On Thu, Jul 06, 2017 at 02:57:34PM +0200, Daniel Vetter wrote:
> It's not accelarated, just system memory. Note we don't even need to
> set the default flag since that's now always 0.
>
> Cc: Dave Airlie
> Cc: Gerd Hoffmann
> Cc: virtualizat...@lists.linux-foundation.org
> Signed-off-by: Daniel
We want to change swap_state to wait indefinitely, but to do this
swap_state should wait interruptibly. This requires propagating
the error to each driver.
Cc: dri-de...@lists.freedesktop.org
Cc: linux-ker...@vger.kernel.org
Cc: intel-gfx@lists.freedesktop.org
Signed-off-by: Maarten Lankhorst
Lin
On Tue, Jul 18, 2017 at 06:00:20PM +0200, Daniel Vetter wrote:
> There's a bunch of reasons why I think we should formalize and enforce
> our review rules for igt patches:
>
> - We have a lot of new engineers joining and review helps enormously
> with mentoring and learning. But right now only p
During the engine reset, there is a race condition which can make the
request submitted to HW twice. This is due to the irq tasklet function
enabled too earliy which is just before init_hw callback. This patch
will move the irq tasklet enabling after init_hw to resolve this race.
Fixes: a1ef70e144
Mahesh Kumar schreef op di 18-07-2017 om 18:12 [+0530]:
> Hi Daniel,
>
> On Monday 17 July 2017 12:56 PM, Daniel Vetter wrote:
> > On Fri, Jun 30, 2017 at 05:40:58PM +0530, Mahesh Kumar wrote:
> > > Gen9+ Interlace fetch mode doesn't support few plane
> > > configurations & pipe scaling.
> > > -
On Tue, 2017-07-18 at 22:21 +0100, Chris Wilson wrote:
> Quoting Paul Kocialkowski (2017-07-18 16:16:26)
> > It may occur that a hotplug uevent is detected at resume, even
> > though it
> > does not indicate that an actual hotplug happened. This is the case
> > when
> > link training fails on any o
On Tue, 2017-07-18 at 12:20 -0700, Manasi Navare wrote:
> On Tue, Jul 18, 2017 at 05:25:36PM +0300, Paul Kocialkowski wrote:
> > This adds the connector name when printing a debug message about the
> > DP
> > link training result. It is useful to figure out what connector is
> > failing when multip
On Wed, 19 Jul 2017, Paul Kocialkowski
wrote:
> On Tue, 2017-07-18 at 12:20 -0700, Manasi Navare wrote:
>> On Tue, Jul 18, 2017 at 05:25:36PM +0300, Paul Kocialkowski wrote:
>> > This adds the connector name when printing a debug message about the
>> > DP
>> > link training result. It is useful t
== Series Details ==
Series: drm/i915: Fix the port submission race during engine reset in execlists
mode
URL : https://patchwork.freedesktop.org/series/27554/
State : success
== Summary ==
Series 27554v1 drm/i915: Fix the port submission race during engine reset in
execlists mode
https://pa
On Wed, Jul 19, 2017 at 08:22:03AM +, Lankhorst, Maarten wrote:
> Mahesh Kumar schreef op di 18-07-2017 om 18:12 [+0530]:
> > Hi Daniel,
> >
> > On Monday 17 July 2017 12:56 PM, Daniel Vetter wrote:
> > > On Fri, Jun 30, 2017 at 05:40:58PM +0530, Mahesh Kumar wrote:
> > > > Gen9+ Interlace fet
On 18/07/2017 16:19, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2017-07-18 15:36:15)
From: Tvrtko Ursulin
Track total time requests have been executing on the hardware.
To make this cheap it is hidden behind a static branch with the
intention that it is only enabled when there is a consumer
On 07/12/2017 04:15 PM, Maarten Lankhorst wrote:
Op 12-07-17 om 11:48 schreef Daniel Vetter:
On Wed, Jul 12, 2017 at 10:13:42AM +0200, Maarten Lankhorst wrote:
for_each_obj_in_state is about to be removed, so convert
to the new iterator macros.
Just like in omap, use crtc_state->active inste
Quoting Chuanxiao Dong (2017-07-19 09:14:14)
> During the engine reset, there is a race condition which can make the
> request submitted to HW twice. This is due to the irq tasklet function
> enabled too earliy which is just before init_hw callback. This patch
> will move the irq tasklet enabling a
On 18/07/2017 16:22, Chris Wilson wrote:
> Quoting Tvrtko Ursulin (2017-07-18 15:36:16)
>> +u64 intel_engine_get_current_busy_ns(struct intel_engine_cs *engine)
>> +{
>> + unsigned long flags;
>> + u64 total;
>> +
>> + spin_lock_irqsave(&engine->stats.lock, flags);
>> +
>> +
Hi Ben,
On 18/07/2017 15:36, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Enables other i915 components to enable and disable
the facility as needed.
Signed-off-by: Tvrtko Ursulin
---
drivers/gpu/drm/i915/intel_engine_cs.c | 53 +
drivers/gpu/drm/i915/intel
On Tue, 18 Jul 2017, Daniel Vetter wrote:
> On Tue, Jul 18, 2017 at 10:34 PM, Lionel Landwerlin
> wrote:
>> Acked-by: Lionel Landwerlin
>>
>> I assume review cannot be provided by someone who doesn't already contribute
>> or has a number of patches in already.
>>
>> What's the criteria to become
> -Original Message-
> From: Chris Wilson [mailto:ch...@chris-wilson.co.uk]
> Sent: Wednesday, July 19, 2017 5:26 PM
> To: Dong, Chuanxiao ; intel-
> g...@lists.freedesktop.org
> Cc: intel-gvt-...@lists.freedesktop.org
> Subject: Re: [Intel-gfx] [PATCH] drm/i915: Fix the port submission rac
Can we reuse calc_residency defined in i915_sysfs.c
On 7/18/2017 8:06 PM, Tvrtko Ursulin wrote:
From: Chris Wilson
The first goal is to be able to measure GPU (and invidual ring) busyness
without having to poll registers from userspace. (Which not only incurs
holding the forcewake lock indefi
I've found a bios with an off-by-one at the other end. There's a pnp
reservation for 0xc540-0xc7fe and we want stolen in 0xc600
through 0xc800.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99872
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98683
Cc: Martin Peres
On 19/07/17 13:00, Daniel Vetter wrote:
I've found a bios with an off-by-one at the other end. There's a pnp
reservation for 0xc540-0xc7fe and we want stolen in 0xc600
through 0xc800.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99872
Bugzilla: https://bugs.freedesktop.
Op 14-07-17 om 17:01 schreef Sean Paul:
> On Tue, Jul 11, 2017 at 04:33:02PM +0200, Maarten Lankhorst wrote:
>> drm_atomic_helper_swap_state could previously never fail, in order to still
>> continue it would set a limit of 10s on waiting for previous updates to
>> complete,
>> then it moved forwa
Op 17-07-17 om 17:21 schreef Daniel Vetter:
> On Mon, Jul 17, 2017 at 11:39:56AM +0200, Maarten Lankhorst wrote:
>> Op 15-07-17 om 11:31 schreef Daniel Vetter:
>>> The legacy plane->fb pointer is refcounted by calling
>>> drm_atomic_clean_old_fb().
>>>
>>> In practice this isn't a real problem beca
With older panels, AUX pin for backlight doesn't work. On some
panels, this causes backlight issues on S3 resume. Enable the
feature only for eDP1.4 or later.
Fix-suggested-by: Puthikorn Voravootivat
Signed-off-by: Jenny TC
---
drivers/gpu/drm/i915/intel_dp_aux_backlight.c | 6 +-
1 file ch
Quoting Tvrtko Ursulin (2017-07-19 10:12:33)
>
> On 18/07/2017 16:19, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2017-07-18 15:36:15)
> >> From: Tvrtko Ursulin
> >>
> >> Track total time requests have been executing on the hardware.
> >>
> >> To make this cheap it is hidden behind a static b
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows
guest VM. This patch is to introduce the format to drm.
v1:
Suggested by Ville to submit this patch to dri-devel.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
include/uapi/drm/drm_fourcc.h | 4
1 file change
v11->v12:
1) add drm_format_mod back. (Gerd and Zhenyu)
2) add region_index. (Gerd)
3) refine the lifecycle of dmabuf.
4) send to dri-de...@lists.freedesktop.org. (Ville)
v10->v11:
1) rename plane_type to drm_plane_type. (Gerd)
2) move fields of vfio_device_query_gfx_plane to
vfio_device_gfx_p
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows 10
guest VM. This patch is to add this pixel format support to gvt device
model. Without this patch, some Apps, e.g. "DXGIGammaVM.exe", will crash
and make guest screen black.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhan
Framebuffer decoder returns guest framebuffer information.
Guest framebuffer includes primary, cursor and sprite plane.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
drivers/gpu/drm/i915/gvt/Makefile | 3 +-
drivers/gpu/drm/i915/gvt/display.c| 2 +-
drivers/gpu/drm/i91
On 18 July 2017 at 18:30, Chris Wilson wrote:
> Supply a pm_domain and its ops for our mock GEM device so that
> device runtime pm doesn't complain even though we only want to mark it
> permanently active!
>
> Signed-off-by: Chris Wilson
Fixes it for me, so:
Tested-by: Matthew Auld
Reviewed-by:
v11->v12:
1) add drm_format_mod back. (Gerd and Zhenyu)
2) add region_index. (Gerd)
3) refine the lifecycle of dmabuf.
4) send to dri-de...@lists.freedesktop.org. (Ville)
v10->v11:
1) rename plane_type to drm_plane_type. (Gerd)
2) move fields of vfio_device_query_gfx_plane to
vfio_device_gfx_p
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows
guest VM. This patch is to introduce the format to drm.
v1:
Suggested by Ville to submit this patch to dri-devel.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
include/uapi/drm/drm_fourcc.h | 4
1 file change
Framebuffer decoder returns guest framebuffer information.
Guest framebuffer includes primary, cursor and sprite plane.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
drivers/gpu/drm/i915/gvt/Makefile | 3 +-
drivers/gpu/drm/i915/gvt/display.c| 2 +-
drivers/gpu/drm/i91
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows 10
guest VM. This patch is to add this pixel format support to gvt device
model. Without this patch, some Apps, e.g. "DXGIGammaVM.exe", will crash
and make guest screen black.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhan
Hi Arek
This time it looks like a bug
I run this test on i7-4770, by command `./igt/tests/kms_frontbuffer_tracking
--run-subtest basic`
I guess it may be caused by I have no monitor connected. But if that is the
case it should be skiped rather than fail.
Output shows below
-
Framebuffer decoder returns guest framebuffer information.
Guest framebuffer includes primary, cursor and sprite plane.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
drivers/gpu/drm/i915/gvt/Makefile | 3 +-
drivers/gpu/drm/i915/gvt/display.c| 2 +-
drivers/gpu/drm/i91
v11->v12:
1) add drm_format_mod back. (Gerd and Zhenyu)
2) add region_index. (Gerd)
3) refine the lifecycle of dmabuf.
4) send to dri-de...@lists.freedesktop.org. (Ville)
v10->v11:
1) rename plane_type to drm_plane_type. (Gerd)
2) move fields of vfio_device_query_gfx_plane to
vfio_device_gfx_p
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows
guest VM. This patch is to introduce the format to drm.
v1:
Suggested by Ville to submit this patch to dri-devel.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhang
---
include/uapi/drm/drm_fourcc.h | 4
1 file change
The RGB 64-bit 16:16:16:16 float pixel format is needed by windows 10
guest VM. This patch is to add this pixel format support to gvt device
model. Without this patch, some Apps, e.g. "DXGIGammaVM.exe", will crash
and make guest screen black.
Signed-off-by: Xiaoguang Chen
Signed-off-by: Tina Zhan
Add VFIO_DEVICE_QUERY_GFX_PLANE ioctl command to let user mode query and
get the plan and its related information.
The dma-buf's life cycle is handled by user mode and tracked by kernel.
The returned fd in struct vfio_device_query_gfx_plane can be a new
fd or an old fd of a re-exported dma-buf. Ho
Windows guest UPT driver can use operegion to configure the setting
for display. Without the opregion support, the display registers won't
be set and this blocks display model to get the correct information
of the guest display plane.
Signed-off-by: Bing Niu
Signed-off-by: Xiaoguang Chen
Signed-
Signed-off-by: Tina Zhang
---
drivers/gpu/drm/i915/i915_gem.c| 26 +-
drivers/gpu/drm/i915/i915_gem_object.h | 9 +
drivers/gpu/drm/i915/i915_gem_tiling.c | 5 +
3 files changed, 39 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/i915_
Quoting Tvrtko Ursulin (2017-07-19 10:30:13)
>
> On 18/07/2017 16:22, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2017-07-18 15:36:16)
> >> +u64 intel_engine_get_current_busy_ns(struct intel_engine_cs *engine)
> >> +{
> >> + unsigned long flags;
> >> + u64 total;
> >> +
> >> +
On Mon, 2017-07-17 at 12:29 -0400, Lyude Paul wrote:
> On Wed, 2017-07-12 at 17:50 +0300, Paul Kocialkowski wrote:
> > This introduces CRC calculation for reference frames, instead of
> > using
> > hardcoded values for them. The rendering of reference frames may
> > differ
> > from machine to machi
Chris Wilson writes:
> Workers on the i915->wq may rearm themselves so for completeness we need
> to replace our flush_workqueue() with a call to drain_workqueue() before
> unloading the device.
>
> v2: Reinforce the drain_workqueue with an preceeding rcu_barrier() as a
> few of the tasks that ne
s/GEM proxy/a GEM proxy object/
Quoting Tina Zhang (2017-07-19 11:59:05)
Rationale goes here.
> Signed-off-by: Tina Zhang
> ---
> drivers/gpu/drm/i915/i915_gem.c| 26 +-
> drivers/gpu/drm/i915/i915_gem_object.h | 9 +
> drivers/gpu/drm/i915/i915_gem_til
Hi~ all,
There is one case will skip and prints "No known gpu found". There are some
test cases will check on this, but they have not skiped.
I think there may be something wrong.
I run the test by command:
./igt/tests/gem_ringfill --run-subtest basic-default-interruptible
./igt/tests/gem_rin
Quoting Mika Kuoppala (2017-07-19 12:18:47)
> Chris Wilson writes:
>
> > Workers on the i915->wq may rearm themselves so for completeness we need
> > to replace our flush_workqueue() with a call to drain_workqueue() before
> > unloading the device.
> >
> > v2: Reinforce the drain_workqueue with a
Quoting Gu, HailinX (2017-07-19 11:58:37)
> Hi Arek
>
> This time it looks like a bug
>
> I run this test on i7-4770, by command `./igt/tests/kms_frontbuffer_tracking
> --run-subtest basic`
>
> I guess it may be caused by I have no monitor connected. But if that is the
> case it should be skiped
== Series Details ==
Series: drm/i915: More stolen quirking
URL : https://patchwork.freedesktop.org/series/27561/
State : success
== Summary ==
Series 27561v1 drm/i915: More stolen quirking
https://patchwork.freedesktop.org/api/1.0/series/27561/revisions/1/mbox/
Test kms_flip:
Subgrou
Quoting Gu, HailinX (2017-07-19 12:28:25)
> Hi~ all,
>
>
>
> There is one case will skip and prints “No known gpu found”. There are some
> test cases will check on this, but they have not skiped.
The test uses the vgem module to measure the ring size on your machine.
The error message is sti
== Series Details ==
Series: drm/i915: Use AUX for backlight only if eDP 1.4 or later
URL : https://patchwork.freedesktop.org/series/27566/
State : success
== Summary ==
Series 27566v1 drm/i915: Use AUX for backlight only if eDP 1.4 or later
https://patchwork.freedesktop.org/api/1.0/series/275
Chris Wilson writes:
> Quoting Mika Kuoppala (2017-07-19 12:18:47)
>> Chris Wilson writes:
>>
>> > Workers on the i915->wq may rearm themselves so for completeness we need
>> > to replace our flush_workqueue() with a call to drain_workqueue() before
>> > unloading the device.
>> >
>> > v2: Rein
On 7/19/2017 11:55 AM, Gerd Hoffmann wrote:
> On Wed, 2017-07-19 at 00:16 +, Zhang, Tina wrote:
>>> -Original Message-
>>> From: Gerd Hoffmann [mailto:kra...@redhat.com]
>>> Sent: Monday, July 17, 2017 7:03 PM
>>> To: Kirti Wankhede ; Zhang, Tina
>>> ; Tian, Kevin ; linux-
>>> ker...@
Op 17-07-17 om 14:06 schreef Mahesh Kumar:
> Reviewed-by: Mahesh Kumar
Thanks, fix pushed. :)
> On Monday 17 July 2017 04:43 PM, Maarten Lankhorst wrote:
>> ddb_allocation && ddb_allocation / blocks_per_line >= 1 is the same
>> as ddb_allocation >= blocks_per_line, so use the latter to simplify
>>
On Wed, Jul 19, 2017 at 01:05:45PM +0300, Martin Peres wrote:
> On 19/07/17 13:00, Daniel Vetter wrote:
> > I've found a bios with an off-by-one at the other end. There's a pnp
> > reservation for 0xc540-0xc7fe and we want stolen in 0xc600
> > through 0xc800.
> >
> > Bugzilla: http
Quoting Tvrtko Ursulin (2017-07-18 15:36:04)
> From: Tvrtko Ursulin
>
> Rough sketch of the idea I mentioned a few times to various people - merging
> the engine busyness tracking with Chris i915 PMU RFC.
>
> First patch is the actual PMU RFC by Chris. It is followed by some cleanup
> patches, t
== Series Details ==
Series: drm/i915/gvt: Dma-buf support for GVT-g
URL : https://patchwork.freedesktop.org/series/27572/
State : success
== Summary ==
Series 27572v1 drm/i915/gvt: Dma-buf support for GVT-g
https://patchwork.freedesktop.org/api/1.0/series/27572/revisions/1/mbox/
Test gem_exe
Quoting Matthew Auld (2017-07-19 11:51:51)
> On 18 July 2017 at 18:30, Chris Wilson wrote:
> > Supply a pm_domain and its ops for our mock GEM device so that
> > device runtime pm doesn't complain even though we only want to mark it
> > permanently active!
> >
> > Signed-off-by: Chris Wilson
> Fi
Quoting Mika Kuoppala (2017-07-19 12:51:04)
> Chris Wilson writes:
>
> > Quoting Mika Kuoppala (2017-07-19 12:18:47)
> >> Chris Wilson writes:
> >>
> >> > Workers on the i915->wq may rearm themselves so for completeness we need
> >> > to replace our flush_workqueue() with a call to drain_workqu
On Mon, 2017-07-17 at 13:35 -0400, Lyude Paul wrote:
> Just one change for this patch below
>
> On Wed, 2017-07-12 at 17:57 +0300, Paul Kocialkowski wrote:
> > This adds support for VGA frame comparison testing with the
> > reference
> > generated from cairo. The retrieved frame from the chamelium
Hi all,
This fixes the dreaded gpu reset vs. modeset deadlocks. And since the defunct
legacy page_flip code is the reason for a bunch of special cases I did remove
that too, on Maarten's request.
Please review, this is currently blocking extended CI runs on our haswell farm
rather badly. The crit
... using the biggest hammer we have. This is essentially a weaponized
version of the timeout-based wedging Chris added in
commit 36703e79a982c8ce5a8e43833291f2719e92d0d1
Author: Chris Wilson
Date: Thu Jun 22 11:56:25 2017 +0100
drm/i915: Break modeset deadlocks on reset
Because defense-i
Just a very minimal patch to nuke that code. Lots of the flip
interrupt handling stuff is still around.
Cc: Maarten Lankhorst
Cc: Ville Syrjälä
Signed-off-by: Daniel Vetter
---
drivers/gpu/drm/i915/i915_drv.h | 5 -
drivers/gpu/drm/i915/intel_display.c | 656
Taking the modeset locks unconditionally isn't the greatest idea,
because atm that part is still broken and times out (and then atomic
keels over). And there's really no reason to do so, the old code
didn't do that either.
To make the patch a bit simpler let's also nuke 2 cases that are only
aroun
There's no reason to entirely wedge the gpu, for the minimal deadlock
bugfix we only need to unbreak/decouple the atomic commit from the gpu
reset. The simplest wait to fix that is by replacing the
unconditional fence wait a the top of commit_tail by a wait which
completes either when the fences ar
This gets rid of all the interactions between the legacy flip code and
the modeset code. Yay!
Cc: Maarten Lankhorst
Cc: Ville Syrjälä
Signed-off-by: Daniel Vetter
---
drivers/gpu/drm/i915/i915_debugfs.c | 70 -
drivers/gpu/drm/i915/i915_drv.c | 1 -
drivers/gpu/drm
Blocking in a worker is ok, that's what the unbound_wq is for. And it
unifies the paths between the blocking and nonblocking commit, giving
me just one path where I have to implement the deadlock avoidance
trickery in the next patch.
Cc: Chris Wilson
Cc: Mika Kuoppala
Cc: Joonas Lahtinen
Signed
All these races and things are now solved through the vblank evasion
trick, plus event handling is done using normal vblank even processing
and drm_crtc_arm_vblank_event. We can get rid of all this complexity.
Cc: Maarten Lankhorst
Cc: Ville Syrjälä
Signed-off-by: Daniel Vetter
---
drivers/gpu
A bit an oversight - the current code did nothing, since only
legacy flips used the unpin_work_count and assorted logic.
Cc: Maarten Lankhorst
Cc: Ville Syrjälä
Signed-off-by: Daniel Vetter
---
drivers/gpu/drm/i915/intel_display.c | 27 ++-
1 file changed, 14 insertions
The core already does this in setup_commit(). With this we can also
remove the unpin_work_count since it's the last user.
Cc: Maarten Lankhorst
Cc: Ville Syrjälä
Signed-off-by: Daniel Vetter
---
drivers/gpu/drm/i915/intel_display.c | 13 +
drivers/gpu/drm/i915/intel_drv.h | 2
Quoting Daniel Vetter (2017-07-19 13:54:57)
> Blocking in a worker is ok,
but needlessly inefficient,
> that's what the unbound_wq is for. And it
> unifies the paths between the blocking and nonblocking commit, giving
> me just one path where I have to implement the deadlock avoidance
> trickery
Quoting Daniel Vetter (2017-07-19 13:55:00)
> A bit an oversight - the current code did nothing, since only
> legacy flips used the unpin_work_count and assorted logic.
>
> Cc: Maarten Lankhorst
> Cc: Ville Syrjälä
> Signed-off-by: Daniel Vetter
There's a fence deadlock testcase for this, kms_
Quoting Daniel Vetter (2017-07-19 13:55:01)
> This gets rid of all the interactions between the legacy flip code and
> the modeset code. Yay!
Including our missed vblank boosting. (That's been dead for a while.)
-Chris
___
Intel-gfx mailing list
Intel-gf
Quoting Daniel Vetter (2017-07-19 13:55:02)
> The core already does this in setup_commit(). With this we can also
> remove the unpin_work_count since it's the last user.
Also note that the loop only existed for the legacy pageflip support,
with that removed it becomes entirely redundant.
-Chris
__
On Wed, Jul 19, 2017 at 02:04:25PM +0100, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:54:57)
> > Blocking in a worker is ok,
>
> but needlessly inefficient,
>
> > that's what the unbound_wq is for. And it
> > unifies the paths between the blocking and nonblocking commit, giving
>
On Wed, Jul 19, 2017 at 02:06:43PM +0100, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:55:00)
> > A bit an oversight - the current code did nothing, since only
> > legacy flips used the unpin_work_count and assorted logic.
> >
> > Cc: Maarten Lankhorst
> > Cc: Ville Syrjälä
> > Sig
Another kernel, another list of failed backports.
The following commits have been marked as Cc: stable or fixing something
in v4.13-rc1 or earlier, but failed to cherry-pick to
drm-intel-fixes. Please see if they are worth backporting, and please do
so if they are.
BR,
Jani.
54d20ed1fff2 ("drm/
On Wed, Jul 19, 2017 at 02:09:02PM +0100, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:55:02)
> > The core already does this in setup_commit(). With this we can also
> > remove the unpin_work_count since it's the last user.
>
> Also note that the loop only existed for the legacy page
On Wed, Jul 19, 2017 at 02:07:36PM +0100, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:55:01)
> > This gets rid of all the interactions between the legacy flip code and
> > the modeset code. Yay!
>
> Including our missed vblank boosting. (That's been dead for a while.)
Hm right, but
Quoting Daniel Vetter (2017-07-19 13:54:56)
> ... using the biggest hammer we have. This is essentially a weaponized
> version of the timeout-based wedging Chris added in
>
> commit 36703e79a982c8ce5a8e43833291f2719e92d0d1
> Author: Chris Wilson
> Date: Thu Jun 22 11:56:25 2017 +0100
>
> d
Quoting Daniel Vetter (2017-07-19 13:54:58)
> There's no reason to entirely wedge the gpu, for the minimal deadlock
> bugfix we only need to unbreak/decouple the atomic commit from the gpu
> reset. The simplest wait to fix that is by replacing the
> unconditional fence wait a the top of commit_tail
On Wed, Jul 19, 2017 at 3:32 PM, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:54:56)
>> ... using the biggest hammer we have. This is essentially a weaponized
>> version of the timeout-based wedging Chris added in
>>
>> commit 36703e79a982c8ce5a8e43833291f2719e92d0d1
>> Author: Chris
This introduces a common FrameDumpPath configuration field, as well as
helper functions in dedicated igt_frame for writing cairo surfaces
to png files.
Signed-off-by: Paul Kocialkowski
---
lib/Makefile.sources | 2 +
lib/igt.h| 1 +
lib/igt_core.c | 12 +
lib/igt_core.
Changes since v4:
* Moved igt_get_cairo_surface out of the thread function to properly
handle assert failure
* Rebased on top of current master
Changes since v3:
* Renamed structure used by async crc calculation for more clarity
* Used const qualifier for untouched buffer when hashing
* Split ac
This introduces a igt_crc_to_string_extended helper that allows
formatting a crc to a string with a given delimiter and size to print
per crc word.
Signed-off-by: Paul Kocialkowski
---
lib/igt_debugfs.c | 27 +++
lib/igt_debugfs.h | 1 +
2 files changed, 24 insertions(+)
This introduces CRC calculation for reference frames, instead of using
hardcoded values for them. The rendering of reference frames may differ
from machine to machine, especially due to font rendering, and the
frame itself may change with subsequent IGT changes.
These differences would cause the C
This merges the two test_display_crc_single and
test_display_crc_multiple functions into one, with a variable number of
frames to capture. This reduces code duplication.
Signed-off-by: Paul Kocialkowski
---
tests/chamelium.c | 72 +++
1 file ch
This adds support for dumping both the frame capture from the chamelium
and the reference frame generated by cairo when the captured crc does
not match the crc calculated from the reference, using common helpers.
Getting a dump of the frames is quite useful in order to compare them.
Signed-off-by
This removes the igt_write_fb_to_png function (that was unused thus far)
and exports the igt_get_cairo_surface function to grab the matching
cairo surface. Writing to a png is now handled by the common frame
handling code in lib/igt_frame.
This also fixes how the surface is retreived in chamelium
This introduces an igt_check_crc_equal function in addition to
igt_assert_crc_equal and makes the CRC comparison logic from the latter
common. In particular, an igt_find_crc_mismatch function indicates
whether there is a mistmatch and at what index, so that the calling
functions can print the diver
This adds support for analog frame comparison check, as used in VGA.
Since VGA uses a DAC-ADC chain, its data cannot be expected to be pixel
perfect. Thus, it is impossible to uses a CRC check and full frames have
to be analyzed instead. Such an analysis is implemented, based on both
an absolute er
Changes since v2:
* Changed analogue in favor of analog
* Integrated analog frame match fixup in the original commit
* Rebased on top of the new revisions of the series this depends on
Changes since v1:
* Split analogue frame comparison to igt_frame, using cairo surfaces
* Added a chamelium helper
This adds support for VGA frame comparison testing with the reference
generated from cairo. The retrieved frame from the chamelium is first
cropped, as it contains the blanking intervals, through a dedicated
helper. Another helper function asserts that the analog frame
matches or dump it to png if
This patch applies on top of: Analogue/VGA frame comparison support
Changes since v2:
* Changed analogue in favor of analog
* Rebased on top of the new revisions of the series this depends on
Changes since v1:
* Rebased on top of the new revisions of the series this depends on
__
Nowadays, many VGA connectors are not actually native VGA but use a
discrete bridge to a digital connector. These bridges usually enforce
their own EDID instead of the one supplied by the chamelium.
Thus, the EDID read test for VGA is not relevant in that case and
should be skipped. Reported modes
Op 19-07-17 om 14:55 schreef Daniel Vetter:
> The core already does this in setup_commit(). With this we can also
> remove the unpin_work_count since it's the last user.
>
> Cc: Maarten Lankhorst
> Cc: Ville Syrjälä
> Signed-off-by: Daniel Vetter
> ---
> drivers/gpu/drm/i915/intel_display.c | 1
We need to unpin the last retired context early in the shutdown sequence
so that its RCU free is done before we try to free the context ida. I
included this in a later patch ("drm/i915: Keep a recent cache of freed
contexts objects for reuse") and so missed that the selftests were broken
in the mea
Op 19-07-17 om 15:17 schreef Jani Nikula:
> Another kernel, another list of failed backports.
>
> The following commits have been marked as Cc: stable or fixing something
> in v4.13-rc1 or earlier, but failed to cherry-pick to
> drm-intel-fixes. Please see if they are worth backporting, and please
On Wed, Jul 19, 2017 at 3:42 PM, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-19 13:54:58)
>> diff --git a/drivers/gpu/drm/i915/intel_display.c
>> b/drivers/gpu/drm/i915/intel_display.c
>> index 5aa7ca1ab592..4762f158032d 100644
>> --- a/drivers/gpu/drm/i915/intel_display.c
>> +++ b/drive
Quoting Daniel Vetter (2017-07-19 14:15:44)
> On Wed, Jul 19, 2017 at 02:06:43PM +0100, Chris Wilson wrote:
> > Quoting Daniel Vetter (2017-07-19 13:55:00)
> > > A bit an oversight - the current code did nothing, since only
> > > legacy flips used the unpin_work_count and assorted logic.
> > >
> >
1 - 100 of 143 matches
Mail list logo
