linux/arch/x86/kernel/cpu
Don Zickus 90ed5b0fa5 perf/x86/p4: Block PMIs on init to prevent a stream of unkown NMIs
A bunch of unknown NMIs have popped up on a Pentium4 recently when booting
into a kdump kernel.  This was exposed because the watchdog timer went
from 60 seconds down to 10 seconds (increasing the ability to reproduce
this problem).

What is happening is on boot up of the second kernel (the kdump one),
the previous nmi_watchdogs were enabled on thread 0 and thread 1.  The
second kernel only initializes one cpu but the perf counter on thread 1
still counts.

Normally in a kdump scenario, the other cpus are blocking in an NMI loop,
but more importantly their local apics have the performance counters disabled
(iow LVTPC is masked).  So any counters that fire are masked and never get
through to the second kernel.

However, on a P4 the local apic is shared by both threads and thread1's PMI
(despite being configured to only interrupt thread1) will generate an NMI on
thread0.  Because thread0 knows nothing about this NMI, it is seen as an
unknown NMI.

This would be fine because it is a kdump kernel, strange things happen
what is the big deal about a single unknown NMI.

Unfortunately, the P4 comes with another quirk: clearing the overflow bit
to prevent a stream of NMIs.  This is the problem.

The kdump kernel can not execute because of the endless NMIs that happen.

To solve this, I instrumented the p4 perf init code, to walk all the counters
and zero them out (just like a normal reset would).

Now when the counters go off, they do not generate anything and no unknown
NMIs are seen.

I tested this on a P4 we have in our lab.  After two or three crashes, I could
normally reproduce the problem.  Now after 10 crashes, everything continues
to boot correctly.

Signed-off-by: Don Zickus <dzickus@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20140120154115.GZ25953@redhat.com
[ Fixed a stylistic detail. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-02-09 13:20:35 +01:00
..
mcheck Merge branch 'x86-ras-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-01-20 12:10:27 -08:00
microcode * Avoid WARN_ON() when mapping BGRT on Baytrail (EFI 32-bit). 2014-02-07 11:27:30 -08:00
mtrr mm, x86: Account for TLB flushes only when debugging 2014-01-25 09:10:41 +01:00
.gitignore
amd.c Merge branch 'linus' into x86/urgent 2014-01-25 09:16:14 +01:00
bugs_64.c
bugs.c x86: Get rid of ->hard_math and all the FPU asm fu 2013-06-06 14:32:04 -07:00
centaur.c x86: Delete non-required instances of include <linux/init.h> 2014-01-06 21:25:18 -08:00
common.c x86, cpu: Detect more TLB configuration 2014-01-03 14:35:42 -08:00
cpu.h x86/cpu: Track legacy CPU model data only on 32-bit kernels 2013-10-26 13:34:39 +02:00
cyrix.c x86: Delete non-required instances of include <linux/init.h> 2014-01-06 21:25:18 -08:00
hypervisor.c x86: Correctly detect hypervisor 2013-08-05 06:35:33 -07:00
intel_cacheinfo.c treewide: Fix common typo in "identify" 2013-10-14 15:31:06 +02:00
intel.c Merge branch 'linus' into x86/urgent 2014-01-25 09:16:14 +01:00
Makefile Merge branch 'x86-microcode-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-01-20 12:07:54 -08:00
match.c x86: Fix typo in MODULE_DEVICE_TABLE example: s/x86_cpu/x86cpu/ 2012-04-16 14:20:19 +02:00
mkcapflags.sh mkcapflags.pl: convert to mkcapflags.sh 2013-04-29 15:54:27 -07:00
mshyperv.c x86, hyperv: Move a variable to avoid an unused variable warning 2013-11-06 10:02:05 -08:00
perf_event_amd_ibs.c perf/x86/amd/ibs: Fix waking up from S3 for AMD family 10h 2014-01-16 09:19:50 +01:00
perf_event_amd_iommu.c perf/x86/amd: Do not print an error when the device is not present 2013-07-05 08:27:15 +02:00
perf_event_amd_iommu.h perf/x86/amd: AMD IOMMU Performance Counter PERF uncore PMU implementation 2013-06-19 13:04:53 +02:00
perf_event_amd_uncore.c x86: delete __cpuinit usage from all x86 files 2013-07-14 19:36:56 -04:00
perf_event_amd.c perf: Convert kmalloc_node(...GFP_ZERO...) to kzalloc_node() 2013-09-02 08:42:49 +02:00
perf_event_intel_ds.c perf: Fix arch_perf_out_copy_user default 2013-11-06 12:34:25 +01:00
perf_event_intel_lbr.c perf: Fix arch_perf_out_copy_user default 2013-11-06 12:34:25 +01:00
perf_event_intel_rapl.c perf/x86/intel: Add Intel RAPL PP1 energy counter support 2014-01-12 10:16:08 +01:00
perf_event_intel_uncore.c perf/x86/intel: Add Ivy Bridge-EP uncore IRP box support 2013-11-06 12:34:31 +01:00
perf_event_intel_uncore.h perf/x86/intel/uncore: Enable EV_SEL_EXT bit for PCU 2013-08-16 17:55:50 +02:00
perf_event_intel.c perf/x86: Suppress duplicated abort LBR records 2013-10-04 10:06:16 +02:00
perf_event_knc.c x86: Constify a few items 2013-03-11 15:11:03 +01:00
perf_event_p4.c perf/x86/p4: Block PMIs on init to prevent a stream of unkown NMIs 2014-02-09 13:20:35 +01:00
perf_event_p6.c perf/x86/intel/p6: Add userspace RDPMC quirk for PPro 2014-02-09 13:08:24 +01:00
perf_event.c perf/x86: Fix Userspace RDPMC switch 2014-02-09 13:08:25 +01:00
perf_event.h perf/x86/intel/p6: Add userspace RDPMC quirk for PPro 2014-02-09 13:08:24 +01:00
perfctr-watchdog.c perf/x86: Add support for Intel Xeon-Phi Knights Corner PMU 2012-10-04 13:32:37 +02:00
powerflags.c update AMD powerflags comments 2013-05-28 12:02:10 +02:00
proc.c x86/cpu: Always print SMP information in /proc/cpuinfo 2013-11-06 08:13:56 +01:00
rdrand.c x86, kaslr: Provide randomness functions 2013-10-13 03:12:12 -07:00
scattered.c treewide: Fix common typo in "identify" 2013-10-14 15:31:06 +02:00
topology.c x86: delete __cpuinit usage from all x86 files 2013-07-14 19:36:56 -04:00
transmeta.c x86: Delete non-required instances of include <linux/init.h> 2014-01-06 21:25:18 -08:00
umc.c x86: Delete non-required instances of include <linux/init.h> 2014-01-06 21:25:18 -08:00
vmware.c x86: Correctly detect hypervisor 2013-08-05 06:35:33 -07:00