[Pixman] [BUG] SIGSEGV in sse2_fill

[Adam Jackson]

continued from pixman@, original thread here:

https://lists.freedesktop.org/archives/pixman/2018-August/004759.html

On Wed, 2018-08-29 at 18:14 +0200, Frédéric Fauberteau wrote:
> Le 2018-08-29 16:33, Adam Jackson a écrit :
> > This is a curious backtrace though. You're crashing while trying to
> > draw the black solid fill for the initial map of the root window. Fine,
> > but you're doing so in software, even though you have glamor enabled,
> > and glamor surely can usually accelerate solid fills. So you're hitting
> > a software fallback for some reason, and if I had to guess...
> 
> The area to fill is 2960x1050 but actually, I have two screens:
> ----------------- -------------
> >               | |           |
> >               | |           |
> >   1680x1050   | | 1280x1024 |
> >               | |           |
> >               | |-----------|
> 
> -----------------
> 
> Do you think it could be a reason to write in an unmapped region...?

Probably not? Typically the allocation for the root window is a
rectangle big enough to bound all the outputs within it, so actually
2960x1050 in your case. The missing corner below the 1280x1024 display
still has memory behind it.

> It is a RS780/RS880 (Radeon HD 3200). If the bug comes from the Mesa 
> driver, it's a big issue since we are totally late with the update (we 
> are on MesaLib 11.2.2)

That should be new enough to not have the problem I was describing.
Unfortunately, that means I'm out of ideas. I think there are probably
at least two bugs here: the fallback shouldn't crash, but neither
should the fallback occur. Probably the second one is a bit easier to
figure out. I'd start by rebuilding xserver with some printf-debugging
in the body of glamor_poly_fill_rect_gl(), to see which 'goto bail'
path is getting hit. (xserver spells printf ErrorF, which will print
into both the X log and stdout.)

- ajax