From: Cyrill Gorcunov <gorcunov@gmail.com>
To: Don Zickus <dzickus@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Jason Wessel <jason.wessel@windriver.com>,
Ingo Molnar <mingo@elte.hu>,
Robert Richter <robert.richter@amd.com>,
ying.huang@intel.com, Andi Kleen <andi@firstfloor.org>,
LKML <linux-kernel@vger.kernel.org>,
Frederic Weisbecker <fweisbec@gmail.com>
Subject: Re: [V2 PATCH 0/6] x86, NMI: give NMI handler a face-lift
Date: Fri, 19 Nov 2010 01:15:54 +0300 [thread overview]
Message-ID: <20101118221554.GA31253@lenovo> (raw)
In-Reply-To: <20101118215650.GL6028@lenovo>
On Fri, Nov 19, 2010 at 12:56:50AM +0300, Cyrill Gorcunov wrote:
...
> ---
> arch/x86/kernel/cpu/perf_event_p4.c | 21 ++++++++++++---------
> 1 file changed, 12 insertions(+), 9 deletions(-)
>
> Index: linux-2.6.git/arch/x86/kernel/cpu/perf_event_p4.c
> =====================================================================
> --- linux-2.6.git.orig/arch/x86/kernel/cpu/perf_event_p4.c
> +++ linux-2.6.git/arch/x86/kernel/cpu/perf_event_p4.c
> @@ -753,19 +753,22 @@ out:
>
> static inline int p4_pmu_clear_cccr_ovf(struct hw_perf_event *hwc)
> {
> - int overflow = 0;
> - u32 low, high;
> + u32 overflow = 0;
> + u32 low, low_cccr, high;
>
> - rdmsr(hwc->config_base + hwc->idx, low, high);
> + /* an official way for overflow indication */
> + rdmsr(hwc->config_base + hwc->idx, low_cccr, high);
> + overflow |= (low_cccr & P4_CCCR_OVF);
> +
> + /* unflagged overflows */
> + rdmsr(hwc->event_base + hwc->idx, low, high);
> + overflow |= high & 0x80000000;
>
> - /* we need to check high bit for unflagged overflows */
> - if ((low & P4_CCCR_OVF) || !(high & (1 << 31))) {
> - overflow = 1;
> + if (overflow)
this should be rather 'if (low_cccr & P4_CCCR_OVF)' otherwise
redundant checking_wrmsrl called, updated patch below
/me seems should not touch code at all today
---
perf, x86: P4 PMU - Fix unflagged overflows handling
Jason pointed out that kgdb no longer works with new
nmi-watchdog. Don found the reason -- P4 PMU reads CCCR
register instead of counter itself, it forces NMIs to
be eaten by perf subsystem.
Fix it by reading a proper register.
v2: Call checking_wrmsrl only if needed
Reported-by: Jason Wessel <jason.wessel@windriver.com>
Reported-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
---
Jason I've removed your Tested-by because the patch differ a bit
arch/x86/kernel/cpu/perf_event_p4.c | 20 ++++++++++++--------
1 file changed, 12 insertions(+), 8 deletions(-)
Index: linux-2.6.git/arch/x86/kernel/cpu/perf_event_p4.c
=====================================================================
--- linux-2.6.git.orig/arch/x86/kernel/cpu/perf_event_p4.c
+++ linux-2.6.git/arch/x86/kernel/cpu/perf_event_p4.c
@@ -753,19 +753,23 @@ out:
static inline int p4_pmu_clear_cccr_ovf(struct hw_perf_event *hwc)
{
- int overflow = 0;
- u32 low, high;
+ u32 overflow = 0;
+ u32 low_cntr, low_cccr, high;
- rdmsr(hwc->config_base + hwc->idx, low, high);
+ /* an official way for overflow indication */
+ rdmsr(hwc->config_base + hwc->idx, low_cccr, high);
+ overflow |= (low_cccr & P4_CCCR_OVF);
- /* we need to check high bit for unflagged overflows */
- if ((low & P4_CCCR_OVF) || !(high & (1 << 31))) {
- overflow = 1;
+ if (overflow) {
(void)checking_wrmsrl(hwc->config_base + hwc->idx,
- ((u64)low) & ~P4_CCCR_OVF);
+ ((u64)low_cccr) & ~P4_CCCR_OVF);
}
- return overflow;
+ /* unflagged overflows */
+ rdmsr(hwc->event_base + hwc->idx, low_cntr, high);
+ overflow |= high & 0x80000000;
+
+ return overflow > 0;
}
static void p4_pmu_disable_pebs(void)
next prev parent reply other threads:[~2010-11-18 22:16 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-12 14:43 [V2 PATCH 0/6] x86, NMI: give NMI handler a face-lift Don Zickus
2010-11-12 14:43 ` [PATCH 1/6] x86, NMI: Add NMI symbol constants and rename memory parity to PCI SERR Don Zickus
2010-11-12 14:43 ` [PATCH 2/6] x86, NMI: Add touch_nmi_watchdog to io_check_error delay Don Zickus
2010-11-12 14:43 ` [PATCH 3/6] x86, NMI: Rewrite NMI handler Don Zickus
2010-11-12 14:43 ` [PATCH 4/6] x86, NMI: Remove DIE_NMI_IPI and add priorties to handlers Don Zickus
2010-11-12 14:43 ` [PATCH 5/6] x86, NMI: Allow NMI reason io port (0x61) to be processed on any CPU Don Zickus
2010-11-12 14:43 ` [PATCH 6/6] x86, NMI: Remove do_nmi_callback logic Don Zickus
2010-11-12 15:05 ` [V2 PATCH 0/6] x86, NMI: give NMI handler a face-lift Jason Wessel
2010-11-12 15:42 ` Don Zickus
2010-11-12 15:55 ` Jason Wessel
2010-11-12 16:11 ` Don Zickus
2010-11-12 16:34 ` Jason Wessel
2010-11-12 17:27 ` Don Zickus
2010-11-16 18:43 ` Don Zickus
2010-11-16 20:04 ` Jason Wessel
2010-11-18 8:05 ` Ingo Molnar
2010-11-18 12:47 ` Jason Wessel
2010-11-18 13:17 ` Peter Zijlstra
2010-11-18 14:32 ` Don Zickus
2010-11-18 15:18 ` Jason Wessel
2010-11-18 15:38 ` Peter Zijlstra
2010-11-18 19:32 ` Don Zickus
2010-11-18 19:51 ` Jason Wessel
2010-11-18 20:04 ` Peter Zijlstra
2010-11-18 20:08 ` Don Zickus
2010-11-18 20:11 ` Cyrill Gorcunov
2010-11-18 20:52 ` Don Zickus
2010-11-18 21:01 ` Cyrill Gorcunov
2010-11-18 21:16 ` Don Zickus
2010-11-18 21:26 ` Cyrill Gorcunov
2010-11-18 20:28 ` Cyrill Gorcunov
2010-11-18 20:39 ` Cyrill Gorcunov
2010-11-18 21:02 ` Don Zickus
2010-11-18 21:19 ` Cyrill Gorcunov
2010-11-18 20:30 ` Peter Zijlstra
2010-11-19 16:59 ` Don Zickus
2010-11-19 18:25 ` Peter Zijlstra
2010-11-19 22:59 ` Don Zickus
2010-11-19 23:09 ` Peter Zijlstra
2010-11-19 23:30 ` Jason Wessel
2010-11-22 14:22 ` Don Zickus
2010-11-22 14:22 ` Don Zickus
2010-11-22 14:29 ` Peter Zijlstra
2010-11-18 20:04 ` Cyrill Gorcunov
2010-11-18 21:56 ` Cyrill Gorcunov
2010-11-18 21:58 ` Cyrill Gorcunov
2010-11-18 22:15 ` Cyrill Gorcunov [this message]
2010-11-18 22:24 ` Jason Wessel
2010-11-18 22:27 ` Cyrill Gorcunov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20101118221554.GA31253@lenovo \
--to=gorcunov@gmail.com \
--cc=andi@firstfloor.org \
--cc=dzickus@redhat.com \
--cc=fweisbec@gmail.com \
--cc=jason.wessel@windriver.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=peterz@infradead.org \
--cc=robert.richter@amd.com \
--cc=ying.huang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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.