From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Stephane Eranian <eranian@google.com>, Ingo Molnar <mingo@elte.hu>
Cc: linux-kernel@vger.kernel.org, Peter Zijlstra <a.p.zijlstra@chello.nl>
Subject: [PATCH 7/7] perf, x86: Use NUMA aware allocations for PEBS/BTS/DS allocations
Date: Tue, 19 Oct 2010 15:43:03 +0200 [thread overview]
Message-ID: <20101019134808.514465326@chello.nl> (raw)
In-Reply-To: 20101019134256.087045503@chello.nl
[-- Attachment #1: perf-perf-event-improve-ds-7.patch --]
[-- Type: text/plain, Size: 1867 bytes --]
For performance reasons its best to use NODE local memory for per-cpu buffers.
This logic comes from a much larger patch proposed by Stephane.
Suggested-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
---
arch/x86/kernel/cpu/perf_event_intel_ds.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
Index: linux-2.6/arch/x86/kernel/cpu/perf_event_intel_ds.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ linux-2.6/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -77,13 +77,14 @@ static void fini_debug_store_on_cpu(int
static int alloc_pebs_buffer(int cpu)
{
struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
+ int node = cpu_to_node(cpu);
int max, thresh = 1; /* always use a single PEBS record */
void *buffer;
if (!x86_pmu.pebs)
return 0;
- buffer = kzalloc(PEBS_BUFFER_SIZE, GFP_KERNEL);
+ buffer = kmalloc_node(PEBS_BUFFER_SIZE, GFP_KERNEL | __GFP_ZERO, node);
if (unlikely(!buffer))
return -ENOMEM;
@@ -114,13 +115,14 @@ static void release_pebs_buffer(int cpu)
static int alloc_bts_buffer(int cpu)
{
struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds;
+ int node = cpu_to_node(cpu);
int max, thresh;
void *buffer;
if (!x86_pmu.bts)
return 0;
- buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL);
+ buffer = kmalloc_node(BTS_BUFFER_SIZE, GFP_KERNEL | __GFP_ZERO, node);
if (unlikely(!buffer))
return -ENOMEM;
@@ -150,9 +152,10 @@ static void release_bts_buffer(int cpu)
static int alloc_ds_buffer(int cpu)
{
+ int node = cpu_to_node(cpu);
struct debug_store *ds;
- ds = kzalloc(sizeof(*ds), GFP_KERNEL);
+ ds = kmalloc_node(sizeof(*ds), GFP_KERNEL | __GFP_ZERO, node);
if (unlikely(!ds))
return -ENOMEM;
next prev parent reply other threads:[~2010-10-19 13:55 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-19 13:42 [PATCH 0/7] perf,x86: PEBS/BTS buffer allocation changes Peter Zijlstra
2010-10-19 13:42 ` [PATCH 1/7] perf, x86: Extract PEBS/BTS buffer free routines Peter Zijlstra
2010-10-22 13:02 ` [tip:perf/urgent] " tip-bot for Peter Zijlstra
2010-10-19 13:42 ` [PATCH 2/7] perf, x86: Extract PEBS/BTS allocation functions Peter Zijlstra
2010-10-22 13:02 ` [tip:perf/urgent] " tip-bot for Peter Zijlstra
2010-10-19 13:42 ` [PATCH 3/7] perf, x86: Extract DS alloc/free functions Peter Zijlstra
2010-10-22 13:03 ` [tip:perf/urgent] " tip-bot for Peter Zijlstra
2010-10-19 13:43 ` [PATCH 4/7] perf, x86: Fixup the precise_ip computation Peter Zijlstra
2010-10-22 13:03 ` [tip:perf/urgent] " tip-bot for Peter Zijlstra
2010-10-22 13:04 ` [tip:perf/urgent] perf, x86: Less disastrous PEBS/BTS buffer allocation failure tip-bot for Peter Zijlstra
2010-10-19 13:43 ` [PATCH 5/7] " Peter Zijlstra
2010-10-19 13:43 ` [PATCH 6/7] perf, x86: Clean up reserve_ds_buffers() signature Peter Zijlstra
2010-10-22 13:04 ` [tip:perf/urgent] " tip-bot for Peter Zijlstra
2010-10-19 13:43 ` Peter Zijlstra [this message]
2010-10-22 13:04 ` [tip:perf/urgent] perf, x86: Use NUMA aware allocations for PEBS/BTS/DS allocations tip-bot for Peter Zijlstra
2010-10-21 11:35 ` [PATCH 0/7] perf,x86: PEBS/BTS buffer allocation changes Peter Zijlstra
2010-10-21 11:45 ` Stephane Eranian
2010-10-21 14:30 ` Stephane Eranian
2010-10-21 14:34 ` Peter Zijlstra
2010-10-21 14:35 ` Stephane Eranian
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=20101019134808.514465326@chello.nl \
--to=a.p.zijlstra@chello.nl \
--cc=eranian@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
/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