display watermark is too large(92/63), disabling

[Colin Law]

On 22 February 2015 at 15:14, Colin Law <clanlaw at gmail.com> wrote:
> On 22 February 2015 at 14:56, Chris <cpollock at embarqmail.com> wrote:
>> ...
>> Colin, just a note, I've switched back to the 3.13.0-45 generic kernel
>> and the 'watermark....' notations in my syslog have disappeared. This
>> tells me that something in the 3.19.* kernel is causing this. I expect
>> the 'hangcheck' lockup issue to reappear at any time now. Probably
>> sometime later today or tomorrow. This whole thing is becoming annoying
>> especially when I keep adding to the two bug reports and no advice is
>> forthcoming.
>
> If the hangcheck problem does not appear with with the later kernel
> then it seems that particular problem has been fixed in that kernel.
> Time to move on and look at the other lockup problem perhaps.  Since
> the original bug is particularly talking about hangcheck then this may
> be a different problem entirely.
>
> The messages you pasted look as if they are from the graphics
> initialisation, were the messages just before the hang?  Were you
> doing anything with the displays at that time (Feb 20 19:17:29)?

Actually I think I am wrong, I think they may be debug output
following the detection of a problem.  If you google for
intel_dump_pipe_config you will find quite a lot of recent discussion
about the problem, though I have not tracked down an actual bug.  Are
you also seeing a message about vblank timeout that is being discussed
along with the dump stuff?  It looks like it is currently being
addressed.  I don't think your problems are being ignored , I think
those looking at it (intel guys I think) have enough data of their own
to be looking at it.  You would not expect much action on the ubuntu
bug as it is nothing that the ubuntu team can do anything about, it is
the intel driver (I think).

Colin

>
> Colin