/[linux-patches]/genpatches-2.6/trunk/2.6.14-pre/1310_amd64-x2-tsc.patch
Gentoo

Contents of /genpatches-2.6/trunk/2.6.14-pre/1310_amd64-x2-tsc.patch

Parent Directory Parent Directory | Revision Log Revision Log


Revision 175 - (show annotations) (download)
Sat Oct 8 10:49:55 2005 UTC (13 years, 2 months ago) by dsd
File size: 1464 byte(s)
Creating 2.6.14-pre branch from 2.6.13
1 From: john stultz <johnstul@us.ibm.com>
2 Date: Wed, 28 Sep 2005 04:45:36 +0000 (-0700)
3 Subject: [PATCH] x86-64: Fix bad assumption that dualcore cpus have synced TSCs
4 X-Git-Tag: v2.6.14-rc3
5 X-Git-Url: http://www.kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=6c132b5fe6579ed4b4892c02fe6c05f7e3afc579
6
7 [PATCH] x86-64: Fix bad assumption that dualcore cpus have synced TSCs
8
9 This should resolve the issue seen in bugme bug #5105, where it is assumed
10 that dualcore x86_64 systems have synced TSCs. This is not the case, and
11 alternate timesources should be used instead.
12
13 For more details, see:
14 http://bugzilla.kernel.org/show_bug.cgi?id=5105
15
16 Andi's earlier concerns that the TSCs should be synced on dualcore systems
17 have been resolved by confirmation from AMD folks that they can be
18 unsynced.
19
20 Acked-by: Andi Kleen <ak@suse.de>
21 Signed-off-by: Andrew Morton <akpm@osdl.org>
22 Signed-off-by: Linus Torvalds <torvalds@osdl.org>
23 ---
24
25 --- a/arch/x86_64/kernel/time.c
26 +++ b/arch/x86_64/kernel/time.c
27 @@ -959,9 +959,6 @@ static __init int unsynchronized_tsc(voi
28 are handled in the OEM check above. */
29 if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL)
30 return 0;
31 - /* All in a single socket - should be synchronized */
32 - if (cpus_weight(cpu_core_map[0]) == num_online_cpus())
33 - return 0;
34 #endif
35 /* Assume multi socket systems are not synchronized */
36 return num_online_cpus() > 1;

  ViewVC Help
Powered by ViewVC 1.1.20