mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Ingo Molnar <mingo@elte.hu>, Paul Mackerras <paulus@samba.org>,
	Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	LKML <linux-kernel@vger.kernel.org>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>
Subject: [PATCH 06/10] perf: Fix wakeup storm for RO mmap()s
Date: Fri, 21 May 2010 11:02:07 +0200	[thread overview]
Message-ID: <20100521090710.684479310@chello.nl> (raw)
In-Reply-To: <20100521090201.326791353@chello.nl>

[-- Attachment #1: perf-fix-RO-wakeups.patch --]
[-- Type: text/plain, Size: 1481 bytes --]

RO mmap()s don't update the tail pointer, so comparing against
it for determining the written data size doesn't really do any good.

Keep track of when we last did a wakeup, and compare against that.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
 include/linux/perf_event.h |    2 +-
 kernel/perf_event.c        |    4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

Index: linux-2.6/include/linux/perf_event.h
===================================================================
--- linux-2.6.orig/include/linux/perf_event.h
+++ linux-2.6/include/linux/perf_event.h
@@ -599,7 +599,7 @@ struct perf_mmap_data {
 	local_t				head;		/* write position    */
 	local_t				nest;		/* nested writers    */
 	local_t				events;		/* event limit       */
-	local_t				wakeup;		/* needs a wakeup    */
+	local_t				wakeup;		/* wakeup stamp      */
 	local_t				lost;		/* nr records lost   */
 
 	long				watermark;	/* wakeup watermark  */
Index: linux-2.6/kernel/perf_event.c
===================================================================
--- linux-2.6.orig/kernel/perf_event.c
+++ linux-2.6/kernel/perf_event.c
@@ -3055,8 +3055,8 @@ int perf_output_begin(struct perf_output
 	handle->offset	= offset;
 	handle->head	= head;
 
-	if (head - tail > data->watermark)
-		local_inc(&data->wakeup);
+	if (head - local_read(&data->wakeup) > data->watermark)
+		local_add(data->watermark, &data->wakeup);
 
 	if (have_lost) {
 		lost_event.header.type = PERF_RECORD_LOST;



  parent reply	other threads:[~2010-05-21  9:12 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-05-21  9:02 [PATCH 00/10] perf tracepoint and output optimizations Peter Zijlstra
2010-05-21  9:02 ` [PATCH 01/10] perf, trace: Remove IRQ-disable from perf/tracepoint interaction Peter Zijlstra
2010-05-21 17:43   ` Frank Ch. Eigler
2010-05-21 17:53     ` Steven Rostedt
2010-05-21 18:07       ` Frank Ch. Eigler
2010-05-23 12:11   ` Paul Mackerras
2010-05-23 18:16     ` Peter Zijlstra
2010-05-24  4:29       ` Paul Mackerras
2010-05-25  8:06         ` [tip:perf/core] perf, trace: Fix IRQ-disable removal " tip-bot for Peter Zijlstra
2010-05-25  9:02           ` Peter Zijlstra
2010-05-25  9:30             ` [tip:perf/core] perf, trace: Fix !x86 build bug tip-bot for Peter Zijlstra
2010-05-24 11:31       ` [PATCH 01/10] perf, trace: Remove IRQ-disable from perf/tracepoint interaction Frederic Weisbecker
2010-05-25  7:30   ` [PATCH 01a/10] perf, trace: Fix !x86 build issue Peter Zijlstra
2010-05-21  9:02 ` [PATCH 02/10] perf, trace: Use per-tracepoint-per-cpu hlist to track events Peter Zijlstra
2010-05-21  9:40   ` Frederic Weisbecker
2010-05-21 10:02     ` Peter Zijlstra
2010-05-21 10:13       ` Frederic Weisbecker
2010-05-21 10:15         ` Peter Zijlstra
2010-05-21 10:19           ` Frederic Weisbecker
2010-05-21 10:38           ` Ingo Molnar
2010-05-21 10:51             ` Ingo Molnar
2010-05-21 10:19         ` Peter Zijlstra
2010-05-21 10:21           ` Frederic Weisbecker
2010-05-21 10:34             ` Peter Zijlstra
2010-05-21 10:38               ` Frederic Weisbecker
2010-05-21 10:41   ` [PATCH 02b/10] perf, trace: Fix probe unregister race Peter Zijlstra
2010-05-21 10:43     ` Frederic Weisbecker
2010-05-31  7:19     ` [tip:perf/urgent] perf_events, " tip-bot for Peter Zijlstra
2010-05-21 11:28   ` [tip:perf/core] perf, trace: Optimize tracepoints by using per-tracepoint-per-cpu hlist to track events tip-bot for Peter Zijlstra
2010-05-21 14:04   ` [PATCH 02/10] perf, trace: Use " Steven Rostedt
2010-05-21 14:18     ` Peter Zijlstra
2010-05-21 14:25       ` Peter Zijlstra
2010-05-31  7:20         ` [tip:perf/urgent] perf_events, trace: Fix perf_trace_destroy(), mutex went missing tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 03/10] perf: Ensure IOC_OUTPUT isnt used to create multi-writer buffers Peter Zijlstra
2010-05-21 11:28   ` [tip:perf/core] perf: Ensure that IOC_OUTPUT isn't " tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 04/10] perf-record: Remove -M Peter Zijlstra
2010-05-21 11:28   ` [tip:perf/core] " tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 05/10] perf-record: Share per-cpu buffers Peter Zijlstra
2010-05-21  9:44   ` Frederic Weisbecker
2010-05-21 10:03     ` Peter Zijlstra
2010-05-21 11:29   ` [tip:perf/core] " tip-bot for Peter Zijlstra
2010-05-21  9:02 ` Peter Zijlstra [this message]
2010-05-21 11:29   ` [tip:perf/core] perf: Fix wakeup storm for RO mmap()s tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 07/10] perf: Optimize perf_output_copy Peter Zijlstra
2010-05-21 11:29   ` [tip:perf/core] perf: Optimize perf_output_copy() tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 08/10] perf: Optimize the !vmalloc backed buffer Peter Zijlstra
2010-05-21 11:29   ` [tip:perf/core] " tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 09/10] perf: Remove more fastpath code Peter Zijlstra
2010-05-21 11:15   ` Steven Rostedt
2010-05-21 11:18     ` Peter Zijlstra
2010-05-21 11:30   ` [tip:perf/core] perf: Remove more code from the fastpath tip-bot for Peter Zijlstra
2010-05-21  9:02 ` [PATCH 10/10] perf: Optimize perf_tp_event_match Peter Zijlstra
2010-05-21 11:30   ` [tip:perf/core] perf: Optimize perf_tp_event_match() tip-bot for Peter Zijlstra

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=20100521090710.684479310@chello.nl \
    --to=a.p.zijlstra@chello.nl \
    --cc=acme@infradead.org \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulus@samba.org \
    --cc=rostedt@goodmis.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome