has the i915 "black screen" boot issue returned?

Chris Wilson chris at chris-wilson.co.uk
Thu Jan 27 02:15:01 PST 2011

On Wed, 26 Jan 2011 06:37:20 -0500 (EST), "Robert P. J. Day" <rpjday at crashcourse.ca> wrote:
>   after suffering through the i915 "black screen" issue for a while,
> it was a relief that it finally seemed to be resolved with 2.6.37 or
> something in tha vicinity.  but i just built and booted a 2.6.38-rc2+
> kernel on my ubuntu 10.10 system and it once again boots to a black
> screen.
>   is this a new known issue?  has anyone else seen this?

It all depends. There is a new ACPI backlight issue causing grief, aside
from that and lingering issues with the legacy backlight combination mode,
I don't think there was a new i915 KMS regression. (We had regressions
elsewhere and finally fixed up a few old regressions for UMS and DRI1 and
various other OOPS-on-boot. And suspend is a bit hit and miss.)

For extra fun, it appears that sched_autocgroup can trigger a leak of
pinned bo eventually hitting a BUG. That's one I'm keen to reproduce.

Without knowing your chipset and seeing a dmesg, I can only guess which
problem you are most likely to have hit.

Chris Wilson, Intel Open Source Technology Centre

More information about the dri-devel mailing list