qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Jan Kiszka <jan.kiszka@web.de>
To: Stefan Hajnoczi <stefanha@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>, qemu-devel <qemu-devel@nongnu.org>
Subject: [Qemu-devel] Re: Guest latency issues due to bdrv_check_byte_request
Date: Sun, 18 Apr 2010 20:05:21 +0200	[thread overview]
Message-ID: <4BCB49E1.8030106@web.de> (raw)
In-Reply-To: <y2hfbd9d3991004171432zdbe02ee9id669a6f2551000b8@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 1352 bytes --]

Stefan Hajnoczi wrote:
> Thanks Christoph.
> 
> Cached getlength with pread/pwrite:
> % time     seconds  usecs/call     calls    errors syscall
> ------ ----------- ----------- --------- --------- ----------------
>  96.97    1.760111       11893       148         4 futex
>   1.61    0.029209           1     46891      2217 select
>   0.28    0.005047           0     64609           timer_gettime
>   0.22    0.004059           0     42745      2578 rt_sigreturn
>   0.22    0.003911           0     46261           timer_settime
>   0.18    0.003280        1093         3           shmdt
>   0.17    0.003095           0     23859           pread  <---
>   0.17    0.003061           0     42800           write
>   0.16    0.002916           0     47759      5151 read
>   0.02    0.000285           0       645           writev
> [...]
>   0.00    0.000000           0        13           lseek
> 
> Note that this is a Tiny Core Linux boot from disk and shutdown; not
> very I/O intensive since it only loads a kernel and ~10 MB initramfs
> without touching the disk much after kernel load.

Nice. Will give this a try tomorrow with "a bit" more load.

We already played with a hack to completely remove the checks from AIO
requests, thus avoiding lseek this way - effect as desired, but fragile
of course.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 257 bytes --]

      parent reply	other threads:[~2010-04-18 18:05 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-15 18:15 [Qemu-devel] Guest latency issues due to bdrv_check_byte_request Jan Kiszka
2010-04-17 19:05 ` Stefan Hajnoczi
2010-04-17 19:40   ` Christoph Hellwig
2010-04-17 21:32     ` Stefan Hajnoczi
2010-04-18 17:37       ` Christoph Hellwig
2010-04-18 18:05       ` Jan Kiszka [this message]

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=4BCB49E1.8030106@web.de \
    --to=jan.kiszka@web.de \
    --cc=hch@lst.de \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@gmail.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).