login
Login
/
Register
Search
Search this site:
Forums
News
Blogs
Features
Site
Home
»
Mailing list archives
»
linux-kernel
»
2010
»
August
»
19
Re: [BUG] perf_events: NMI watchdog event cannot be throttled
view
thread
Previous message: [
thread
] [
date
] [
author
]
Next message: [
thread
] [
date
] [
author
]
[view in full thread]
From: Stephane Eranian
Subject:
Re: [BUG] perf_events: NMI watchdog event cannot be throttled
Date: Thursday, August 19, 2010 - 6:01 am
On Thu, Aug 19, 2010 at 1:24 PM, Stephane Eranian <eranian@google.com> wrote:
quoted text
> Yeah, that should probably fix it. Let me try it out. >
Works for me. Thanks.
quoted text
> > On Thu, Aug 19, 2010 at 1:05 PM, Peter Zijlstra <peterz@infradead.org> wrote: >> On Wed, 2010-08-18 at 22:26 +0200, Stephane Eranian wrote: >>> Hi, >>> >>> I ran into some issue with the NMI watchdog not firing in a deadlock >>> situation. After some debugging I found the source of the problem. >>> >>> The NMI watchdog is currently subject, like any other events, to interrupt >>> throttling. The heart of the problem is that if you are deadlocked on a CPU >>> with interrupts masked, the timer interrupt won't fire, therefore the >>> hwc->interrupts >>> field won't be reset. Then, depending on the max sampling rate, you >>> could eventually >>> fail the max interrupt rate test in __pfm_overflow_handler() and >>> perf_events would >>> throttle, i.e., stop, the NMI watchdog event before the 5s delay to panic. >>> Thus, you would never get the panic. I ran into this problem myself. >>> >>> This is a serious issue because perf_events must ensure the watchdog can >>> always fire, regardless of the interrupt masking situation. >>> >>> Look like one way of solving the problem would be to mark the NMI watchdog >>> event as immune to throttling. The event being internal to the kernel we could >>> trust the event setup from perf_event_create_kernel_counter(). >> >> Something like so? >> >> --- >> kernel/watchdog.c | 3 +++ >> 1 files changed, 3 insertions(+), 0 deletions(-) >> >> diff --git a/kernel/watchdog.c b/kernel/watchdog.c >> index 613bc1f..e0fe6e4 100644 >> --- a/kernel/watchdog.c >> +++ b/kernel/watchdog.c >> @@ -206,6 +206,9 @@ void watchdog_overflow_callback(struct perf_event *event, int nmi, >> struct perf_sample_data *data, >> struct pt_regs *regs) >> { >> + /* Ensure the watchdog never gets throttled. */ >> + event->hw.interrupts = 0; >> + >> if (__get_cpu_var(watchdog_nmi_touch) == true) { >> __get_cpu_var(watchdog_nmi_touch) = false; >> return; >> >> >
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to
majordomo@vger.kernel.org
More majordomo info at
http://vger.kernel.org/majordomo-info.html
Please read the FAQ at
http://www.tux.org/lkml/
Previous message: [
thread
] [
date
] [
author
]
Next message: [
thread
] [
date
] [
author
]
Messages in current thread:
[BUG] perf_events: NMI watchdog event cannot be throttled
, Stephane Eranian
, (Wed Aug 18, 1:26 pm)
Re: [BUG] perf_events: NMI watchdog event cannot be throttled
, Peter Zijlstra
, (Thu Aug 19, 4:05 am)
Re: [BUG] perf_events: NMI watchdog event cannot be throttled
, Stephane Eranian
, (Thu Aug 19, 4:24 am)
Re: [BUG] perf_events: NMI watchdog event cannot be throttled
, Stephane Eranian
, (Thu Aug 19, 6:01 am)
[tip:perf/urgent] watchdog: Don't throttle the watchdog
, tip-bot for Peter Zi ...
, (Fri Aug 20, 7:18 am)
[tip:perf/urgent] watchdog: Don't throttle the watchdog
, tip-bot for Peter Zi ...
, (Mon Aug 23, 1:51 am)
Navigation
Mailing list archives
Recent posts
Popular discussions
linux-kernel
:
Paul Turner
[tg_shares_up rewrite v4 11/11] sched: update tg->shares after cpu.shares write
Matthew Garrett
Re: [PATCH] Enable speedstep for sonoma processors.
Mauro Carvalho Chehab
Re: [PATCH 1/2] media: Add timberdale video-in driver
Peter Zijlstra
[PATCH 23/30] netvm: skb processing
Greg Kroah-Hartman
[PATCH 21/28] cgroupfs: create /sys/fs/cgroup to mount cgroupfs on
git
:
Jan Hudec
Re: GIT push to sftp (feature request)
Steffen Prohaska
[PATCH 0/4] core.ignorecase
Johannes Schindelin
Re: Git checkout preserve timestamp?
Linus Torvalds
[PATCH 1/7] Make unpack_trees_options bit flags actual bitfields
Johan Herland
Re: What's cooking in git.git (Oct 2010, #01; Wed, 13)
linux-netdev
:
David Miller
Re: [PATCH 1/3] f_phonet: dev_kfree_skb instead of dev_kfree_skb_any in TX callback
Richard Cochran
Re: [PATCH v3 3/3] ptp: Added a clock that uses the eTSEC found on the MPC85xx.
Jan Engelhardt
Re: [PATCH] Fix netfilter xt_time's time_mt()'s use of do_div()
Herbert Xu
Re: [RFC PATCH 00/17] virtual-bus
Jeff Kirsher
Re: [net-next-2.6 PATCH] e1000e: don't inadvertently re-set INTX_DISABLE
git-commits-head
:
Linux Kernel Mailing List
ALSA: hda - Enable beep on Realtek codecs with PCI SSID override
Linux Kernel Mailing List
Use path_put() in a few places instead of {mnt,d}put()
Linux Kernel Mailing List
mv643xx_eth: use sw csum for big packets
Linux Kernel Mailing List
arm: fix HAVE_CLK merge goof
Linux Kernel Mailing List
arm: convert pcm037 platform to use smsc911x
freebsd-current
:
David Wolfskill
"interrupt storm..."; seems associated with an0 NIC
Andriy Gapon
Re: letting glabel recognise a media change
Garrett Cooper
Re: Only display ACPI bootmenu key if ACPI is present
Pyun YongHyeon
CFT: msk(4) Rx checksum offloading support
FreeBSD Tinderbox
[head tinderbox] failure on sparc64/sparc64
Colocation donated by:
Syndicate