[PATCH 0/2] [lucid] LP#538377 -- Radeon HDMI output turns off randomly

Bryce Harrington bryce at canonical.com
Thu Apr 8 23:25:30 UTC 2010

On Thu, Apr 08, 2010 at 03:32:35PM -0400, Chase Douglas wrote:
> On Thu, Apr 8, 2010 at 3:19 PM, Bryce Harrington <bryce at canonical.com> wrote:
> > On Thu, Apr 08, 2010 at 01:13:04PM +0100, Andy Whitcroft wrote:
> >> This seems to be triggered by PLL setup issues.  There is a somewhat
> >> larger than I would like rewrite of this whole algorithm in mainline.
> >> This seems to fix the issue for the reporters.
> >>
> >> I think we want this for Lucid, but it is rather large.  Comments?
> >
> > There's been a ton of these bug reports come over the wire, which appear
> > to need pll quirking, so despite the size of the patch it sounds like it
> > could enable addressing issues on a broad range of hardware.
> Bryce, what do you think of the proposal to leave the default as is,
> and allow people to enable it afterwards? Is that good enough, or does
> this hit so many people that we're better off enabling it now and
> backing it out if it catastrophically breaks things?

Bugs 537640, 543045, 543265, and 544662 (and their dupes) are marked as
needing the pll quirk, which makes me think this affects a lot of
people.  This seems to be an R5xx issue, and R5xx cards are pretty
widespread (recall that in karmic R5xx was the newest radeon card where
you could get 3D acceleration with the open driver).

So, I think in this case putting the patch in and watching for errors is
probably the right way to go.  I've gotten really well caught up with
-radeon bug reports currently, and plan to keep a tight watch from here
to release, so if this patch were to lead to regressions getting
reported, we should be able to spot it swiftly.


More information about the kernel-team mailing list