powerpc: Optimise smp_rmb
authorNick Piggin <npiggin@suse.de>
Tue, 11 Nov 2008 17:51:18 +0000 (17:51 +0000)
committerPaul Mackerras <paulus@samba.org>
Wed, 19 Nov 2008 05:04:57 +0000 (16:04 +1100)
commit957ab07b44d839ee8267e827fc4e8f1853798f57
treed282112a07dfc82dfc44522ad9e5a0baffdd6888
parent46d075be585eae2b74265e4e64ca38dde16a09c6
powerpc: Optimise smp_rmb

After commit 598056d5af8fef1dbe8f96f5c2b641a528184e5a ("[POWERPC] Fix
rmb to order cacheable vs. noncacheable"), rmb() becomes a sync
instruction, which is needed to order cacheable vs noncacheable loads.
However smp_rmb() is #defined to rmb(), and smp_rmb() can be an
lwsync.

This restores smp_rmb() performance by using lwsync there and updates
the comments.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Paul Mackerras <paulus@samba.org>
arch/powerpc/include/asm/system.h