android_kernel_motorola_sm6375/lib
Repository files (latest commit first)
Filename Latest commit message Latest commit date
Mateusz Guzik a58c144b9e lockref: stop doing cpu_relax in the cmpxchg loop
[ Upstream commit f5fe24ef17b5fbe6db49534163e77499fb10ae8c ]

On the x86-64 architecture even a failing cmpxchg grants exclusive
access to the cacheline, making it preferable to retry the failed op
immediately instead of stalling with the pause instruction.

To illustrate the impact, below are benchmark results obtained by
running various will-it-scale tests on top of the 6.2-rc3 kernel and
Cascade Lake (2 sockets * 24 cores * 2 threads) CPU.

All results in ops/s.  Note there is some variance in re-runs, but the
code is consistently faster when contention is present.

  open3 ("Same file open/close"):
  proc          stock       no-pause
     1         805603         814942       (+%1)
     2        1054980        1054781       (-0%)
     8        1544802        1822858      (+18%)
    24        1191064        2199665      (+84%)
    48         851582        1469860      (+72%)
    96         609481        1427170     (+134%)

  fstat2 ("Same file fstat"):
  proc          stock       no-pause
     1        3013872        3047636       (+1%)
     2        4284687        4400421       (+2%)
     8        3257721        5530156      (+69%)
    24        2239819        5466127     (+144%)
    48        1701072        5256609     (+209%)
    96        1269157        6649326     (+423%)

Additionally, a kernel with a private patch to help access() scalability:
access2 ("Same file access"):

  proc          stock        patched      patched
                                         +nopause
    24        2378041        2005501      5370335  (-15% / +125%)

That is, fixing the problems in access itself *reduces* scalability
after the cacheline ping-pong only happens in lockref with the pause
instruction.

Note that fstat and access benchmarks are not currently integrated into
will-it-scale, but interested parties can find them in pull requests to
said project.

Code at hand has a rather tortured history.  First modification showed
up in commit d472d9d98b ("lockref: Relax in cmpxchg loop"), written
with Itanium in mind.  Later it got patched up to use an arch-dependent
macro to stop doing it on s390 where it caused a significant regression.
Said macro had undergone revisions and was ultimately eliminated later,
going back to cpu_relax.

While I intended to only remove cpu_relax for x86-64, I got the
following comment from Linus:

    I would actually prefer just removing it entirely and see if
    somebody else hollers. You have the numbers to prove it hurts on
    real hardware, and I don't think we have any numbers to the
    contrary.

    So I think it's better to trust the numbers and remove it as a
    failure, than say "let's just remove it on x86-64 and leave
    everybody else with the potentially broken code"

Additionally, Will Deacon (maintainer of the arm64 port, one of the
architectures previously benchmarked):

    So, from the arm64 side of the fence, I'm perfectly happy just
    removing the cpu_relax() calls from lockref.

As such, come back full circle in history and whack it altogether.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/all/CAGudoHHx0Nqg6DE70zAVA75eV-HXfWyhVMWZ-aSeOofkA_=WdA@mail.gmail.com/
Acked-by: Tony Luck <tony.luck@intel.com> # ia64
Acked-by: Nicholas Piggin <npiggin@gmail.com> # powerpc
Acked-by: Will Deacon <will@kernel.org> # arm64
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-02-06 07:52:42 +01:00
..
842
crypto lib/crypto: blake2s: move hmac construction into wireguard 2022-06-22 14:11:02 +02:00
dim dim: initialize all struct fields 2022-05-18 09:47:25 +02:00
fonts lib/fonts: fix undefined behavior in bit shift for get_default_font 2023-01-18 11:40:54 +01:00
livepatch
lz4 lz4: fix LZ4_decompress_safe_partial read out of bound 2022-04-15 14:18:39 +02:00
lzo
math
mpi lib/mpi: use kcalloc in mpi_resize 2021-09-15 09:47:29 +02:00
raid6 lib/raid6/test/Makefile: Use $(pound) instead of \# for Make 4.3 2022-04-15 14:18:23 +02:00
reed_solomon
vdso lib/vdso: use "grep -E" instead of "egrep" 2022-12-08 11:22:59 +01:00
xz lib/xz: Validate the value before assigning it to an enum variable 2021-11-17 09:48:31 +01:00
zlib_deflate
zlib_inflate lib/zlib: remove outdated and incorrect pre-increment optimization 2020-06-24 17:50:39 +02:00
zstd
.gitignore
argv_split.c
ashldi3.c
ashrdi3.c
asn1_decoder.c
assoc_array.c assoc_array: Fix BUG_ON during garbage collect 2022-06-06 08:33:50 +02:00
atomic64.c
atomic64_test.c
audit.c
bcd.c
bch.c
bitmap.c
bitrev.c
bsearch.c
btree.c
bucket_locks.c
bug.c bug: Remove redundant condition check in report_bug 2021-05-14 09:44:27 +02:00
build_OID_registry
bust_spinlocks.c
chacha.c
check_signature.c
checksum.c
clz_ctz.c
clz_tab.c
cmdline.c
cmpdi2.c
compat_audit.c
cpu_rmap.c
cpumask.c
crc-ccitt.c
crc-itu-t.c
crc-t10dif.c crc-t10dif: Fix potential crypto notify dead-lock 2020-08-19 08:15:54 +02:00
crc4.c
crc7.c
crc8.c
crc16.c
crc32.c lib/crc32.c: fix trivial typo in preprocessor condition 2020-10-29 09:57:52 +01:00
crc32defs.h
crc32test.c lib/crc32test: remove extra local_irq_disable/enable 2020-11-10 12:37:27 +01:00
crc64.c
ctype.c
debug_info.c
debug_locks.c
debugobjects.c
dec_and_lock.c
decompress.c
decompress_bunzip2.c
decompress_inflate.c
decompress_unlz4.c lib/decompress_unlz4.c: correctly handle zero-padding around initrds. 2021-07-20 16:10:46 +02:00
decompress_unlzma.c
decompress_unlzo.c
decompress_unxz.c lib/xz: Avoid overlapping memcpy() with invalid input with in-place decompression 2021-11-17 09:48:31 +01:00
devres.c devres: keep both device name and resource name in pretty name 2020-08-21 13:05:32 +02:00
digsig.c
dump_stack.c
dynamic_debug.c dyndbg: let query-modname override actual module name 2022-10-26 13:22:37 +02:00
dynamic_queue_limits.c
earlycpio.c
error-inject.c
errseq.c
extable.c
fault-inject.c
fdt.c
fdt_empty_tree.c
fdt_ro.c
fdt_rw.c
fdt_strerror.c
fdt_sw.c
fdt_wip.c
find_bit.c
find_bit_benchmark.c
flex_proportions.c
gen_crc32table.c
gen_crc64table.c
genalloc.c lib/genalloc: fix the overflow when size is too big 2021-01-12 20:16:10 +01:00
generic-radix-tree.c
glob.c
globtest.c
hexdump.c hex2bin: fix access beyond string end 2022-05-09 09:03:22 +02:00
hweight.c
idr.c ida: don't use BUG_ON() for debugging 2022-07-12 16:30:49 +02:00
inflate.c
interval_tree.c
interval_tree_test.c
iomap.c
iomap_copy.c
iommu-helper.c
ioremap.c
iov_iter.c mm/highmem: Lift memcpy_[to|from]_page to core 2023-01-18 11:41:55 +01:00
irq_poll.c
irq_regs.c
is_single_threaded.c
kasprintf.c
Kconfig ARM: 9178/1: fix unmet dependency on BITREVERSE for HAVE_ARCH_BITREVERSE 2022-03-19 13:40:16 +01:00
Kconfig.debug Kconfig.debug: provide a little extra FRAME_WARN leeway when KASAN is enabled 2022-12-08 11:23:05 +01:00
Kconfig.kasan
Kconfig.kgdb
Kconfig.ubsan
kfifo.c
klist.c
kobject.c kobject: Restore old behaviour of kobject_del(NULL) 2020-09-17 13:47:52 +02:00
kobject_uevent.c kobject_uevent: remove warning in init_uevent_argv() 2021-05-19 10:08:33 +02:00
kstrtox.c lib: vsprintf: Fix handling of number field widths in vsscanf 2021-07-14 16:53:16 +02:00
kstrtox.h lib: vsprintf: Fix handling of number field widths in vsscanf 2021-07-14 16:53:16 +02:00
libcrc32c.c
list_debug.c lib/list_debug.c: Detect uninitialized lists 2022-08-25 11:18:36 +02:00
list_sort.c
llist.c
locking-selftest-hardirq.h
locking-selftest-mutex.h
locking-selftest-rlock-hardirq.h
locking-selftest-rlock-softirq.h
locking-selftest-rlock.h
locking-selftest-rsem.h
locking-selftest-rtmutex.h
locking-selftest-softirq.h
locking-selftest-spin-hardirq.h
locking-selftest-spin-softirq.h
locking-selftest-spin.h
locking-selftest-wlock-hardirq.h
locking-selftest-wlock-softirq.h
locking-selftest-wlock.h
locking-selftest-wsem.h
locking-selftest.c
lockref.c lockref: stop doing cpu_relax in the cmpxchg loop 2023-02-06 07:52:42 +01:00
logic_pio.c PCI: Fix pci_register_io_range() memory leak 2021-03-17 17:03:44 +01:00
lru_cache.c
lshrdi3.c
Makefile avoid __memcat_p link failure 2021-05-07 10:51:37 +02:00
memcat_p.c
memory-notifier-error-inject.c
memweight.c
muldi3.c
net_utils.c
netdev-notifier-error-inject.c
nlattr.c net: fix nla_strcmp to handle more then one trailing null character 2021-05-19 10:08:27 +02:00
nmi_backtrace.c
nodemask.c nodemask: Fix return values to be unsigned 2022-06-14 18:12:02 +02:00
notifier-error-inject.c lib/notifier-error-inject: fix error when writing -errno to debugfs file 2023-01-18 11:40:55 +01:00
notifier-error-inject.h
objagg.c
of-reconfig-notifier-error-inject.c
oid_registry.c
once.c once: add DO_ONCE_SLOW() for sleepable contexts 2022-10-26 13:22:27 +02:00
packing.c
parman.c
parser.c
pci_iomap.c
percpu-refcount.c
percpu_counter.c
percpu_test.c
plist.c
pm-notifier-error-inject.c
radix-tree.c
random32.c random: replace custom notifier chain with standard one 2022-06-22 14:11:13 +02:00
ratelimit.c ratelimit: Fix data-races in ___ratelimit(). 2022-09-05 10:27:42 +02:00
rbtree.c
rbtree_test.c
refcount.c locking/refcount: Consolidate implementations of refcount_t 2022-07-29 17:14:17 +02:00
rhashtable.c
sbitmap.c
scatterlist.c sgl_alloc_order: fix memory leak 2020-11-05 11:43:22 +01:00
seq_buf.c seq_buf: Fix overflow in seq_buf_putmem_hex() 2021-07-19 08:53:16 +02:00
sg_pool.c
sg_split.c
sha1.c lib/crypto: sha1: re-roll loops to reduce code size 2022-06-22 14:11:03 +02:00
show_mem.c
siphash.c siphash: use one source of truth for siphash permutations 2022-06-22 14:11:16 +02:00
smp_processor_id.c
sort.c
stackdepot.c lib: stackdepot: turn depot_lock spinlock to raw_spinlock 2021-05-22 11:38:30 +02:00
stmp_device.c
string.c lib/string.c: implement stpcpy 2020-10-01 13:18:23 +02:00
string_helpers.c
strncpy_from_user.c
strnlen_user.c
syscall.c lib/syscall: fix syscall registers retrieval on 32-bit platforms 2020-12-11 13:23:32 +01:00
test-kstrtox.c
test-string_helpers.c
test_bitfield.c
test_bitmap.c
test_blackhole_dev.c
test_bpf.c bpf/tests: Do not PASS tests without actually testing the result 2021-09-22 12:26:29 +02:00
test_debug_virtual.c
test_firmware.c test_firmware: fix memory leak in test_firmware_init() 2023-01-18 11:41:23 +01:00
test_hash.c
test_hexdump.c
test_ida.c
test_kasan.c kasan: stop tests being eliminated as dead code with FORTIFY_SOURCE 2020-06-22 09:31:12 +02:00
test_kmod.c lib/test: use after free in register_test_dev_kmod() 2022-04-15 14:18:21 +02:00
test_list_sort.c
test_memcat_p.c
test_meminit.c lib/test_meminit: destroy cache in kmem_cache_alloc_bulk() test 2022-01-27 09:19:55 +01:00
test_module.c
test_objagg.c test_objagg: Fix potential memory leak in error handling 2020-06-30 15:37:04 -04:00
test_overflow.c
test_parman.c
test_printf.c
test_rhashtable.c
test_siphash.c
test_sort.c
test_stackinit.c lib/test_stackinit: Fix static initializer test 2021-09-22 12:26:38 +02:00
test_static_key_base.c
test_static_keys.c
test_string.c
test_strscpy.c
test_sysctl.c
test_ubsan.c
test_user_copy.c
test_uuid.c
test_vmalloc.c
test_xarray.c XArray: Fix xas_create_range() when multi-order entry present 2022-04-15 14:18:28 +02:00
textsearch.c
timerqueue.c
ts_bm.c
ts_fsm.c
ts_kmp.c
ubsan.c
ubsan.h
ucmpdi2.c
ucs2_string.c
usercopy.c
uuid.c
vsprintf.c random: replace custom notifier chain with standard one 2022-06-22 14:11:13 +02:00
win_minmax.c
xarray.c XArray: Update the LRU list in xas_split() 2022-04-15 14:18:28 +02:00
xxhash.c