> -Original Message-
> From: Chris Wilson [mailto:ch...@chris-wilson.co.uk]
> Sent: Wednesday, July 19, 2017 7:20 PM
> To: Zhang, Tina ; intel-gfx@lists.freedesktop.org;
> intel-
> gvt-...@lists.freedesktop.org; dri-de...@lists.freedesktop.org;
>
On 17-07-07 09:28:08, Jason Ekstrand wrote:
On Thu, Jul 6, 2017 at 4:27 PM, Ben Widawsky wrote:
We don't yet have optimal MOCS settings, but we have enough to know how
to at least determine when we might have non-optimal settings within our
driver.
Signed-off-by: Ben
On 17-07-14 22:10:15, Ville Syrjälä wrote:
On Fri, Jul 14, 2017 at 11:41:49AM -0700, Ben Widawsky wrote:
On 17-06-29 22:49:44, Ville Syrjälä wrote:
[snip]
>
>... but here it's ALIGN(formats_offset+formats_size). I think we should
>be aligning the same thing in both cases, or we add a
Hi all,
Today's linux-next merge of the drm tree got a conflict in:
drivers/gpu/drm/i915/intel_pm.c
between commit:
9cc5bb18bd0a ("drm/i915: Fix bad comparison in skl_compute_plane_wm")
from the drm-intel-fixes tree and commit:
eac2cb81fb87 ("drm/i915: cleanup fixed-point wrappers
Hi all,
The following conflict now exists between the drm and drm-intel trees.
On Thu, 20 Jul 2017 11:23:33 +1000 Stephen Rothwell
wrote:
>
> Today's linux-next merge of the drm-misc tree got a conflict in:
>
> drivers/gpu/drm/i915/i915_drv.c
>
> between commit:
>
>
Hi Dave,
The following is now applicable to the drm and staging.current trees ...
On Wed, 19 Jul 2017 11:46:57 +1000 Stephen Rothwell
wrote:
>
> After merging the drm-misc tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
>
>
Hi Dave,
The conflict below now exists between the drm-misc-fixes tree and the
drm tree.
On Tue, 18 Jul 2017 11:39:46 +1000 Stephen Rothwell
wrote:
>
> Today's linux-next merge of the drm-misc tree got a conflict in:
>
> drivers/gpu/drm/vc4/vc4_crtc.c
>
> between
Hi Dave,
Here's the first -misc-fixes pull for the current cycle. Note that I've based it
off Linus' tree, so you might want to ff drm-fixes before applying.
Pretty self-explanatory set of patches, I'm sure you have a good handle on the
mst group. The vc4 fix comes from drm-misc-next-fixes after
On Thu, Jul 20, 2017 at 07:57:54PM +0200, Daniel Vetter wrote:
> The core already does this in setup_commit(). With this we can also
> remove the unpin_work_count since it's the last user, and also remove
> the loop since that was only used for stalling against legacy flips.
>
> v2: Amend commit
== Series Details ==
Series: drm/i915: Avoid the gpu reset vs. modeset deadlock
URL : https://patchwork.freedesktop.org/series/27670/
State : failure
== Summary ==
Series 27670v1 drm/i915: Avoid the gpu reset vs. modeset deadlock
On Thu, Jul 20, 2017 at 02:48:19PM +0100, Chris Wilson wrote:
> If the request has been completed before the reset took effect, we don't
> need to mark it up as being a victim. Touching fence->error after the
> fence has been signaled is detected by dma_fence_set_error() and
> triggers a BUG:
>
>
== Series Details ==
Series: drm/i915: Don't set fence errors on non-guilty requests (rev2)
URL : https://patchwork.freedesktop.org/series/27665/
State : success
== Summary ==
Series 27665v2 drm/i915: Don't set fence errors on non-guilty requests
On Thu, 2017-07-20 at 10:06 +, Tc, Jenny wrote:
> > >> With older panels, AUX pin for backlight doesn't work.
> >
> > What evidence do you have to back up that claim?
>
> Debugging further it's found that the panel I am having doesn't support AUX
> Backlight.
>
> cat
Oops, sent out the wrong patch, pls disregard.
-Daniel
On Thu, Jul 20, 2017 at 10:14 PM, Daniel Vetter wrote:
> ... at least not in that place, because this can also fire for
> requests which have completed. Which then hits the assert in
> dma_fence_set_error.
>
> Since
... 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 Thu, 2017-07-20 at 12:25 +0300, Jani Nikula wrote:
> This reverts commit 560a758d39c616f83ac25ff6e0816a49ebe6401c.
>
> The DPCD backlight commits regress a Thinkpad X1 Carbon 4th Gen and a
> BXT-P (in CI). Enabling dynamic backlight boots to a black screen, and
> enabling DPCD backlight
On Thu, Jul 20, 2017 at 10:16 PM, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-20 21:04:50)
>> On Thu, Jul 20, 2017 at 9:47 PM, Chris Wilson
>> wrote:
>> >> diff --git a/drivers/gpu/drm/i915/intel_display.c
>> >>
... at least not in that place, because this can also fire for
requests which have completed. Which then hits the assert in
dma_fence_set_error.
Since I'm not really clear what should be done here I pinged Chris,
and he suggested we simply remove this for now.
Found while stress-testing gpu
Quoting Daniel Vetter (2017-07-20 21:04:50)
> On Thu, Jul 20, 2017 at 9:47 PM, Chris Wilson
> wrote:
> >> diff --git a/drivers/gpu/drm/i915/intel_display.c
> >> b/drivers/gpu/drm/i915/intel_display.c
> >> index 02b1f4966049..995522e40ec1 100644
> >> ---
On Thu, Jul 20, 2017 at 9:47 PM, Chris Wilson wrote:
>> diff --git a/drivers/gpu/drm/i915/intel_display.c
>> b/drivers/gpu/drm/i915/intel_display.c
>> index 02b1f4966049..995522e40ec1 100644
>> --- a/drivers/gpu/drm/i915/intel_display.c
>> +++
Quoting Daniel Vetter (2017-07-20 18:57:48)
> ... 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
Quoting Patchwork (2017-07-20 19:14:57)
> == Series Details ==
>
> Series: gpu reset and page_flip removal, take 2
> URL : https://patchwork.freedesktop.org/series/27664/
> State : success
>
> == Summary ==
>
> Series 27664v1 gpu reset and page_flip removal, take 2
>
Somehow a few formatting issues crept in that rst complained about in
commit 7b414ba9e7aa75fc4dff6597036ff758031b9bdf (HEAD -> maintainer-tools)
Author: Daniel Vetter
Date: Thu Jul 20 21:28:12 2017 +0200
drm-intel.rst: Fix formatting
Cc: Rodrigo Vivi
== Series Details ==
Series: drm/i915: Don't set fence errors on non-guilty requests
URL : https://patchwork.freedesktop.org/series/27665/
State : success
== Summary ==
Series 27665v1 drm/i915: Don't set fence errors on non-guilty requests
Quoting Daniel Vetter (2017-07-20 18:59:01)
> ... at least not in that place, because this can also fire for
> requests which have completed. Which then hits the assert in
> dma_fence_set_error.
>
> Since I'm not really clear what should be done here I pinged Chris,
> and he suggested we simply
On Thu, 2017-07-20 at 18:35 +0300, David Weinehall wrote:
> On 2017-07-20 12:33, Jani Nikula wrote:
> > On Wed, 19 Jul 2017, "Pandiyan, Dhinakaran"
> > wrote:
> >> On Wed, 2017-07-19 at 15:59 +0530, Jenny TC wrote:
> >>> With older panels, AUX pin for backlight
== Series Details ==
Series: gpu reset and page_flip removal, take 2
URL : https://patchwork.freedesktop.org/series/27664/
State : success
== Summary ==
Series 27664v1 gpu reset and page_flip removal, take 2
https://patchwork.freedesktop.org/api/1.0/series/27664/revisions/1/mbox/
Test
On Fri, Jul 07, 2017 at 09:28:11AM -0400, Sean Paul wrote:
> On Thu, Jul 6, 2017 at 9:00 AM, Daniel Vetter wrote:
> > From: Thierry Reding
> >
> > The FB helper core now supports deferred setup, so the driver's custom
> > implementation can be removed.
... at least not in that place, because this can also fire for
requests which have completed. Which then hits the assert in
dma_fence_set_error.
Since I'm not really clear what should be done here I pinged Chris,
and he suggested we simply remove this for now.
Found while stress-testing gpu
This gets rid of all the interactions between the legacy flip code and
the modeset code. Yay!
This highlights an ommission in the atomic paths, where we fail to
apply a boost to the pending rendering when we miss the target vblank.
But the existing code is still dead and can be removed.
v2: Note
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
Reviewed-by: Maarten Lankhorst
The core already does this in setup_commit(). With this we can also
remove the unpin_work_count since it's the last user, and also remove
the loop since that was only used for stalling against legacy flips.
v2: Amend commit message a bit (Chris).
Cc: Maarten Lankhorst
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 way 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
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
... 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
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.
I first tried to implement the following patch without this
Hi all,
I landed the locking change, which fixes the gpu reset deadlocks at least for
anything modern (and non-igt). Would be great if we can get the first patch
landed fast, just to address the regression in CI and get rid of the dmesg-warn.
Currently we're blind to dmesg noise in a lot of gpu
For the whole series:
Reviewed-by: Lyude
And pushed
On Thu, 2017-07-20 at 17:11 +0300, Paul Kocialkowski wrote:
> This moves all the pieces related to config parsing to the dedicated
> function for this purpose, renamed common_init_config for
> consistency.
>
> It allows
For the whole series:
Reviewed-by: Lyude
I've also just pushed them, cheers!
On Thu, 2017-07-20 at 18:13 +0300, Paul Kocialkowski wrote:
> Changes since v3:
> * Squashed configure.ac changes in this series
>
> Changes since v2:
> * Changed analogue in favor of analog
> *
On 7/20/2017 5:51 AM, Chris Wilson wrote:
Quoting Michel Thierry (2017-07-18 01:15:00)
On 17/07/17 02:11, Chris Wilson wrote:
We should only ever do nop_submit_request when the machine is wedged, so
assert it is so.
Signed-off-by: Chris Wilson
---
On Thu, Jul 20, 2017 at 1:27 AM, Daniel Vetter wrote:
> On Wed, Jul 19, 2017 at 10:39:28AM -0700, Matthias Kaehlcke wrote:
>> Commit a21960339c8c ("drm/i915: Consistently use enum pipe for PCH
>> transcoders") misses some pieces, due to a problem with the patch
>> format, this
Op 16-03-17 om 09:47 schreef Daniel Vetter:
> On Wed, Mar 15, 2017 at 09:23:30AM +0100, Maarten Lankhorst wrote:
>> Op 14-03-17 om 07:50 schreef Daniel Vetter:
>>> On Mon, Mar 13, 2017 at 12:08:19PM +0100, Maarten Lankhorst wrote:
Op 13-03-17 om 11:15 schreef Daniel Vetter:
> On Thu, Mar
On Mon, Jul 17, 2017 at 08:06:25PM +0530, Shashank Sharma wrote:
> To support ycbcr output, we need a pipe CSC block to do
> RGB->YCBCR conversion.
>
> Current Intel platforms have only one pipe CSC unit, so
> we can either do color correction using it, or we can perform
> RGB->YCBCR conversion.
On Mon, Jul 17, 2017 at 08:06:24PM +0530, Shashank Sharma wrote:
> To get HDMI YCBCR420 output, the PIPEMISC register should be
> programmed to:
> - Generate YCBCR output (bit 11)
> - In case of YCBCR420 outputs, it should be programmed in full
> blend mode to use the scaler in 5x3 ratio (bits
Op 20-07-17 om 15:15 schreef Maarten Lankhorst:
> The watermarks it should calculate against are the old optimal watermarks.
> The currently active crtc watermarks are pure fiction, and are invalid in
> case of a nonblocking modeset, page flip enabling/disabling planes or any
> other reason.
>
>
Hi everyone,
As some of you may already know, we have made great strides in making
our CI system usable, especially in the last 6 months when everything
started clicking together.
The CI team is no longer overwhelmed with fires and bug reports, so we
started working on increasing the
== Series Details ==
Series: drm/i915: Don't touch fence->error when resetting an innocent request
(rev2)
URL : https://patchwork.freedesktop.org/series/27641/
State : success
== Summary ==
Series 27641v2 drm/i915: Don't touch fence->error when resetting an innocent
request
== Series Details ==
Series: drm/i915: Stop using get_existing_state.
URL : https://patchwork.freedesktop.org/series/27640/
State : success
== Summary ==
Series 27640v1 drm/i915: Stop using get_existing_state.
https://patchwork.freedesktop.org/api/1.0/series/27640/revisions/1/mbox/
Test
On Mon, Jul 17, 2017 at 08:06:23PM +0530, Shashank Sharma wrote:
> To get a YCBCR420 output from intel platforms, we need one
> scaler to scale down YCBCR444 samples to YCBCR420 samples.
>
> This patch:
> - Does scaler allocation for HDMI ycbcr420 outputs.
> - Programs PIPE_MISC register for
On 2017-07-20 12:33, Jani Nikula wrote:
On Wed, 19 Jul 2017, "Pandiyan, Dhinakaran"
wrote:
On Wed, 2017-07-19 at 15:59 +0530, Jenny TC wrote:
With older panels, AUX pin for backlight doesn't work.
What evidence do you have to back up that claim?
Chapter 10.1
== Series Details ==
Series: drm/i915: Calculate vlv/chv intermediate watermarks correctly, v3.
URL : https://patchwork.freedesktop.org/series/27639/
State : failure
== Summary ==
CHK include/config/kernel.release
CHK include/generated/uapi/linux/version.h
CHK
On Wed, 2017-07-19 at 13:50 -0400, Lyude Paul wrote:
> For both patches once you squash the configure.ac changes into the
> first one:
>
> Reviewed-by: Lyude
>
> Once you post the new versions I'll just go ahead and push them
I think it makes more sense to split the
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
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
Changes since v3:
* Squashed configure.ac changes in this series
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
On Tue, Jul 18, 2017 at 11:56:23AM -0700, Puthikorn Voravootivat wrote:
> May be the older panel might not work well with this feature.
> David/Jani, what do you think about adding check that the panel is eDP
> 1.4 or later in the heuristic?
While this patch might seem correct on the surface
Acked-by: Shashank Sharma
Regards
Shashank
On 7/20/2017 4:20 AM, Imre Deak wrote:
The scaler allocation code depends on a non-zero default value for the
crtc scaler_id, so make sure we initialize the scaler state accordingly
even if the crtc is off. This fixes at
On Wed, 2017-07-19 at 23:11 -0700, Manasi Navare wrote:
> On Tue, Jul 18, 2017 at 03:11:42PM +0300, Paul Kocialkowski wrote:
> > On Mon, 2017-06-26 at 15:32 +0300, Paul Kocialkowski wrote:
> > > After detecting an IRQ storm, hotplug detection will switch from
> > > irq-based detection to
On Thu, 2017-07-20 at 01:04 +, Pandiyan, Dhinakaran wrote:
> On Mon, 2017-06-26 at 15:32 +0300, Paul Kocialkowski wrote:
> > After detecting an IRQ storm, hotplug detection will switch from
> > irq-based detection to poll-based detection. After a short delay or
> > when resetting storm
Moving to intel-gfx list, intel-gfx-bugs is mainly just for the
automated emails from bugzilla. And please let's keep the debugging in
the bug.
BR,
Jani.
On Thu, 20 Jul 2017, sdrb wrote:
> Hello,
>
> I have got some problems with intel drm frame buffer on Linux virtual
>
On Wed, 19 Jul 2017, Maarten Lankhorst
wrote:
> 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
emon_crash should skip if the debugfs file could not be opened the first
time, and debugfs_test.read_all_entries should skip files that could not
be opened, instead of returning an error.
This is because in a typical IGT run there may be more debugfs files
registered than can be opened.
This is
Every time we add something to debugfs, we test on the new platform
but forget to test that old platforms still work.
The test adds at most 200 ms extra time, which is worth it considering
how often we break debugfs.
Signed-off-by: Maarten Lankhorst
---
This patch adds a flowchart to the drm-misc documentation to
help committers decide which branch is most appropriate for a
given patch.
Signed-off-by: Sean Paul
---
.gitignore | 1 +
Makefile | 2 +-
drm-misc-commit-flow.dot | 22
Chris Wilson writes:
> Quoting Mika Kuoppala (2017-07-20 14:31:31)
>> Chris Wilson writes:
>>
>> > When we wedge the device, we clear out the in-flight requests and
>> > advance the breadcrumb to indicate they are complete. However, the
>> >
On Thu, 2017-07-20 at 09:24 +0200, Daniel Vetter wrote:
> On Wed, Jul 19, 2017 at 04:46:07PM +0300, Paul Kocialkowski wrote:
> > This introduces a common FrameDumpPath configuration field, as well
> > as
> > helper functions in dedicated igt_frame for writing cairo surfaces
> > to png files.
> >
This moves all the pieces related to config parsing to the dedicated
function for this purpose, renamed common_init_config for consistency.
It allows making the common_init function less big and more readable.
Signed-off-by: Paul Kocialkowski
---
This moves the parts of the code doing env-related handling from
common_init to a new dedicated common_init_env function, making
common_init a bit more readable.
Signed-off-by: Paul Kocialkowski
---
lib/igt_core.c | 46
This adds missing entries for documentation generation, both for tests
and the API reference.
The list of tests is made complete and ordered alphabetically, with
modified descriptions for consistency.
More files are added to the API reference, with a minimalistic
description block added to them
Am Donnerstag, den 20.07.2017, 14:51 +0200 schrieb Daniel Vetter:
> It makes debugging a massive pain.
It is also considered very bad style to BUG the kernel on anything other
than filesystem eating catastrophic failures.
Reviewed-by: Lucas Stach
> Signed-off-by: Daniel
Quoting Mika Kuoppala (2017-07-20 14:31:31)
> Chris Wilson writes:
>
> > When we wedge the device, we clear out the in-flight requests and
> > advance the breadcrumb to indicate they are complete. However, the
> > breadcrumb advance includes an assert that the engine is
If the request has been completed before the reset took effect, we don't
need to mark it up as being a victim. Touching fence->error after the
fence has been signaled is detected by dma_fence_set_error() and
triggers a BUG:
[ 231.743133] kernel BUG at ./include/linux/dma-fence.h:434!
[
Chris Wilson writes:
> After setting the WEDGED bit, make sure that we do wake up waiters as
> they may not be waiting for a request completion yet, just for its
> execution.
>
> Signed-off-by: Chris Wilson
Reviewed-by: Mika Kuoppala
== Series Details ==
Series: dma-fence: Don't BUG_ON when not absolutely needed
URL : https://patchwork.freedesktop.org/series/27636/
State : success
== Summary ==
Series 27636v1 dma-fence: Don't BUG_ON when not absolutely needed
Chris Wilson writes:
> The engine also provides a mirror of the CSB write pointer in the HWSP,
> but not of our read pointer. To take advantage of this we need to
> remember where we read up to on the last interrupt and continue off from
> there. This poses a problem
Quoting Chris Wilson (2017-07-20 14:24:29)
> If the request has been completed before the reset took effect, we don't
> need to mark it up as being a victim. Touching fence->error after the
> fence has been signaled is detected by dma_fence_set_error() and
> triggers a BUG:
>
> [ 231.743133]
Chris Wilson writes:
> When we wedge the device, we clear out the in-flight requests and
> advance the breadcrumb to indicate they are complete. However, the
> breadcrumb advance includes an assert that the engine is idle, so that
> advancement needs to be the last step
If the request has been completed before the reset took effect, we don't
need to mark it up as being a victim. Touching fence->error after the
fence has been signaled is detected by dma_fence_set_error() and
triggers a BUG:
[ 231.743133] kernel BUG at ./include/linux/dma-fence.h:434!
[
== Series Details ==
Series: drm/i915: Only mark the execobject as pinned on success
URL : https://patchwork.freedesktop.org/series/27634/
State : failure
== Summary ==
Series 27634v1 drm/i915: Only mark the execobject as pinned on success
Chris Wilson writes:
> intel_engine_init_globa_seqno() may be called from an uncontrolled
> set-wedged path where we have given up waiting for broken hw and declare
> it defunct. Along that path, any sanity checks that the hw is idle
> before we adjust its state will
All the references to get_existing_state can be converted to
get_new_state or get_old_state, which means that i915 is now
get_existing_state free.
Signed-off-by: Maarten Lankhorst
---
drivers/gpu/drm/i915/intel_display.c | 57
get_existing_crtc_state is currently unused, but the next commit
needs to access the old intel state. Rename this and use this
as convenience wrapper.
Signed-off-by: Maarten Lankhorst
---
drivers/gpu/drm/i915/intel_drv.h | 6 +++---
1 file changed, 3
Remove the use of plane->state and drm_atomic_get_existing_state,
instead use the new helpers, and also add
intel_atomic_get_new_crtc_state as it's needed.
Signed-off-by: Maarten Lankhorst
---
drivers/gpu/drm/i915/intel_atomic_plane.c | 22
The previous commit added intel_atomic_get_new_crtc_state,
convert intel_fbc.c to the new helper.
Signed-off-by: Maarten Lankhorst
---
drivers/gpu/drm/i915/intel_fbc.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git
The watermarks it should calculate against are the old optimal watermarks.
The currently active crtc watermarks are pure fiction, and are invalid in
case of a nonblocking modeset, page flip enabling/disabling planes or any
other reason.
When the crtc is disabled or during a modeset the
FBC has been converted to atomic, so updating
some legacy variables is no longer needed.
drm_atomic_helper_update_legacy_modeset_state
does update crtc->x/y anyway, but we shouldn't need it.
Signed-off-by: Maarten Lankhorst
---
The watermarks it should calculate against are the old optimal watermarks.
The currently active crtc watermarks are pure fiction, and are invalid in
case of a nonblocking modeset, page flip enabling/disabling planes or any
other reason.
When the crtc is disabled or during a modeset the
The get_existing macros are deprecated and should be replaced by
get_old/new_state for clarity.
Signed-off-by: Maarten Lankhorst
---
drivers/gpu/drm/i915/intel_atomic.c | 4 ++--
drivers/gpu/drm/i915/intel_drv.h| 4 ++--
drivers/gpu/drm/i915/intel_pm.c
drm_atomic_get_existing_state should be removed, so stop using it in i915.
Fortunately all places can be converted to use the new state or old state,
even removing some dereferncing of obj->state in the process.
i915 no longer depends on plane->fb, so patch 6 removes the assignment.
The first 6
Chris Wilson writes:
> When doing a GPU reset, the CSB register will be trashed and we will
> lose any context-switch notifications that happened since the tasklet
> was disabled. If we find that all requests on this engine were
> completed, we want to make sure that
On Thu, Jul 06, 2017 at 05:40:29PM +0300, Imre Deak wrote:
> Follow-up patches will add new fields to the i915_power_well struct that
> are specific to the hsw_power_well_ops helpers. Prepare for this by
> changing the generic 'data' field to a union of platform specific
> structs.
>
>
On Tue, Jul 11, 2017 at 11:42:33PM +0300, Imre Deak wrote:
> Check that all the power well IDs are unique on the given platform.
>
> v2:
> - Fix using BIT_ULL() instead of BIT() for 64 bit mask.
> v3:
> - Move the check to a separate function. (Ville)
>
> Signed-off-by: Imre Deak
The watermarks it should calculate against are the old optimal watermarks.
The currently active crtc watermarks are pure fiction, and are invalid in
case of a nonblocking modeset, page flip enabling/disabling planes or any
other reason.
When the crtc is disabled or during a modeset the
== Series Details ==
Series: drm/i915: Remove assertion from raw __i915_vma_unpin()
URL : https://patchwork.freedesktop.org/series/27632/
State : success
== Summary ==
Series 27632v1 drm/i915: Remove assertion from raw __i915_vma_unpin()
Quoting Michel Thierry (2017-07-18 01:22:28)
> On 17/07/17 02:11, Chris Wilson wrote:
> > Although a banned context will be told to -EIO off if they try to submit
> > more requests, we have a discrepancy between whole device resets and
> > per-engine resets where we report the GPU reset but not
It makes debugging a massive pain.
Signed-off-by: Daniel Vetter
Cc: Sumit Semwal
Cc: Gustavo Padovan
Cc: linux-me...@vger.kernel.org
Cc: linaro-mm-...@lists.linaro.org
---
drivers/dma-buf/dma-fence.c | 4 ++--
== Series Details ==
Series: series starting with [1/7] drm/i915: Avoid the gpu reset vs. modeset
deadlock
URL : https://patchwork.freedesktop.org/series/27631/
State : success
== Summary ==
Series 27631v1 Series without cover letter
Quoting Michel Thierry (2017-07-18 01:15:00)
> On 17/07/17 02:11, Chris Wilson wrote:
> > We should only ever do nop_submit_request when the machine is wedged, so
> > assert it is so.
> >
> > Signed-off-by: Chris Wilson
> > ---
> > drivers/gpu/drm/i915/i915_gem.c | 1
On Thu, 2017-07-20 at 14:41 +0200, Daniel Vetter wrote:
> On Thu, Jul 20, 2017 at 1:27 PM, Paul Kocialkowski
> wrote:
> > On Thu, 2017-07-20 at 12:39 +0300, Jani Nikula wrote:
> > > For future reference, please post new versions of the entire
> > > series as
> >
On Thu, Jul 20, 2017 at 1:27 PM, Paul Kocialkowski
wrote:
> On Thu, 2017-07-20 at 12:39 +0300, Jani Nikula wrote:
>> For future reference, please post new versions of the entire series as
>> new threads. When posting new versions of just some individual
>>
1 - 100 of 152 matches
Mail list logo