<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - Segmentation Fault in sna_put_zpixmap_blt (X server crash)" href="https://bugs.freedesktop.org/show_bug.cgi?id=92751#c20">Comment # 20</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - Segmentation Fault in sna_put_zpixmap_blt (X server crash)" href="https://bugs.freedesktop.org/show_bug.cgi?id=92751">bug 92751</a> from <a class="email" href="mailto:joe@wildlava.com" title="Joe Peterson <joe@wildlava.com>"> Joe Peterson</a> <pre>(In reply to Joe Peterson from <a href="show_bug.cgi?id=92751#c19">comment #19</a>) > (In reply to Joe Peterson from <a href="show_bug.cgi?id=92751#c18">comment #18</a>) > > (In reply to Chris Wilson from <a href="show_bug.cgi?id=92751#c17">comment #17</a>) > > > I made a mistake in adding that assertion. Darn it. > > > > Ok, I'm building a branch with a cherry-pick of your latest commit, skipping > > commit 1982462 that caused the assertion; I'll let you know if that works... > > Ok, with this (omitting the kgem commit), I get this backtrace (similar to > my original reported trace): > > Backtrace: > 0: /usr/lib/xorg/modules/drivers/intel_drv.so > (__kgem_retire_requests_upto+0x71) [0x7f3876671f91] > 1: /usr/lib/xorg/modules/drivers/intel_drv.so > (sna_put_zpixmap_blt.isra.128+0x1b2) [0x7f38766c03e2] > 2: /usr/lib/xorg/modules/drivers/intel_drv.so (sna_put_image+0x2f8) > [0x7f38766c0f18] > > Perhaps this is expected, since that commit can fix this (?), now commenting > our the assert in question but including the commit otherwise... Ok, just comfirming that I still get above backtrace even with all your commits (through b1015d4, including kgem one (1982462)) minus the two asserts that kept X from starting (diff from your b1015d4 is: diff --git a/src/sna/kgem.c b/src/sna/kgem.c index 7acc69e..0e2f830 100644 --- a/src/sna/kgem.c +++ b/src/sna/kgem.c @@ -4064,13 +4064,13 @@ void _kgem_submit(struct kgem *kgem) #endif rq = kgem->next_request; - assert(rq->bo == NULL); + //assert(rq->bo == NULL); rq->bo = kgem_create_batch(kgem); if (rq->bo) { struct drm_i915_gem_execbuffer2 execbuf; - assert(rq->bo->refcnt == 1); + //assert(rq->bo->refcnt == 1); assert(!rq->bo->needs_flush); i = kgem->nexec++; So the new code could have fixed the TearFree issue (I'll let you know if I see that again), but original issue seems to still exist. I'll try to get a gdb backtrace of that and/or a debug=full log...</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the QA Contact for the bug.</li> </ul> </body> </html>