Min Li [Mon, 16 May 2022 14:47:06 +0000 (10:47 -0400)]
ptp: ptp_clockmatrix: Add PTP_CLK_REQ_EXTTS support
Use TOD_READ_SECONDARY for extts to keep TOD_READ_PRIMARY
for gettime and settime exclusively. Before this change,
TOD_READ_PRIMARY was used for both extts and gettime/settime,
which would result in changing TOD read/write triggers between
operations. Using TOD_READ_SECONDARY would make extts
independent of gettime/settime operation
====================
net/smc: send and write inline optimization for smc
Send cdc msgs and write data inline if qp has sufficent inline
space, helps latency reducing.
In my test environment, which are 2 VMs running on the same
physical host and whose NICs(ConnectX-4Lx) are working on
SR-IOV mode, qperf shows 0.4us-1.3us improvement in latency.
The results shown below:
msgsize before after
1B 11.9 us 10.6 us (-1.3 us)
2B 11.7 us 10.7 us (-1.0 us)
4B 11.7 us 10.7 us (-1.0 us)
8B 11.6 us 10.6 us (-1.0 us)
16B 11.7 us 10.7 us (-1.0 us)
32B 11.7 us 10.6 us (-1.1 us)
64B 11.7 us 11.2 us (-0.5 us)
128B 11.6 us 11.2 us (-0.4 us)
256B 11.8 us 11.2 us (-0.6 us)
512B 11.8 us 11.3 us (-0.5 us)
1KB 11.9 us 11.5 us (-0.4 us)
2KB 12.1 us 11.5 us (-0.6 us)
====================
Guangguan Wang [Mon, 16 May 2022 05:51:37 +0000 (13:51 +0800)]
net/smc: rdma write inline if qp has sufficient inline space
Rdma write with inline flag when sending small packages,
whose length is shorter than the qp's max_inline_data, can
help reducing latency.
In my test environment, which are 2 VMs running on the same
physical host and whose NICs(ConnectX-4Lx) are working on
SR-IOV mode, qperf shows 0.5us-0.7us improvement in latency.
The results shown below:
msgsize before after
1B 11.2 us 10.6 us (-0.6 us)
2B 11.2 us 10.7 us (-0.5 us)
4B 11.3 us 10.7 us (-0.6 us)
8B 11.2 us 10.6 us (-0.6 us)
16B 11.3 us 10.7 us (-0.6 us)
32B 11.3 us 10.6 us (-0.7 us)
64B 11.2 us 11.2 us (0 us)
128B 11.2 us 11.2 us (0 us)
256B 11.2 us 11.2 us (0 us)
512B 11.4 us 11.3 us (-0.1 us)
1KB 11.4 us 11.5 us (0.1 us)
2KB 11.5 us 11.5 us (0 us)
Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com> Reviewed-by: Tony Lu <tonylu@linux.alibaba.com> Tested-by: kernel test robot <lkp@intel.com> Acked-by: Karsten Graul <kgraul@linux.ibm.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Guangguan Wang [Mon, 16 May 2022 05:51:36 +0000 (13:51 +0800)]
net/smc: send cdc msg inline if qp has sufficient inline space
As cdc msg's length is 44B, cdc msgs can be sent inline in
most rdma devices, which can help reducing sending latency.
In my test environment, which are 2 VMs running on the same
physical host and whose NICs(ConnectX-4Lx) are working on
SR-IOV mode, qperf shows 0.4us-0.7us improvement in latency.
The results shown below:
msgsize before after
1B 11.9 us 11.2 us (-0.7 us)
2B 11.7 us 11.2 us (-0.5 us)
4B 11.7 us 11.3 us (-0.4 us)
8B 11.6 us 11.2 us (-0.4 us)
16B 11.7 us 11.3 us (-0.4 us)
32B 11.7 us 11.3 us (-0.4 us)
64B 11.7 us 11.2 us (-0.5 us)
128B 11.6 us 11.2 us (-0.4 us)
256B 11.8 us 11.2 us (-0.6 us)
512B 11.8 us 11.4 us (-0.4 us)
1KB 11.9 us 11.4 us (-0.5 us)
2KB 12.1 us 11.5 us (-0.6 us)
Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com> Reviewed-by: Tony Lu <tonylu@linux.alibaba.com> Tested-by: kernel test robot <lkp@intel.com> Acked-by: Karsten Graul <kgraul@linux.ibm.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
test_bit() tests if one bit is set or not.
Here the logic seems to check of bit QL_RESET_PER_SCSI (i.e. 4) OR bit
QL_RESET_START (i.e. 3) is set.
In fact, it checks if bit 7 (4 | 3 = 7) is set, that is to say
QL_ADAPTER_UP.
This looks harmless, because this bit is likely be set, and when the
ql_reset_work() delayed work is scheduled in ql3xxx_isr() (the only place
that schedule this work), QL_RESET_START or QL_RESET_PER_SCSI is set.
These interconnects are modeled as clks, not interconnects, therefore
remove the compatibles from the binding as they're unused.
Cc: Alex Elder <elder@linaro.org> Cc: Taniya Das <quic_tdas@quicinc.com> Cc: Mike Tipton <quic_mdtipton@quicinc.com> Signed-off-by: Stephen Boyd <swboyd@chromium.org> Link: https://lore.kernel.org/r/20220415005828.1980055-3-swboyd@chromium.org Signed-off-by: Georgi Djakov <djakov@kernel.org>
Georgi Djakov [Wed, 18 May 2022 00:03:30 +0000 (03:03 +0300)]
Merge branch 'icc-rpm' into icc-next
This patch set is to address two clock rate setting issues.
The first patch is to fix a potential cached clock rate mismatching
issue, the issue can lead to the clock rate is missed to be set. Note,
since this potential issue requires specific time window and certain
condition (consumers need to request the same bandwidth) to produce,
the patch is based on analysis but not a real trace log.
The second patch is an extension to cache clock rates for active and
sleep clocks separately, with this change it gives us possibility to set
active and sleep clock with different clock rates.
* icc-rpm
interconnect: qcom: icc-rpm: Fix for cached clock rate
interconnect: qcom: icc-rpm: Cache every clock rate
Bjorn Andersson [Tue, 3 May 2022 22:53:00 +0000 (15:53 -0700)]
interconnect: qcom: sc8180x: Reformat node and bcm definitions
The Qualcomm interconnect providers started off defining nodes and BCMs
using the DEFINE_QNODE() and DEFINE_QBCM() macros. Unfortunately this
results in a block of long lines that are hard to read, a transition to
explicitly stated definition has been made for newly introduced
platforms.
Transition the SC8180X interconnect provider driver to this style as
well, to make it easier to read while debugging interconnect related
issues.
Bjorn Andersson [Tue, 3 May 2022 21:19:24 +0000 (14:19 -0700)]
interconnect: qcom: sc8180x: Fix QUP0 nodes
The QUP0 BCM relates to some internal property of the QUPs, and should
be configured independently of the path to the QUP. In line with other
platforms expose QUP_CORE endpoints in order allow this configuration.
The introduction of the Qualcomm SC8180X provider raced with the
refactoring of the RPMh common code and SC8180X was left with the old
style of duplicating the probe function in each provider driver.
Linus Torvalds [Tue, 17 May 2022 23:46:22 +0000 (13:46 -1000)]
Merge tag 'pci-v5.18-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull PCI fixes from Bjorn Helgaas:
- Avoid putting Elo i2 PCIe Ports in D3cold because downstream devices
are inaccessible after going back to D0 (Rafael J. Wysocki)
- Qualcomm SM8250 has a ddrss_sf_tbu clock but SC8180X does not; make a
SC8180X-specific config without the clock so it probes correctly
(Bjorn Andersson)
- Revert aardvark chained IRQ handler rewrite because it broke
interrupt affinity (Pali Rohár)
* tag 'pci-v5.18-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
Revert "PCI: aardvark: Rewrite IRQ code to chained IRQ handler"
PCI: qcom: Remove ddrss_sf_tbu clock from SC8180X
PCI/PM: Avoid putting Elo i2 PCIe Ports in D3cold
Linus Torvalds [Tue, 17 May 2022 23:40:44 +0000 (13:40 -1000)]
Merge tag 'thermal-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull thermal control fix from Rafael Wysocki:
"Fix up a recent change in the int340x thermal driver that
inadvertently broke thermal zone handling on some systems
(Srinivas Pandruvada)"
* tag 'thermal-5.18-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
thermal: int340x: Mode setting with new OS handshake
Now, you can use native gdb on riscv64 for rv32 app debugging.
$ uname -a
Linux buildroot 5.16.0-rc4-00036-gbef6b82fdf23-dirty #53 SMP Mon Dec 20 23:06:53 CST 2021 riscv64 GNU/Linux
$ cat /proc/cpuinfo
processor : 0
hart : 0
isa : rv64imafdcsuh
mmu : sv48
$ file /bin/busybox
/bin/busybox: setuid ELF 32-bit LSB shared object, UCB RISC-V, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-riscv32-ilp32d.so.1, for GNU/Linux 5.15.0, stripped
$ file /usr/bin/gdb
/usr/bin/gdb: ELF 32-bit LSB shared object, UCB RISC-V, version 1 (GNU/Linux), dynamically linked, interpreter /lib/ld-linux-riscv32-ilp32d.so.1, for GNU/Linux 5.15.0, stripped
$ /usr/bin/gdb /bin/busybox
GNU gdb (GDB) 10.2
Copyright (C) 2021 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
...
Reading symbols from /bin/busybox...
(No debugging symbols found in /bin/busybox)
(gdb) b main
Breakpoint 1 at 0x8ddc
(gdb) r
Starting program: /bin/busybox
Failed to read a valid object file image from memory.
Breakpoint 1, 0x555a8ddc in main ()
(gdb) i r
ra 0x77df0b74 0x77df0b74
sp 0x7fdd3d10 0x7fdd3d10
gp 0x5567e800 0x5567e800 <bb_common_bufsiz1+160>
tp 0x77f64280 0x77f64280
t0 0x0 0
t1 0x555a6fac 1431990188
t2 0x77dd8db4 2011008436
fp 0x7fdd3e34 0x7fdd3e34
s1 0x7fdd3e34 2145205812
a0 0xffffffff -1
a1 0x2000 8192
a2 0x7fdd3e3c 2145205820
a3 0x0 0
a4 0x7fdd3d30 2145205552
a5 0x555a8dc0 1431997888
a6 0x77f2c170 2012397936
a7 0x6a7c7a2f 1786542639
s2 0x0 0
s3 0x0 0
s4 0x555a8dc0 1431997888
s5 0x77f8a3a8 2012783528
s6 0x7fdd3e3c 2145205820
s7 0x5567cecc 1432866508
--Type <RET> for more, q to quit, c to continue without paging--
s8 0x1 1
s9 0x0 0
s10 0x55634448 1432568904
s11 0x0 0
t3 0x77df0bb8 2011106232
t4 0x42fc 17148
t5 0x0 0
t6 0x40 64
pc 0x555a8ddc 0x555a8ddc <main+28>
(gdb) si
0x555a78f0 in mallopt@plt ()
(gdb) c
Continuing.
BusyBox v1.34.1 (2021-12-19 22:39:48 CST) multi-call binary.
BusyBox is copyrighted by many authors between 1998-2015.
Licensed under GPLv2. See source distribution for detailed
copyright notices.
Implement compat_setup_rt_frame for sigcontext save & restore. The
main process is the same with signal, but the rv32 pt_regs' size
is different from rv64's, so we needs convert them.
Ondrej Mosnacek [Tue, 17 May 2022 12:08:16 +0000 (14:08 +0200)]
selinux: fix bad cleanup on error in hashtab_duplicate()
The code attempts to free the 'new' pointer using kmem_cache_free(),
which is wrong because this function isn't responsible of freeing it.
Instead, the function should free new->htable and clear the contents of
*new (to prevent double-free).
Cc: stable@vger.kernel.org Fixes: c7c556f1e81b ("selinux: refactor changing booleans") Reported-by: Wander Lairson Costa <wander@redhat.com> Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com> Signed-off-by: Paul Moore <paul@paul-moore.com>
Stephen Boyd [Tue, 17 May 2022 22:16:22 +0000 (15:16 -0700)]
Merge tag 'v5.19-rockchip-clk2' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into clk-rockchip
Pull Rockchip clk driver updates from Heiko Stuebner:
Conversion from txt to Yaml for a number of Rockchip
clock bindings.
Some fixes for recent yaml conversion of clock bindinds
and making the hclk_vo critical for rk3568.
* tag 'v5.19-rockchip-clk2' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
dt-bindings: clock: convert rockchip,rk3368-cru.txt to YAML
dt-bindings: clock: convert rockchip,rk3228-cru.txt to YAML
dt-bindings: clock: convert rockchip,rk3036-cru.txt to YAML
dt-bindings: clock: convert rockchip,rk3308-cru.txt to YAML
dt-bindings: clock: convert rockchip,px30-cru.txt to YAML
dt-bindings: clock: convert rockchip,rk3188-cru.txt to YAML
dt-bindings: clock: convert rockchip,rk3288-cru.txt to YAML
dt-bindings: clock: convert rockchip,rv1108-cru.txt to YAML
dt-binding: clock: Add missing rk3568 cru bindings
clk: rockchip: Mark hclk_vo as critical on rk3568
dt-bindings: clock: fix rk3399 cru clock issues
dt-bindings: clock: use generic node name for pmucru example in rockchip,rk3399-cru.yaml
dt-bindings: clock: replace a maintainer for rockchip,rk3399-cru.yaml
dt-bindings: clock: fix some conversion style issues for rockchip,rk3399-cru.yaml
where a move and a compare after cmpxchg is saved. The improvements
for 32-bit targets are even more noticeable, because dual-word compare
after cmpxchg8b gets eliminated.
Peter Zijlstra [Tue, 10 May 2022 19:22:04 +0000 (21:22 +0200)]
perf/x86/amd: Fix AMD BRS period adjustment
There's two problems with the current amd_brs_adjust_period() code:
- it isn't in fact AMD specific and wil always adjust the period;
- it adjusts the period, while it should only adjust the event count,
resulting in repoting a short period.
Fix this by using x86_pmu.limit_period, this makes it specific to the
AMD BRS case and ensures only the event count is adjusted while the
reported period is unmodified.
Fixes: ba2fe7500845 ("perf/x86/amd: Add AMD branch sampling period adjustment") Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Kees Cook [Tue, 17 May 2022 21:44:20 +0000 (14:44 -0700)]
lkdtm/heap: Hide allocation size from -Warray-bounds
With the kmalloc() size annotations, GCC is smart enough to realize that
LKDTM is intentionally writing past the end of the buffer. This is on
purpose, of course, so hide the buffer from the optimizer. Silences:
../drivers/misc/lkdtm/heap.c: In function 'lkdtm_SLAB_LINEAR_OVERFLOW':
../drivers/misc/lkdtm/heap.c:59:13: warning: array subscript 256 is outside array bounds of 'void[1020]' [-Warray-bounds]
59 | data[1024 / sizeof(u32)] = 0x12345678;
| ~~~~^~~~~~~~~~~~~~~~~~~~
In file included from ../drivers/misc/lkdtm/heap.c:7:
In function 'kmalloc',
inlined from 'lkdtm_SLAB_LINEAR_OVERFLOW' at ../drivers/misc/lkdtm/heap.c:54:14:
../include/linux/slab.h:581:24: note: at offset 1024 into object of size 1020 allocated by 'kmem_cache_alloc_trace'
581 | return kmem_cache_alloc_trace(
| ^~~~~~~~~~~~~~~~~~~~~~~
582 | kmalloc_caches[kmalloc_type(flags)][index],
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
583 | flags, size);
| ~~~~~~~~~~~~
selftests/lkdtm: Add configs for stackleak and "after free" tests
Add config options which are needed for LKDTM sub-tests:
STACKLEAK_ERASING test needs GCC_PLUGIN_STACKLEAK config.
READ_AFTER_FREE and READ_BUDDY_AFTER_FREE tests need
INIT_ON_FREE_DEFAULT_ON config.
When a Root Port or Root Complex Event Collector receives an error Message
e.g., ERR_COR, it sets PCI_ERR_ROOT_COR_RCV in the Root Error Status
register and logs the Requester ID in the Error Source Identification
register. If it receives a second ERR_COR Message before software clears
PCI_ERR_ROOT_COR_RCV, hardware sets PCI_ERR_ROOT_MULTI_COR_RCV and the
Requester ID is lost.
In the following scenario, PCI_ERR_ROOT_MULTI_COR_RCV was never cleared:
- hardware receives ERR_COR message
- hardware sets PCI_ERR_ROOT_COR_RCV
- aer_irq() entered
- aer_irq(): status = pci_read_config_dword(PCI_ERR_ROOT_STATUS)
- aer_irq(): now status == PCI_ERR_ROOT_COR_RCV
- hardware receives second ERR_COR message
- hardware sets PCI_ERR_ROOT_MULTI_COR_RCV
- aer_irq(): pci_write_config_dword(PCI_ERR_ROOT_STATUS, status)
- PCI_ERR_ROOT_COR_RCV is cleared; PCI_ERR_ROOT_MULTI_COR_RCV is set
- aer_irq() entered again
- aer_irq(): status = pci_read_config_dword(PCI_ERR_ROOT_STATUS)
- aer_irq(): now status == PCI_ERR_ROOT_MULTI_COR_RCV
- aer_irq() exits because PCI_ERR_ROOT_COR_RCV not set
- PCI_ERR_ROOT_MULTI_COR_RCV is still set
The same problem occurred with ERR_NONFATAL/ERR_FATAL Messages and
PCI_ERR_ROOT_UNCOR_RCV and PCI_ERR_ROOT_MULTI_UNCOR_RCV.
Fix the problem by queueing an AER event and clearing the Root Error Status
bits when any of these bits are set:
Miquel Raynal [Mon, 16 May 2022 08:25:03 +0000 (10:25 +0200)]
rtc: rzn1: Add oscillator offset support
The RZN1 RTC can compensate the imprecision of the oscillator up to
approximately 190ppm.
Seconds can last slightly shorter or longer depending on the
configuration.
Below ~65ppm of correction, we can change the time spent in a second
every minute, which is the most accurate compensation that the RTC can
offer. Above, the compensation will be active every 20s.
Samuel Holland [Wed, 13 Apr 2022 23:17:30 +0000 (18:17 -0500)]
rtc: sun6i: Add NVMEM provider
The sun6i RTC provides 32 bytes of general-purpose data registers.
They can be used to save data in the always-on RTC power domain.
The registers are writable via 32-bit MMIO accesses only.
Expose them with a NVMEM provider so they can be used by other drivers.
1) The flags variable is not initialized. Always use raw_spin_lock_irqsave
and raw_spin_unlock_irqrestore to serialize patching.
2) flush_kernel_vmap_range is primarily intended for DMA flushes.
The whole cache flush in flush_kernel_vmap_range is only possible
when interrupts are enabled on SMP machines. Since __patch_text_multiple
calls flush_kernel_vmap_range with interrupts disabled, it is better
to directly call flush_kernel_dcache_range_asm and
flush_kernel_icache_range_asm.
3) The final call to flush_icache_range is unnecessary.
Tested with `[PATCH, V3] parisc: Rewrite cache flush code for
PA8800/PA8900' change on rp3440, c8000 and c3750 (32 and 64-bit).
Note by Helge:
This patch had been temporarily reverted shortly before v5.18-rc6 in order
to fix boot issues. Now it can be re-applied.
Signed-off-by: John David Anglin <dave.anglin@bell.net> Signed-off-by: Helge Deller <deller@gmx.de>
parisc: Rewrite cache flush code for PA8800/PA8900
Originally, I was convinced that we needed to use tmpalias flushes
everwhere, for both user and kernel flushes. However, when I modified
flush_kernel_dcache_page_addr, to use a tmpalias flush, my c8000
would crash quite early when booting.
The PDC returns alias values of 0 for the icache and dcache. This
indicates that either the alias boundary is greater than 16MB or
equivalent aliasing doesn't work. I modified the tmpalias code to
make it easy to try alternate boundaries. I tried boundaries up to
128MB but still kernel tmpalias flushes didn't work on c8000.
This led me to conclude that tmpalias flushes don't work on PA8800
and PA8900 machines, and that we needed to flush directly using the
virtual address of user and kernel pages. This is likely the major
cause of instability on the c8000 and rp34xx machines.
Flushing user pages requires doing a temporary context switch as we
have to flush pages that don't belong to the current context. Further,
we have to deal with pages that aren't present. If a page isn't
present, the flush instructions fault on every line.
Other code has been rearranged and simplified based on testing. For
example, I introduced a flush_cache_dup_mm routine. flush_cache_mm
and flush_cache_dup_mm differ in that flush_cache_mm calls
purge_cache_pages and flush_cache_dup_mm calls flush_cache_pages.
In some implementations, pdc is more efficient than fdc. Based on
my testing, I don't believe there's any performance benefit on the
c8000.
Signed-off-by: John David Anglin <dave.anglin@bell.net> Signed-off-by: Helge Deller <deller@gmx.de>
parisc: Disable debug code regarding cache flushes in handle_nadtlb_fault()
Change the "BUG" to "WARNING" and disable the message because it triggers
occasionally in spite of the check in flush_cache_page_if_present.
The pte value extracted for the "from" page in copy_user_highpage is racy and
occasionally the pte is cleared before the flush is complete. I assume that
the page is simultaneously flushed by flush_cache_mm before the pte is cleared
as nullifying the fdc doesn't seem to cause problems.
I investigated various locking scenarios but I wasn't able to find a way to
sequence the flushes. This code is called for every COW break and locks impact
performance.
This patch is related to the bigger cache flush patch because we need the pte
on PA8800/PA8900 to flush using the vma context.
I have also seen this from copy_to_user_page and copy_from_user_page.
The messages appear infrequently when enabled.
Signed-off-by: John David Anglin <dave.anglin@bell.net> Signed-off-by: Helge Deller <deller@gmx.de>
clk: at91: generated: consider range when calculating best rate
clk_generated_best_diff() helps in finding the parent and the divisor to
compute a rate closest to the required one. However, it doesn't take into
account the request's range for the new rate. Make sure the new rate
is within the required range.
Fixes: 8a8f4bf0c480 ("clk: at91: clk-generated: create function to find best_diff") Signed-off-by: Codrin Ciubotariu <codrin.ciubotariu@microchip.com> Link: https://lore.kernel.org/r/20220413071318.244912-1-codrin.ciubotariu@microchip.com Reviewed-by: Claudiu Beznea <claudiu.beznea@microchip.com> Signed-off-by: Stephen Boyd <sboyd@kernel.org>
NFS: Pass i_size to fscache_unuse_cookie() when a file is released
Pass updated i_size in fscache_unuse_cookie() when called
from nfs_fscache_release_file(), which ensures the size of
an fscache object gets written to the cache storage. Failing
to do so results in unnessary reads from the NFS server, even
when the data is cached, due to a cachefiles object coherency
check failing with a trace similar to the following:
cachefiles_coherency: o=0000000e BAD osiz B=afbb3 c=0
This problem can be reproduced as follows:
#!/bin/bash
v=4.2; NFS_SERVER=127.0.0.1
set -e; trap cleanup EXIT; rc=1
function cleanup {
umount /mnt/nfs > /dev/null 2>&1
RC_STR="TEST PASS"
[ $rc -eq 1 ] && RC_STR="TEST FAIL"
echo "$RC_STR on $(uname -r) with NFSv$v and server $NFS_SERVER"
}
mount -o vers=$v,fsc $NFS_SERVER:/export /mnt/nfs
rm -f /mnt/nfs/file1.bin > /dev/null 2>&1
dd if=/dev/zero of=/mnt/nfs/file1.bin bs=4096 count=1 > /dev/null 2>&1
echo 3 > /proc/sys/vm/drop_caches
echo Read file 1st time from NFS server into fscache
dd if=/mnt/nfs/file1.bin of=/dev/null > /dev/null 2>&1
umount /mnt/nfs && mount -o vers=$v,fsc $NFS_SERVER:/export /mnt/nfs
echo 3 > /proc/sys/vm/drop_caches
echo Read file 2nd time from fscache
dd if=/mnt/nfs/file1.bin of=/dev/null > /dev/null 2>&1
echo Check mountstats for NFS read
grep -q "READ: 0" /proc/self/mountstats # (1st number) == 0
[ $? -eq 0 ] && rc=0
Fixes: a6b5a28eb56c "nfs: Convert to new fscache volume/cookie API" Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Tested-by: Daire Byrne <daire@dneg.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Schspa Shi [Mon, 16 May 2022 03:02:51 +0000 (11:02 +0800)]
cpufreq: make interface functions and lock holding state clear
cpufreq_offline() calls offline() and exit() under the policy rwsem
But they are called outside the rwsem in cpufreq_online().
Make cpufreq_online() call offline() and exit() as well as online() and
init() under the policy rwsem to achieve a clear lock relationship.
All of the init() and online() implementations in the tree only
initialize the policy object without attempting to acquire the policy
rwsem and they won't call cpufreq APIs attempting to acquire it.
Signed-off-by: Schspa Shi <schspa@gmail.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[ rjw: Changelog edits ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Schspa Shi [Mon, 16 May 2022 03:02:50 +0000 (11:02 +0800)]
cpufreq: Abort show()/store() for half-initialized policies
If policy initialization fails after the sysfs files are created,
there is a possibility to end up running show()/store() callbacks
for half-initialized policies, which may have unpredictable
outcomes.
Abort show()/store() in such a case by making sure the policy is active.
Also dectivate the policy on such failures.
Signed-off-by: Schspa Shi <schspa@gmail.com> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
[ rjw: Subject and changelog edits ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Chuck Lever [Mon, 18 Apr 2022 15:58:50 +0000 (11:58 -0400)]
Documentation: Add an explanation of NFSv4 client identifiers
To enable NFSv4 to work correctly, NFSv4 client identifiers have
to be globally unique and persistent over client reboots. We
believe that in many cases, a good default identifier can be
chosen and set when a client system is imaged.
Because there are many different ways a system can be imaged,
provide an explanation of how NFSv4 client identifiers and
principals can be set by install scripts and imaging tools.
Additional cases, such as NFSv4 clients running in containers, also
need unique and persistent identifiers. The Linux NFS community
sets forth this explanation to aid those who create and manage
container environments.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
NeilBrown [Mon, 21 Mar 2022 05:20:43 +0000 (16:20 +1100)]
NFS: update documentation for the nfs4_unique_id parameter
The documentation for nfs4_unique_id is out-of-date. In particular it
claim that when nfs4_unique_id is set, the host name is not used. since
Commit 55b592933b7d ("NFSv4: Fix nfs4_init_uniform_client_string for net
namespaces") both the unique_id AND the host name are used.
Update the documentation to match the code.
Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
is generated, possibly once for each lock (though rate limited).
This is potentially misleading as is can be read as suggesting that lock
reclaim was attempted. However the default behaviour is to not attempt
to recover locks (except due to server report).
This patch changes the reporting to produce at most one message for each
attempt to recover all state from a given server. The message reports
the server name and the number of locks lost if that number is non-zero.
It reports that locks were lost and give no suggestion as to whether
there was an attempt or not.
Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Jason Gunthorpe [Mon, 16 May 2022 23:41:21 +0000 (20:41 -0300)]
vfio: Simplify the life cycle of the group FD
Once userspace opens a group FD it is prevented from opening another
instance of that same group FD until all the prior group FDs and users of
the container are done.
The first is done trivially by checking the group->opened during group FD
open.
However, things get a little weird if userspace creates a device FD and
then closes the group FD. The group FD still cannot be re-opened, but this
time it is because the group->container is still set and container_users
is elevated by the device FD.
Due to this mismatched lifecycle we have the
vfio_group_try_dissolve_container() which tries to auto-free a container
after the group FD is closed but the device FD remains open.
Instead have the device FD hold onto a reference to the single group
FD. This directly prevents vfio_group_fops_release() from being called
when any device FD exists and makes the lifecycle model more
understandable.
vfio_group_try_dissolve_container() is removed as the only place a
container is auto-deleted is during vfio_group_fops_release(). At this
point the container_users is either 1 or 0 since all device FDs must be
closed.
Change group->opened to group->opened_file which points to the single
struct file * that is open for the group. If the group->open_file is
NULL then group->container == NULL.
If all device FDs have closed then the group's notifier list must be
empty.
Reviewed-by: Kevin Tian <kevin.tian@intel.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Tested-by: Nicolin Chen <nicolinc@nvidia.com> Tested-by: Matthew Rosato <mjrosato@linux.ibm.com> Link: https://lore.kernel.org/r/5-v2-d035a1842d81+1bf-vfio_group_locking_jgg@nvidia.com Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
down_write(&container->group_lock);
group->container = NULL;
up_write(&container->group_lock);
vfio_container_put(container);
/* woops we lost/leaked the new container */
This can then go on to NULL pointer deref since container == 0 and
container_users == 1.
Wrap all touches of container, except those on a performance path with a
known open device, with the group_rwsem.
The only user of vfio_group_add_container_user() holds the user count for
a simple operation, change it to just hold the group_lock over the
operation and delete vfio_group_add_container_user(). Containers now only
gain a user when a device FD is opened.
Reviewed-by: Kevin Tian <kevin.tian@intel.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Tested-by: Nicolin Chen <nicolinc@nvidia.com> Tested-by: Matthew Rosato <mjrosato@linux.ibm.com> Link: https://lore.kernel.org/r/4-v2-d035a1842d81+1bf-vfio_group_locking_jgg@nvidia.com Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
intel_vgpu_group_notifier()
vdev->kvm = data
[..]
kvm_get_kvm(vgpu->kvm);
// UAF!
Add a simple rwsem in the group to protect the kvm while the notifier is
using it.
Note this doesn't fix the race internal to i915 where userspace can
trigger two VFIO_GROUP_NOTIFY_SET_KVM's before we reach a consumer of
vgpu->kvm and trigger this same UAF, it just makes the notifier
self-consistent.
Fixes: ccd46dbae77d ("vfio: support notifier chain in vfio_group") Reviewed-by: Kevin Tian <kevin.tian@intel.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Tested-by: Nicolin Chen <nicolinc@nvidia.com> Tested-by: Matthew Rosato <mjrosato@linux.ibm.com> Link: https://lore.kernel.org/r/1-v2-d035a1842d81+1bf-vfio_group_locking_jgg@nvidia.com Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Wan Jiabing [Tue, 17 May 2022 02:34:41 +0000 (10:34 +0800)]
kvm/vfio: Fix potential deadlock problem in vfio
Fix following coccicheck warning:
./virt/kvm/vfio.c:258:1-7: preceding lock on line 236
If kvm_vfio_file_iommu_group() failed, code would goto err_fdput with
mutex_lock acquired and then return ret. It might cause potential
deadlock. Move mutex_unlock bellow err_fdput tag to fix it.
Fixes: d55d9e7a45721 ("kvm/vfio: Store the struct file in the kvm_vfio_group") Signed-off-by: Wan Jiabing <wanjiabing@vivo.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Link: https://lore.kernel.org/r/20220517023441.4258-1-wanjiabing@vivo.com Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Julian Orth [Tue, 17 May 2022 10:32:53 +0000 (12:32 +0200)]
audit,io_uring,io-wq: call __audit_uring_exit for dummy contexts
Not calling the function for dummy contexts will cause the context to
not be reset. During the next syscall, this will cause an error in
__audit_syscall_entry:
WARN_ON(context->context != AUDIT_CTX_UNUSED);
WARN_ON(context->name_count);
if (context->context != AUDIT_CTX_UNUSED || context->name_count) {
audit_panic("unrecoverable error in audit_syscall_entry()");
return;
}
These problematic dummy contexts are created via the following call
chain:
Cc: stable@vger.kernel.org Fixes: 5bd2182d58e9 ("audit,io_uring,io-wq: add some basic audit support to io_uring") Signed-off-by: Julian Orth <ju.orth@gmail.com>
[PM: subject line tweaks] Signed-off-by: Paul Moore <paul@paul-moore.com>
If the device is a detachable (and therefore lacks full keyboard), we may
still want to load this driver because the device might have some other
buttons or switches (e.g. volume and power buttons or a tablet mode
switch). In such case we do not want to register the "main" keyboard device
to allow userspace detect when the detachable keyboard is disconnected and
adjust the system behavior for the tablet mode.
Originally it was suggested to simply skip keyboard registration if row and
columns properties didn't exist, but that approach did not convey the
intent strongly enough and also had a slight problem for migrating existing
DTBs without updating the kernel first, so it was decided to introduce new
google,cros-ec-keyb-switches to explicitly mark devices that only have
axillary buttons and switches.
Stephen Boyd [Mon, 16 May 2022 18:55:31 +0000 (11:55 -0700)]
dt-bindings: google,cros-ec-keyb: Introduce switches only compatible
If the ChromeOS board is a detachable, this cros-ec-keyb device won't
have a matrix keyboard but it may have some button switches, e.g. volume
buttons and power buttons. The driver still registers a keyboard though
and that leads to userspace confusion around where the keyboard is.
We tried to work around this by only registering the keyboard device when
rows/columns properties were specified for the device, but that led to
another problem where removing the rows/columns properties breaks the
existing binding. Technically before that commit the rows/columns
properties were required, otherwise the driver would fail to probe.
Removing the properties from devicetrees makes the driver fail to probe
unless the corresponding driver patch is present. Furthermore, this makes
requiring matrix keyboard properties for devices that really have a
keyboard impossible because the compatible drives the schema and now the
properties are optional.
Add a more specific compatible for this type of device that indicates to
the OS that there are only switches and no matrix keyboard present.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Stephen Boyd <swboyd@chromium.org> Link: https://lore.kernel.org/r/20220516183452.942008-2-swboyd@chromium.org Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Johan Jonker [Tue, 29 Mar 2022 11:13:22 +0000 (13:13 +0200)]
dt-bindings: clock: convert rockchip,rk3188-cru.txt to YAML
Current dts files with RK3188/RK3066 'cru' nodes are manually verified.
In order to automate this process rockchip,rk3188-cru.txt has to be
converted to YAML.
Changed:
Add properties to fix notifications by clocks.yaml for example:
clocks
clock-names
Johan Jonker [Tue, 29 Mar 2022 11:36:56 +0000 (13:36 +0200)]
dt-bindings: clock: convert rockchip,rk3288-cru.txt to YAML
Current dts files with RK3288 'cru' nodes are manually verified.
In order to automate this process rockchip,rk3288-cru.txt has to be
converted to YAML.
Changed:
Add properties to fix notifications by clocks.yaml for example:
clocks
clock-names
Steve French [Thu, 12 May 2022 15:18:00 +0000 (10:18 -0500)]
SMB3: EBADF/EIO errors in rename/open caused by race condition in smb2_compound_op
There is a race condition in smb2_compound_op:
after_close:
num_rqst++;
if (cfile) {
cifsFileInfo_put(cfile); // sends SMB2_CLOSE to the server
cfile = NULL;
This is triggered by smb2_query_path_info operation that happens during
revalidate_dentry. In smb2_query_path_info, get_readable_path is called to
load the cfile, increasing the reference counter. If in the meantime, this
reference becomes the very last, this call to cifsFileInfo_put(cfile) will
trigger a SMB2_CLOSE request sent to the server just before sending this compound
request – and so then the compound request fails either with EBADF/EIO depending
on the timing at the server, because the handle is already closed.
In the first scenario, the race seems to be happening between smb2_query_path_info
triggered by the rename operation, and between “cleanup” of asynchronous writes – while
fsync(fd) likely waits for the asynchronous writes to complete, releasing the writeback
structures can happen after the close(fd) call. So the EBADF/EIO errors will pop up if
the timing is such that:
1) There are still outstanding references after close(fd) in the writeback structures
2) smb2_query_path_info successfully fetches the cfile, increasing the refcounter by 1
3) All writeback structures release the same cfile, reducing refcounter to 1
4) smb2_compound_op is called with that cfile
In the second scenario, the race seems to be similar – here open triggers the
smb2_query_path_info operation, and if all other threads in the meantime decrease the
refcounter to 1 similarly to the first scenario, again SMB2_CLOSE will be sent to the
server just before issuing the compound request. This case is harder to reproduce.
See https://bugzilla.samba.org/show_bug.cgi?id=15051
Cc: stable@vger.kernel.org Fixes: 8de9e86c67ba ("cifs: create a helper to find a writeable handle by path name") Signed-off-by: Ondrej Hubsch <ohubsch@purestorage.com> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
Jens Axboe [Tue, 17 May 2022 18:32:05 +0000 (12:32 -0600)]
io_uring: don't attempt to IOPOLL for MSG_RING requests
We gate whether to IOPOLL for a request on whether the opcode is allowed
on a ring setup for IOPOLL and if it's got a file assigned. MSG_RING
is the only one that allows a file yet isn't pollable, it's merely
supported to allow communication on an IOPOLL ring, not because we can
poll for completion of it.
Put the assigned file early and clear it, so we don't attempt to poll
for it.
drm/mgag200: Split up connector's mode_valid helper
Split up the connector's mode_valid helper into a simple-pipe and a
mode-config helper. The simple-pipe helper tests for display-size
limits while the mode-config helper tests for memory-bandwidth limits.
Also add the mgag200_ prefix to mga_vga_calculate_mode_bandwidth() and
comment on the function's purpose.
The memory-bandwidth tests assume that the display uses 4 bytes per
pixel. The first models of G200SE-A only had 1.75 MiB of VRAM, which
limits these devices to 640x480-32.
v2:
* note the memory constraints on early G200SE-A
drm/mgag200: Test memory requirements in drm_mode_config_funcs.mode_valid
Test for a mode's memory requirements in the device-wide mode_valid
helper. For simplicify, always assume a 32-bit color format. While
some rejected modes would work with less colors, implementing this
is probably not worth the effort.
Also remove the memory-related test from the connector's mode_valid
helper. The test uses the bpp value that users can specify on the
kernel's command line. This value is unrelated and the test would
belong into atomic_check.
struct mga_connector has outlived its purpose. Inline the rsp init
helper into the mode-config code and remove the data structure. No
functional changes.
Store the I2C state within struct mga_device and switch I2C to
managed release. Simplifies the related code and lets us remove
mga_connector_destroy().
drm/mgag200: Implement connector's get_modes with helper
Provide drm_connector_helper_get_modes_from_ddc() to implement the
connector's get_modes callback. The new helper updates the connector
from DDC-provided EDID data.
Initialization of the I2C adapter was allowed to fail. The mgag200
driver would have continued without DDC support. Had this happened in
practice, it would have led to segmentation faults in the connector
code. Resolve this problem by failing driver initialization on I2C-
related errors.
v2:
* initialize 'ret' before drm_err() (kernel test robot)
DDC operation conflicts with concurrent mode setting. Acquire the
driver's I/O lock in get_modes to prevent this. This change should
have been part of commit 931e3f3a0e99 ("drm/mgag200: Protect
concurrent access to I/O registers with lock"), but apparently got
lost somewhere.
v3:
* fix commit message to say 'drm/mgag200' (Jocelyn)
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Fixes: 931e3f3a0e99 ("drm/mgag200: Protect concurrent access to I/O registers with lock") Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com> Tested-by: Jocelyn Falempe <jfalempe@redhat.com> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Jocelyn Falempe <jfalempe@redhat.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Dave Airlie <airlied@redhat.com> Cc: dri-devel@lists.freedesktop.org Link: https://patchwork.freedesktop.org/patch/msgid/20220516134343.6085-2-tzimmermann@suse.de
Eric Biggers [Sat, 14 May 2022 17:59:29 +0000 (10:59 -0700)]
f2fs: don't use casefolded comparison for "." and ".."
Tryng to rename a directory that has all following properties fails with
EINVAL and triggers the 'WARN_ON_ONCE(!fscrypt_has_encryption_key(dir))'
in f2fs_match_ci_name():
- The directory is casefolded
- The directory is encrypted
- The directory's encryption key is not yet set up
- The parent directory is *not* encrypted
The problem is incorrect handling of the lookup of ".." to get the
parent reference to update. fscrypt_setup_filename() treats ".." (and
".") specially, as it's never encrypted. It's passed through as-is, and
setting up the directory's key is not attempted. As the name isn't a
no-key name, f2fs treats it as a "normal" name and attempts a casefolded
comparison. That breaks the assumption of the WARN_ON_ONCE() in
f2fs_match_ci_name() which assumes that for encrypted directories,
casefolded comparisons only happen when the directory's key is set up.
We could just remove this WARN_ON_ONCE(). However, since casefolding is
always a no-op on "." and ".." anyway, let's instead just not casefold
these names. This results in the standard bytewise comparison.
Fixes: 7ad08a58bf67 ("f2fs: Handle casefolding with Encryption") Cc: <stable@vger.kernel.org> # v5.11+ Signed-off-by: Eric Biggers <ebiggers@google.com> Reviewed-by: Gabriel Krisman Bertazi <krisman@collabora.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Jaegeuk Kim [Fri, 6 May 2022 20:34:41 +0000 (13:34 -0700)]
f2fs: do not stop GC when requiring a free section
The f2fs_gc uses a bitmap to indicate pinned sections, but when disabling
chckpoint, we call f2fs_gc() with NULL_SEGNO which selects the same dirty
segment as a victim all the time, resulting in checkpoint=disable failure,
for example. Let's pick another one, if we fail to collect it.
Reviewed-by: Chao Yu <chao@kernel.org> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The lock held by ext4_convert_inline_data is xattr_sem, but the lock
held by generic_perform_write is i_rwsem. Therefore, the two locks can
be concurrent.
To solve above issue, we add inode_lock() for ext4_convert_inline_data().
At the same time, move ext4_convert_inline_data() in front of
ext4_punch_hole(), remove similar handling from ext4_punch_hole().
Fixes: 0c8d414f163f ("ext4: let fallocate handle inline data correctly") Cc: stable@vger.kernel.org Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Baokun Li <libaokun1@huawei.com> Reviewed-by: Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20220428134031.4153381-1-libaokun1@huawei.com Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Zhang Yi [Sun, 24 Apr 2022 14:09:36 +0000 (22:09 +0800)]
ext4: convert symlink external data block mapping to bdev
Symlink's external data block is one kind of metadata block, and now
that almost all ext4 metadata block's page cache (e.g. directory blocks,
quota blocks...) belongs to bdev backing inode except the symlink. It
is essentially worked in data=journal mode like other regular file's
data block because probably in order to make it simple for generic VFS
code handling symlinks or some other historical reasons, but the logic
of creating external data block in ext4_symlink() is complicated. and it
also make things confused if user do not want to let the filesystem
worked in data=journal mode. This patch convert the final exceptional
case and make things clean, move the mapping of the symlink's external
data block to bdev like any other metadata block does.
Zhang Yi [Sun, 24 Apr 2022 14:09:35 +0000 (22:09 +0800)]
ext4: add nowait mode for ext4_getblk()
Current ext4_getblk() might sleep if some resources are not valid or
could be race with a concurrent extents modifing procedure. So we
cannot call ext4_getblk() and ext4_map_blocks() to get map blocks in
the atomic context in some fast path (e.g. the upcoming procedure of
getting symlink external block in the RCU context), even if the map
extents have already been check and cached.
In __ext4_super() we always overwrote the user specified journal_ioprio
value with a default value, expecting parse_apply_sb_mount_options() to
later correctly set ctx->journal_ioprio to the user specified value.
However, if parse_apply_sb_mount_options() returned early because of
empty sbi->es_s->s_mount_opts, the correct journal_ioprio value was
never set.
This patch fixes __ext4_super() to only use the default value if the
user has not specified any value for journal_ioprio.
Similarly, the remount behavior was to either use journal_ioprio
value specified during initial mount, or use the default value
irrespective of the journal_ioprio value specified during remount.
This patch modifies this to first check if a new value for ioprio
has been passed during remount and apply it. If no new value is
passed, use the value specified during initial mount.
ext4: mark group as trimmed only if it was fully scanned
Otherwise nonaligned fstrim calls will works inconveniently for iterative
scanners, for example:
// trim [0,16MB] for group-1, but mark full group as trimmed
fstrim -o $((1024*1024*128)) -l $((1024*1024*16)) ./m
// handle [16MB,16MB] for group-1, do nothing because group already has the flag.
fstrim -o $((1024*1024*144)) -l $((1024*1024*16)) ./m
[ Update function documentation for ext4_trim_all_free -- TYT ]
Reason is first directory entry which 'rec_len' is 34478, then will get illegal
parent entry. Now, we do not check directory entry after read directory block
in 'ext4_get_first_dir_block'.
To solve this issue, check directory entry in 'ext4_get_first_dir_block'.
[ Trigger an ext4_error() instead of just warning if the directory is
missing a '.' or '..' entry. Also make sure we return an error code
if the file system is corrupted. -TYT ]
btrfs: zoned: introduce a minimal zone size 4M and reject mount
Zoned devices are expected to have zone sizes in the range of 1-2GB for
ZNS SSDs and SMR HDDs have zone sizes of 256MB, so there is no need to
allow arbitrarily small zone sizes on btrfs.
But for testing purposes with emulated devices it is sometimes desirable
to create devices with as small as 4MB zone size to uncover errors.
So use 4MB as the smallest possible zone size and reject mounts of devices
with a smaller zone size.
Reviewed-by: Nikolay Borisov <nborisov@suse.com> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
Qu Wenruo [Mon, 9 May 2022 12:00:53 +0000 (20:00 +0800)]
btrfs: allow defrag to convert inline extents to regular extents
Btrfs defaults to max_inline=2K to make small writes inlined into
metadata.
The default value is always a win, as even DUP/RAID1/RAID10 doubles the
metadata usage, it should still cause less physical space used compared
to a 4K regular extents.
But since the introduction of RAID1C3 and RAID1C4 it's no longer the case,
users may find inlined extents causing too much space wasted, and want
to convert those inlined extents back to regular extents.
Unfortunately defrag will unconditionally skip all inline extents, no
matter if the user is trying to converting them back to regular extents.
So this patch will add a small exception for defrag_collect_targets() to
allow defragging inline extents, if and only if the inlined extents are
larger than max_inline, allowing users to convert them to regular ones.
This also allows us to defrag extents like the following:
item 6 key (257 EXTENT_DATA 0) itemoff 15794 itemsize 69
generation 7 type 0 (inline)
inline extent data size 48 ram_bytes 4096 compression 1 (zlib)
item 7 key (257 EXTENT_DATA 4096) itemoff 15741 itemsize 53
generation 7 type 1 (regular)
extent data disk byte 13631488 nr 4096
extent data offset 0 nr 16384 ram 16384
extent compression 1 (zlib)
Previously we're unable to do any defrag, since the first extent is
inlined, and the second one has no extent to merge.
Now we can defrag it to just one single extent, saving 48 bytes metadata
space.
item 6 key (257 EXTENT_DATA 0) itemoff 15810 itemsize 53
generation 8 type 1 (regular)
extent data disk byte 13635584 nr 4096
extent data offset 0 nr 20480 ram 20480
extent compression 1 (zlib)
Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Qu Wenruo <wqu@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
projects / people / ms / linux.git / log
