Re: [PATCH] Revert "drm/i915: fix vblank wait test condition"

From: Pekka Enberg
Date: Sun Sep 05 2010 - 04:08:10 EST


On 5.9.2010 1.44, Sitsofe Wheeler wrote:
On Sat, Sep 04, 2010 at 10:42:46PM +0200, Sven Joachim wrote:
On 2010-09-04 21:54 +0200, Hugh Dickins wrote:

On Sat, 4 Sep 2010, Pekka Enberg wrote:
On Sat, 4 Sep 2010 11:17:14 +0300, Pekka Enberg<penberg@xxxxxxxxxx> wrote:
This reverts commit 9559fcdbff4f93d29af04478bbc48294519424f5. The commit causes
wrong screen resolution on some i915 machines such as MacBook which is a
regression from 2.6.35. The issue has been reported multiple times but no fix
has emerged:
On Sat, Sep 4, 2010 at 12:21 PM, Chris Wilson<chris@xxxxxxxxxxxxxxxxxx> wrote:
NAK. You can't simply hide one bug by introducing a much bigger one.
Dude, it's a nasty regression from 2.6.35. What exactly are you NAK'ing here?

As you have an affected machine, find which wait_for_vblank() needs to be
replaced with a msleep(20|50).
The attached patch fixes things here. I'm not a drm developer so I
don't know if it makes technical sense.
Many thanks to you, Pekka: I also don't know if it makes technical sense,
but it solves the same issue for me, so I can at last use 2.6.36-rc on
this laptop.
Helps for me as well, but I cannot judge whether it makes technical
sense either. Thanks, Pekka!
Sadly this doesn't help my EeePC 900 with an i915 when it is booted with
a battery inside or after doing a suspend to ram (running
modetest/xrandr in these situations says VGA is connected when there
isn't even a cable attached).
I haven't heard of this problem. Is it reported on LKML? Is it a regression from 2.6.35?

Pekka
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/