diff options
| author | Andy Lutomirski <luto@MIT.EDU> | 2011-05-23 09:31:26 -0400 |
|---|---|---|
| committer | Thomas Gleixner <tglx@linutronix.de> | 2011-05-24 14:51:28 +0200 |
| commit | 3729db5ca2b2000c660e5a5d0eb68b1053212cab (patch) | |
| tree | a8b038145eadd067f92798adb4292532eccedfba /arch/x86/kernel/tsc.c | |
| parent | 057e6a8c660e95c3f4e7162e00e2fee1fc90c50d (diff) | |
x86-64: Don't generate cmov in vread_tsc
vread_tsc checks whether rdtsc returns something less than
cycle_last, which is an extremely predictable branch. GCC likes
to generate a cmov anyway, which is several cycles slower than
a predicted branch. This saves a couple of nanoseconds.
Signed-off-by: Andy Lutomirski <luto@mit.edu>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Borislav Petkov <bp@amd64.org>
Link: http://lkml.kernel.org/r/%3C561280649519de41352fcb620684dfb22bad6bac.1306156808.git.luto%40mit.edu%3E
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Diffstat (limited to 'arch/x86/kernel/tsc.c')
| -rw-r--r-- | arch/x86/kernel/tsc.c | 18 |
1 files changed, 16 insertions, 2 deletions
diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 1e624420261..24249a5360b 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -767,6 +767,7 @@ static cycle_t read_tsc(struct clocksource *cs) static cycle_t __vsyscall_fn vread_tsc(void) { cycle_t ret; + u64 last; /* * Empirically, a fence (of type that depends on the CPU) @@ -778,8 +779,21 @@ static cycle_t __vsyscall_fn vread_tsc(void) rdtsc_barrier(); ret = (cycle_t)vget_cycles(); - return ret >= VVAR(vsyscall_gtod_data).clock.cycle_last ? - ret : VVAR(vsyscall_gtod_data).clock.cycle_last; + last = VVAR(vsyscall_gtod_data).clock.cycle_last; + + if (likely(ret >= last)) + return ret; + + /* + * GCC likes to generate cmov here, but this branch is extremely + * predictable (it's just a funciton of time and the likely is + * very likely) and there's a data dependence, so force GCC + * to generate a branch instead. I don't barrier() because + * we don't actually need a barrier, and if this function + * ever gets inlined it will generate worse code. + */ + asm volatile (""); + return last; } #endif |