Bug 73437

Summary:

[ivb] Tried to flip to an unpinned bo

Product:

DRI

Reporter:

Alexandru DAMIAN <alexandru.damian>

Component:

DRM/Intel

Assignee:

Damien Lespiau <damien.lespiau>

Status:

CLOSED DUPLICATE

QA Contact:

Intel GFX Bugs mailing list <intel-gfx-bugs>

Severity:

normal

Priority:

low

CC:

intel-gfx-bugs

Version:

XOrg git

Hardware:

Other

OS:

All

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
content of /sys/class/drm/card0/error	none

Description Alexandru DAMIAN 2014-01-09 14:28:05 UTC

Created attachment 91754 [details]
content of /sys/class/drm/card0/error

I got this message in the dmesg after a brief freeze.

Ubuntu 13.10, mainline kernel: uname -a
Linux adamian-desktop 3.12.0-997-generic #201311300431 SMP Sat Nov 30 09:32:04 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

dmesg lines:
[261690.496128] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[261690.496129] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[261690.496129] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[261690.496130] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[261690.496131] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.

The card is Ivybridge i965, with SNA enabled.

VGA PCI identifier:
00:02.0 0300: 8086:0162 (rev 09) (prog-if 00 [VGA controller])
	Subsystem: 103c:2ad5
	Flags: bus master, fast devsel, latency 0, IRQ 45
	Memory at f7800000 (64-bit, non-prefetchable) [size=4M]
	Memory at e0000000 (64-bit, prefetchable) [size=256M]
	I/O ports at f000 [size=64]
	Expansion ROM at <unassigned> [disabled]
	Capabilities: <access denied>
	Kernel driver in use: i915

Running X.org driver for i965, version:
intel(0): SNA compiled: xserver-xorg-video-intel 2:2.99.906+git20131202.9a8478da-0ubuntu0sarvatt (Robert Hooker <sarvatt@ubuntu.com>)

Comment 1 Chris Wilson 2014-01-09 14:59:53 UTC

The command that fails is

0x00006190:      0x0a000001: MI_DISPLAY_BUFFER_INFO
0x00006194:      0x00003c01:    dword 1
0x00006198:      0x056bc000:    dword 2

Since we have not yet flipped we cannot have unpinned the future fb, and so it should appear in the list of pinned bo. It does not. Nor does it pin a fence register.

Comment 2 Daniel Vetter 2014-01-09 15:27:12 UTC

Is this readily reproducible or just a freak hang?

If you can reproduce it it'll be interesting to figure out when this happens. Most likely we have a stray MMIO write somewhere that completes the flip too early. Otoh how that manged to kill ivb is beyond me, the scratch page ptes should saves us.

Comment 3 Alexandru DAMIAN 2014-01-10 11:02:33 UTC

Seen only once, I don't have a repro rule.

I'll keep an eye open and report if it happens again.

Comment 4 Chris Wilson 2014-02-28 12:06:10 UTC

Let's go out on a limb and assume that this was a mysterious instance of bug 74053.

*** This bug has been marked as a duplicate of bug 74053 ***

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.