Chris Wilson writes:
> Similar to the staging around handling of engine->submit_request, we
> need to stop adding to the execlists->queue prior to calling
> engine->cancel_requests. cancel_requests will move requests from the
> queue onto the timeline, so if we add a
For some reason, I've reviewed these from the middle of the series
(maybe transport delay?). Are the rest still applicable or refreshed
somewhere?
Regards, Joonas
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
On Mon, Mar 05, 2018 at 12:39:58PM +0530, Sagar Arun Kamble wrote:
> Overall change looks good. Could you please clarify on below:
>
> intel_uc_log_register|unregister are removed in patch later in the series.
> Should we just stay with inner functions then to minimize changes?
I've done this to
== Series Details ==
Series: drm/i915/guc: Removed unused GuC parameters. (rev2)
URL : https://patchwork.freedesktop.org/series/39154/
State : failure
== Summary ==
Series 39154v2 drm/i915/guc: Removed unused GuC parameters.
== Series Details ==
Series: drm/i915: Handle changing enable_fbc parameter at runtime better.
URL : https://patchwork.freedesktop.org/series/39375/
State : success
== Summary ==
Series 39375v1 drm/i915: Handle changing enable_fbc parameter at runtime better.
It looks that GuC does not actively use GUC_CTL_DEVICE_INFO parameter
where we are passing GT type and Core family values.
Lets stop setup this parameter and remove related definitions.
v2: (this time without squashed HAX)
- New title and description
- Remove also GUC_CORE_FAMILY_*
On Thu, Mar 01, 2018 at 11:17:50PM +, Chris Wilson wrote:
> Quoting Arkadiusz Hiler (2018-03-01 09:47:06)
> > Hey all,
> >
> > Since not so long ago our CI is running and reporting sparse and
> > checkpatch. Sparse is doing just fine but I had to disable checkpatch
> > for the time being -
On Mon, Mar 05, 2018 at 01:10:21PM +0200, Jani Nikula wrote:
> On Mon, 05 Mar 2018, Daniel Vetter wrote:
> > I'd recommend not making checkpatch ever fail CI, but at most warning.
>
> Agreed. But we want the automated warnings on the list, neutrally from a
> bot instead of a
From: Tvrtko Ursulin
Request tracepoints have been renames so update the tool to be able to
find them.
Also support falling back to the old name for compatibility.
Signed-off-by: Tvrtko Ursulin
---
Compile tested only so also looking for
If i915.enable_fbc is cleared at runtime, but FBC was previously enabled
then we don't disable FBC until the next time the crtc is disabled.
Make sure that if the module param is changed, we disable FBC in
intel_fbc_post_update so we never have to worry about disabling.
Signed-off-by: Maarten
On 05/03/2018 11:21, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-03-05 11:12:45)
On 05/03/2018 10:41, Chris Wilson wrote:
After we call dma_fence_signal(), confirm that the request was indeed
complete.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
On 05/03/2018 11:18, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2018-03-05 11:06:23)
On 24/02/2018 02:59, Weinan Li wrote:
There is one corner case missing schedule out notification of the preempted
request. The preempted request is just completed when preemption happen,
then it will be
Quoting Ville Syrjälä (2018-03-05 11:12:12)
> On Mon, Mar 05, 2018 at 10:33:12AM +, Chris Wilson wrote:
> > If we fail to acquire a fence when we must, we must unwind before
> > reporting the error. Otherwise, we lose tracking of the vma pinning and
> > eventually hit a bug like
> >
> > <3>[
On Mon, Mar 05, 2018 at 05:28:33PM +0530, Sagar Arun Kamble wrote:
>
>
> On 2/27/2018 6:22 PM, Michał Winiarski wrote:
> > Having both guc_flush_logs and guc_log_flush functions is confusing.
> > While we could just rename things, guc_flush_logs implementation is
> > quite simple. Let's get rid
== Series Details ==
Series: drm/i915: Assert that the request is indeed complete when signaled from
irq
URL : https://patchwork.freedesktop.org/series/39369/
State : success
== Summary ==
Series 39369v1 drm/i915: Assert that the request is indeed complete when
signaled from irq
On 2/27/2018 6:22 PM, Michał Winiarski wrote:
Having both guc_flush_logs and guc_log_flush functions is confusing.
While we could just rename things, guc_flush_logs implementation is
quite simple. Let's get rid of it and move its content to unregister.
Signed-off-by: Michał Winiarski
== Series Details ==
Series: drm/i915: Unwind vma pinning for intel_pin_and_fence_fb_obj error path
URL : https://patchwork.freedesktop.org/series/39367/
State : success
== Summary ==
Series 39367v1 drm/i915: Unwind vma pinning for intel_pin_and_fence_fb_obj
error path
On 2/27/2018 6:22 PM, Michał Winiarski wrote:
We have all the information we need at relay_open call time.
Since there's no reason to split the process into relay_open and
relay_late_setup_files, let's remove the extra code.
Signed-off-by: Michał Winiarski
Cc:
Quoting Tvrtko Ursulin (2018-03-05 11:12:45)
>
> On 05/03/2018 10:41, Chris Wilson wrote:
> > After we call dma_fence_signal(), confirm that the request was indeed
> > complete.
> >
> > Signed-off-by: Chris Wilson
> > Cc: Tvrtko Ursulin
> >
Quoting Tvrtko Ursulin (2018-03-05 11:06:23)
>
> On 24/02/2018 02:59, Weinan Li wrote:
> > There is one corner case missing schedule out notification of the preempted
> > request. The preempted request is just completed when preemption happen,
> > then it will be canceled and won't be resubmitted
On 05/03/2018 10:41, Chris Wilson wrote:
After we call dma_fence_signal(), confirm that the request was indeed
complete.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
drivers/gpu/drm/i915/i915_irq.c | 1 +
1 file changed, 1
On Fri, 2018-03-02 at 12:53 +0530, Sagar Arun Kamble wrote:
>
> On 3/2/2018 12:44 AM, John Spotswood wrote:
> > On Thu, 2018-03-01 at 17:35 +0530, Sagar Arun Kamble wrote:
> > > On 3/1/2018 1:32 PM, Chris Wilson wrote:
> > > > Quoting Michel Thierry (2018-02-28 22:07:51)
> > > > > On 28/02/18
On Mon, Mar 05, 2018 at 10:33:12AM +, Chris Wilson wrote:
> If we fail to acquire a fence when we must, we must unwind before
> reporting the error. Otherwise, we lose tracking of the vma pinning and
> eventually hit a bug like
>
> <3>[ 46.163202] i915_vma_unpin:333
On Mon, 05 Mar 2018, Daniel Vetter wrote:
> I'd recommend not making checkpatch ever fail CI, but at most warning.
Agreed. But we want the automated warnings on the list, neutrally from a
bot instead of a developer spending time nitpicking this stuff. And the
committers should
On 24/02/2018 02:59, Weinan Li wrote:
There is one corner case missing schedule out notification of the preempted
request. The preempted request is just completed when preemption happen,
then it will be canceled and won't be resubmitted later, GVT-g will lost
the schedule out notification.
It appears that waiting for a 100us period whereby we are unable to
submit another batch and proclaim the ring full, may have the false
positive where the scheduler intervenes and we are signalled twice
before having slept on ring space. Increasing the interval reduces the
likelihood of the
On Mon, 05 Mar 2018, "Kumar, Mahesh" wrote:
> Please review.
Pushed to drm-intel-next-queued, thanks for the patch.
Personally, I would have split this into 2-3 patches: 1) code movement
to allow 2) abstraction of the function and 3) functional change of the
limit on
After we call dma_fence_signal(), confirm that the request was indeed
complete.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
drivers/gpu/drm/i915/i915_irq.c | 1 +
1 file changed, 1 insertion(+)
diff --git
If we fail to acquire a fence when we must, we must unwind before
reporting the error. Otherwise, we lose tracking of the vma pinning and
eventually hit a bug like
<3>[ 46.163202] i915_vma_unpin:333 GEM_BUG_ON(!i915_vma_is_pinned(vma))
<4>[ 46.163424] [ cut here ]
<2>[
On 2/27/2018 6:22 PM, Michał Winiarski wrote:
Currently, we're treating relay and mapping of GuC log as a separate
concepts. We're also using inconsistent locking, sometimes using
relay_lock, sometimes using struct mutex.
Let's correct that. Anything touching the runtime is now serialized
Actually check the error state exists (!"No error state captured") and
that it contains the expected engine dump.
v2: Throw in some debug clues.
Signed-off-by: Chris Wilson
---
tests/drv_hangman.c | 12
1 file changed, 12 insertions(+)
diff --git
On Wed, Feb 28, 2018 at 03:51:35PM +, Chris Wilson wrote:
> If we do a global wait while trying to execute spinners in parallel,
> it ends badly with a GPU hang.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104352
> Signed-off-by: Chris Wilson
Op 21-02-18 om 20:39 schreef Harry Wentland:
> On 2018-02-21 01:36 PM, Daniel Vetter wrote:
>> On Wed, Feb 21, 2018 at 04:23:31PM +0100, Maarten Lankhorst wrote:
>>> After we acquired all generic modeset locks in drm_modeset_lock_all, it's
>>> unsafe acquire any other so just mark acquisition as
On 2/27/2018 6:22 PM, Michał Winiarski wrote:
The GuC log contains a separate space used for crash dump.
We even get a separate notification for it. While we're not handling
crash differently yet, it makes sense to decouple the two right now to
simplify the following patches.
Signed-off-by:
On Thu, Mar 01, 2018 at 11:47:06AM +0200, Arkadiusz Hiler wrote:
> Hey all,
>
> Since not so long ago our CI is running and reporting sparse and
> checkpatch. Sparse is doing just fine but I had to disable checkpatch
> for the time being - too much "false" positives causing people to
> complain.
On Fri, Mar 02, 2018 at 04:22:15PM -0500, Sean Paul wrote:
> On Wed, Feb 28, 2018 at 3:34 PM, Sean Paul wrote:
> >
> > Hi Dave,
> > Here's this weeks pull, relatively small when you pull out the trivial
> > fixes.
> >
> > drm-misc-next-2018-02-28:
> > drm-misc-next for
101 - 136 of 136 matches
Mail list logo