This CPU family check is not required -- existence of the NodeId MSR is indicated by a CPUID feature flag which is already checked in amd_fixup_dcm() -- and it needlessly prevents amd_fixup_dcm() to be called for newer AMD CPUs.
In worst case this can lead to a panic in the scheduler code for AMD family 0x15 multi-node AMD CPUs. I just have a picture of VGA console output so I can't copy-and-paste it herein, but the call stack of such a panic looked like: do_divide_error ... find_busiest_group run_rebalance_domains ... apic_timer_interrupt ... cpu_idle The mainline kernel doesn't have this problem. Commit "(23588c3) x86, amd: Add support for CPUID topology extension of AMD CPUs" removed the family check. But 2.6.32.y needs to be fixed. Signed-off-by: Andreas Herrmann <[email protected]> --- arch/x86/kernel/cpu/amd.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index a932b9d..4e430b9 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -305,8 +305,7 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c) /* use socket ID also for last level cache */ per_cpu(cpu_llc_id, cpu) = c->phys_proc_id; /* fixup topology information on multi-node processors */ - if ((c->x86 == 0x10) && (c->x86_model == 9)) - amd_fixup_dcm(c); + amd_fixup_dcm(c); #endif } -- 1.7.3.2 _______________________________________________ stable mailing list [email protected] http://linux.kernel.org/mailman/listinfo/stable
