This also happens on betty (overnight canary v9820). M-60 works, haven't tried M-61

$ ~/chromeos/chroot/usr/bin/qemu-system-x86_64 -smp 4 --enable-kvm -vnc 127.0.0.1:0,ipv4 -m 4096 -usbdevice tablet -vga virtio -net nic,model=virtio -net user,hostfwd=tcp:127.0.0.1:9222-:22 -drive format=raw,file=chromiumos_qemu_image.bin


2017-08-08T18:27:40.665791+00:00 INFO kernel: [   96.027042] DrmThread[932]: segfault at 8 ip 00007e141a054600 sp 00007e1411a007f8 error 4 in libminigbm.so.1.0.0[7e141a052000+6000]
2017-08-08T18:27:40.692861+00:00 WARNING crash_reporter[4242]: [user] Received crash notification for chrome[913] sig 11, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly)
2017-08-08T18:27:40.720735+00:00 WARNING kernel: [   96.082514] ------------[ cut here ]------------
2017-08-08T18:27:40.720755+00:00 WARNING kernel: [   96.082578] WARNING: CPU: 1 PID: 39 at /mnt/host/source/src/third_party/kernel/v4.4/drivers/gpu/drm/drm_atomic_helper.c:1645 drm_atomic_helper_commit_hw_done+0x63/0x91()
2017-08-08T18:27:40.720760+00:00 WARNING kernel: [   96.082582] Modules linked in: uinput xt_nat bluetooth bridge stp llc ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 zram nf_nat xt_mark sr_mod cdrom virtio_net i2c_piix4 fuse cfg80211 ip6table_filter joydev
2017-08-08T18:27:40.720762+00:00 WARNING kernel: [   96.082686] CPU: 1 PID: 39 Comm: kworker/1:1 Tainted: G        W       4.4.70-11300-gea5cce42d069 #1
2017-08-08T18:27:40.720796+00:00 WARNING kernel: [   96.082689] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.9.1-0-gb3ef39f-prebuilt.qemu-project.org 04/01/2014
2017-08-08T18:27:40.720799+00:00 WARNING kernel: [   96.082708] Workqueue: events drm_mode_rmfb_work_fn
2017-08-08T18:27:40.720807+00:00 WARNING kernel: [   96.082711]  0000000000000286 0000000081d7b621 ffff88013a80bc10 ffffffffa0a93a91
2017-08-08T18:27:40.720810+00:00 WARNING kernel: [   96.082716]  0000000000000000 0000000000000009 ffff88013a80bc48 ffffffffa08697ff
2017-08-08T18:27:40.720811+00:00 WARNING kernel: [   96.082720]  ffffffffa0b5d37d ffff8800bb710700 0000000000000000 ffff8800996b50c0
2017-08-08T18:27:40.720817+00:00 WARNING kernel: [   96.082724] Call Trace:
2017-08-08T18:27:40.720819+00:00 WARNING kernel: [   96.082759]  [<ffffffffa0a93a91>] dump_stack+0x4d/0x63
2017-08-08T18:27:40.720821+00:00 WARNING kernel: [   96.082789]  [<ffffffffa08697ff>] warn_slowpath_common+0x9f/0xb8
2017-08-08T18:27:40.720823+00:00 WARNING kernel: [   96.082799]  [<ffffffffa0b5d37d>] ? drm_atomic_helper_commit_hw_done+0x63/0x91
2017-08-08T18:27:40.720829+00:00 WARNING kernel: [   96.082803]  [<ffffffffa0869911>] warn_slowpath_null+0x1a/0x1c
2017-08-08T18:27:40.720831+00:00 WARNING kernel: [   96.082807]  [<ffffffffa0b5d37d>] drm_atomic_helper_commit_hw_done+0x63/0x91
2017-08-08T18:27:40.720833+00:00 WARNING kernel: [   96.082818]  [<ffffffffa0b5ff28>] drm_atomic_helper_commit_tail+0x3e/0x59
2017-08-08T18:27:40.720835+00:00 WARNING kernel: [   96.082822]  [<ffffffffa0b5ff89>] commit_tail+0x46/0x5b
2017-08-08T18:27:40.720858+00:00 WARNING kernel: [   96.082827]  [<ffffffffa0b6004d>] drm_atomic_helper_commit+0xaf/0xb8
2017-08-08T18:27:40.720859+00:00 WARNING kernel: [   96.082856]  [<ffffffffa0b77a11>] drm_atomic_commit+0x4b/0x50
2017-08-08T18:27:40.720861+00:00 WARNING kernel: [   96.082860]  [<ffffffffa0b604c5>] drm_atomic_helper_set_config+0x53/0x98
2017-08-08T18:27:40.720863+00:00 WARNING kernel: [   96.082866]  [<ffffffffa0b6db63>] drm_mode_set_config_internal+0x5e/0xff
2017-08-08T18:27:40.720865+00:00 WARNING kernel: [   96.082869]  [<ffffffffa0b6dc3b>] drm_crtc_force_disable+0x37/0x4d
2017-08-08T18:27:40.720866+00:00 WARNING kernel: [   96.082873]  [<ffffffffa0b79647>] drm_framebuffer_remove+0x89/0x105
2017-08-08T18:27:40.720872+00:00 WARNING kernel: [   96.082876]  [<ffffffffa0b796fd>] drm_mode_rmfb_work_fn+0x3a/0x41
2017-08-08T18:27:40.720877+00:00 WARNING kernel: [   96.082895]  [<ffffffffa087e596>] process_one_work+0x1bd/0x2fd
2017-08-08T18:27:40.720879+00:00 WARNING kernel: [   96.082900]  [<ffffffffa087f8d3>] worker_thread+0x251/0x2ee
2017-08-08T18:27:40.720881+00:00 WARNING kernel: [   96.082904]  [<ffffffffa087f682>] ? rescuer_thread+0x37c/0x37c
2017-08-08T18:27:40.720882+00:00 WARNING kernel: [   96.082908]  [<ffffffffa087f682>] ? rescuer_thread+0x37c/0x37c
2017-08-08T18:27:40.720883+00:00 WARNING kernel: [   96.082916]  [<ffffffffa0883b05>] kthread+0x12c/0x134
2017-08-08T18:27:40.720884+00:00 WARNING kernel: [   96.082920]  [<ffffffffa08839d9>] ? kthread_parkme+0x24/0x24
2017-08-08T18:27:40.720885+00:00 WARNING kernel: [   96.082949]  [<ffffffffa0fb6d2f>] ret_from_fork+0x3f/0x70
2017-08-08T18:27:40.720886+00:00 WARNING kernel: [   96.082953]  [<ffffffffa08839d9>] ? kthread_parkme+0x24/0x24
2017-08-08T18:27:40.720891+00:00 WARNING kernel: [   96.082956] ---[ end trace 53ef6ab440ad8e06 ]---
2017-08-08T18:27:40.747604+00:00 INFO crash_reporter[4244]: Processing kernel warning: always collect from developer builds
2017-08-08T18:27:40.748665+00:00 INFO crash_reporter[4244]: Stored kernel warning into /var/spool/crash/kernel_warning.20170808.112740.0.kcrash
2017-08-08T18:27:40.918742+00:00 INFO kernel: [   96.280148] DrmThread[4258]: segfault at 8 ip 00007c99b9259600 sp 00007c99b0c057f8 error 4 in libminigbm.so.1.0.0[7c99b9257000+6000]
2017-08-08T18:27:40.939503+00:00 WARNING crash_reporter[4259]: [user] Received crash notification for chrome[4245] sig 11, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly)
2017-08-08T18:27:40.959191+00:00 WARNING crash_reporter[4261]: Received crash notification for chrome[4245] user 1000 (called directly)
2017-08-08T18:27:41.126683+00:00 INFO kernel: [   96.488817] DrmThread[4293]: segfault at 8 ip 00007c5ce80a8600 sp 00007c5cdfa547f8 error 4 in libminigbm.so.1.0.0[7c5ce80a6000+6000]
2017-08-08T18:27:41.144163+00:00 WARNING crash_reporter[4294]: [user] Received crash notification for chrome[4262] sig 11, user 1000 (ignoring call by kernel - chrome crash; waiting for chrome to call us directly)
2017-08-08T18:27:41.158196+00:00 WARNING crash_reporter[4296]: Received crash notification for chrome[4262] user 1000 (called directly)
2017-08-08T18:27:41.298712+00:00 INFO kernel: [   96.660634] DrmThread[4328]: segfault at 8 ip 00007e49033af600 sp 00007e48fad5b7f8 error 4 in libminigbm.so.1.0.0[7e49033ad000+6000]


Stephane, can you recommend an owner?

I think this is fixed (was lack of locking in minigbm). Can you retry on tot?

I just tried ToT and the bug is still there, exactly the same crash.

Can you try without vga virtio? It doesn't crash on cirrus (but the video doesn't play either).

As a side note, we should have a builder with virtio.

I'll try cirrus tomorrow. VMs (betty/novato/newbie) are actually tested with virtio in the CQ, but it looks like we're not browsing Youtube in those tests.

The crash seems gone (I think the minigbm fix really did fix this), but the video stack is in a "I can't initialize" loop:

[26068:26081:0809/192215.951588:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26068:26068:0809/192215.951883:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26082:26082:0809/192216.012911:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.
[26082:26095:0809/192216.015556:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26082:26082:0809/192216.015853:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26096:26096:0809/192216.074895:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.
[26096:26109:0809/192216.077487:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26096:26096:0809/192216.077781:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26110:26110:0809/192216.140009:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.
[26110:26123:0809/192216.142268:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26110:26110:0809/192216.142757:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26124:26124:0809/192216.203906:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.
[26124:26137:0809/192216.205725:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26124:26124:0809/192216.205802:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26138:26138:0809/192216.267914:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.
[26138:26138:0809/192216.270153:WARNING:vaapi_wrapper.cc(1198)] vaInitialize failed: unknown libva error
[26138:26151:0809/192216.270664:VERBOSE1:drm_device_manager.cc(54)] Could not initialize DRM device for /sys/devices/platform/vgem/drm/card0
[26152:26152:0809/192216.330900:ERROR:sandbox_linux.cc(344)] InitializeSandbox() called with multiple threads in process gpu-process. This error can be safely ignored in VMTests.

The libminigbm crash is still present for me in betty canary R62-9827.0.0. It happens both with '-vga cirrus' and '-vga virtio', using qemu-2.6 (from the Chrome OS chroot).


~/chromeos/chroot/usr/bin/qemu-system-x86_64 -smp 4 --enable-kvm -vnc 127.0.0.1:0,ipv4 -m 4096 -usbdevice tablet -vga cirrus -net nic,model=virtio -net user,hostfwd=tcp:127.0.0.1:9222-:22 -drive file=chromiumos_qemu_image.bin

Apparently there's a Chrome CL that could fix the crash, I think that one: https://chromium-review.googlesource.com/c/609189 ?
But that won't fix the vaapi error. I'll try again after the next successful Chrome roll.

 Issue 754106  has been merged into this issue.