[Bug 106642] X server crashes in i965 on desktop startup when DRI3 v1.2 / modifier support is enabled

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=106642

--- Comment #19 from Eero Tamminen <eero.t.tamminen at intel.com> ---
(In reply to Lionel Landwerlin from comment #14)
> Created attachment 140055 [details] [review]
> dri: add missing 16bits formats mapping
> 
> Here is a v2 for the new dri_format (G16R16 this time).
> I've gone through all the dri image formats, so hopefully we're not missing
> anything anymore.

A quick try of this patch alone -> still get the same crash?

---------------------------------------------------
Thread 1 "X" received signal SIGSEGV, Segmentation fault.
0x00007f4666cd5c2a in isl_format_supports_ccs_e
(devinfo=devinfo at entry=0x1aa72e8, 
    format=ISL_FORMAT_UNSUPPORTED) at isl/isl_format.c:558
558     isl/isl_format.c: No such file or directory.
(gdb) bt
#0  0x00007f4666cd5c2a in isl_format_supports_ccs_e
(devinfo=devinfo at entry=0x1aa72e8, 
    format=ISL_FORMAT_UNSUPPORTED) at isl/isl_format.c:558
#1  0x00007f4666ba1fac in modifier_is_supported
(devinfo=devinfo at entry=0x1aa72e8, 
    fmt=fmt at entry=0x7f4666ed6f98 <intel_image_formats+216>, dri_format=4112,
dri_format at entry=0, 
    modifier=modifier at entry=72057594037927940) at intel_screen.c:345
#2  0x00007f4666ba212a in intel_query_dma_buf_modifiers (_screen=<optimized
out>, 
    fourcc=<optimized out>, max=0, modifiers=0x0, external_only=0x0,
count=count at entry=0x7fff9c783a64)
    at intel_screen.c:1304
#3  0x00007f4664e05050 in dri2_query_dma_buf_modifiers (drv=<optimized out>,
disp=<optimized out>, 
    format=<optimized out>, max=<optimized out>, modifiers=<optimized out>, 
    external_only=<optimized out>, count=0x7fff9c783a64) at
drivers/dri2/egl_dri2.c:2394
#4  0x00007f4664df9f67 in eglQueryDmaBufModifiersEXT (dpy=0x1bc90a0,
format=808665688, max_modifiers=0, 
    modifiers=0x0, external_only=0x0, num_modifiers=0x7fff9c783a64) at
main/eglapi.c:2519
#5  0x00007f4667c21396 in glamor_get_modifiers (screen=<optimized out>,
format=808665688, 
    num_modifiers=0x7fff9c783b04, modifiers=0x7fff9c783b10) at
../../../glamor/glamor_egl.c:647
#6  0x0000000000564416 in cache_formats_and_modifiers (screen=<optimized out>)
at dri3_screen.c:196
#7  dri3_get_supported_modifiers (screen=<optimized out>, drawable=0x1da5f80,
depth=<optimized out>, 
    bpp=<optimized out>,
num_intersect_modifiers=num_intersect_modifiers at entry=0x7fff9c783b70, 
    intersect_modifiers=intersect_modifiers at entry=0x7fff9c783b80,
num_screen_modifiers=0x7fff9c783b74, 
    screen_modifiers=0x7fff9c783b88) at dri3_screen.c:234
#8  0x0000000000563700 in proc_dri3_get_supported_modifiers (client=0x203e750)
at dri3_request.c:373
#9  0x000000000043e6cb in Dispatch () at dispatch.c:478
---------------------------------------------------


Btw. With Weston/Xwayland this is even easier to trigger.  It's enough to run
any X program (like xrandr or xhost), something that doesn't even use Mesa
itself, and both Xwayland and Weston come crashing down.

I was unable to get backtrace out of Xwayland (with gdb server, Gdb wasn't able
to get sync for some reason), but the same one-liner fix in comment 1 gets rid
of that crash too.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20180606/4b27cf73/attachment.html>