diff options
| author | Andi Kleen <ak@suse.de> | 2006-02-03 21:51:02 +0100 |
|---|---|---|
| committer | Linus Torvalds <torvalds@g5.osdl.org> | 2006-02-04 16:43:13 -0800 |
| commit | 7bcd3f34e262bbebffa954d80eab3a84f053da31 (patch) | |
| tree | f0765da9eaa8024a2b1d67d3e43730cb32f99fa7 /arch/x86_64/kernel/setup.c | |
| parent | 6bca52b544489b626c7d0db801df6b4aa3d5adb5 (diff) | |
[PATCH] x86_64: Undo the earlier changes to remove unrolled copy/memset functions
They cause quite bad performance regressions on Netburst
This is temporary until we can get new optimized functions
for these CPUs.
This undoes changes that were done in 2.6.15 and in 2.6.16-rc1,
essentially bringing the code back to 2.6.14 level. Only change
is I renamed the X86_FEATURE_K8_C flag to X86_FEATURE_REP_GOOD
and fixed the check for the flag and also fixed some comments.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch/x86_64/kernel/setup.c')
| -rw-r--r-- | arch/x86_64/kernel/setup.c | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 28895c03cb1..506f152c238 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -877,6 +877,7 @@ static void __init amd_detect_cmp(struct cpuinfo_x86 *c) static int __init init_amd(struct cpuinfo_x86 *c) { int r; + unsigned level; #ifdef CONFIG_SMP unsigned long value; @@ -899,6 +900,11 @@ static int __init init_amd(struct cpuinfo_x86 *c) 3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */ clear_bit(0*32+31, &c->x86_capability); + /* On C+ stepping K8 rep microcode works well for copy/memset */ + level = cpuid_eax(1); + if (c->x86 == 15 && ((level >= 0x0f48 && level < 0x0f50) || level >= 0x0f58)) + set_bit(X86_FEATURE_REP_GOOD, &c->x86_capability); + r = get_model_name(c); if (!r) { switch (c->x86) { |