lib/find_bit.c: micro-optimise find_next_*_bit
authorMatthew Wilcox <mawilcox@microsoft.com>
Fri, 24 Feb 2017 23:00:58 +0000 (15:00 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sat, 25 Feb 2017 01:46:57 +0000 (17:46 -0800)
commite4afd2e5567fc5d59988025f7528f9b4794d86a5
treed00cf7c2409d3fdeb4dbb81ec04d6cb9c0e7965d
parent55ded9551f9a64f2872df77a954d4c30f8958e82
lib/find_bit.c: micro-optimise find_next_*_bit

This saves 32 bytes on my x86-64 build, mostly due to alignment
considerations and sharing more code between find_next_bit and
find_next_zero_bit, but it does save a couple of instructions.

There's really two parts to this commit:
 - First, the first half of the test: (!nbits || start >= nbits) is
   trivially a subset of the second half, since nbits and start are both
   unsigned
 - Second, while looking at the disassembly, I noticed that GCC was
   predicting the branch taken. Since this is a failure case, it's
   clearly the less likely of the two branches, so add an unlikely() to
   override GCC's heuristics.

[mawilcox@microsoft.com: v2]
Link: http://lkml.kernel.org/r/1483709016-1834-1-git-send-email-mawilcox@linuxonhyperv.com
Link: http://lkml.kernel.org/r/1483709016-1834-1-git-send-email-mawilcox@linuxonhyperv.com
Signed-off-by: Matthew Wilcox <mawilcox@microsoft.com>
Acked-by: Yury Norov <ynorov@caviumnetworks.com>
Acked-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
lib/find_bit.c
tools/lib/find_bit.c