* + lib-add-fast-path-for-find_next__bit.patch added to -mm tree
@ 2021-04-01 0:54 akpm
0 siblings, 0 replies; only message in thread
From: akpm @ 2021-04-01 0:54 UTC (permalink / raw)
To: aklimov, andriy.shevchenko, arnd, dalias, dennis, dsterba, geert,
glaubitz, jianpeng.ma, joe, jpoimboe, linux, mm-commits,
richard.weiyang, sbrivio, wsa+renesas, ysato, yury.norov
The patch titled
Subject: lib: add fast path for find_next_*_bit()
has been added to the -mm tree. Its filename is
lib-add-fast-path-for-find_next__bit.patch
This patch should soon appear at
https://ozlabs.org/~akpm/mmots/broken-out/lib-add-fast-path-for-find_next__bit.patch
and later at
https://ozlabs.org/~akpm/mmotm/broken-out/lib-add-fast-path-for-find_next__bit.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Yury Norov <yury.norov@gmail.com>
Subject: lib: add fast path for find_next_*_bit()
Similarly to bitmap functions, find_next_*_bit() users will benefit if
we'll handle a case of bitmaps that fit into a single word inline. In the
very best case, the compiler may replace a function call with a few
instructions.
This is the quite typical find_next_bit() user:
unsigned int cpumask_next(int n, const struct cpumask *srcp)
{
/* -1 is a legal arg here. */
if (n != -1)
cpumask_check(n);
return find_next_bit(cpumask_bits(srcp), nr_cpumask_bits, n + 1);
}
EXPORT_SYMBOL(cpumask_next);
Currently, on ARM64 the generated code looks like this:
0000000000000000 <cpumask_next>:
0: a9bf7bfd stp x29, x30, [sp, #-16]!
4: 11000402 add w2, w0, #0x1
8: aa0103e0 mov x0, x1
c: d2800401 mov x1, #0x40 // #64
10: 910003fd mov x29, sp
14: 93407c42 sxtw x2, w2
18: 94000000 bl 0 <find_next_bit>
1c: a8c17bfd ldp x29, x30, [sp], #16
20: d65f03c0 ret
24: d503201f nop
After applying this patch:
0000000000000140 <cpumask_next>:
140: 11000400 add w0, w0, #0x1
144: 93407c00 sxtw x0, w0
148: f100fc1f cmp x0, #0x3f
14c: 54000168 b.hi 178 <cpumask_next+0x38> // b.pmore
150: f9400023 ldr x3, [x1]
154: 92800001 mov x1, #0xffffffffffffffff // #-1
158: 9ac02020 lsl x0, x1, x0
15c: 52800802 mov w2, #0x40 // #64
160: 8a030001 and x1, x0, x3
164: dac00020 rbit x0, x1
168: f100003f cmp x1, #0x0
16c: dac01000 clz x0, x0
170: 1a800040 csel w0, w2, w0, eq // eq = none
174: d65f03c0 ret
178: 52800800 mov w0, #0x40 // #64
17c: d65f03c0 ret
find_next_bit() call is replaced with 6 instructions. find_next_bit()
itself is 41 instructions plus function call overhead.
Despite inlining, the scripts/bloat-o-meter report smaller .text size
after applying the series:
add/remove: 11/9 grow/shrink: 233/176 up/down: 5780/-6768 (-988)
Link: https://lkml.kernel.org/r/20210401003153.97325-10-yury.norov@gmail.com
Signed-off-by: Yury Norov <yury.norov@gmail.com>
Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Alexey Klimov <aklimov@redhat.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: David Sterba <dsterba@suse.com>
Cc: Dennis Zhou <dennis@kernel.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jianpeng Ma <jianpeng.ma@intel.com>
Cc: Joe Perches <joe@perches.com>
Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Rich Felker <dalias@libc.org>
Cc: Stefano Brivio <sbrivio@redhat.com>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Wolfram Sang <wsa+renesas@sang-engineering.com>
Cc: Yoshinori Sato <ysato@users.osdn.me>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/asm-generic/bitops/find.h | 30 ++++++++++++++++++++++++++++
include/asm-generic/bitops/le.h | 21 +++++++++++++++++++
2 files changed, 51 insertions(+)
--- a/include/asm-generic/bitops/find.h~lib-add-fast-path-for-find_next__bit
+++ a/include/asm-generic/bitops/find.h
@@ -20,6 +20,16 @@ static inline
unsigned long find_next_bit(const unsigned long *addr, unsigned long size,
unsigned long offset)
{
+ if (small_const_nbits(size)) {
+ unsigned long val;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ val = *addr & GENMASK(size - 1, offset);
+ return val ? __ffs(val) : size;
+ }
+
return _find_next_bit(addr, NULL, size, offset, 0UL, 0);
}
#endif
@@ -40,6 +50,16 @@ unsigned long find_next_and_bit(const un
const unsigned long *addr2, unsigned long size,
unsigned long offset)
{
+ if (small_const_nbits(size)) {
+ unsigned long val;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ val = *addr1 & *addr2 & GENMASK(size - 1, offset);
+ return val ? __ffs(val) : size;
+ }
+
return _find_next_bit(addr1, addr2, size, offset, 0UL, 0);
}
#endif
@@ -58,6 +78,16 @@ static inline
unsigned long find_next_zero_bit(const unsigned long *addr, unsigned long size,
unsigned long offset)
{
+ if (small_const_nbits(size)) {
+ unsigned long val;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ val = *addr | ~GENMASK(size - 1, offset);
+ return val == ~0UL ? size : ffz(val);
+ }
+
return _find_next_bit(addr, NULL, size, offset, ~0UL, 0);
}
#endif
--- a/include/asm-generic/bitops/le.h~lib-add-fast-path-for-find_next__bit
+++ a/include/asm-generic/bitops/le.h
@@ -5,6 +5,7 @@
#include <asm-generic/bitops/find.h>
#include <asm/types.h>
#include <asm/byteorder.h>
+#include <linux/swab.h>
#if defined(__LITTLE_ENDIAN)
@@ -37,6 +38,16 @@ static inline
unsigned long find_next_zero_bit_le(const void *addr, unsigned
long size, unsigned long offset)
{
+ if (small_const_nbits(size)) {
+ unsigned long val = *(const unsigned long *)addr;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ val = swab(val) | ~GENMASK(size - 1, offset);
+ return val == ~0UL ? size : ffz(val);
+ }
+
return _find_next_bit(addr, NULL, size, offset, ~0UL, 1);
}
#endif
@@ -46,6 +57,16 @@ static inline
unsigned long find_next_bit_le(const void *addr, unsigned
long size, unsigned long offset)
{
+ if (small_const_nbits(size)) {
+ unsigned long val = *(const unsigned long *)addr;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ val = swab(val) & GENMASK(size - 1, offset);
+ return val ? __ffs(val) : size;
+ }
+
return _find_next_bit(addr, NULL, size, offset, 0UL, 1);
}
#endif
_
Patches currently in -mm which might be from yury.norov@gmail.com are
tools-disable-wno-type-limits.patch
tools-bitmap-sync-function-declarations-with-the-kernel.patch
tools-sync-bitmap_last_word_mask-macro-with-the-kernel.patch
arch-rearrange-headers-inclusion-order-in-asm-bitops-for-m68k-and-sh.patch
lib-extend-the-scope-of-small_const_nbits-macro.patch
tools-sync-small_const_nbits-macro-with-the-kernel.patch
lib-inline-_find_next_bit-wrappers.patch
tools-sync-find_next_bit-implementation.patch
lib-add-fast-path-for-find_next__bit.patch
lib-add-fast-path-for-find_first__bit-and-find_last_bit.patch
tools-sync-lib-find_bit-implementation.patch
maintainers-add-entry-for-the-bitmap-api.patch
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2021-04-01 0:54 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-04-01 0:54 + lib-add-fast-path-for-find_next__bit.patch added to -mm tree akpm
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).