Mesa CI reports a low error rate (2/700k), however the number of intermittent failures is consistently nonzero. This is worse than our historical results.
The rarity of the failures makes it difficult to pinpoint the regression, however there are several repeating errors:
i965: Failed to submit batchbuffer: Bad address
piglit.spec.!opengl 1_1.copypixels-draw-sync ivb
piglit.spec.!opengl 1_3.gl-1_3-texture-env snb
Mesa CI reports a low error rate (2/700k), however the number of intermittent failures is consistently nonzero. This is worse than our historical results.
The rarity of the failures makes it difficult to pinpoint the regression, however there are several repeating errors:
i965: Failed to submit batchbuffer: Bad address spec.!opengl 1_1.copypixels- draw-sync ivb spec.!opengl 1_3.gl- 1_3-texture- env snb
piglit.
piglit.
intel_batchbuff er.c:937: submit_batch: Assertion `entry->handle == batch-> batch.bo- >gem_handle' failed. spec.!opengl 1_3.gl- 1_3-texture- env.bdwm64 shaders. glsl-fs- raytrace-bug27060 skl
piglit.
piglit.
HSW tesselation failures
deqp-gles31': corrupted double-linked list: 0x0000561e9518be50 *** GLES31. functional. debug.error_ filters. case_0. bdwm64
dEQP-