ARM: 7983/1: atomics: implement a better __atomic_add_unless for v6+

Browse source

Looking at perf profiles of multi-threaded hackbench runs, a significant
performance hit appears to manifest from the cmpxchg loop used to
implement the 32-bit atomic_add_unless function. This can be mitigated
by writing a direct implementation of __atomic_add_unless which doesn't
require iteration outside of the atomic operation.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

...

This commit is contained in:

Will Deacon 2014-02-21 17:01:48 +01:00 • committed by Russell King

parent d98b90ea22

commit db38ee874c

1 changed files with 31 additions and 4 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

35

arch/arm/include/asm/atomic.h

View file