Hi Inki,

On 27.02.2014 05:43, Inki Dae wrote:

Hi Tomasz,


2014-02-08 11:48 GMT+09:00 Tomasz Figa <tomasz.figa@gmail.com>:

On 06.02.2014 20:54, Olof Johansson wrote:

On Thu, Jan 30, 2014 at 1:18 PM, Sean Paul <seanpaul@chromium.org> wrote:

This patchset refactors parts of the exynos driver to move it closer to a
proper
drm driver (rather than just implementing a drm layer on top of the
hardware
drivers). The hope is to get to a point where the dp/hdmi drivers can
implement
drm_connector/drm_encoder directly, and fimd/mixer can directly implement
drm_crtc.

The notable changes in this set:
          - drm_encoder funcs no longer route through the crtc
          - DP driver has been moved from video into the drm driver
          - fimd no longer implements encoder callbacks
          - exynos_drm_hdmi is removed in favor of generic manager/display
            handling

At a glance, differences between v1 and v2:
          - Passing manager/display in callbacks instead of ctx
          - Tacked on some dpms patches on the end to handle
suspend/resume
            through the dpms path

At a glance, differences between v2 and v3:
          - Fixed vidi issues
          - Moved exynos_drm_hdmi.c removal to the right place
          - Added the exynos_drm_connector removal patches on the end

At a glance, differences between v3 and v4:
          - Rebased on top of exynos-drm-next
          - Addressed review comments, no major functional changes

Sean

Even if there is room for even more cleanups and refactorings on top
of this, it's good enough as a base and keeping it out of tree is
hurting productivity and usefulness for everybody. Let's get this
merged for 3.15 so that we can make progress with the rest of the
platform!

I wouldn't be overly excited about merging this series yet...

If I have applied it correctly (and I believe so, as reading the code tells
me the same), it introduces a regression due to removal of support of
standard parallel displays from FIMD driver.

I had mentioned this in my review comments for previous version, but
apparently I got ignored. Please fix this.

Right, and also some my comments. Sean should have commented his
opinions even if our comments aren't  reasonable to him.

I sent a email about the progress of this refactoring patch series to
Sean and Olof personally because there was no any follow-up action for
about two months since v3 had been posted, and I found out that Sean
was busy with other works.

And there are no any his comments until now so I'd like to believe
Sean would still be busy. So I already started to clean up and enhance
Exynos drm framework using the component framework for super device
posted by Russell like below,
         https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2a41e6070dd7ef539d0f3b1652b4839d04378e11

Last month, Russell had posted this framework, and it has been merged
to mainline. This provides a generic way to gather up the individual
sub devices together using super node?? (not sure if it's a proper
expression) of device tree.
With this, we could remove existing exynos_drm_core.c and I guess we
could also resolve your concern about dt broken this time.

The patch series I'm preparing will go to on top of Sean's refactoring
patch series.

Andrzej Hajda from our team is already working on addressing remaining issues with Sean's series, including fixing removed parallel display support. Earlier this month he sent a series restoring support for DSI panels: http://thread.gmane.org/gmane.

linux.kernel.samsung-soc/27044 . Now he's also trying to make the design more modular. He will give you more information on Monday, as he's on holiday for next two days.

Also Tomasz Stanislawski will have some patches to enable HDMI support using Exynos DRM on Exynos 4 SoCs on top of this series, along with reworking Sean's and Daniel's patch for HDMI driver's DT bindings as he mentioned in reply to other message of this thread.