From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752997AbYE2GPD (ORCPT ); Thu, 29 May 2008 02:15:03 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751165AbYE2GOy (ORCPT ); Thu, 29 May 2008 02:14:54 -0400 Received: from smtp1.linux-foundation.org ([140.211.169.13]:37554 "EHLO smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751154AbYE2GOx (ORCPT ); Thu, 29 May 2008 02:14:53 -0400 Date: Wed, 28 May 2008 23:13:51 -0700 From: Andrew Morton To: Jens Axboe Cc: linux-kernel@vger.kernel.org Subject: Re: block: make blktrace use per-cpu buffers for message notes Message-Id: <20080528231351.a35001bc.akpm@linux-foundation.org> In-Reply-To: <200805281559.m4SFx7b9022392@hera.kernel.org> References: <200805281559.m4SFx7b9022392@hera.kernel.org> X-Mailer: Sylpheed 2.4.8 (GTK+ 2.12.5; x86_64-redhat-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 28 May 2008 15:59:07 GMT Linux Kernel Mailing List wrote: > Gitweb: http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=64565911cdb57c2f512a9715b985b5617402cc67 > Commit: 64565911cdb57c2f512a9715b985b5617402cc67 > Parent: 4722dc52a891ab6cb2d637ddb87233e0ce277827 > Author: Jens Axboe > AuthorDate: Wed May 28 14:45:33 2008 +0200 > Committer: Jens Axboe > CommitDate: Wed May 28 14:49:27 2008 +0200 please try to avoid merging unreviewed changes. > block: make blktrace use per-cpu buffers for message notes > > Currently it uses a single static char array, but that risks > being corrupted when multiple users issue message notes at the > same time. Make the buffers dynamically allocated when the trace > is setup and make them per-cpu instead. > > The default max message size of 1k is also very large, the > interface is mainly for small text notes. So shrink it to 128 bytes. > > Signed-off-by: Jens Axboe > --- > block/blktrace.c | 15 ++++++++++++--- > include/linux/blktrace_api.h | 3 ++- > 2 files changed, 14 insertions(+), 4 deletions(-) > > diff --git a/block/blktrace.c b/block/blktrace.c > index 20e11f3..7ae87cc 100644 > --- a/block/blktrace.c > +++ b/block/blktrace.c > @@ -79,13 +79,16 @@ void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) > { > int n; > va_list args; > - static char bt_msg_buf[BLK_TN_MAX_MSG]; > + char *buf; > > + preempt_disable(); > + buf = per_cpu_ptr(bt->msg_data, smp_processor_id()); get_cpu()/put_cpu() is the standard way of doing this. > @@ -172,7 +173,7 @@ extern void __trace_note_message(struct blk_trace *, const char *fmt, ...); > if (unlikely(bt)) \ > __trace_note_message(bt, fmt, ##__VA_ARGS__); \ > } while (0) > -#define BLK_TN_MAX_MSG 1024 > +#define BLK_TN_MAX_MSG 128 It seems a bit strange to do this right when we've taken this _off_ the stack. But I suppose nothing will break. I cannot work out why 9d5f09a424a67ddb959829894efb4c71cbf6d600 ("Added in MESSAGE notes for blktraces") was -rc material.