Poor performance

[Robert Rojas]

After an indeterminate amount of time, the refresh rate drops to about 15 fps until reboot.

The problem appeared with 17.2.x. It also happens with 17.3.3.

Initially reported with ubuntu-bug: https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/1752830. Decided to report against mesa because of the dmesg log.

When the refresh rate drops, dmesg logs the following repeatedly:

[40447.264213] WARNING: CPU: 0 PID: 11771 at /build/linux-hwe-qx9Tq0/linux-hwe-4.13.0/drivers/gpu/drm/i915/intel_display.c:12848 intel_atomic_commit_tail+0x1081/0x10b0 [i915]
[40447.264215] Modules linked in: hid_generic usbhid hid ccm pci_stub vboxpci(OE) vboxnetadp(OE) vboxnetflt(OE) vboxdrv(OE) crypto_simd glue_helper cryptd aes_x86_64 dm_crypt arc4 snd_hda_codec_conexant uvcvideo snd_hda_codec_generic videobuf2_vmalloc videobuf2_memops snd_hda_intel snd_hda_codec snd_hda_core videobuf2_v4l2 snd_hwdep ath9k videobuf2_core videodev media ath9k_common snd_pcm snd_seq_midi snd_seq_midi_event ath9k_hw ath snd_rawmidi mac80211 snd_seq snd_seq_device coretemp snd_timer snd soundcore input_leds joydev wmi_bmof serio_raw cfg80211 shpchp lpc_ich mac_hid binfmt_misc parport_pc ppdev lp parport autofs4 i915 psmouse ahci libahci i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm atl1c wmi video
[40447.264273] CPU: 0 PID: 11771 Comm: kworker/u4:0 Tainted: G W OE 4.13.0-37-generic #42~16.04.1-Ubuntu
[40447.264274] Hardware name: Acer Aspire 4332 /HM41 , BIOS V1.10 03/29/2010
[40447.264322] Workqueue: events_unbound intel_atomic_commit_work [i915]
[40447.264324] task: ffff98612e4fac80 task.stack: ffffae1441d30000
[40447.264366] RIP: 0010:intel_atomic_commit_tail+0x1081/0x10b0 [i915]
[40447.264367] RSP: 0000:ffffae1441d33d98 EFLAGS: 00010282
[40447.264369] RAX: 000000000000001c RBX: 0000000000000000 RCX: 0000000000000000
[40447.264371] RDX: 0000000000000000 RSI: ffff9861ffc16578 RDI: ffff9861ffc16578
[40447.264373] RBP: ffffae1441d33e50 R08: 0000000000000001 R09: 000000000000037d
[40447.264374] R10: ffffae1441d33d98 R11: 000000000000037d R12: ffff9861f0af8000
[40447.264375] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001
[40447.264378] FS: 0000000000000000(0000) GS:ffff9861ffc00000(0000) knlGS:0000000000000000
[40447.264379] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[40447.264381] CR2: 00007f11caf57000 CR3: 0000000137670000 CR4: 00000000000406f0
[40447.264382] Call Trace:
[40447.264392] ? wait_woken+0x80/0x80
[40447.264435] intel_atomic_commit_work+0x12/0x20 [i915]
[40447.264440] process_one_work+0x15b/0x410
[40447.264443] worker_thread+0x4b/0x460
[40447.264446] kthread+0x10c/0x140
[40447.264448] ? process_one_work+0x410/0x410
[40447.264450] ? kthread_create_on_node+0x70/0x70
[40447.264454] ret_from_fork+0x35/0x40
[40447.264456] Code: ff ff ff 48 83 c7 08 e8 ce 50 78 d3 4c 8b 85 68 ff ff ff 4d 85 c0 0f 85 00 fe ff ff 8d 73 41 48 c7 c7 58 87 41 c0 e8 80 05 7a d3 <0f> ff e9 ea fd ff ff 8d 70 41 48 c7 c7 28 87 41 c0 e8 6a 05 7a
[40447.264500] ---[ end trace e2a13a3e2dcf2ae1 ]---
[40447.376082] pipe A vblank wait timed out
[40447.480123] ------------[ cut here ]------------

Using Ubuntu 16.04.3, Intel GM45.

[Timo Aaltonen (tjaalton)]

that's an old kernel, if you still can reproduce this on 4.15 or newer, open a new bug against 'linux'