All of lore.kernel.org
 help / color / mirror / Atom feed
* [perf] 3.0 overflow test locks up system
@ 2011-07-22 16:13 Vince Weaver
  2011-07-27 18:51 ` [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels Vince Weaver
  0 siblings, 1 reply; 9+ messages in thread
From: Vince Weaver @ 2011-07-22 16:13 UTC (permalink / raw)
  To: linux-kernel
  Cc: Peter Zijlstra, Paul Mackerras, Ingo Molnar, Arnaldo Carvalho de Melo

Hello

With 3.0.0 the PAPI "overflow_allcounters" test reliably locks up my 
Nehalem system.  This definitely does not happen in 2.6.39 and possibly 
doesn't in 3.0-rc4 either.

I won't have time to bisect or to come up with a simpler test case until 
next week sometime.

A warning is printed just before the machine becomes unresponsive, it is 
included below.

Thanks,

Vince
vweaver1@eecs.utk.edu

[  392.504845] ------------[ cut here ]------------
[  392.504962] WARNING: at kernel/smp.c:320 smp_call_function_single+0x6c/0xf2()
[  392.505074] Hardware name: Precision M4500
[  392.505181] Modules linked in: acpi_cpufreq cpufreq_conservative mperf cpufreq_powersave cpufreq_userspace cpufreq_stats uinput nouveau snd_hda_codec_hdmi ttm drm_kms_helper mxm_wmi snd_hda_codec_idt iwlagn mac80211 snd_hda_intel snd_hda_codec cfg80211 dell_laptop snd_hwdep video processor ehci_hcd dell_wmi sparse_keymap psmouse sdhci_pci rfkill snd_pcm sdhci thermal_sys pcspkr ac battery wmi serio_raw snd_timer snd_page_alloc evdev i2c_i801 dcdbas button
[  392.509709] Pid: 2310, comm: overflow_allcou Not tainted 3.0.0 #43
[  392.509819] Call Trace:
[  392.509925]  <IRQ>  [<ffffffff81041bb0>] ? warn_slowpath_common+0x78/0x8c
[  392.510144]  [<ffffffff810a4d19>] ? perf_exclude_event.part.23+0x31/0x31
[  392.510257]  [<ffffffff8106b7c5>] ? smp_call_function_single+0x6c/0xf2
[  392.510369]  [<ffffffff810a38aa>] ? task_function_call+0x42/0x4c
[  392.510476]  [<ffffffff810a50e4>] ? update_cgrp_time_from_event+0x2c/0x2c
[  392.510589]  [<ffffffff810a5a9d>] ? perf_event_disable+0x45/0x8c
[  392.510700]  [<ffffffff810a8e89>] ? __perf_event_overflow+0xf1/0x1a3
[  392.510812]  [<ffffffff8103c7bb>] ? select_task_rq_fair+0x349/0x574
[  392.510924]  [<ffffffff810a7f0a>] ? perf_ctx_adjust_freq+0x42/0xe6
[  392.511038]  [<ffffffff8105f152>] ? sched_clock_cpu+0xb/0xc3
[  392.511152]  [<ffffffff8100ded5>] ? paravirt_read_tsc+0x5/0x8
[  392.511262]  [<ffffffff8100e320>] ? native_sched_clock+0x27/0x2f
[  392.511366]  [<ffffffff810a9500>] ? perf_event_overflow+0x10/0x10
[  392.511476]  [<ffffffff810a959f>] ? perf_swevent_hrtimer+0x9f/0xda
[  392.511599]  [<ffffffff8105ca04>] ? run_posix_cpu_timers+0x23/0x346
[  392.511721]  [<ffffffff8131c2ef>] ? rb_insert_color+0xb1/0xd9
[  392.511841]  [<ffffffff8105d373>] ? __run_hrtimer+0xac/0x135
[  392.511960]  [<ffffffff8105daa3>] ? hrtimer_interrupt+0xdb/0x195
[  392.512083]  [<ffffffff8108d040>] ? check_for_new_grace_period.isra.32+0x99/0xa4
[  392.512220]  [<ffffffff8108d201>] ? __rcu_process_callbacks+0x72/0x2b7
[  392.512345]  [<ffffffff8102437c>] ? hpet_interrupt_handler+0x23/0x2b
[  392.512469]  [<ffffffff81089446>] ? handle_irq_event_percpu+0x50/0x180
[  392.512592]  [<ffffffff81046f76>] ? __do_softirq+0x13e/0x177
[  392.512713]  [<ffffffff810f680c>] ? send_sigio+0x95/0xab
[  392.512832]  [<ffffffff810895aa>] ? handle_irq_event+0x34/0x52
[  392.512952]  [<ffffffff8108b433>] ? handle_edge_irq+0x9f/0xc6
[  392.513072]  [<ffffffff8100a831>] ? handle_irq+0x1d/0x21
[  392.513192]  [<ffffffff8100a561>] ? do_IRQ+0x42/0x98
[  392.513314]  [<ffffffff81593053>] ? common_interrupt+0x13/0x13
[  392.513438]  <EOI> 
[  392.513542] ---[ end trace 12f3f913316a2866 ]---

^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-22 16:13 [perf] 3.0 overflow test locks up system Vince Weaver
@ 2011-07-27 18:51 ` Vince Weaver
  2011-07-28  1:54   ` Lin Ming
  2011-07-28 18:47   ` Peter Zijlstra
  0 siblings, 2 replies; 9+ messages in thread
From: Vince Weaver @ 2011-07-27 18:51 UTC (permalink / raw)
  To: linux-kernel
  Cc: Peter Zijlstra, Paul Mackerras, Ingo Molnar, Arnaldo Carvalho de Melo

[-- Attachment #1: Type: TEXT/PLAIN, Size: 3482 bytes --]

Hello

> With 3.0.0 the PAPI "overflow_allcounters" test reliably locks up my 
> Nehalem system.

I finally managed to narrow this down to a small test, which is attached.

Basically measuring overflow on the perf::perf_count_sw_cpu_clock
event will potentially *lock up* your system from user-space.

This seems to be a long standing bug.  It will quickly lock solid
my Nehalem test box on 3.0, 2.6.39 and 2.6.38.

On a Core2 2.6.32 box the crash testing program will wedge and become 
unkillable, but it doesn't actually kill the machine.

As mentioned before, on the Nehalem machine the following warning happens
before the machine becomes unusable:

[  392.504845] ------------[ cut here ]------------
[  392.504962] WARNING: at kernel/smp.c:320 smp_call_function_single+0x6c/0xf2()
[  392.505074] Hardware name: Precision M4500
[  392.505181] Modules linked in: acpi_cpufreq cpufreq_conservative mperf cpufreq_powersave cpufreq_userspace cpufreq_stats uinput nouveau snd_hda_codec_hdmi ttm drm_kms_helper mxm_wmi snd_hda_codec_idt iwlagn mac80211 snd_hda_intel snd_hda_codec cfg80211 dell_laptop snd_hwdep video processor ehci_hcd dell_wmi sparse_keymap psmouse sdhci_pci rfkill snd_pcm sdhci thermal_sys pcspkr ac battery wmi serio_raw snd_timer snd_page_alloc evdev i2c_i801 dcdbas button
[  392.509709] Pid: 2310, comm: overflow_allcou Not tainted 3.0.0 #43
[  392.509819] Call Trace:
[  392.509925]  <IRQ>  [<ffffffff81041bb0>] ? warn_slowpath_common+0x78/0x8c
[  392.510144]  [<ffffffff810a4d19>] ? perf_exclude_event.part.23+0x31/0x31
[  392.510257]  [<ffffffff8106b7c5>] ? smp_call_function_single+0x6c/0xf2
[  392.510369]  [<ffffffff810a38aa>] ? task_function_call+0x42/0x4c
[  392.510476]  [<ffffffff810a50e4>] ? update_cgrp_time_from_event+0x2c/0x2c
[  392.510589]  [<ffffffff810a5a9d>] ? perf_event_disable+0x45/0x8c
[  392.510700]  [<ffffffff810a8e89>] ? __perf_event_overflow+0xf1/0x1a3
[  392.510812]  [<ffffffff8103c7bb>] ? select_task_rq_fair+0x349/0x574
[  392.510924]  [<ffffffff810a7f0a>] ? perf_ctx_adjust_freq+0x42/0xe6
[  392.511038]  [<ffffffff8105f152>] ? sched_clock_cpu+0xb/0xc3
[  392.511152]  [<ffffffff8100ded5>] ? paravirt_read_tsc+0x5/0x8
[  392.511262]  [<ffffffff8100e320>] ? native_sched_clock+0x27/0x2f
[  392.511366]  [<ffffffff810a9500>] ? perf_event_overflow+0x10/0x10
[  392.511476]  [<ffffffff810a959f>] ? perf_swevent_hrtimer+0x9f/0xda
[  392.511599]  [<ffffffff8105ca04>] ? run_posix_cpu_timers+0x23/0x346
[  392.511721]  [<ffffffff8131c2ef>] ? rb_insert_color+0xb1/0xd9
[  392.511841]  [<ffffffff8105d373>] ? __run_hrtimer+0xac/0x135
[  392.511960]  [<ffffffff8105daa3>] ? hrtimer_interrupt+0xdb/0x195
[  392.512083]  [<ffffffff8108d040>] ? check_for_new_grace_period.isra.32+0x99/0xa4
[  392.512220]  [<ffffffff8108d201>] ? __rcu_process_callbacks+0x72/0x2b7
[  392.512345]  [<ffffffff8102437c>] ? hpet_interrupt_handler+0x23/0x2b
[  392.512469]  [<ffffffff81089446>] ? handle_irq_event_percpu+0x50/0x180
[  392.512592]  [<ffffffff81046f76>] ? __do_softirq+0x13e/0x177
[  392.512713]  [<ffffffff810f680c>] ? send_sigio+0x95/0xab
[  392.512832]  [<ffffffff810895aa>] ? handle_irq_event+0x34/0x52
[  392.512952]  [<ffffffff8108b433>] ? handle_edge_irq+0x9f/0xc6
[  392.513072]  [<ffffffff8100a831>] ? handle_irq+0x1d/0x21
[  392.513192]  [<ffffffff8100a561>] ? do_IRQ+0x42/0x98
[  392.513314]  [<ffffffff81593053>] ? common_interrupt+0x13/0x13
[  392.513438]  <EOI> 
[  392.513542] ---[ end trace 12f3f913316a2866 ]---

Thanks,

Vince

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Type: TEXT/x-csrc; name=oflo_sw_cpu_clock_crash.c, Size: 3065 bytes --]

/* Error with overflows and perf::perf_count_sw_cpu_clock                    */
/* This test will crash Linux 3.0.0                                          */
/* compile with gcc -O2 -o oflo_sw_cpu_clock_crash oflo_sw_cpu_clock_crash.c */

/* by Vince Weaver <vweaver1 _at_ eecs.utk.edu>                              */

#define _GNU_SOURCE 1

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include <fcntl.h>


#include <linux/perf_event.h>
#include <sys/syscall.h>
#include <unistd.h>
#include <asm/unistd.h>
#include <sys/ioctl.h>

#include <sys/mman.h>

#include <signal.h>


#include <sys/prctl.h>

#define MATRIX_SIZE 512
static double a[MATRIX_SIZE][MATRIX_SIZE];
static double b[MATRIX_SIZE][MATRIX_SIZE];
static double c[MATRIX_SIZE][MATRIX_SIZE];

static void naive_matrix_multiply(int quiet) {

  double s;
  int i,j,k;

  for(i=0;i<MATRIX_SIZE;i++) {
    for(j=0;j<MATRIX_SIZE;j++) {
      a[i][j]=(double)i*(double)j;
      b[i][j]=(double)i/(double)(j+5);
    }
  }

  for(j=0;j<MATRIX_SIZE;j++) {
     for(i=0;i<MATRIX_SIZE;i++) {
        s=0;
        for(k=0;k<MATRIX_SIZE;k++) {
	   s+=a[i][k]*b[k][j];
	}
        c[i][j] = s;
     }
  }

  s=0.0;
  for(i=0;i<MATRIX_SIZE;i++) {
    for(j=0;j<MATRIX_SIZE;j++) {
      s+=c[i][j];
    }
  }

  if (!quiet) printf("Matrix multiply sum: s=%lf\n",s);

  return;
}


static int total=0;

void our_handler(int signum,siginfo_t *oh, void *blah) {

  int fd=oh->si_fd;

  ioctl(fd , PERF_EVENT_IOC_DISABLE,0);
  total++;
  ioctl(fd , PERF_EVENT_IOC_REFRESH,1);
}

int perf_event_open(struct perf_event_attr *hw_event_uptr,
		    pid_t pid, int cpu, int group_fd, unsigned long flags) {

  return syscall(__NR_perf_event_open,hw_event_uptr,pid,cpu,group_fd,flags);

}

int main( int argc, char **argv ) {

	int fd;
	void *blargh;

	struct perf_event_attr pe;

	struct sigaction sa;

	memset(&sa, 0, sizeof(struct sigaction));
	sa.sa_sigaction=our_handler;
	sa.sa_flags=SA_SIGINFO;

	if (sigaction(SIGIO,&sa,NULL)<0) {
	  fprintf(stderr,"Error setting up signal handler\n");
	  exit(1);
	}
	      
        memset(&pe,0,sizeof(struct perf_event_attr));	
	pe.type=PERF_TYPE_SOFTWARE;
	pe.size=sizeof(struct perf_event_attr);
        pe.config=PERF_COUNT_SW_CPU_CLOCK;
	pe.sample_period=100000;
	pe.sample_type=PERF_SAMPLE_IP;
	pe.read_format=PERF_FORMAT_GROUP|PERF_FORMAT_ID;
	pe.disabled=1;
	pe.pinned=1;
	pe.exclude_kernel=1;
	pe.exclude_hv=1;
	pe.wakeup_events=1;

	fd=perf_event_open(&pe,0,-1,-1,0);
	if (fd<0) {
	   printf("Error opening\n");
	}

	blargh=mmap(NULL,(1+2)*4096,PROT_READ|PROT_WRITE,MAP_SHARED,fd,0);
	fcntl(fd,F_SETFL,O_RDWR|O_NONBLOCK|O_ASYNC);
	fcntl(fd,F_SETSIG,SIGIO);
	fcntl(fd,F_SETOWN,getpid());

	ioctl(fd,PERF_EVENT_IOC_RESET,0);
	ioctl(fd,PERF_EVENT_IOC_ENABLE,0);

	naive_matrix_multiply(0);

	ioctl(fd,PERF_EVENT_IOC_DISABLE,0);
	munmap(blargh,(1+2)*4096);
	close(fd);

	printf("Total overflows: %d\n",total);

	return 0;
}

^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-27 18:51 ` [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels Vince Weaver
@ 2011-07-28  1:54   ` Lin Ming
  2011-07-28  2:39     ` Lin Ming
  2011-07-28 18:47   ` Peter Zijlstra
  1 sibling, 1 reply; 9+ messages in thread
From: Lin Ming @ 2011-07-28  1:54 UTC (permalink / raw)
  To: Vince Weaver
  Cc: linux-kernel, Peter Zijlstra, Paul Mackerras, Ingo Molnar,
	Arnaldo Carvalho de Melo, ming.m.lin

On Thu, Jul 28, 2011 at 2:51 AM, Vince Weaver <vweaver1@eecs.utk.edu> wrote:
> Hello
>
>> With 3.0.0 the PAPI "overflow_allcounters" test reliably locks up my
>> Nehalem system.
>
> I finally managed to narrow this down to a small test, which is attached.
>
> Basically measuring overflow on the perf::perf_count_sw_cpu_clock
> event will potentially *lock up* your system from user-space.
>
> This seems to be a long standing bug.  It will quickly lock solid
> my Nehalem test box on 3.0, 2.6.39 and 2.6.38.
>
> On a Core2 2.6.32 box the crash testing program will wedge and become
> unkillable, but it doesn't actually kill the machine.

Tried on my Nehalem machine, the testing program also becomes unkillable.

[147868.734111] WARNING: at
/home/zhyan/sources/linux-2.6/kernel/smp.c:320
smp_call_function_single+0x68/0x104()
[147868.734113] Hardware name: Studio XPS 8000
[147868.734115] Modules linked in: ebtable_nat ebtables ipt_MASQUERADE
iptable_nat nf_nat bridge stp llc rmd160 crypto_null camellia lzo
cast6 cast5 deflate zlib_deflate cts ctr gcm ccm serpent blowfish
twofish_generic twofish_i586 twofish_common ecb xcbc cbc
sha256_generic sha512_generic des_generic aes_i586 geode_aes
aes_generic ah6 ah4 esp6 esp4 xfrm4_mode_beet xfrm4_tunnel tunnel4
xfrm4_mode_tunnel xfrm4_mode_transport xfrm6_mode_transport
xfrm6_mode_ro xfrm6_mode_beet xfrm6_mode_tunnel ipcomp ipcomp6
xfrm_ipcomp xfrm6_tunnel tunnel6 af_key sunrpc ip6t_REJECT
nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables ipv6
kvm_intel kvm uinput snd_hda_codec_hdmi snd_hda_codec_realtek
snd_hda_intel snd_hda_codec snd_hwdep snd_seq snd_seq_device radeon
snd_pcm snd_timer ttm drm_kms_helper drm snd broadcom tg3
firewire_ohci firewire_core crc_itu_t i2c_i801 i2c_algo_bit soundcore
serio_raw i2c_core snd_page_alloc usb_storage iTCO_wdt
iTCO_vendor_support joydev dcdbas microcod
e pcspkr [last unloaded: mperf]
[147868.734175] Pid: 12152, comm: oflo_sw_cpu_clo Not tainted 3.0.0 #38
[147868.734176] Call Trace:
[147868.734180]  [<c043b29f>] warn_slowpath_common+0x6a/0x7f
[147868.734183]  [<c0464440>] ? smp_call_function_single+0x68/0x104
[147868.734186]  [<c04a5bc8>] ? perf_event_tid+0x21/0x21
[147868.734189]  [<c043b2c8>] warn_slowpath_null+0x14/0x18
[147868.734191]  [<c0464440>] smp_call_function_single+0x68/0x104
[147868.734193]  [<c04a5c3c>] task_function_call+0x37/0x40
[147868.734196]  [<c04a68e7>] ? __perf_event_exit_context+0x7e/0x7e
[147868.734199]  [<c04a5cf0>] perf_event_disable+0x38/0x75
[147868.734201]  [<c04a9e13>] __perf_event_overflow+0x15a/0x209
[147868.734205]  [<c0432f5e>] ? get_parent_ip+0xb/0x31
[147868.734209]  [<c0457fc8>] ? local_clock+0x22/0x2b
[147868.734211]  [<c04aa461>] perf_event_overflow+0x11/0x13
[147868.734214]  [<c04aa4f8>] perf_swevent_hrtimer+0x95/0xed
[147868.734218]  [<c05d1291>] ? timerqueue_del+0x49/0x56
[147868.734221]  [<c0455a8c>] ? __remove_hrtimer+0x58/0x75
[147868.734223]  [<c0455cea>] __run_hrtimer+0xb0/0x12f
[147868.734226]  [<c04aa463>] ? perf_event_overflow+0x13/0x13
[147868.734228]  [<c0456691>] hrtimer_interrupt+0xe9/0x1ce
[147868.734233]  [<c04198cb>] smp_apic_timer_interrupt+0x5d/0x70
[147868.734236]  [<c07cf1ad>] apic_timer_interrupt+0x31/0x38
[147868.734238] ---[ end trace a164f652f8bfd400 ]---
[147928.569073] INFO: rcu_preempt_state detected stalls on CPUs/tasks:
{ 6} (detected by 7, t=60002 jiffies)
[147930.784928] INFO: rcu_bh_state detected stalls on CPUs/tasks: { 6}
(detected by 4, t=60002 jiffies)

Lin Ming

^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-28  1:54   ` Lin Ming
@ 2011-07-28  2:39     ` Lin Ming
  0 siblings, 0 replies; 9+ messages in thread
From: Lin Ming @ 2011-07-28  2:39 UTC (permalink / raw)
  To: Vince Weaver
  Cc: linux-kernel, Peter Zijlstra, Paul Mackerras, Ingo Molnar,
	Arnaldo Carvalho de Melo

On Thu, 2011-07-28 at 09:54 +0800, Lin Ming wrote:
> On Thu, Jul 28, 2011 at 2:51 AM, Vince Weaver <vweaver1@eecs.utk.edu> wrote:
> > Hello
> >
> >> With 3.0.0 the PAPI "overflow_allcounters" test reliably locks up my
> >> Nehalem system.
> >
> > I finally managed to narrow this down to a small test, which is attached.
> >
> > Basically measuring overflow on the perf::perf_count_sw_cpu_clock
> > event will potentially *lock up* your system from user-space.
> >
> > This seems to be a long standing bug.  It will quickly lock solid
> > my Nehalem test box on 3.0, 2.6.39 and 2.6.38.
> >
> > On a Core2 2.6.32 box the crash testing program will wedge and become
> > unkillable, but it doesn't actually kill the machine.

Hi, Vince

I re-tested current -tip tree(commit 0931941) and it works OK now.

$ ./oflo_sw_cpu_clock_crash 
Matrix multiply sum: s=27665734022509.746094
Total overflows: 3460

Could you also have a try the current -tip tree?

Lin Ming



^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-27 18:51 ` [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels Vince Weaver
  2011-07-28  1:54   ` Lin Ming
@ 2011-07-28 18:47   ` Peter Zijlstra
  2011-07-29  5:05     ` [stable] " Greg KH
  1 sibling, 1 reply; 9+ messages in thread
From: Peter Zijlstra @ 2011-07-28 18:47 UTC (permalink / raw)
  To: Vince Weaver
  Cc: linux-kernel, Paul Mackerras, Ingo Molnar,
	Arnaldo Carvalho de Melo, stable

On Wed, 2011-07-27 at 14:51 -0400, Vince Weaver wrote:
> [  392.504845] ------------[ cut here ]------------
> [  392.504962] WARNING: at kernel/smp.c:320 smp_call_function_single+0x6c/0xf2()
> [  392.505074] Hardware name: Precision M4500
> [  392.505181] Modules linked in: acpi_cpufreq cpufreq_conservative mperf cpufreq_powersave cpufreq_userspace cpufreq_stats uinput nouveau snd_hda_codec_hdmi ttm drm_kms_helper mxm_wmi snd_hda_codec_idt iwlagn mac80211 snd_hda_intel snd_hda_codec cfg80211 dell_laptop snd_hwdep video processor ehci_hcd dell_wmi sparse_keymap psmouse sdhci_pci rfkill snd_pcm sdhci thermal_sys pcspkr ac battery wmi serio_raw snd_timer snd_page_alloc evdev i2c_i801 dcdbas button
> [  392.509709] Pid: 2310, comm: overflow_allcou Not tainted 3.0.0 #43
> [  392.509819] Call Trace:
> [  392.509925]  <IRQ>  [<ffffffff81041bb0>] ? warn_slowpath_common+0x78/0x8c
> [  392.510144]  [<ffffffff810a4d19>] ? perf_exclude_event.part.23+0x31/0x31
> [  392.510257]  [<ffffffff8106b7c5>] ? smp_call_function_single+0x6c/0xf2
> [  392.510369]  [<ffffffff810a38aa>] ? task_function_call+0x42/0x4c
> [  392.510476]  [<ffffffff810a50e4>] ? update_cgrp_time_from_event+0x2c/0x2c
> [  392.510589]  [<ffffffff810a5a9d>] ? perf_event_disable+0x45/0x8c
> [  392.510700]  [<ffffffff810a8e89>] ? __perf_event_overflow+0xf1/0x1a3
> [  392.510812]  [<ffffffff8103c7bb>] ? select_task_rq_fair+0x349/0x574
> [  392.510924]  [<ffffffff810a7f0a>] ? perf_ctx_adjust_freq+0x42/0xe6
> [  392.511038]  [<ffffffff8105f152>] ? sched_clock_cpu+0xb/0xc3
> [  392.511152]  [<ffffffff8100ded5>] ? paravirt_read_tsc+0x5/0x8
> [  392.511262]  [<ffffffff8100e320>] ? native_sched_clock+0x27/0x2f
> [  392.511366]  [<ffffffff810a9500>] ? perf_event_overflow+0x10/0x10
> [  392.511476]  [<ffffffff810a959f>] ? perf_swevent_hrtimer+0x9f/0xda
> [  392.511599]  [<ffffffff8105ca04>] ? run_posix_cpu_timers+0x23/0x346
> [  392.511721]  [<ffffffff8131c2ef>] ? rb_insert_color+0xb1/0xd9
> [  392.511841]  [<ffffffff8105d373>] ? __run_hrtimer+0xac/0x135
> [  392.511960]  [<ffffffff8105daa3>] ? hrtimer_interrupt+0xdb/0x195
> [  392.512083]  [<ffffffff8108d040>] ? check_for_new_grace_period.isra.32+0x99/0xa4
> [  392.512220]  [<ffffffff8108d201>] ? __rcu_process_callbacks+0x72/0x2b7
> [  392.512345]  [<ffffffff8102437c>] ? hpet_interrupt_handler+0x23/0x2b
> [  392.512469]  [<ffffffff81089446>] ? handle_irq_event_percpu+0x50/0x180
> [  392.512592]  [<ffffffff81046f76>] ? __do_softirq+0x13e/0x177
> [  392.512713]  [<ffffffff810f680c>] ? send_sigio+0x95/0xab
> [  392.512832]  [<ffffffff810895aa>] ? handle_irq_event+0x34/0x52
> [  392.512952]  [<ffffffff8108b433>] ? handle_edge_irq+0x9f/0xc6
> [  392.513072]  [<ffffffff8100a831>] ? handle_irq+0x1d/0x21
> [  392.513192]  [<ffffffff8100a561>] ? do_IRQ+0x42/0x98
> [  392.513314]  [<ffffffff81593053>] ? common_interrupt+0x13/0x13
> [  392.513438]  <EOI> 
> [  392.513542] ---[ end trace 12f3f913316a2866 ]--- 

The below patch is for -stable only, upstream has a much larger patch
that contains the below hunk.

---
Subject: perf: Fix software event overflow

Vince found that under certain circumstances software event overflows
go wrong and deadlock. Avoid trying to delete a timer from the timer
callback.

Cc: stable@kernel.org
Reported-by: Vince Weaver <vweaver1@eecs.utk.edu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
 kernel/events/core.c |    7 ++-----
 1 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 9efe710..32a6151 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -5016,11 +5016,8 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
 	if (events && atomic_dec_and_test(&event->event_limit)) {
 		ret = 1;
 		event->pending_kill = POLL_HUP;
-		if (nmi) {
-			event->pending_disable = 1;
-			irq_work_queue(&event->pending);
-		} else
-			perf_event_disable(event);
+		event->pending_disable = 1;
+		irq_work_queue(&event->pending);
 	}
 
 	if (event->overflow_handler)



^ permalink raw reply related	[flat|nested] 9+ messages in thread
* Re: [stable] [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-28 18:47   ` Peter Zijlstra
@ 2011-07-29  5:05     ` Greg KH
  2011-07-29 11:59       ` Peter Zijlstra
  0 siblings, 1 reply; 9+ messages in thread
From: Greg KH @ 2011-07-29  5:05 UTC (permalink / raw)
  To: Peter Zijlstra
  Cc: Vince Weaver, Ingo Molnar, Paul Mackerras, linux-kernel,
	Arnaldo Carvalho de Melo, stable

On Thu, Jul 28, 2011 at 08:47:10PM +0200, Peter Zijlstra wrote:
> On Wed, 2011-07-27 at 14:51 -0400, Vince Weaver wrote:
> > [  392.504845] ------------[ cut here ]------------
> > [  392.504962] WARNING: at kernel/smp.c:320 smp_call_function_single+0x6c/0xf2()
> > [  392.505074] Hardware name: Precision M4500
> > [  392.505181] Modules linked in: acpi_cpufreq cpufreq_conservative mperf cpufreq_powersave cpufreq_userspace cpufreq_stats uinput nouveau snd_hda_codec_hdmi ttm drm_kms_helper mxm_wmi snd_hda_codec_idt iwlagn mac80211 snd_hda_intel snd_hda_codec cfg80211 dell_laptop snd_hwdep video processor ehci_hcd dell_wmi sparse_keymap psmouse sdhci_pci rfkill snd_pcm sdhci thermal_sys pcspkr ac battery wmi serio_raw snd_timer snd_page_alloc evdev i2c_i801 dcdbas button
> > [  392.509709] Pid: 2310, comm: overflow_allcou Not tainted 3.0.0 #43
> > [  392.509819] Call Trace:
> > [  392.509925]  <IRQ>  [<ffffffff81041bb0>] ? warn_slowpath_common+0x78/0x8c
> > [  392.510144]  [<ffffffff810a4d19>] ? perf_exclude_event.part.23+0x31/0x31
> > [  392.510257]  [<ffffffff8106b7c5>] ? smp_call_function_single+0x6c/0xf2
> > [  392.510369]  [<ffffffff810a38aa>] ? task_function_call+0x42/0x4c
> > [  392.510476]  [<ffffffff810a50e4>] ? update_cgrp_time_from_event+0x2c/0x2c
> > [  392.510589]  [<ffffffff810a5a9d>] ? perf_event_disable+0x45/0x8c
> > [  392.510700]  [<ffffffff810a8e89>] ? __perf_event_overflow+0xf1/0x1a3
> > [  392.510812]  [<ffffffff8103c7bb>] ? select_task_rq_fair+0x349/0x574
> > [  392.510924]  [<ffffffff810a7f0a>] ? perf_ctx_adjust_freq+0x42/0xe6
> > [  392.511038]  [<ffffffff8105f152>] ? sched_clock_cpu+0xb/0xc3
> > [  392.511152]  [<ffffffff8100ded5>] ? paravirt_read_tsc+0x5/0x8
> > [  392.511262]  [<ffffffff8100e320>] ? native_sched_clock+0x27/0x2f
> > [  392.511366]  [<ffffffff810a9500>] ? perf_event_overflow+0x10/0x10
> > [  392.511476]  [<ffffffff810a959f>] ? perf_swevent_hrtimer+0x9f/0xda
> > [  392.511599]  [<ffffffff8105ca04>] ? run_posix_cpu_timers+0x23/0x346
> > [  392.511721]  [<ffffffff8131c2ef>] ? rb_insert_color+0xb1/0xd9
> > [  392.511841]  [<ffffffff8105d373>] ? __run_hrtimer+0xac/0x135
> > [  392.511960]  [<ffffffff8105daa3>] ? hrtimer_interrupt+0xdb/0x195
> > [  392.512083]  [<ffffffff8108d040>] ? check_for_new_grace_period.isra.32+0x99/0xa4
> > [  392.512220]  [<ffffffff8108d201>] ? __rcu_process_callbacks+0x72/0x2b7
> > [  392.512345]  [<ffffffff8102437c>] ? hpet_interrupt_handler+0x23/0x2b
> > [  392.512469]  [<ffffffff81089446>] ? handle_irq_event_percpu+0x50/0x180
> > [  392.512592]  [<ffffffff81046f76>] ? __do_softirq+0x13e/0x177
> > [  392.512713]  [<ffffffff810f680c>] ? send_sigio+0x95/0xab
> > [  392.512832]  [<ffffffff810895aa>] ? handle_irq_event+0x34/0x52
> > [  392.512952]  [<ffffffff8108b433>] ? handle_edge_irq+0x9f/0xc6
> > [  392.513072]  [<ffffffff8100a831>] ? handle_irq+0x1d/0x21
> > [  392.513192]  [<ffffffff8100a561>] ? do_IRQ+0x42/0x98
> > [  392.513314]  [<ffffffff81593053>] ? common_interrupt+0x13/0x13
> > [  392.513438]  <EOI> 
> > [  392.513542] ---[ end trace 12f3f913316a2866 ]--- 
> 
> The below patch is for -stable only, upstream has a much larger patch
> that contains the below hunk.

What is that commit id so it can be properly referenced in the patch
below?  And what tree should this patch go into?  .39?  3.0?  Anything
newer?

thanks,

greg k-h

> 
> ---
> Subject: perf: Fix software event overflow
> 
> Vince found that under certain circumstances software event overflows
> go wrong and deadlock. Avoid trying to delete a timer from the timer
> callback.
> 
> Cc: stable@kernel.org
> Reported-by: Vince Weaver <vweaver1@eecs.utk.edu>
> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
> ---
>  kernel/events/core.c |    7 ++-----
>  1 files changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 9efe710..32a6151 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -5016,11 +5016,8 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
>  	if (events && atomic_dec_and_test(&event->event_limit)) {
>  		ret = 1;
>  		event->pending_kill = POLL_HUP;
> -		if (nmi) {
> -			event->pending_disable = 1;
> -			irq_work_queue(&event->pending);
> -		} else
> -			perf_event_disable(event);
> +		event->pending_disable = 1;
> +		irq_work_queue(&event->pending);
>  	}
>  
>  	if (event->overflow_handler)
> 
> 
> _______________________________________________
> stable mailing list
> stable@linux.kernel.org
> http://linux.kernel.org/mailman/listinfo/stable

^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [stable] [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-29  5:05     ` [stable] " Greg KH
@ 2011-07-29 11:59       ` Peter Zijlstra
  2011-08-02 16:35         ` Greg KH
  0 siblings, 1 reply; 9+ messages in thread
From: Peter Zijlstra @ 2011-07-29 11:59 UTC (permalink / raw)
  To: Greg KH
  Cc: Vince Weaver, Ingo Molnar, Paul Mackerras, linux-kernel,
	Arnaldo Carvalho de Melo, stable

On Thu, 2011-07-28 at 22:05 -0700, Greg KH wrote:
> 
> What is that commit id so it can be properly referenced in the patch
> below?  And what tree should this patch go into?  .39?  3.0?  Anything
> newer? 

.32+ in as far as anybody is still maintaining them, if it fails to
apply to any of them (quite possible) do tell and I can provide
backports etc..

Upstream commit (which doesn't mention this issue because it wasn't
known at the time and thus thinks its a cleanup only):

---
>From a8b0ca17b80e92faab46ee7179ba9e99ccb61233 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
Date: Mon, 27 Jun 2011 14:41:57 +0200
Subject: [PATCH] perf: Remove the nmi parameter from the swevent and overflow
 interface

The nmi parameter indicated if we could do wakeups from the current
context, if not, we would set some state and self-IPI and let the
resulting interrupt do the wakeup.

For the various event classes:

  - hardware: nmi=0; PMI is in fact an NMI or we run irq_work_run from
    the PMI-tail (ARM etc.)
  - tracepoint: nmi=0; since tracepoint could be from NMI context.
  - software: nmi=[0,1]; some, like the schedule thing cannot
    perform wakeups, and hence need 0.

As one can see, there is very little nmi=1 usage, and the down-side of
not using it is that on some platforms some software events can have a
jiffy delay in wakeup (when arch_irq_work_raise isn't implemented).

The up-side however is that we can remove the nmi parameter and save a
bunch of conditionals in fast paths.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Michael Cree <mcree@orcon.net.nz>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
Cc: Anton Blanchard <anton@samba.org>
Cc: Eric B Munson <emunson@mgebm.net>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jason Wessel <jason.wessel@windriver.com>
Cc: Don Zickus <dzickus@redhat.com>
Link: http://lkml.kernel.org/n/tip-agjev8eu666tvknpb3iaj0fg@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/alpha/kernel/perf_event.c            |    2 
 arch/arm/kernel/perf_event_v6.c           |    2 
 arch/arm/kernel/perf_event_v7.c           |    2 
 arch/arm/kernel/perf_event_xscale.c       |    4 -
 arch/arm/kernel/ptrace.c                  |    2 
 arch/arm/kernel/swp_emulate.c             |    2 
 arch/arm/mm/fault.c                       |    6 +-
 arch/mips/kernel/perf_event.c             |    2 
 arch/mips/kernel/traps.c                  |    8 +--
 arch/mips/kernel/unaligned.c              |    5 --
 arch/mips/math-emu/cp1emu.c               |    3 -
 arch/mips/mm/fault.c                      |    8 +--
 arch/powerpc/include/asm/emulated_ops.h   |    4 -
 arch/powerpc/kernel/perf_event.c          |    6 +-
 arch/powerpc/kernel/perf_event_fsl_emb.c  |    6 +-
 arch/powerpc/kernel/ptrace.c              |    2 
 arch/powerpc/mm/fault.c                   |    6 +-
 arch/s390/mm/fault.c                      |    6 +-
 arch/sh/kernel/ptrace_32.c                |    2 
 arch/sh/kernel/traps_32.c                 |    2 
 arch/sh/kernel/traps_64.c                 |    8 +--
 arch/sh/math-emu/math.c                   |    2 
 arch/sh/mm/fault_32.c                     |    6 +-
 arch/sh/mm/tlbflush_64.c                  |    6 +-
 arch/sparc/kernel/perf_event.c            |    2 
 arch/sparc/kernel/unaligned_32.c          |    4 -
 arch/sparc/kernel/unaligned_64.c          |   12 ++---
 arch/sparc/kernel/visemul.c               |    2 
 arch/sparc/math-emu/math_32.c             |    2 
 arch/sparc/math-emu/math_64.c             |    2 
 arch/sparc/mm/fault_32.c                  |    8 +--
 arch/sparc/mm/fault_64.c                  |    8 +--
 arch/x86/kernel/cpu/perf_event.c          |    2 
 arch/x86/kernel/cpu/perf_event_intel.c    |    2 
 arch/x86/kernel/cpu/perf_event_intel_ds.c |    4 -
 arch/x86/kernel/cpu/perf_event_p4.c       |    2 
 arch/x86/kernel/kgdb.c                    |    2 
 arch/x86/kernel/ptrace.c                  |    2 
 arch/x86/mm/fault.c                       |    6 +-
 include/linux/perf_event.h                |   18 +++-----
 kernel/events/core.c                      |   63 +++++++++++++-----------------
 kernel/events/internal.h                  |    1 
 kernel/events/ring_buffer.c               |   10 +---
 kernel/sched.c                            |    2 
 kernel/watchdog.c                         |    2 
 samples/hw_breakpoint/data_breakpoint.c   |    2 
 46 files changed, 119 insertions(+), 141 deletions(-)

diff --git a/arch/alpha/kernel/perf_event.c b/arch/alpha/kernel/perf_event.c
index 90561c4..8e47709 100644
--- a/arch/alpha/kernel/perf_event.c
+++ b/arch/alpha/kernel/perf_event.c
@@ -847,7 +847,7 @@ static void alpha_perf_event_irq_handler(unsigned long la_ptr,
 	data.period = event->hw.last_period;
 
 	if (alpha_perf_event_set_period(event, hwc, idx)) {
-		if (perf_event_overflow(event, 1, &data, regs)) {
+		if (perf_event_overflow(event, &data, regs)) {
 			/* Interrupts coming too quickly; "throttle" the
 			 * counter, i.e., disable it for a little while.
 			 */
diff --git a/arch/arm/kernel/perf_event_v6.c b/arch/arm/kernel/perf_event_v6.c
index f1e8dd9..38dc4da 100644
--- a/arch/arm/kernel/perf_event_v6.c
+++ b/arch/arm/kernel/perf_event_v6.c
@@ -479,7 +479,7 @@ armv6pmu_handle_irq(int irq_num,
 		if (!armpmu_event_set_period(event, hwc, idx))
 			continue;
 
-		if (perf_event_overflow(event, 0, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			armpmu->disable(hwc, idx);
 	}
 
diff --git a/arch/arm/kernel/perf_event_v7.c b/arch/arm/kernel/perf_event_v7.c
index 4960686..6e5f875 100644
--- a/arch/arm/kernel/perf_event_v7.c
+++ b/arch/arm/kernel/perf_event_v7.c
@@ -787,7 +787,7 @@ static irqreturn_t armv7pmu_handle_irq(int irq_num, void *dev)
 		if (!armpmu_event_set_period(event, hwc, idx))
 			continue;
 
-		if (perf_event_overflow(event, 0, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			armpmu->disable(hwc, idx);
 	}
 
diff --git a/arch/arm/kernel/perf_event_xscale.c b/arch/arm/kernel/perf_event_xscale.c
index 39affbe..99b6b85 100644
--- a/arch/arm/kernel/perf_event_xscale.c
+++ b/arch/arm/kernel/perf_event_xscale.c
@@ -251,7 +251,7 @@ xscale1pmu_handle_irq(int irq_num, void *dev)
 		if (!armpmu_event_set_period(event, hwc, idx))
 			continue;
 
-		if (perf_event_overflow(event, 0, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			armpmu->disable(hwc, idx);
 	}
 
@@ -583,7 +583,7 @@ xscale2pmu_handle_irq(int irq_num, void *dev)
 		if (!armpmu_event_set_period(event, hwc, idx))
 			continue;
 
-		if (perf_event_overflow(event, 0, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			armpmu->disable(hwc, idx);
 	}
 
diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c
index 9726006..0c9b105 100644
--- a/arch/arm/kernel/ptrace.c
+++ b/arch/arm/kernel/ptrace.c
@@ -396,7 +396,7 @@ static long ptrace_hbp_idx_to_num(int idx)
 /*
  * Handle hitting a HW-breakpoint.
  */
-static void ptrace_hbptriggered(struct perf_event *bp, int unused,
+static void ptrace_hbptriggered(struct perf_event *bp,
 				     struct perf_sample_data *data,
 				     struct pt_regs *regs)
 {
diff --git a/arch/arm/kernel/swp_emulate.c b/arch/arm/kernel/swp_emulate.c
index 40ee7e5..5f452f8 100644
--- a/arch/arm/kernel/swp_emulate.c
+++ b/arch/arm/kernel/swp_emulate.c
@@ -183,7 +183,7 @@ static int swp_handler(struct pt_regs *regs, unsigned int instr)
 	unsigned int address, destreg, data, type;
 	unsigned int res = 0;
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, regs->ARM_pc);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, regs->ARM_pc);
 
 	if (current->pid != previous_pid) {
 		pr_debug("\"%s\" (%ld) uses deprecated SWP{B} instruction\n",
diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c
index bc0e1d8..9ea4f7d 100644
--- a/arch/arm/mm/fault.c
+++ b/arch/arm/mm/fault.c
@@ -318,11 +318,11 @@ do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs)
 	fault = __do_page_fault(mm, addr, fsr, tsk);
 	up_read(&mm->mmap_sem);
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, addr);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
 	if (fault & VM_FAULT_MAJOR)
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0, regs, addr);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, addr);
 	else if (fault & VM_FAULT_MINOR)
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0, regs, addr);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, addr);
 
 	/*
 	 * Handle the "normal" case first - VM_FAULT_MAJOR / VM_FAULT_MINOR
diff --git a/arch/mips/kernel/perf_event.c b/arch/mips/kernel/perf_event.c
index a824485..d0deaab 100644
--- a/arch/mips/kernel/perf_event.c
+++ b/arch/mips/kernel/perf_event.c
@@ -527,7 +527,7 @@ handle_associated_event(struct cpu_hw_events *cpuc,
 	if (!mipspmu_event_set_period(event, hwc, idx))
 		return;
 
-	if (perf_event_overflow(event, 0, data, regs))
+	if (perf_event_overflow(event, data, regs))
 		mipspmu->disable_event(idx);
 }
 
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index e9b3af2..b7517e3 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -578,12 +578,12 @@ static int simulate_llsc(struct pt_regs *regs, unsigned int opcode)
 {
 	if ((opcode & OPCODE) == LL) {
 		perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-				1, 0, regs, 0);
+				1, regs, 0);
 		return simulate_ll(regs, opcode);
 	}
 	if ((opcode & OPCODE) == SC) {
 		perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-				1, 0, regs, 0);
+				1, regs, 0);
 		return simulate_sc(regs, opcode);
 	}
 
@@ -602,7 +602,7 @@ static int simulate_rdhwr(struct pt_regs *regs, unsigned int opcode)
 		int rd = (opcode & RD) >> 11;
 		int rt = (opcode & RT) >> 16;
 		perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-				1, 0, regs, 0);
+				1, regs, 0);
 		switch (rd) {
 		case 0:		/* CPU number */
 			regs->regs[rt] = smp_processor_id();
@@ -640,7 +640,7 @@ static int simulate_sync(struct pt_regs *regs, unsigned int opcode)
 {
 	if ((opcode & OPCODE) == SPEC0 && (opcode & FUNC) == SYNC) {
 		perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-				1, 0, regs, 0);
+				1, regs, 0);
 		return 0;
 	}
 
diff --git a/arch/mips/kernel/unaligned.c b/arch/mips/kernel/unaligned.c
index cfea1ad..eb319b5 100644
--- a/arch/mips/kernel/unaligned.c
+++ b/arch/mips/kernel/unaligned.c
@@ -111,8 +111,7 @@ static void emulate_load_store_insn(struct pt_regs *regs,
 	unsigned long value;
 	unsigned int res;
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-		      1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 	/*
 	 * This load never faults.
@@ -517,7 +516,7 @@ asmlinkage void do_ade(struct pt_regs *regs)
 	mm_segment_t seg;
 
 	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS,
-			1, 0, regs, regs->cp0_badvaddr);
+			1, regs, regs->cp0_badvaddr);
 	/*
 	 * Did we catch a fault trying to load an instruction?
 	 * Or are we running in MIPS16 mode?
diff --git a/arch/mips/math-emu/cp1emu.c b/arch/mips/math-emu/cp1emu.c
index d32cb05..dbf2f93 100644
--- a/arch/mips/math-emu/cp1emu.c
+++ b/arch/mips/math-emu/cp1emu.c
@@ -272,8 +272,7 @@ static int cop1Emulate(struct pt_regs *xcp, struct mips_fpu_struct *ctx,
 	}
 
       emul:
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,
-			1, 0, xcp, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, xcp, 0);
 	MIPS_FPU_EMU_INC_STATS(emulated);
 	switch (MIPSInst_OPCODE(ir)) {
 	case ldc1_op:{
diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c
index 137ee76..937cf33 100644
--- a/arch/mips/mm/fault.c
+++ b/arch/mips/mm/fault.c
@@ -145,7 +145,7 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, unsigned long writ
 	 * the fault.
 	 */
 	fault = handle_mm_fault(mm, vma, address, write ? FAULT_FLAG_WRITE : 0);
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 	if (unlikely(fault & VM_FAULT_ERROR)) {
 		if (fault & VM_FAULT_OOM)
 			goto out_of_memory;
@@ -154,12 +154,10 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs, unsigned long writ
 		BUG();
 	}
 	if (fault & VM_FAULT_MAJOR) {
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ,
-				1, 0, regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address);
 		tsk->maj_flt++;
 	} else {
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN,
-				1, 0, regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address);
 		tsk->min_flt++;
 	}
 
diff --git a/arch/powerpc/include/asm/emulated_ops.h b/arch/powerpc/include/asm/emulated_ops.h
index 4592167..2cc41c7 100644
--- a/arch/powerpc/include/asm/emulated_ops.h
+++ b/arch/powerpc/include/asm/emulated_ops.h
@@ -78,14 +78,14 @@ extern void ppc_warn_emulated_print(const char *type);
 #define PPC_WARN_EMULATED(type, regs)					\
 	do {								\
 		perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS,		\
-			1, 0, regs, 0);					\
+			1, regs, 0);					\
 		__PPC_WARN_EMULATED(type);				\
 	} while (0)
 
 #define PPC_WARN_ALIGNMENT(type, regs)					\
 	do {								\
 		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS,		\
-			1, 0, regs, regs->dar);				\
+			1, regs, regs->dar);				\
 		__PPC_WARN_EMULATED(type);				\
 	} while (0)
 
diff --git a/arch/powerpc/kernel/perf_event.c b/arch/powerpc/kernel/perf_event.c
index 822f630..14967de 100644
--- a/arch/powerpc/kernel/perf_event.c
+++ b/arch/powerpc/kernel/perf_event.c
@@ -1207,7 +1207,7 @@ struct pmu power_pmu = {
  * here so there is no possibility of being interrupted.
  */
 static void record_and_restart(struct perf_event *event, unsigned long val,
-			       struct pt_regs *regs, int nmi)
+			       struct pt_regs *regs)
 {
 	u64 period = event->hw.sample_period;
 	s64 prev, delta, left;
@@ -1258,7 +1258,7 @@ static void record_and_restart(struct perf_event *event, unsigned long val,
 		if (event->attr.sample_type & PERF_SAMPLE_ADDR)
 			perf_get_data_addr(regs, &data.addr);
 
-		if (perf_event_overflow(event, nmi, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			power_pmu_stop(event, 0);
 	}
 }
@@ -1346,7 +1346,7 @@ static void perf_event_interrupt(struct pt_regs *regs)
 		if ((int)val < 0) {
 			/* event has overflowed */
 			found = 1;
-			record_and_restart(event, val, regs, nmi);
+			record_and_restart(event, val, regs);
 		}
 	}
 
diff --git a/arch/powerpc/kernel/perf_event_fsl_emb.c b/arch/powerpc/kernel/perf_event_fsl_emb.c
index b0dc8f7..0a6d2a9 100644
--- a/arch/powerpc/kernel/perf_event_fsl_emb.c
+++ b/arch/powerpc/kernel/perf_event_fsl_emb.c
@@ -568,7 +568,7 @@ static struct pmu fsl_emb_pmu = {
  * here so there is no possibility of being interrupted.
  */
 static void record_and_restart(struct perf_event *event, unsigned long val,
-			       struct pt_regs *regs, int nmi)
+			       struct pt_regs *regs)
 {
 	u64 period = event->hw.sample_period;
 	s64 prev, delta, left;
@@ -616,7 +616,7 @@ static void record_and_restart(struct perf_event *event, unsigned long val,
 		perf_sample_data_init(&data, 0);
 		data.period = event->hw.last_period;
 
-		if (perf_event_overflow(event, nmi, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			fsl_emb_pmu_stop(event, 0);
 	}
 }
@@ -644,7 +644,7 @@ static void perf_event_interrupt(struct pt_regs *regs)
 			if (event) {
 				/* event has overflowed */
 				found = 1;
-				record_and_restart(event, val, regs, nmi);
+				record_and_restart(event, val, regs);
 			} else {
 				/*
 				 * Disabled counter is negative,
diff --git a/arch/powerpc/kernel/ptrace.c b/arch/powerpc/kernel/ptrace.c
index cb22024..3177617 100644
--- a/arch/powerpc/kernel/ptrace.c
+++ b/arch/powerpc/kernel/ptrace.c
@@ -882,7 +882,7 @@ void user_disable_single_step(struct task_struct *task)
 }
 
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
-void ptrace_triggered(struct perf_event *bp, int nmi,
+void ptrace_triggered(struct perf_event *bp,
 		      struct perf_sample_data *data, struct pt_regs *regs)
 {
 	struct perf_event_attr attr;
diff --git a/arch/powerpc/mm/fault.c b/arch/powerpc/mm/fault.c
index 54f4fb9..dbc4825 100644
--- a/arch/powerpc/mm/fault.c
+++ b/arch/powerpc/mm/fault.c
@@ -173,7 +173,7 @@ int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
 		die("Weird page fault", regs, SIGSEGV);
 	}
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	/* When running in the kernel we expect faults to occur only to
 	 * addresses in user space.  All other faults represent errors in the
@@ -319,7 +319,7 @@ int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
 	}
 	if (ret & VM_FAULT_MAJOR) {
 		current->maj_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				     regs, address);
 #ifdef CONFIG_PPC_SMLPAR
 		if (firmware_has_feature(FW_FEATURE_CMO)) {
@@ -330,7 +330,7 @@ int __kprobes do_page_fault(struct pt_regs *regs, unsigned long address,
 #endif
 	} else {
 		current->min_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				     regs, address);
 	}
 	up_read(&mm->mmap_sem);
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index fe103e8..095f782 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -299,7 +299,7 @@ static inline int do_exception(struct pt_regs *regs, int access,
 		goto out;
 
 	address = trans_exc_code & __FAIL_ADDR_MASK;
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 	flags = FAULT_FLAG_ALLOW_RETRY;
 	if (access == VM_WRITE || (trans_exc_code & store_indication) == 0x400)
 		flags |= FAULT_FLAG_WRITE;
@@ -345,11 +345,11 @@ static inline int do_exception(struct pt_regs *regs, int access,
 	if (flags & FAULT_FLAG_ALLOW_RETRY) {
 		if (fault & VM_FAULT_MAJOR) {
 			tsk->maj_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				      regs, address);
 		} else {
 			tsk->min_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				      regs, address);
 		}
 		if (fault & VM_FAULT_RETRY) {
diff --git a/arch/sh/kernel/ptrace_32.c b/arch/sh/kernel/ptrace_32.c
index 3d7b209..8051976 100644
--- a/arch/sh/kernel/ptrace_32.c
+++ b/arch/sh/kernel/ptrace_32.c
@@ -63,7 +63,7 @@ static inline int put_stack_long(struct task_struct *task, int offset,
 	return 0;
 }
 
-void ptrace_triggered(struct perf_event *bp, int nmi,
+void ptrace_triggered(struct perf_event *bp,
 		      struct perf_sample_data *data, struct pt_regs *regs)
 {
 	struct perf_event_attr attr;
diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c
index b51a171..d9006f8 100644
--- a/arch/sh/kernel/traps_32.c
+++ b/arch/sh/kernel/traps_32.c
@@ -393,7 +393,7 @@ int handle_unaligned_access(insn_size_t instruction, struct pt_regs *regs,
 	 */
 	if (!expected) {
 		unaligned_fixups_notify(current, instruction, regs);
-		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1,
 			      regs, address);
 	}
 
diff --git a/arch/sh/kernel/traps_64.c b/arch/sh/kernel/traps_64.c
index 6713ca9..67110be 100644
--- a/arch/sh/kernel/traps_64.c
+++ b/arch/sh/kernel/traps_64.c
@@ -434,7 +434,7 @@ static int misaligned_load(struct pt_regs *regs,
 		return error;
 	}
 
-	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, address);
 
 	destreg = (opcode >> 4) & 0x3f;
 	if (user_mode(regs)) {
@@ -512,7 +512,7 @@ static int misaligned_store(struct pt_regs *regs,
 		return error;
 	}
 
-	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, address);
 
 	srcreg = (opcode >> 4) & 0x3f;
 	if (user_mode(regs)) {
@@ -588,7 +588,7 @@ static int misaligned_fpu_load(struct pt_regs *regs,
 		return error;
 	}
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, address);
 
 	destreg = (opcode >> 4) & 0x3f;
 	if (user_mode(regs)) {
@@ -665,7 +665,7 @@ static int misaligned_fpu_store(struct pt_regs *regs,
 		return error;
 	}
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, address);
 
 	srcreg = (opcode >> 4) & 0x3f;
 	if (user_mode(regs)) {
diff --git a/arch/sh/math-emu/math.c b/arch/sh/math-emu/math.c
index f76a509..9771952 100644
--- a/arch/sh/math-emu/math.c
+++ b/arch/sh/math-emu/math.c
@@ -620,7 +620,7 @@ int do_fpu_inst(unsigned short inst, struct pt_regs *regs)
 	struct task_struct *tsk = current;
 	struct sh_fpu_soft_struct *fpu = &(tsk->thread.xstate->softfpu);
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 	if (!(task_thread_info(tsk)->status & TS_USEDFPU)) {
 		/* initialize once. */
diff --git a/arch/sh/mm/fault_32.c b/arch/sh/mm/fault_32.c
index d4c34d7..7bebd04 100644
--- a/arch/sh/mm/fault_32.c
+++ b/arch/sh/mm/fault_32.c
@@ -160,7 +160,7 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs,
 	if ((regs->sr & SR_IMASK) != SR_IMASK)
 		local_irq_enable();
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	/*
 	 * If we're in an interrupt, have no user context or are running
@@ -210,11 +210,11 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs,
 	}
 	if (fault & VM_FAULT_MAJOR) {
 		tsk->maj_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				     regs, address);
 	} else {
 		tsk->min_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				     regs, address);
 	}
 
diff --git a/arch/sh/mm/tlbflush_64.c b/arch/sh/mm/tlbflush_64.c
index 7f5810f..e3430e0 100644
--- a/arch/sh/mm/tlbflush_64.c
+++ b/arch/sh/mm/tlbflush_64.c
@@ -116,7 +116,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long writeaccess,
 	/* Not an IO address, so reenable interrupts */
 	local_irq_enable();
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	/*
 	 * If we're in an interrupt or have no user
@@ -200,11 +200,11 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long writeaccess,
 
 	if (fault & VM_FAULT_MAJOR) {
 		tsk->maj_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				     regs, address);
 	} else {
 		tsk->min_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				     regs, address);
 	}
 
diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c
index 2cb0e1c..0b32f2e 100644
--- a/arch/sparc/kernel/perf_event.c
+++ b/arch/sparc/kernel/perf_event.c
@@ -1277,7 +1277,7 @@ static int __kprobes perf_event_nmi_handler(struct notifier_block *self,
 		if (!sparc_perf_event_set_period(event, hwc, idx))
 			continue;
 
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			sparc_pmu_stop(event, 0);
 	}
 
diff --git a/arch/sparc/kernel/unaligned_32.c b/arch/sparc/kernel/unaligned_32.c
index 4491f4c..7efbb2f 100644
--- a/arch/sparc/kernel/unaligned_32.c
+++ b/arch/sparc/kernel/unaligned_32.c
@@ -247,7 +247,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
 		unsigned long addr = compute_effective_address(regs, insn);
 		int err;
 
-		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
+		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr);
 		switch (dir) {
 		case load:
 			err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f),
@@ -338,7 +338,7 @@ asmlinkage void user_unaligned_trap(struct pt_regs *regs, unsigned int insn)
 		}
 
 		addr = compute_effective_address(regs, insn);
-		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
+		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr);
 		switch(dir) {
 		case load:
 			err = do_int_load(fetch_reg_addr(((insn>>25)&0x1f),
diff --git a/arch/sparc/kernel/unaligned_64.c b/arch/sparc/kernel/unaligned_64.c
index b2b019e..35cff16 100644
--- a/arch/sparc/kernel/unaligned_64.c
+++ b/arch/sparc/kernel/unaligned_64.c
@@ -317,7 +317,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
 
 		addr = compute_effective_address(regs, insn,
 						 ((insn >> 25) & 0x1f));
-		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, addr);
+		perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, addr);
 		switch (asi) {
 		case ASI_NL:
 		case ASI_AIUPL:
@@ -384,7 +384,7 @@ int handle_popc(u32 insn, struct pt_regs *regs)
 	int ret, i, rd = ((insn >> 25) & 0x1f);
 	int from_kernel = (regs->tstate & TSTATE_PRIV) != 0;
 	                        
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 	if (insn & 0x2000) {
 		maybe_flush_windows(0, 0, rd, from_kernel);
 		value = sign_extend_imm13(insn);
@@ -431,7 +431,7 @@ int handle_ldf_stq(u32 insn, struct pt_regs *regs)
 	int asi = decode_asi(insn, regs);
 	int flag = (freg < 32) ? FPRS_DL : FPRS_DU;
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 	save_and_clear_fpu();
 	current_thread_info()->xfsr[0] &= ~0x1c000;
@@ -554,7 +554,7 @@ void handle_ld_nf(u32 insn, struct pt_regs *regs)
 	int from_kernel = (regs->tstate & TSTATE_PRIV) != 0;
 	unsigned long *reg;
 	                        
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 	maybe_flush_windows(0, 0, rd, from_kernel);
 	reg = fetch_reg_addr(rd, regs);
@@ -586,7 +586,7 @@ void handle_lddfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr
 
 	if (tstate & TSTATE_PRIV)
 		die_if_kernel("lddfmna from kernel", regs);
-	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar);
+	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, sfar);
 	if (test_thread_flag(TIF_32BIT))
 		pc = (u32)pc;
 	if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
@@ -647,7 +647,7 @@ void handle_stdfmna(struct pt_regs *regs, unsigned long sfar, unsigned long sfsr
 
 	if (tstate & TSTATE_PRIV)
 		die_if_kernel("stdfmna from kernel", regs);
-	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, 0, regs, sfar);
+	perf_sw_event(PERF_COUNT_SW_ALIGNMENT_FAULTS, 1, regs, sfar);
 	if (test_thread_flag(TIF_32BIT))
 		pc = (u32)pc;
 	if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
diff --git a/arch/sparc/kernel/visemul.c b/arch/sparc/kernel/visemul.c
index 3635771..32b626c 100644
--- a/arch/sparc/kernel/visemul.c
+++ b/arch/sparc/kernel/visemul.c
@@ -802,7 +802,7 @@ int vis_emul(struct pt_regs *regs, unsigned int insn)
 
 	BUG_ON(regs->tstate & TSTATE_PRIV);
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 	if (test_thread_flag(TIF_32BIT))
 		pc = (u32)pc;
diff --git a/arch/sparc/math-emu/math_32.c b/arch/sparc/math-emu/math_32.c
index a3fccde..aa4d55b 100644
--- a/arch/sparc/math-emu/math_32.c
+++ b/arch/sparc/math-emu/math_32.c
@@ -164,7 +164,7 @@ int do_mathemu(struct pt_regs *regs, struct task_struct *fpt)
 	int retcode = 0;                               /* assume all succeed */
 	unsigned long insn;
 
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 
 #ifdef DEBUG_MATHEMU
 	printk("In do_mathemu()... pc is %08lx\n", regs->pc);
diff --git a/arch/sparc/math-emu/math_64.c b/arch/sparc/math-emu/math_64.c
index 56d2c44..e575bd2 100644
--- a/arch/sparc/math-emu/math_64.c
+++ b/arch/sparc/math-emu/math_64.c
@@ -184,7 +184,7 @@ int do_mathemu(struct pt_regs *regs, struct fpustate *f)
 
 	if (tstate & TSTATE_PRIV)
 		die_if_kernel("unfinished/unimplemented FPop from kernel", regs);
-	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, 0, regs, 0);
+	perf_sw_event(PERF_COUNT_SW_EMULATION_FAULTS, 1, regs, 0);
 	if (test_thread_flag(TIF_32BIT))
 		pc = (u32)pc;
 	if (get_user(insn, (u32 __user *) pc) != -EFAULT) {
diff --git a/arch/sparc/mm/fault_32.c b/arch/sparc/mm/fault_32.c
index 7543ddb..aa1c1b1 100644
--- a/arch/sparc/mm/fault_32.c
+++ b/arch/sparc/mm/fault_32.c
@@ -251,7 +251,7 @@ asmlinkage void do_sparc_fault(struct pt_regs *regs, int text_fault, int write,
         if (in_atomic() || !mm)
                 goto no_context;
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	down_read(&mm->mmap_sem);
 
@@ -301,12 +301,10 @@ asmlinkage void do_sparc_fault(struct pt_regs *regs, int text_fault, int write,
 	}
 	if (fault & VM_FAULT_MAJOR) {
 		current->maj_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
-			      regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address);
 	} else {
 		current->min_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
-			      regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address);
 	}
 	up_read(&mm->mmap_sem);
 	return;
diff --git a/arch/sparc/mm/fault_64.c b/arch/sparc/mm/fault_64.c
index f92ce56..504c062 100644
--- a/arch/sparc/mm/fault_64.c
+++ b/arch/sparc/mm/fault_64.c
@@ -325,7 +325,7 @@ asmlinkage void __kprobes do_sparc64_fault(struct pt_regs *regs)
 	if (in_atomic() || !mm)
 		goto intr_or_no_mm;
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	if (!down_read_trylock(&mm->mmap_sem)) {
 		if ((regs->tstate & TSTATE_PRIV) &&
@@ -433,12 +433,10 @@ asmlinkage void __kprobes do_sparc64_fault(struct pt_regs *regs)
 	}
 	if (fault & VM_FAULT_MAJOR) {
 		current->maj_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
-			      regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, regs, address);
 	} else {
 		current->min_flt++;
-		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
-			      regs, address);
+		perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, regs, address);
 	}
 	up_read(&mm->mmap_sem);
 
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 8a57f9a..5b86ec5 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1339,7 +1339,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
 		if (!x86_perf_event_set_period(event))
 			continue;
 
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index 41178c8..d38b002 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -1003,7 +1003,7 @@ static int intel_pmu_handle_irq(struct pt_regs *regs)
 
 		data.period = event->hw.last_period;
 
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index bab491b..0941f93 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -340,7 +340,7 @@ static int intel_pmu_drain_bts_buffer(void)
 	 */
 	perf_prepare_sample(&header, &data, event, &regs);
 
-	if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1))
+	if (perf_output_begin(&handle, event, header.size * (top - at), 1))
 		return 1;
 
 	for (; at < top; at++) {
@@ -616,7 +616,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
 	else
 		regs.flags &= ~PERF_EFLAGS_EXACT;
 
-	if (perf_event_overflow(event, 1, &data, &regs))
+	if (perf_event_overflow(event, &data, &regs))
 		x86_pmu_stop(event, 0);
 }
 
diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c
index f76fddf..d6e6a67 100644
--- a/arch/x86/kernel/cpu/perf_event_p4.c
+++ b/arch/x86/kernel/cpu/perf_event_p4.c
@@ -970,7 +970,7 @@ static int p4_pmu_handle_irq(struct pt_regs *regs)
 
 		if (!x86_perf_event_set_period(event))
 			continue;
-		if (perf_event_overflow(event, 1, &data, regs))
+		if (perf_event_overflow(event, &data, regs))
 			x86_pmu_stop(event, 0);
 	}
 
diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c
index 5f9ecff..98da6a7 100644
--- a/arch/x86/kernel/kgdb.c
+++ b/arch/x86/kernel/kgdb.c
@@ -608,7 +608,7 @@ int kgdb_arch_init(void)
 	return register_die_notifier(&kgdb_notifier);
 }
 
-static void kgdb_hw_overflow_handler(struct perf_event *event, int nmi,
+static void kgdb_hw_overflow_handler(struct perf_event *event,
 		struct perf_sample_data *data, struct pt_regs *regs)
 {
 	struct task_struct *tsk = current;
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 807c2a2..11db2e9 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -528,7 +528,7 @@ static int genregs_set(struct task_struct *target,
 	return ret;
 }
 
-static void ptrace_triggered(struct perf_event *bp, int nmi,
+static void ptrace_triggered(struct perf_event *bp,
 			     struct perf_sample_data *data,
 			     struct pt_regs *regs)
 {
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 2dbf6bf..4d09df0 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -1059,7 +1059,7 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	if (unlikely(error_code & PF_RSVD))
 		pgtable_bad(regs, error_code, address);
 
-	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, 0, regs, address);
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
 	/*
 	 * If we're in an interrupt, have no user context or are running
@@ -1161,11 +1161,11 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	if (flags & FAULT_FLAG_ALLOW_RETRY) {
 		if (fault & VM_FAULT_MAJOR) {
 			tsk->maj_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ, 1,
 				      regs, address);
 		} else {
 			tsk->min_flt++;
-			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1, 0,
+			perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN, 1,
 				      regs, address);
 		}
 		if (fault & VM_FAULT_RETRY) {
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 2f7b5d4..0946a8b 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -682,7 +682,7 @@ enum perf_event_active_state {
 struct file;
 struct perf_sample_data;
 
-typedef void (*perf_overflow_handler_t)(struct perf_event *, int,
+typedef void (*perf_overflow_handler_t)(struct perf_event *,
 					struct perf_sample_data *,
 					struct pt_regs *regs);
 
@@ -925,7 +925,6 @@ struct perf_output_handle {
 	unsigned long			size;
 	void				*addr;
 	int				page;
-	int				nmi;
 	int				sample;
 };
 
@@ -993,7 +992,7 @@ extern void perf_prepare_sample(struct perf_event_header *header,
 				struct perf_event *event,
 				struct pt_regs *regs);
 
-extern int perf_event_overflow(struct perf_event *event, int nmi,
+extern int perf_event_overflow(struct perf_event *event,
 				 struct perf_sample_data *data,
 				 struct pt_regs *regs);
 
@@ -1012,7 +1011,7 @@ static inline int is_software_event(struct perf_event *event)
 
 extern struct jump_label_key perf_swevent_enabled[PERF_COUNT_SW_MAX];
 
-extern void __perf_sw_event(u32, u64, int, struct pt_regs *, u64);
+extern void __perf_sw_event(u32, u64, struct pt_regs *, u64);
 
 #ifndef perf_arch_fetch_caller_regs
 static inline void perf_arch_fetch_caller_regs(struct pt_regs *regs, unsigned long ip) { }
@@ -1034,7 +1033,7 @@ static inline void perf_fetch_caller_regs(struct pt_regs *regs)
 }
 
 static __always_inline void
-perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr)
+perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr)
 {
 	struct pt_regs hot_regs;
 
@@ -1043,7 +1042,7 @@ perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr)
 			perf_fetch_caller_regs(&hot_regs);
 			regs = &hot_regs;
 		}
-		__perf_sw_event(event_id, nr, nmi, regs, addr);
+		__perf_sw_event(event_id, nr, regs, addr);
 	}
 }
 
@@ -1057,7 +1056,7 @@ static inline void perf_event_task_sched_in(struct task_struct *task)
 
 static inline void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next)
 {
-	perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, NULL, 0);
+	perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, NULL, 0);
 
 	__perf_event_task_sched_out(task, next);
 }
@@ -1119,7 +1118,7 @@ extern void perf_bp_event(struct perf_event *event, void *data);
 
 extern int perf_output_begin(struct perf_output_handle *handle,
 			     struct perf_event *event, unsigned int size,
-			     int nmi, int sample);
+			     int sample);
 extern void perf_output_end(struct perf_output_handle *handle);
 extern void perf_output_copy(struct perf_output_handle *handle,
 			     const void *buf, unsigned int len);
@@ -1143,8 +1142,7 @@ static inline int perf_event_task_disable(void)				{ return -EINVAL; }
 static inline int perf_event_task_enable(void)				{ return -EINVAL; }
 
 static inline void
-perf_sw_event(u32 event_id, u64 nr, int nmi,
-		     struct pt_regs *regs, u64 addr)			{ }
+perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr)	{ }
 static inline void
 perf_bp_event(struct perf_event *event, void *data)			{ }
 
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 270e32f..dbd1ca7 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -3972,7 +3972,7 @@ void perf_prepare_sample(struct perf_event_header *header,
 	}
 }
 
-static void perf_event_output(struct perf_event *event, int nmi,
+static void perf_event_output(struct perf_event *event,
 				struct perf_sample_data *data,
 				struct pt_regs *regs)
 {
@@ -3984,7 +3984,7 @@ static void perf_event_output(struct perf_event *event, int nmi,
 
 	perf_prepare_sample(&header, data, event, regs);
 
-	if (perf_output_begin(&handle, event, header.size, nmi, 1))
+	if (perf_output_begin(&handle, event, header.size, 1))
 		goto exit;
 
 	perf_output_sample(&handle, &header, data, event);
@@ -4024,7 +4024,7 @@ perf_event_read_event(struct perf_event *event,
 	int ret;
 
 	perf_event_header__init_id(&read_event.header, &sample, event);
-	ret = perf_output_begin(&handle, event, read_event.header.size, 0, 0);
+	ret = perf_output_begin(&handle, event, read_event.header.size, 0);
 	if (ret)
 		return;
 
@@ -4067,7 +4067,7 @@ static void perf_event_task_output(struct perf_event *event,
 	perf_event_header__init_id(&task_event->event_id.header, &sample, event);
 
 	ret = perf_output_begin(&handle, event,
-				task_event->event_id.header.size, 0, 0);
+				task_event->event_id.header.size, 0);
 	if (ret)
 		goto out;
 
@@ -4204,7 +4204,7 @@ static void perf_event_comm_output(struct perf_event *event,
 
 	perf_event_header__init_id(&comm_event->event_id.header, &sample, event);
 	ret = perf_output_begin(&handle, event,
-				comm_event->event_id.header.size, 0, 0);
+				comm_event->event_id.header.size, 0);
 
 	if (ret)
 		goto out;
@@ -4351,7 +4351,7 @@ static void perf_event_mmap_output(struct perf_event *event,
 
 	perf_event_header__init_id(&mmap_event->event_id.header, &sample, event);
 	ret = perf_output_begin(&handle, event,
-				mmap_event->event_id.header.size, 0, 0);
+				mmap_event->event_id.header.size, 0);
 	if (ret)
 		goto out;
 
@@ -4546,7 +4546,7 @@ static void perf_log_throttle(struct perf_event *event, int enable)
 	perf_event_header__init_id(&throttle_event.header, &sample, event);
 
 	ret = perf_output_begin(&handle, event,
-				throttle_event.header.size, 1, 0);
+				throttle_event.header.size, 0);
 	if (ret)
 		return;
 
@@ -4559,7 +4559,7 @@ static void perf_log_throttle(struct perf_event *event, int enable)
  * Generic event overflow handling, sampling.
  */
 
-static int __perf_event_overflow(struct perf_event *event, int nmi,
+static int __perf_event_overflow(struct perf_event *event,
 				   int throttle, struct perf_sample_data *data,
 				   struct pt_regs *regs)
 {
@@ -4602,34 +4602,28 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
 	if (events && atomic_dec_and_test(&event->event_limit)) {
 		ret = 1;
 		event->pending_kill = POLL_HUP;
-		if (nmi) {
-			event->pending_disable = 1;
-			irq_work_queue(&event->pending);
-		} else
-			perf_event_disable(event);
+		event->pending_disable = 1;
+		irq_work_queue(&event->pending);
 	}
 
 	if (event->overflow_handler)
-		event->overflow_handler(event, nmi, data, regs);
+		event->overflow_handler(event, data, regs);
 	else
-		perf_event_output(event, nmi, data, regs);
+		perf_event_output(event, data, regs);
 
 	if (event->fasync && event->pending_kill) {
-		if (nmi) {
-			event->pending_wakeup = 1;
-			irq_work_queue(&event->pending);
-		} else
-			perf_event_wakeup(event);
+		event->pending_wakeup = 1;
+		irq_work_queue(&event->pending);
 	}
 
 	return ret;
 }
 
-int perf_event_overflow(struct perf_event *event, int nmi,
+int perf_event_overflow(struct perf_event *event,
 			  struct perf_sample_data *data,
 			  struct pt_regs *regs)
 {
-	return __perf_event_overflow(event, nmi, 1, data, regs);
+	return __perf_event_overflow(event, 1, data, regs);
 }
 
 /*
@@ -4678,7 +4672,7 @@ static u64 perf_swevent_set_period(struct perf_event *event)
 }
 
 static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
-				    int nmi, struct perf_sample_data *data,
+				    struct perf_sample_data *data,
 				    struct pt_regs *regs)
 {
 	struct hw_perf_event *hwc = &event->hw;
@@ -4692,7 +4686,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
 		return;
 
 	for (; overflow; overflow--) {
-		if (__perf_event_overflow(event, nmi, throttle,
+		if (__perf_event_overflow(event, throttle,
 					    data, regs)) {
 			/*
 			 * We inhibit the overflow from happening when
@@ -4705,7 +4699,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
 }
 
 static void perf_swevent_event(struct perf_event *event, u64 nr,
-			       int nmi, struct perf_sample_data *data,
+			       struct perf_sample_data *data,
 			       struct pt_regs *regs)
 {
 	struct hw_perf_event *hwc = &event->hw;
@@ -4719,12 +4713,12 @@ static void perf_swevent_event(struct perf_event *event, u64 nr,
 		return;
 
 	if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq)
-		return perf_swevent_overflow(event, 1, nmi, data, regs);
+		return perf_swevent_overflow(event, 1, data, regs);
 
 	if (local64_add_negative(nr, &hwc->period_left))
 		return;
 
-	perf_swevent_overflow(event, 0, nmi, data, regs);
+	perf_swevent_overflow(event, 0, data, regs);
 }
 
 static int perf_exclude_event(struct perf_event *event,
@@ -4812,7 +4806,7 @@ find_swevent_head(struct swevent_htable *swhash, struct perf_event *event)
 }
 
 static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
-				    u64 nr, int nmi,
+				    u64 nr,
 				    struct perf_sample_data *data,
 				    struct pt_regs *regs)
 {
@@ -4828,7 +4822,7 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
 
 	hlist_for_each_entry_rcu(event, node, head, hlist_entry) {
 		if (perf_swevent_match(event, type, event_id, data, regs))
-			perf_swevent_event(event, nr, nmi, data, regs);
+			perf_swevent_event(event, nr, data, regs);
 	}
 end:
 	rcu_read_unlock();
@@ -4849,8 +4843,7 @@ inline void perf_swevent_put_recursion_context(int rctx)
 	put_recursion_context(swhash->recursion, rctx);
 }
 
-void __perf_sw_event(u32 event_id, u64 nr, int nmi,
-			    struct pt_regs *regs, u64 addr)
+void __perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr)
 {
 	struct perf_sample_data data;
 	int rctx;
@@ -4862,7 +4855,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi,
 
 	perf_sample_data_init(&data, addr);
 
-	do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs);
+	do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, &data, regs);
 
 	perf_swevent_put_recursion_context(rctx);
 	preempt_enable_notrace();
@@ -5110,7 +5103,7 @@ void perf_tp_event(u64 addr, u64 count, void *record, int entry_size,
 
 	hlist_for_each_entry_rcu(event, node, head, hlist_entry) {
 		if (perf_tp_event_match(event, &data, regs))
-			perf_swevent_event(event, count, 1, &data, regs);
+			perf_swevent_event(event, count, &data, regs);
 	}
 
 	perf_swevent_put_recursion_context(rctx);
@@ -5203,7 +5196,7 @@ void perf_bp_event(struct perf_event *bp, void *data)
 	perf_sample_data_init(&sample, bp->attr.bp_addr);
 
 	if (!bp->hw.state && !perf_exclude_event(bp, regs))
-		perf_swevent_event(bp, 1, 1, &sample, regs);
+		perf_swevent_event(bp, 1, &sample, regs);
 }
 #endif
 
@@ -5232,7 +5225,7 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
 
 	if (regs && !perf_exclude_event(event, regs)) {
 		if (!(event->attr.exclude_idle && current->pid == 0))
-			if (perf_event_overflow(event, 0, &data, regs))
+			if (perf_event_overflow(event, &data, regs))
 				ret = HRTIMER_NORESTART;
 	}
 
diff --git a/kernel/events/internal.h b/kernel/events/internal.h
index 114f27f..09097dd 100644
--- a/kernel/events/internal.h
+++ b/kernel/events/internal.h
@@ -27,7 +27,6 @@ struct ring_buffer {
 	void				*data_pages[0];
 };
 
-
 extern void rb_free(struct ring_buffer *rb);
 extern struct ring_buffer *
 rb_alloc(int nr_pages, long watermark, int cpu, int flags);
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index fc2701c..8b3b736 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -38,11 +38,8 @@ static void perf_output_wakeup(struct perf_output_handle *handle)
 {
 	atomic_set(&handle->rb->poll, POLL_IN);
 
-	if (handle->nmi) {
-		handle->event->pending_wakeup = 1;
-		irq_work_queue(&handle->event->pending);
-	} else
-		perf_event_wakeup(handle->event);
+	handle->event->pending_wakeup = 1;
+	irq_work_queue(&handle->event->pending);
 }
 
 /*
@@ -102,7 +99,7 @@ static void perf_output_put_handle(struct perf_output_handle *handle)
 
 int perf_output_begin(struct perf_output_handle *handle,
 		      struct perf_event *event, unsigned int size,
-		      int nmi, int sample)
+		      int sample)
 {
 	struct ring_buffer *rb;
 	unsigned long tail, offset, head;
@@ -127,7 +124,6 @@ int perf_output_begin(struct perf_output_handle *handle,
 
 	handle->rb	= rb;
 	handle->event	= event;
-	handle->nmi	= nmi;
 	handle->sample	= sample;
 
 	if (!rb->nr_pages)
diff --git a/kernel/sched.c b/kernel/sched.c
index 3f2e502..d08d110 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -2220,7 +2220,7 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu)
 
 	if (task_cpu(p) != new_cpu) {
 		p->se.nr_migrations++;
-		perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, 1, NULL, 0);
+		perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, NULL, 0);
 	}
 
 	__set_task_cpu(p, new_cpu);
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 752b75b..a6708e6 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -211,7 +211,7 @@ static struct perf_event_attr wd_hw_attr = {
 };
 
 /* Callback function for perf event subsystem */
-static void watchdog_overflow_callback(struct perf_event *event, int nmi,
+static void watchdog_overflow_callback(struct perf_event *event,
 		 struct perf_sample_data *data,
 		 struct pt_regs *regs)
 {
diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c
index 0636539..7b164d3 100644
--- a/samples/hw_breakpoint/data_breakpoint.c
+++ b/samples/hw_breakpoint/data_breakpoint.c
@@ -41,7 +41,7 @@ module_param_string(ksym, ksym_name, KSYM_NAME_LEN, S_IRUGO);
 MODULE_PARM_DESC(ksym, "Kernel symbol to monitor; this module will report any"
 			" write operations on the kernel symbol");
 
-static void sample_hbp_handler(struct perf_event *bp, int nmi,
+static void sample_hbp_handler(struct perf_event *bp,
 			       struct perf_sample_data *data,
 			       struct pt_regs *regs)
 {


^ permalink raw reply related	[flat|nested] 9+ messages in thread
* Re: [stable] [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-07-29 11:59       ` Peter Zijlstra
@ 2011-08-02 16:35         ` Greg KH
  2011-08-03 11:49           ` Peter Zijlstra
  0 siblings, 1 reply; 9+ messages in thread
From: Greg KH @ 2011-08-02 16:35 UTC (permalink / raw)
  To: Peter Zijlstra
  Cc: Vince Weaver, Ingo Molnar, Paul Mackerras, linux-kernel,
	Arnaldo Carvalho de Melo, stable

On Fri, Jul 29, 2011 at 01:59:10PM +0200, Peter Zijlstra wrote:
> On Thu, 2011-07-28 at 22:05 -0700, Greg KH wrote:
> > 
> > What is that commit id so it can be properly referenced in the patch
> > below?  And what tree should this patch go into?  .39?  3.0?  Anything
> > newer? 
> 
> .32+ in as far as anybody is still maintaining them, if it fails to
> apply to any of them (quite possible) do tell and I can provide
> backports etc..

I've applied it to 3.0, and it didn't apply to any older kernel due to
file movements, so if you could, can you provide a version for
.32-stable?

thanks,

greg k-h

^ permalink raw reply	[flat|nested] 9+ messages in thread
* Re: [stable] [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels
  2011-08-02 16:35         ` Greg KH
@ 2011-08-03 11:49           ` Peter Zijlstra
  0 siblings, 0 replies; 9+ messages in thread
From: Peter Zijlstra @ 2011-08-03 11:49 UTC (permalink / raw)
  To: Greg KH
  Cc: Vince Weaver, Ingo Molnar, Paul Mackerras, linux-kernel,
	Arnaldo Carvalho de Melo, stable

On Tue, 2011-08-02 at 09:35 -0700, Greg KH wrote:
> On Fri, Jul 29, 2011 at 01:59:10PM +0200, Peter Zijlstra wrote:
> > On Thu, 2011-07-28 at 22:05 -0700, Greg KH wrote:
> > > 
> > > What is that commit id so it can be properly referenced in the patch
> > > below?  And what tree should this patch go into?  .39?  3.0?  Anything
> > > newer? 
> > 
> > .32+ in as far as anybody is still maintaining them, if it fails to
> > apply to any of them (quite possible) do tell and I can provide
> > backports etc..
> 
> I've applied it to 3.0, and it didn't apply to any older kernel due to
> file movements, so if you could, can you provide a version for
> .32-stable?


---
From: Peter Zijlstra <a.p.zijlstra@chello.nl>

The below patch is for -stable only, upstream has a much larger patch
that contains the below hunk in commit a8b0ca17b80e92faab46ee7179ba9e99ccb61233

Vince found that under certain circumstances software event overflows
go wrong and deadlock. Avoid trying to delete a timer from the timer
callback.

Reported-by: Vince Weaver <vweaver1@eecs.utk.edu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
 kernel/perf_event.c |    8 ++------
 1 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/kernel/perf_event.c b/kernel/perf_event.c
index 183d437..8f97aea 100644
--- a/kernel/perf_event.c
+++ b/kernel/perf_event.c
@@ -3693,12 +3693,8 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
 	if (events && atomic_dec_and_test(&event->event_limit)) {
 		ret = 1;
 		event->pending_kill = POLL_HUP;
-		if (nmi) {
-			event->pending_disable = 1;
-			perf_pending_queue(&event->pending,
-					   perf_pending_event);
-		} else
-			perf_event_disable(event);
+		event->pending_disable = 1;
+		perf_pending_queue(&event->pending, perf_pending_event);
 	}
 
 	perf_event_output(event, nmi, data, regs);


^ permalink raw reply related	[flat|nested] 9+ messages in thread
end of thread, other threads:[~2011-08-03 11:50 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-07-22 16:13 [perf] 3.0 overflow test locks up system Vince Weaver
2011-07-27 18:51 ` [perf] overflow/perf_count_sw_cpu_clock crashes recent kernels Vince Weaver
2011-07-28  1:54   ` Lin Ming
2011-07-28  2:39     ` Lin Ming
2011-07-28 18:47   ` Peter Zijlstra
2011-07-29  5:05     ` [stable] " Greg KH
2011-07-29 11:59       ` Peter Zijlstra
2011-08-02 16:35         ` Greg KH
2011-08-03 11:49           ` Peter Zijlstra

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.