<html> <head> <base href="https://bugs.freedesktop.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - VulkanCTS 1.1.2 Kernel error: dEQP-VK.pipeline.render_to_image.core.cube_array.huge.width_height_layers" href="https://bugs.freedesktop.org/show_bug.cgi?id=108456#c4">Comment # 4</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - VulkanCTS 1.1.2 Kernel error: dEQP-VK.pipeline.render_to_image.core.cube_array.huge.width_height_layers" href="https://bugs.freedesktop.org/show_bug.cgi?id=108456">bug 108456</a> from <a class="email" href="mailto:daniel@ffwll.ch" title="Daniel Vetter <daniel@ffwll.ch>"> Daniel Vetter</a> <pre>Smells like deadlock in userptr. Tons of backtraces stuck in places where most likely they want to acquire the dev->struct_mutex, except one: [<0>] 0xffffffffffffffff /proc/16502 /tmp/build_root/m64/opt/deqp/modules/vulkan/deqp-vk�--deqp-log-images=disable�--deqp-gl-config-name=rgba8888d24s8�--deqp-surface-width=400�--deqp-surface-height=300�--deqp-visibility=hidden�--deqp-surface-type=fbo�--deqp-shadercache=disable�--deqp-log-filename=TestResults-6.qpa�--deqp-caselist-file=mesa-ci-caselist-6.txt�[<0>] flush_workqueue+0x132/0x420 [<0>] i915_gem_userptr_mn_invalidate_range_start+0x174/0x1d0 [i915] [<0>] __mmu_notifier_invalidate_range_start+0x4f/0x70 [<0>] try_to_unmap_one+0x894/0x960 [<0>] rmap_walk_file+0xec/0x240 [<0>] try_to_unmap+0x93/0xf0 [<0>] migrate_pages+0x7be/0xaa0 [<0>] compact_zone+0x5d9/0x860 [<0>] compact_zone_order+0xa0/0xd0 [<0>] try_to_compact_pages+0xa5/0x250 [<0>] __alloc_pages_direct_compact+0x4a/0xf0 [<0>] __alloc_pages_slowpath+0x462/0xd80 [<0>] __alloc_pages_nodemask+0x233/0x250 [<0>] setup_scratch_page+0xfd/0x240 [i915] [<0>] i915_ppgtt_create+0xad/0xa10 [i915] [<0>] i915_gem_create_context+0x254/0x440 [i915] [<0>] i915_gem_context_create_ioctl+0x55/0x100 [i915] [<0>] drm_ioctl_kernel+0x59/0xb0 [drm] [<0>] drm_ioctl+0x2cb/0x380 [drm] [<0>] do_vfs_ioctl+0xa1/0x610 [<0>] SyS_ioctl+0x74/0x80 [<0>] entry_SYSCALL_64_fastpath+0x24/0x87 [<0>] 0xffffffffffffffff Well I lied, there's more than one stuck on that flush_workqueue(): [<0>] 0xffffffffffffffff /proc/65 [<0>] flush_workqueue+0x132/0x420 [<0>] i915_gem_userptr_mn_invalidate_range_start+0x174/0x1d0 [i915] [<0>] __mmu_notifier_invalidate_range_start+0x4f/0x70 [<0>] try_to_unmap_one+0x894/0x960 [<0>] rmap_walk_file+0xec/0x240 [<0>] try_to_unmap+0x93/0xf0 [<0>] migrate_pages+0x7be/0xaa0 [<0>] compact_zone+0x5d9/0x860 [<0>] compact_zone_order+0xa0/0xd0 [<0>] try_to_compact_pages+0xa5/0x250 [<0>] __alloc_pages_direct_compact+0x4a/0xf0 [<0>] __alloc_pages_slowpath+0x462/0xd80 [<0>] __alloc_pages_nodemask+0x233/0x250 [<0>] khugepaged+0xb4d/0x2060 [<0>] kthread+0x111/0x130 [<0>] ret_from_fork+0x35/0x40 I think this is one for Tvrtko and Chris.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the QA Contact for the bug.</li> </ul> </body> </html>