memory freeing of opt->iocharset into
isofs_free_fc function.
Fixes: 1b17a46c9243 ("isofs: convert isofs to use the new mount API") Signed-off-by: Hao Ge <gehao@kylinos.cn> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Jan Kara <jack@suse.cz> Link: https://patch.msgid.link/20241106082841.51773-1-hao.ge@linux.dev Signed-off-by: Sasha Levin <sashal@kernel.org>
projects / thirdparty / kernel / stable.git / log
