GTT full but LRU empty

GTT full but LRU empty

[Jacek Luczak]

Hi Eric,

Are those errors known?

[drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[drm:i915_gem_object_pin] *ERROR* Failure to bind:
-12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
flushing empty 1
[drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[drm:i915_gem_object_pin] *ERROR* Failure to bind:
-12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
flushing empty 1
[drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[drm:i915_gem_object_pin] *ERROR* Failure to bind:
-12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12
[drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[drm:i915_gem_object_pin] *ERROR* Failure to bind:
-12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
flushing empty 1
[drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
[drm:i915_gem_object_pin] *ERROR* Failure to bind:
-12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12

They appeared once and no more, but I've changed my configuration in between
(switched to UXA, removed IntelFB from kernel).

Also I've got those messages with KMS enabled, but AFAIK those are only
harmless_spamming:

Jan 25 12:01:49 difrost kernel: WARNING: at drivers/gpu/drm/drm_crtc.c:213
drm_mode_object_get+0x36/0x8e()
Jan 25 12:01:49 difrost kernel: Hardware name: AMILO Pro Edition V3505
Jan 25 12:01:49 difrost kernel: drm_mode_object_get called w/o mode_config lock
Jan 25 12:01:49 difrost kernel: Modules linked in:
Jan 25 12:01:49 difrost kernel: Pid: 1, comm: swapper Not tainted
2.6.29-rc2-00013-gf3b8436-dirty #3
Jan 25 12:01:49 difrost kernel: Call Trace:
Jan 25 12:01:49 difrost kernel:  [<c011f310>] warn_slowpath+0x80/0xb6
Jan 25 12:01:49 difrost kernel:  [<c011755c>] dequeue_task+0x60/0x6e
Jan 25 12:01:49 difrost kernel:  [<c033c038>] schedule+0x6e0/0x71c
Jan 25 12:01:49 difrost kernel:  [<c0126894>] lock_timer_base+0x19/0x35
Jan 25 12:01:49 difrost kernel:  [<c01268f4>] try_to_del_timer_sync+0x44/0x4a
Jan 25 12:01:49 difrost kernel:  [<c025c63a>] drm_mode_object_get+0x36/0x8e
Jan 25 12:01:49 difrost kernel:  [<c025c766>]
drm_mode_connector_update_edid_property+0xd4/0x144
Jan 25 12:01:49 difrost kernel:  [<c026d500>] intel_ddc_get_modes+0x24/0x3c
Jan 25 12:01:49 difrost kernel:  [<c026bcc9>] intel_lvds_init+0xe3/0x296
Jan 25 12:01:49 difrost kernel:  [<c026aadc>] intel_modeset_init+0x307/0x5ce
Jan 25 12:01:49 difrost kernel:  [<c026274f>] i915_driver_irq_postinstall+0xa8/0xac
Jan 25 12:01:49 difrost kernel:  [<c0261688>] i915_driver_load+0x6ef/0x73e
Jan 25 12:01:49 difrost kernel:  [<c0259560>] drm_get_dev+0x2be/0x358
Jan 25 12:01:49 difrost kernel:  [<c025552d>] drm_init+0x5a/0x89
Jan 25 12:01:49 difrost kernel:  [<c04574bd>] i915_init+0x0/0x43
Jan 25 12:01:49 difrost kernel:  [<c010112c>] _stext+0x44/0x108
Jan 25 12:01:49 difrost kernel:  [<c0142b38>] register_irq_proc+0x79/0x92
Jan 25 12:01:49 difrost kernel:  [<c0441297>] kernel_init+0x101/0x14f
Jan 25 12:01:49 difrost kernel:  [<c0441196>] kernel_init+0x0/0x14f
Jan 25 12:01:49 difrost kernel:  [<c01034db>] kernel_thread_helper+0x7/0x10
Jan 25 12:01:49 difrost kernel: ---[ end trace 1e62cbea2b78759c ]---

Some details:
	1) Kernel: 2.6.29-rc2-00013-gf3b8436-dirty (marked dirty due to not-yet-in-tree
XFS patch).
	2) xf86-video-intel: 2.6.1
	3) Xorg-server: 1.5.99.901
	4) Mesa: 7.3
	4) DRI2 enabled, now also UXA (with strange screen flickering effect).

-Jacek

Re: GTT full but LRU empty

[Eric Anholt]

[-- Attachment #1: Type: text/plain, Size: 2053 bytes --]

On Mon, 2009-01-26 at 19:10 +0100, Jacek Luczak wrote:
> Hi Eric,
> 
> Are those errors known?
> 
> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
> flushing empty 1
> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
> flushing empty 1
> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
> -12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12
> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
> flushing empty 1
> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
> -12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12
> 
> They appeared once and no more, but I've changed my configuration in between
> (switched to UXA, removed IntelFB from kernel).

Those messages occur when a client asks the kernel to exec a batchbuffer
that the kernel can't fit into the unpinned aperture space.  Userland is
supposed to be tracking how much space its stuff uses and flushing
before that time, but certain bad things could break that by violating
the assumption that the amount of pinned space doesn't change.  Our XV
code today, for example.  I've been working on cleaning up the remaining
bad userland bits, but it's painful work.  Luckily, other than dmesg
spam warning that some rendering just go dropped on the floor, and some
rendering not occurring, nothing else harmful should be happening.

-- 
Eric Anholt
eric@anholt.net                         eric.anholt@intel.com



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: GTT full but LRU empty

[Jacek Luczak]

Eric Anholt pisze:
> On Mon, 2009-01-26 at 19:10 +0100, Jacek Luczak wrote:
>> Hi Eric,
>>
>> Are those errors known?
>>
>> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
>> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
>> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
>> flushing empty 1
>> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
>> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
>> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
>> flushing empty 1
>> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
>> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
>> -12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12
>> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
>> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
>> -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1
>> flushing empty 1
>> [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
>> [drm:i915_gem_object_pin] *ERROR* Failure to bind:
>> -12<3>[drm:i915_gem_execbuffer] *ERROR* Failed to pin buffers -12
>>
>> They appeared once and no more, but I've changed my configuration in between
>> (switched to UXA, removed IntelFB from kernel).
> 
> Those messages occur when a client asks the kernel to exec a batchbuffer
> that the kernel can't fit into the unpinned aperture space.  Userland is
> supposed to be tracking how much space its stuff uses and flushing
> before that time, but certain bad things could break that by violating
> the assumption that the amount of pinned space doesn't change.  Our XV
> code today, for example.  I've been working on cleaning up the remaining
> bad userland bits, but it's painful work.  Luckily, other than dmesg
> spam warning that some rendering just go dropped on the floor, and some
> rendering not occurring, nothing else harmful should be happening.

Thanks Eric. So it's mostly broken userspace. Pity that I don't remember what I
was running while this error occurred.

-Jacek