Submitter : coreym Status : open
Assigned Engineer : nb Priority : 3
*Modified Date : 01/17/01 *Modified User : raa
*Modified User Domain : sgi.com *Description :
I ran rwtest on a xfs-linux filesystem on the machine permit:
rwtest 100000000:/mnt1/file_1 >/tmp/file_1.out 2>&1 &
This caused df, ls -l, fdisk, and top all to hang.
Permit is running Redhat 6.2 with a 2.4.0-test5 kernel
==========================
ADDITIONAL INFORMATION (ADD)
.....
==========================
ADDITIONAL INFORMATION (ADD)
From: "bob albers jr." <raa@xxxxxxx>
Date: Jan 17 2001 07:10:04PM
[pvnews version: 1.71]
==========================
>From axboe@xxxxxxxxxxxxxxx Wed Jan 17 09:16:29 2001
Date: Wed, 17 Jan 2001 16:16:10 +0100
From: Jens Axboe <axboe@xxxxxxx>
To: Rajagopal Ananthanarayanan <ananth@xxxxxxx>
Cc: srn@xxxxxxx, chait@xxxxxxx, raa@xxxxxxx
Subject: Re: Elevator stall fix
On Wed, Jan 17 2001, Rajagopal Ananthanarayanan wrote:
>
> Hi Jens,
>
> Could you please send us a pointer to your latest
> patch that fixes a problem where some requests were always
> by-passing other requests in the elevator queue? Since
> ELEVATOR_LINUS had tuneables set so high, the by-pass
> was percieved as a virtual stall.
The patch is blk-13C, before integration in 2.4.0-acX it
was renamed to blk-14 because I added some low mem hysterisis.
*.kernel.org/pub/linux/kernel/people/axboe/patches/2.4.0-ac4/blk-13C.bz2
is the latest non-integrated patch.
> This patch is the one that Linus specifically said is
> being considered for inclusion in 2.4.1. The problem
> was reported by someone from Japan a couple of months
> back and even LWN (lwn.com) ran an article about it
> at that time.
Yes, Linus integrated blk-13C in 2.4.1-pre5, and pre6 went to blk-14.
So the patch is in the main tree now.
--
* Jens Axboe <axboe@xxxxxxx>
* SuSE Labs
|