Bug#752156: xorg: X server crashes in fbBresSolid+0x1c0 (intel driver) only with KiCad

[Angus Gratton <gus@projectgus.com>]

Hi,

I experienced this same bug regularly running Kicad on Debian Jessie
with xorg 1:7.7+7. I can supply more bug report info & stack traces if
required, but I'm also pretty sure I know the cause and a potential
fix.

The original upstream bug was this one #24274, which has the same
symptoms (segfault in fbBresSolid):
https://bugs.freedesktop.org/show_bug.cgi?id=24274

The fix for that bug caused this regression reported upstream:
https://bugs.freedesktop.org/show_bug.cgi?id=54168

That regression was also reported in Debian as Debian bug #688908
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=688908

The Debian fix for the regression was to revert the upstream patch
that fixed the original bug. In Debian this patch is
06_Revert-fb-reorder-Bresenham-error-correction-to-avoi.diff -
regression is fixed, but the segfault comes back.

Recently a commit was applied upstream (1b94fd7779) that fixes the
regression (while also keeping the original segfault fixed).

I've been running custom xserver-xorg 1:7.7+7 packages with the
"06_Revert-fb-reorder" patch replaced by the patch from 1b94fd7779
instead, without any problems.

If it's useful for the maintainers then I can attempt to send a patch
to the Debian package that makes this same change. I'm pretty
unfamiliar with Debian packaging which is why I'm not sending one now,
but there's only one way I'll ever learn... :)

Thanks everyone for their time looking into this. Please let me know
if I can supply any more info at all.


Angus