On Fri, Apr 10, 2020 at 7:55 AM Olivier Fourdan <ofourdan(a)redhat.com> wrote:
The backtrace I get with the current Fedora build is the same every
time, but does not involve `sna_accel_flush()` (it crashes in `kgem_buffer_release()`
down from `sna_accel_block()`)
That is because the debug build triggers the the assertion rather than
progressing the segmentation fault
Hmmm. So far so good but why???