[0/2] drm: connector firmware nodes

Message

Heikki Krogerus March 25, 2019, 1:16 p.m. UTC

Hi,

If the system firmware supplies device nodes representing the
display connectors on integrated video hardware, those nodes should
probable always be associated with the display connector entries
(struct drm_connector).

With the USB Type-C DisplayPort alt mode, we will need a way to inform
the correct drm connector entry about HPD IRQ, lane counts and other
details. In ACPI (and most likely in DT too) the device node
representing a DisplayPort behind USB Type-C connector should have a
reference to the device node representing the USB Type-C connector (or
vise versa). Once we have associated the DP connector device nodes
with the drm connector entries, we can use those references to find
the correct drm connector that the information the USB Type-C drivers
are sending is meant for.

Because I think the connector firmware nodes should be associated with
the connector entries in any case (those nodes do seem to be supplying
the connectors all kinds of resources, not only references to other
components), I'm proposing this now instead of waiting for the USB
Type-C patches.

thanks,

Heikki Krogerus (2):
  drm: Add fwnode member to the struct drm_connector
  drm/i915: Associate ACPI connector nodes with connector entries

 drivers/gpu/drm/drm_sysfs.c          | 49 +++++++++++++++++++---------
 drivers/gpu/drm/i915/intel_display.c | 40 +++++++++++++++++++++++
 include/drm/drm_connector.h          |  2 ++
 3 files changed, 76 insertions(+), 15 deletions(-)

Comments

Heikki Krogerus March 26, 2019, 8:32 a.m. UTC | #1

On Mon, Mar 25, 2019 at 08:57:44PM -0000, Patchwork wrote:
> == Series Details ==
> 
> Series: drm: connector firmware nodes
> URL   : https://patchwork.freedesktop.org/series/58531/
> State : failure
> 
> == Summary ==
> 
> CI Bug Log - changes from CI_DRM_5810_full -> Patchwork_12592_full
> ====================================================
> 
> Summary
> -------
> 
>   **FAILURE**
> 
>   Serious unknown changes coming with Patchwork_12592_full absolutely need to be
>   verified manually.
>   
>   If you think the reported changes have nothing to do with the changes
>   introduced in Patchwork_12592_full, please notify your bug team to allow them
>   to document this new failure mode, which will reduce false positives in CI.
> 
>   
> 
> Possible new issues
> -------------------
> 
>   Here are the unknown changes that may have been introduced in Patchwork_12592_full:
> 
> ### IGT changes ###
> 
> #### Possible regressions ####
> 
>   * igt@perf_pmu@idle-no-semaphores-bcs0:
>     - shard-iclb:         PASS -> INCOMPLETE


I have not idea what this is about? Is there something I need to do?


thanks,

Saarinen, Jani March 26, 2019, 3:16 p.m. UTC | #2

> -----Original Message-----
> From: Intel-gfx [mailto:[e-mail]] On Behalf Of Heikki
> Krogerus
> Sent: tiistai 26. maaliskuuta 2019 10.32
> To: [e-mail]
> Subject: Re: [Intel-gfx] ✗ Fi.CI.IGT: failure for drm: connector firmware nodes
> 
> On Mon, Mar 25, 2019 at 08:57:44PM -0000, Patchwork wrote:
> > == Series Details ==
> >
> > Series: drm: connector firmware nodes
> > URL   : https://patchwork.freedesktop.org/series/58531/
> > State : failure
> >
> > == Summary ==
> >
> > CI Bug Log - changes from CI_DRM_5810_full -> Patchwork_12592_full
> > ====================================================
> >
> > Summary
> > -------
> >
> >   **FAILURE**
> >
> >   Serious unknown changes coming with Patchwork_12592_full absolutely need to
> be
> >   verified manually.
> >
> >   If you think the reported changes have nothing to do with the changes
> >   introduced in Patchwork_12592_full, please notify your bug team to allow them
> >   to document this new failure mode, which will reduce false positives in CI.
> >
> >
> >
> > Possible new issues
> > -------------------
> >
> >   Here are the unknown changes that may have been introduced in
> Patchwork_12592_full:
> >
> > ### IGT changes ###
> >
> > #### Possible regressions ####
> >
> >   * igt@perf_pmu@idle-no-semaphores-bcs0:
> >     - shard-iclb:         PASS -> INCOMPLETE
> 
> 
> I have not idea what this is about? Is there something I need to do?
Do you see from logs if relevant to your patche(s) ? 
If not read above. But also copying here:
> >   If you think the reported changes have nothing to do with the changes
> >   introduced in Patchwork_12592_full, please notify your bug team to allow them
> >   to document this new failure mode, which will reduce false positives in CI.


> 
> thanks,
> 
> --
> heikki
> _______________________________________________
> Intel-gfx mailing list
> [e-mail]
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Heikki Krogerus March 27, 2019, 3:38 p.m. UTC | #3

Hi Jani,

On Tue, Mar 26, 2019 at 03:16:29PM +0000, Saarinen, Jani wrote:
> > -----Original Message-----
> > From: Intel-gfx [mailto:[e-mail]] On Behalf Of Heikki
> > Krogerus
> > Sent: tiistai 26. maaliskuuta 2019 10.32
> > To: [e-mail]
> > Subject: Re: [Intel-gfx] ✗ Fi.CI.IGT: failure for drm: connector firmware nodes
> > 
> > On Mon, Mar 25, 2019 at 08:57:44PM -0000, Patchwork wrote:
> > > == Series Details ==
> > >
> > > Series: drm: connector firmware nodes
> > > URL   : https://patchwork.freedesktop.org/series/58531/
> > > State : failure
> > >
> > > == Summary ==
> > >
> > > CI Bug Log - changes from CI_DRM_5810_full -> Patchwork_12592_full
> > > ====================================================
> > >
> > > Summary
> > > -------
> > >
> > >   **FAILURE**
> > >
> > >   Serious unknown changes coming with Patchwork_12592_full absolutely need to
> > be
> > >   verified manually.
> > >
> > >   If you think the reported changes have nothing to do with the changes
> > >   introduced in Patchwork_12592_full, please notify your bug team to allow them
> > >   to document this new failure mode, which will reduce false positives in CI.
> > >
> > >
> > >
> > > Possible new issues
> > > -------------------
> > >
> > >   Here are the unknown changes that may have been introduced in
> > Patchwork_12592_full:
> > >
> > > ### IGT changes ###
> > >
> > > #### Possible regressions ####
> > >
> > >   * igt@perf_pmu@idle-no-semaphores-bcs0:
> > >     - shard-iclb:         PASS -> INCOMPLETE
> > 
> > 
> > I have not idea what this is about? Is there something I need to do?
> Do you see from logs if relevant to your patche(s) ? 

Sorry, but I can't really say, as I don't know how to interpret the
logs. I don't understand what that test is doing.

I'm assuming this log shows the failure:
https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_12592/shard-iclb8/run7.log

The failure happens when running test (I'm guessing):
"[30/82] (859s left) gem_tiled_swapping (non-threaded)"

In the log there is a line that says:
"FATAL: Unable to delete script file /tmp/jenkins7657231043286309557.sh"

That certainly does not sound like anything that these patches would
affect, but since I really have no idea what the test is doing, I
can't really be sure.


thanks,

Tvrtko Ursulin March 27, 2019, 3:59 p.m. UTC | #4

On 27/03/2019 15:38, Heikki Krogerus wrote:
> Hi Jani,
> 
> On Tue, Mar 26, 2019 at 03:16:29PM +0000, Saarinen, Jani wrote:
>>> -----Original Message-----
>>> From: Intel-gfx [mailto:[e-mail]] On Behalf Of Heikki
>>> Krogerus
>>> Sent: tiistai 26. maaliskuuta 2019 10.32
>>> To: [e-mail]
>>> Subject: Re: [Intel-gfx] ✗ Fi.CI.IGT: failure for drm: connector firmware nodes
>>>
>>> On Mon, Mar 25, 2019 at 08:57:44PM -0000, Patchwork wrote:
>>>> == Series Details ==
>>>>
>>>> Series: drm: connector firmware nodes
>>>> URL   : https://patchwork.freedesktop.org/series/58531/
>>>> State : failure
>>>>
>>>> == Summary ==
>>>>
>>>> CI Bug Log - changes from CI_DRM_5810_full -> Patchwork_12592_full
>>>> ====================================================
>>>>
>>>> Summary
>>>> -------
>>>>
>>>>    **FAILURE**
>>>>
>>>>    Serious unknown changes coming with Patchwork_12592_full absolutely need to
>>> be
>>>>    verified manually.
>>>>
>>>>    If you think the reported changes have nothing to do with the changes
>>>>    introduced in Patchwork_12592_full, please notify your bug team to allow them
>>>>    to document this new failure mode, which will reduce false positives in CI.
>>>>
>>>>
>>>>
>>>> Possible new issues
>>>> -------------------
>>>>
>>>>    Here are the unknown changes that may have been introduced in
>>> Patchwork_12592_full:
>>>>
>>>> ### IGT changes ###
>>>>
>>>> #### Possible regressions ####
>>>>
>>>>    * igt@perf_pmu@idle-no-semaphores-bcs0:
>>>>      - shard-iclb:         PASS -> INCOMPLETE
>>>
>>>
>>> I have not idea what this is about? Is there something I need to do?
>> Do you see from logs if relevant to your patche(s) ?
> 
> Sorry, but I can't really say, as I don't know how to interpret the
> logs. I don't understand what that test is doing.
> 
> I'm assuming this log shows the failure:
> https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_12592/shard-iclb8/run7.log
> 
> The failure happens when running test (I'm guessing):
> "[30/82] (859s left) gem_tiled_swapping (non-threaded)"
> 
> In the log there is a line that says:
> "FATAL: Unable to delete script file /tmp/jenkins7657231043286309557.sh"
> 
> That certainly does not sound like anything that these patches would
> affect, but since I really have no idea what the test is doing, I
> can't really be sure.

gem_tiled_swapping caused OOM killer to trigger which first killed the 
test, then killed java, and then CI put the blame on 
igt@perf_pmu@idle-no-semaphores-bcs0 which AFAICS did not even get to run.

Regards,

Tvrtko