From: "Émeric Maschino" <emeric.maschino@gmail.com>
To: dri-devel@lists.sourceforge.net
Subject: Re: 3D OpenGL applications eat CPU ressources
Date: Thu, 4 Feb 2010 15:37:58 +0100 [thread overview]
Message-ID: <d72eeaf61002040637pa328d65j57f91638fe7ea079@mail.gmail.com> (raw)
In-Reply-To: <6a89f9d51002030253rd39eb7fu8ce850bba0ac2f69@mail.gmail.com>
2010/2/3 Stephane Marchesin <stephane.marchesin@gmail.com>:
> No, you are right they don't trigger MCA. Hmm I didn't have any of
> those back then, my lockups came from the bus mostly...
Thank you for clarifying this point.
> Really if you have such lockups they may also happen on x86, did you
> try the card there?
Yes, I have no problem with this (AGP Pro 4x) graphics adapter (ATI
FireGL X1) in x86 hardware.
> At this point your best bet is probably replay the crashing sequence
> until you can reduce it to the offending couple of commands.
OK. Are the commands you're talking about the argument passed in the
various ioctl() calls logged when stracing the offending OpenGL
application? For example, strace glxgears gives lines like:
ioctl(4, 0xc0106451, 0x60000fffffd52d30) = 0
ioctl(4, 0xc0186419, 0x60000fffffd52d30) = 0
ioctl(4, 0x40106459, 0x60000fffffd52d58) = 0
where 4 is the file descriptor of /dev/dri/card0. Are 0xc0106451,
0xc0186419 or 0x40106459 the commands passed to the GPU?
I don't know if it's related to GPU lockup or not (I mean, being the
cause or a consequence), but I've also noticed in the strace glxgears
logs (or even simple application like glxinfo) that most of the read()
calls to /tmp/.X11-unix/X0 fail, whereas the writev() calls seem to
succeed:
poll([{fd=3, events=POLLIN|POLLOUT}], 1, -1) = 1 ([{fd=3, revents=POLLOUT}])
writev(3, [{"\222\0\3\0\4\0\0\0\0\0\0\0", 12}, {NULL, 0}, {"", 0}], 3) = 12
poll([{fd=3, events=POLLIN}], 1, -1) = 1 ([{fd=3, revents=POLLIN}])
read(3, "\1\0*\0\0\0\0\0\4\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0",
4096) = 32
read(3, 0x60000000000093e4, 4096) = -1 EAGAIN (Resource
temporarily unavailable)
where 3 is the file descriptor of /tmp/.X11-unix/X0
Émeric
------------------------------------------------------------------------------
The Planet: dedicated and managed hosting, cloud storage, colocation
Stay online with enterprise data centers and the best network in the business
Choose flexible plans and management services without long-term contracts
Personal 24x7 support from experience hosting pros just a phone call away.
http://p.sf.net/sfu/theplanet-com
--
next prev parent reply other threads:[~2010-02-04 14:37 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-01-31 13:28 3D OpenGL applications eat CPU ressources Émeric Maschino
2010-01-31 16:42 ` Jerome Glisse
2010-02-01 21:17 ` Émeric Maschino
2010-02-01 22:00 ` Stephane Marchesin
2010-02-02 21:19 ` Émeric Maschino
2010-02-03 10:53 ` Stephane Marchesin
2010-02-04 14:37 ` Émeric Maschino [this message]
2010-02-04 15:35 ` Jerome Glisse
2010-02-06 19:47 ` Émeric Maschino
2010-02-06 20:04 ` Alex Deucher
2010-02-06 23:42 ` Dave Airlie
2010-02-07 18:17 ` Émeric Maschino
2010-02-07 0:45 ` Stephane Marchesin
2010-02-07 17:18 ` Émeric Maschino
2010-02-07 18:44 ` Émeric Maschino
2010-02-08 6:15 ` Alex Deucher
2010-02-08 20:24 ` Émeric Maschino
2010-02-11 20:25 ` Émeric Maschino
2010-02-16 22:37 ` Émeric Maschino
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=d72eeaf61002040637pa328d65j57f91638fe7ea079@mail.gmail.com \
--to=emeric.maschino@gmail.com \
--cc=dri-devel@lists.sourceforge.net \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.