Bug 746929

Created attachment 528765 [details]
Screen of what happens at boot with NVS 300

Created attachment 528836 [details]
with drm.debug=14 as kernel option

Do we have a chance how to record this messages during boot time (booting from live).

Can you boot with "drm.debug=14 log_buf_len=1M", and try to ssh in and save the dmesg output somehow?

Also, this particular monitor is known working on other cards without EDID issues?

(In reply to comment #4)
> Can you boot with "drm.debug=14 log_buf_len=1M", and try to ssh in and save the
> dmesg output somehow?
> 

Ben I don't think this will work, since the sshd will not start, this issue precedes the start of sshd.

Do we have any other way how to get longer log?

> Also, this particular monitor is known working on other cards without EDID
> issues?

We can definitely try older monitor.

(In reply to comment #5)
> (In reply to comment #4)
> > Can you boot with "drm.debug=14 log_buf_len=1M", and try to ssh in and save the
> > dmesg output somehow?
> > 
> 
> Ben I don't think this will work, since the sshd will not start, this issue
> precedes the start of sshd.
You can boot with "drm.debug=14 log_buf_len=1M nomodeset 3", wait for sshd to start, then "modprobe -r nouveau; modprobe nouveau modeset=1".

If you still can't ssh in after this (if the system is hung hard), you can use netconsole, and start it before the modprobe nouveau step.

> 
> Do we have any other way how to get longer log?
> 
> > Also, this particular monitor is known working on other cards without EDID
> > issues?
> 
> We can definitely try older monitor.
Well, if this monitor works without errors with other cards/drivers, it should work with nouveau too :)  In any case nouveau shouldn't be taking the whole system down!

Thanks, those parameters work, for dmesg output see attachment (dmesg.out)

Created attachment 529016 [details]
dmesg output

Created attachment 529023 [details]
Probably the more useful output

Forgot to try loading nouveau before, sorry about that -- the hopefully actually useful output

Created attachment 529025 [details]
mistake, too many tabs, wrong bug

The previous output wasn't very useful (didn't even try to load nouveau drivers)

Hmm, interesting.  From the logs it doesn't look like the system should have been hung, just no proper screen output after nouveau has initialised.

Nouveau is failing to detect any displays due to, it looks like, the monitor's EDID being broken.  Are you able to test with another known-working NVIDIA GPU (and perhaps ATI or something if that fails still), and confirm the monitor works correctly there?

The next big issue I notice is that this card can somehow manage to drive DP, VGA and DVI natively all from the same connector.  This is something I haven't come across yet, and would very much like to get my hands on this hw to implement it.

The system isn't completely hung, ssh works just fine, the only thing that seems to be frozen after loading the driver is the screen.

The monitor works well (doesn't crash, displays what's required) with Quadro 2000,  Quadro FX 380 and Intel HD graphics (didn't test other cards). 

Results of having a different monitor connected (which usually works fine) were the same (didn't boot, EDID checksum is invalid, remainder is 128 RAW EDID...).

There is also a similar bug with NVS 450 ( https://bugzilla.redhat.com/show_bug.cgi?id=747034 ). Except that nvs 450 doesn't fail so early and doesn't have the connector at all. 

Both boot properly with nomodeset parameter (should have probably mentioned earlier)

(In reply to comment #12)
> The system isn't completely hung, ssh works just fine, the only thing that
> seems to be frozen after loading the driver is the screen.
Yep, nouveau thinks nothing's connected, so never does a modeset, leaving whatever was there before - which is VGA text mode in this case.

> 
> The monitor works well (doesn't crash, displays what's required) with Quadro
> 2000,  Quadro FX 380 and Intel HD graphics (didn't test other cards). 
> 
> Results of having a different monitor connected (which usually works fine) were
> the same (didn't boot, EDID checksum is invalid, remainder is 128 RAW EDID...).
Okay, it definitely looks like something's wrong driver-side in that case.  How was the monitor connected?  I believe that card has DMS-59 connectors on the back, but what does your DMS-59 adaptor provide on the other end?

Ben.

> 
> There is also a similar bug with NVS 450 (
> https://bugzilla.redhat.com/show_bug.cgi?id=747034 ). Except that nvs 450
> doesn't fail so early and doesn't have the connector at all. 
> 
> Both boot properly with nomodeset parameter (should have probably mentioned
> earlier)

You are right, the card only has a DMS-59 and the adaptor provides 2 DVI outputs

Since RHEL 6.2 External Beta has begun, and this bug remains
unresolved, it has been rejected as it is not proposed as
exception or blocker.

Red Hat invites you to ask your support representative to
propose this request, if appropriate and relevant, in the
next release of Red Hat Enterprise Linux.

*** Bug 769391 has been marked as a duplicate of this bug. ***

*** Bug 696119 has been marked as a duplicate of this bug. ***

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHSA-2012-0862.html