On Sat, Jul 26, 2014 at 12:52:03AM +0530, Ajay Kumar wrote: > Most of the panels need an init sequence as mentioned below: > -- poweron LCD unit/LCD_EN > -- start video data > -- poweron LED unit/BACKLIGHT_EN > And, a de-init sequence as mentioned below: > -- poweroff LED unit/BACKLIGHT_EN > -- stop video data > -- poweroff LCD unit/LCD_EN > With existing callbacks for drm panel, we cannot accomodate such panels, > since only two callbacks, i.e "panel_enable" and panel_disable are supported. > > This patch adds: > -- "prepare" callback which can be called before > the actual video data is on, and then call the "enable" > callback after the video data is available. > > -- "unprepare" callback which can be called after > the video data is off, and use "disable" callback > to do something before switching off the video data. > > Now, we can easily map the above scenario as shown below: > poweron LCD unit/LCD_EN = "prepare" callback > poweron LED unit/BACKLIGHT_EN = "enable" callback > poweroff LED unit/BACKLIGHT_EN = "disable" callback > poweroff LCD unit/LCD_EN = "unprepare" callback > > Also, a helper function for get_modes is added. > > Signed-off-by: Ajay Kumar <ajaykumar.rs@xxxxxxxxxxx> > --- > include/drm/drm_panel.h | 26 ++++++++++++++++++++++++++ > 1 file changed, 26 insertions(+) I had already decided to apply an earlier version that you posted (I think the only difference was that it didn't have the inline implementation for drm_panel_get_modes() yet), reworded the commit message a little (to be less specific about LED, LCD and backlight units) and added some kerneldoc for struct drm_panel_funcs. What I propose is to merge that patch (I'll hopefully push it today so that it shows up in tomorrow's linux-next) and then apply a separate patch to add drm_panel_get_modes(). Do you have any objections to that? Thierry
Attachment:
pgpLUv2YfnXoT.pgp
Description: PGP signature
