mbox series

[RFC,0/3] i915 writeback private framework

Message ID 20220421050756.2036609-1-suraj.kandpal@intel.com (mailing list archive)
Headers show
Series i915 writeback private framework | expand

Message

Suraj Kandpal April 21, 2022, 5:07 a.m. UTC 
A patch series was floated in the drm mailing list which aimed to change
the drm_connector and drm_encoder fields to pointer in the
drm_connector_writeback structure, this received a huge pushback from
the community but since i915 expects each connector present in the
drm_device list to be a intel_connector but drm_writeback framework.
[1] https://patchwork.kernel.org/project/dri-devel/patch/20220202081702.22119-1-suraj.kandpal@intel.com/
[2] https://patchwork.kernel.org/project/dri-devel/patch/20220202085429.22261-6-suraj.kandpal@intel.com/
This forces us to use a drm_connector which is not embedded in
intel_connector the current drm_writeback framework becomes very
unfeasible to us as it would mean a lot of checks at a lot of places
to take into account the above issue.Since no one had an issue with
encoder field being changed into a pointer it was decided to break the
connector and encoder pointer changes into two different series.The
encoder field changes is currently being worked upon by Abhinav Kumar
[3]https://patchwork.kernel.org/project/dri-devel/list/?series=633565
In the meantime for i915 to start using the writeback functionality we
came up with a interim solution to own writeback pipeline bypassing one
provided by drm which is what these patches do.
Note: these are temp patches till we figure out how we can either change
drm core writeback to work with our intel_connector structure or find a
different solution which allows us to work with the current
drm_writeback framework

Suraj Kandpal (3):
  drm/i915: Creating writeback pipeline to bypass drm_writeback
    framework
  drm/i915: Define WD trancoder for i915
  drm/i915: Enabling WD Transcoder

 drivers/gpu/drm/i915/Makefile                 |   2 +
 drivers/gpu/drm/i915/display/intel_acpi.c     |   1 +
 drivers/gpu/drm/i915/display/intel_display.c  |  89 +-
 drivers/gpu/drm/i915/display/intel_display.h  |  15 +
 .../drm/i915/display/intel_display_types.h    |  18 +
 drivers/gpu/drm/i915/display/intel_dpll.c     |   3 +
 drivers/gpu/drm/i915/display/intel_opregion.c |   3 +
 .../gpu/drm/i915/display/intel_wb_connector.c | 296 ++++++
 .../gpu/drm/i915/display/intel_wb_connector.h |  99 ++
 drivers/gpu/drm/i915/display/intel_wd.c       | 978 ++++++++++++++++++
 drivers/gpu/drm/i915/display/intel_wd.h       |  82 ++
 drivers/gpu/drm/i915/i915_drv.h               |   5 +
 drivers/gpu/drm/i915/i915_irq.c               |   8 +-
 drivers/gpu/drm/i915/i915_pci.c               |   7 +-
 drivers/gpu/drm/i915/i915_reg.h               | 139 +++
 15 files changed, 1742 insertions(+), 3 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/display/intel_wb_connector.c
 create mode 100644 drivers/gpu/drm/i915/display/intel_wb_connector.h
 create mode 100644 drivers/gpu/drm/i915/display/intel_wd.c
 create mode 100644 drivers/gpu/drm/i915/display/intel_wd.h

Comments

Suraj Kandpal April 28, 2022, 5:51 a.m. UTC | #1 
++Laurent ,Dmitry, and Abhinav

Hi,
Can you have a look at the private implementation i915 is currently going with till
we can figure out how  to work with drm core .

Regards,
Suraj Kandpal
> A patch series was floated in the drm mailing list which aimed to change the
> drm_connector and drm_encoder fields to pointer in the
> drm_connector_writeback structure, this received a huge pushback from the
> community but since i915 expects each connector present in the drm_device
> list to be a intel_connector but drm_writeback framework.
> [1] https://patchwork.kernel.org/project/dri-
> devel/patch/20220202081702.22119-1-suraj.kandpal@intel.com/
> [2] https://patchwork.kernel.org/project/dri-
> devel/patch/20220202085429.22261-6-suraj.kandpal@intel.com/
> This forces us to use a drm_connector which is not embedded in
> intel_connector the current drm_writeback framework becomes very
> unfeasible to us as it would mean a lot of checks at a lot of places to take into
> account the above issue.Since no one had an issue with encoder field being
> changed into a pointer it was decided to break the connector and encoder
> pointer changes into two different series.The encoder field changes is
> currently being worked upon by Abhinav Kumar
> [3]https://patchwork.kernel.org/project/dri-devel/list/?series=633565
> In the meantime for i915 to start using the writeback functionality we came
> up with a interim solution to own writeback pipeline bypassing one provided
> by drm which is what these patches do.
> Note: these are temp patches till we figure out how we can either change
> drm core writeback to work with our intel_connector structure or find a
> different solution which allows us to work with the current drm_writeback
> framework
> 
> Suraj Kandpal (3):
>   drm/i915: Creating writeback pipeline to bypass drm_writeback
>     framework
>   drm/i915: Define WD trancoder for i915
>   drm/i915: Enabling WD Transcoder
> 
>  drivers/gpu/drm/i915/Makefile                 |   2 +
>  drivers/gpu/drm/i915/display/intel_acpi.c     |   1 +
>  drivers/gpu/drm/i915/display/intel_display.c  |  89 +-
> drivers/gpu/drm/i915/display/intel_display.h  |  15 +
>  .../drm/i915/display/intel_display_types.h    |  18 +
>  drivers/gpu/drm/i915/display/intel_dpll.c     |   3 +
>  drivers/gpu/drm/i915/display/intel_opregion.c |   3 +
>  .../gpu/drm/i915/display/intel_wb_connector.c | 296 ++++++
> .../gpu/drm/i915/display/intel_wb_connector.h |  99 ++
>  drivers/gpu/drm/i915/display/intel_wd.c       | 978 ++++++++++++++++++
>  drivers/gpu/drm/i915/display/intel_wd.h       |  82 ++
>  drivers/gpu/drm/i915/i915_drv.h               |   5 +
>  drivers/gpu/drm/i915/i915_irq.c               |   8 +-
>  drivers/gpu/drm/i915/i915_pci.c               |   7 +-
>  drivers/gpu/drm/i915/i915_reg.h               | 139 +++
>  15 files changed, 1742 insertions(+), 3 deletions(-)  create mode 100644
> drivers/gpu/drm/i915/display/intel_wb_connector.c
>  create mode 100644 drivers/gpu/drm/i915/display/intel_wb_connector.h
>  create mode 100644 drivers/gpu/drm/i915/display/intel_wd.c
>  create mode 100644 drivers/gpu/drm/i915/display/intel_wd.h
> 
> --
> 2.35.1
Laurent Pinchart April 28, 2022, 12:02 p.m. UTC | #2 
Hi Suraj,

On Thu, Apr 28, 2022 at 05:51:47AM +0000, Kandpal, Suraj wrote:
> ++Laurent ,Dmitry, and Abhinav
> 
> Hi,
> Can you have a look at the private implementation i915 is currently going with till
> we can figure out how  to work with drm core .

No, sorry, I barely have time to follow up on core DRM changes, I can't
help with private i915 topics in my spare time.

> > A patch series was floated in the drm mailing list which aimed to change the
> > drm_connector and drm_encoder fields to pointer in the
> > drm_connector_writeback structure, this received a huge pushback from the
> > community but since i915 expects each connector present in the drm_device
> > list to be a intel_connector but drm_writeback framework.
> > [1] https://patchwork.kernel.org/project/dri-
> > devel/patch/20220202081702.22119-1-suraj.kandpal@intel.com/
> > [2] https://patchwork.kernel.org/project/dri-
> > devel/patch/20220202085429.22261-6-suraj.kandpal@intel.com/
> > This forces us to use a drm_connector which is not embedded in
> > intel_connector the current drm_writeback framework becomes very
> > unfeasible to us as it would mean a lot of checks at a lot of places to take into
> > account the above issue.Since no one had an issue with encoder field being
> > changed into a pointer it was decided to break the connector and encoder
> > pointer changes into two different series.The encoder field changes is
> > currently being worked upon by Abhinav Kumar
> > [3]https://patchwork.kernel.org/project/dri-devel/list/?series=633565
> > In the meantime for i915 to start using the writeback functionality we came
> > up with a interim solution to own writeback pipeline bypassing one provided
> > by drm which is what these patches do.
> > Note: these are temp patches till we figure out how we can either change
> > drm core writeback to work with our intel_connector structure or find a
> > different solution which allows us to work with the current drm_writeback
> > framework
> > 
> > Suraj Kandpal (3):
> >   drm/i915: Creating writeback pipeline to bypass drm_writeback
> >     framework
> >   drm/i915: Define WD trancoder for i915
> >   drm/i915: Enabling WD Transcoder
> > 
> >  drivers/gpu/drm/i915/Makefile                 |   2 +
> >  drivers/gpu/drm/i915/display/intel_acpi.c     |   1 +
> >  drivers/gpu/drm/i915/display/intel_display.c  |  89 +-
> > drivers/gpu/drm/i915/display/intel_display.h  |  15 +
> >  .../drm/i915/display/intel_display_types.h    |  18 +
> >  drivers/gpu/drm/i915/display/intel_dpll.c     |   3 +
> >  drivers/gpu/drm/i915/display/intel_opregion.c |   3 +
> >  .../gpu/drm/i915/display/intel_wb_connector.c | 296 ++++++
> > .../gpu/drm/i915/display/intel_wb_connector.h |  99 ++
> >  drivers/gpu/drm/i915/display/intel_wd.c       | 978 ++++++++++++++++++
> >  drivers/gpu/drm/i915/display/intel_wd.h       |  82 ++
> >  drivers/gpu/drm/i915/i915_drv.h               |   5 +
> >  drivers/gpu/drm/i915/i915_irq.c               |   8 +-
> >  drivers/gpu/drm/i915/i915_pci.c               |   7 +-
> >  drivers/gpu/drm/i915/i915_reg.h               | 139 +++
> >  15 files changed, 1742 insertions(+), 3 deletions(-)  create mode 100644
> > drivers/gpu/drm/i915/display/intel_wb_connector.c
> >  create mode 100644 drivers/gpu/drm/i915/display/intel_wb_connector.h
> >  create mode 100644 drivers/gpu/drm/i915/display/intel_wd.c
> >  create mode 100644 drivers/gpu/drm/i915/display/intel_wd.h