Bug 48090 - Spam of system log with EDID messages
Summary: Spam of system log with EDID messages
Status: RESOLVED DUPLICATE of bug 64858
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-03-30 05:08 UTC by mieszcz
Modified: 2013-08-31 01:49 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Part of system log (34.83 KB, text/plain)
2012-03-30 05:08 UTC, mieszcz
no flags Details
EDID binary file (128 bytes, application/octet-stream)
2012-03-30 05:09 UTC, mieszcz
no flags Details
Edid File retreived using nvidia settings from blob driver (256 bytes, text/plain)
2012-11-06 20:00 UTC, mieszcz
no flags Details
Edid file from nouveau retreived from /sys/class/drm/card0-DVI-I-1 (128 bytes, text/plain)
2012-11-06 20:06 UTC, mieszcz
no flags Details

Description mieszcz 2012-03-30 05:08:34 UTC
Created attachment 59271 [details]
Part of system log

When started kernel with kms-nouveau compiled in, the kernel logi is spammed every 10 seconds with unknown EDID messages:
Comment 1 mieszcz 2012-03-30 05:09:42 UTC
Created attachment 59272 [details]
EDID binary file

Edid data file for eDP-1 output
Comment 2 Emil Velikov 2012-11-01 19:08:42 UTC
Apart from the annoying messages are you experiencing any issues with nouveau ?

A few thing to help out meanwhile 

Try with output polling turned off
$echo 0 > /sys/module/drm_kms_helper/parameters/poll

or provide a customer EDID - see linux/Documentation/EDID/HOWTO.txt and $modinfo drm_kms_helper
Comment 3 mieszcz 2012-11-04 09:23:14 UTC
Emil thank you for the tip, it acts good as workaround.
Now I just compiled new kernel with support of nouveau.
And except of above described log spamming it works perfectly.

But I don't understand one thing. Why the driver is mentioning different EDID data in the logs, and different in 
/sys/class/drm/card0/card0-eDP-1

While status of other video outputs is disabled
Comment 4 Emil Velikov 2012-11-04 16:52:33 UTC
You're welcome, can you let me know are you loading the edid from file/firmware?

And here is a bit of background about edid(s)

The edid has header (block 0) and extensions (block 1-...), each 128 bytes in size

drm core tries to retrieve the whole edid, thus in your case is failing to fetch extension #1 (block 1).

Can you do the following
* attach the edid retrieved from both the blob and nouveau
* clearly indicate which one is which

I assume the file attached was fetched with the blob running
Comment 5 mieszcz 2012-11-06 20:00:21 UTC
Created attachment 69635 [details]
Edid File retreived using nvidia settings from blob driver
Comment 6 mieszcz 2012-11-06 20:06:07 UTC
Created attachment 69636 [details]
Edid file from nouveau retreived from /sys/class/drm/card0-DVI-I-1
Comment 7 mieszcz 2012-11-06 20:11:11 UTC
OK.

Now I understand.
Attached file was from nouveau.
Now I attached both: from blob, and from nouveau with right description.
If necessary, I could retreive and send bios from the video card, but
I need to know how to find it.

Right now I switched off polling, and the system works well. I tried to make it with external edid file, but system don't use it. 
I put file 1920x1080.bin in /lib/firmware, but without success.


Thanks 

Mirek
Comment 8 Ilia Mirkin 2013-08-31 01:49:32 UTC
This seems identical to the issue in the other bug, feel free to undup it if you think I'm wrong. Unfortunately I can't tell the linux kernel version you used from your logs... but please try out linux 3.5 or 3.6 -- it should work fine there, and also try with the latest kernel (e.g. 3.11-rc7) and report back on whether things work again now (all without setting poll=0).

*** This bug has been marked as a duplicate of bug 64858 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.