<html> <head> <base href="https://bugs.freedesktop.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - [G965] piglit.spec.!opengl 1_2.tex3d-maxsize spins forever" href="https://bugs.freedesktop.org/show_bug.cgi?id=108630#c5">Comment # 5</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - [G965] piglit.spec.!opengl 1_2.tex3d-maxsize spins forever" href="https://bugs.freedesktop.org/show_bug.cgi?id=108630">bug 108630</a> from <a class="email" href="mailto:andrey.simiklit@gmail.com" title="asimiklit <andrey.simiklit@gmail.com>"> asimiklit</a> <pre>(In reply to Mark Janes from <a href="show_bug.cgi?id=108630#c0">comment #0</a>) > bisected to: > > a9031bf9b55602d93cccef6c926e2179c23205b4 > Author: Andrii Simiklit <<a href="mailto:andrii.simiklit@globallogic.com">andrii.simiklit@globallogic.com</a>> > > i965/batch: avoid reverting batch buffer if saved state is an empty > > There's no point reverting to the last saved point if that save point is > the empty batch, we will just repeat ourselves. > > CC: Chris Wilson <<a href="mailto:chris@chris-wilson.co.uk">chris@chris-wilson.co.uk</a>> > Fixes: 3faf56ffbdeb "intel: Add an interface for saving/restoring > the batchbuffer state." > Bugzilla: <a class="bz_bug_link bz_status_NEW " title="NEW - [SNB] The graphical corruption and GPU hang occur sometimes on the piglit test "arb_texture_multisample-large-float-texture" with parameter --fp16" href="show_bug.cgi?id=107626">https://bugs.freedesktop.org/show_bug.cgi?id=107626</a> > Reviewed-by: Jordan Justen <<a href="mailto:jordan.l.justen@intel.com">jordan.l.justen@intel.com</a>> > Reviewed-by: Kenneth Graunke <<a href="mailto:kenneth@whitecape.org">kenneth@whitecape.org</a>> > > > Previously, this test failed within 3s with: > i965: Failed to submit batchbuffer: No space left on device Yes you are right. It my mistake :( Sorry for that. Thanks a lot for your work. (In reply to Jordan Justen from <a href="show_bug.cgi?id=108630#c3">comment #3</a>) > Andrii, > > I found the issue. It is not specific to g965. In the blorp code, on a > retry, the check_aperture_failed_once variable will be reset back to > false since we clear things before the retry. This small patch added > to your change fixes it. > > Another idea could be to preset check_aperture_failed_once before > the retry label. > > Can you re-submit the patch with this, or a similar change? > > diff --git a/src/mesa/drivers/dri/i965/genX_blorp_exec.c > b/src/mesa/drivers/dri/i965/genX_blorp_exec.c > index fd9ce93c6c..7b0f77af25 100644 > --- a/src/mesa/drivers/dri/i965/genX_blorp_exec.c > +++ b/src/mesa/drivers/dri/i965/genX_blorp_exec.c > @@ -309,7 +309,7 @@ retry: > intel_batchbuffer_require_space(brw, 1400); > brw_require_statebuffer_space(brw, 600); > intel_batchbuffer_save_state(brw); > - check_aperture_failed_once = intel_batchbuffer_saved_state_is_empty(brw); > + check_aperture_failed_once |= > intel_batchbuffer_saved_state_is_empty(brw); > brw->batch.no_wrap = true; > > #if GEN_GEN == 6 You are right. Thanks a lot for your fix. I will send updated patch shortly.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the QA Contact for the bug.</li> </ul> </body> </html>