Message ID | 20200601065207.492614-1-sam@ravnborg.org (mailing list archive) |
---|---|
Headers | show |
Series | backlight updates | expand |
Hi Sam, On Mon, 1 Jun 2020 at 07:52, Sam Ravnborg <sam@ravnborg.org> wrote: > > v3: > - Dropped video patch that was reviewd and thus applied > - Updated kernel-doc so all fields now have a short intro > - Improved readability in a lot of places, thanks to review > feedback from Daniel - thanks! > - Added better intro to backlight > - Added acks > > Several other smaller changes documented in the > patches. > I left out patches to make functions static as > there are dependencies to drm-misc-next for these. > When this is landed I have a pile of follow-up patches waiting, > mostly introducing backlight_is_blank() all over. > What happened with my suggestion to use backlight_is_blank() in fbdev core itself? It effectively makes 13/13 and the above mentioned follow-up series obsolete. That said, series look spot on. With the documentation fixed (pointer by Daniel) patches 1-12 are: Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com> -Emil
Hi Emil. Long overdue feedback, I did not find time to go back to this patch-set until now. On Tue, Jun 02, 2020 at 03:04:39PM +0100, Emil Velikov wrote: > Hi Sam, > > On Mon, 1 Jun 2020 at 07:52, Sam Ravnborg <sam@ravnborg.org> wrote: > > > > v3: > > - Dropped video patch that was reviewd and thus applied > > - Updated kernel-doc so all fields now have a short intro > > - Improved readability in a lot of places, thanks to review > > feedback from Daniel - thanks! > > - Added better intro to backlight > > - Added acks > > > > Several other smaller changes documented in the > > patches. > > I left out patches to make functions static as > > there are dependencies to drm-misc-next for these. > > When this is landed I have a pile of follow-up patches waiting, > > mostly introducing backlight_is_blank() all over. > > > What happened with my suggestion to use backlight_is_blank() in fbdev > core itself? Following your suggestion I implemented: +static inline int backlight_get_brightness(struct backlight_device *bd) +{ + if (backlight_is_blank(bd)) + return 0; + else + return bd->props.brightness; +} This results in code like this: static int adp8870_bl_update_status(struct backlight_device *bl) { return adp8870_bl_set(bl, backlight_get_brightness(bl)); } Compare that with the original code: static int adp8870_bl_update_status(struct backlight_device *bl) { int brightness = bl->props.brightness; if (bl->props.power != FB_BLANK_UNBLANK) brightness = 0; if (bl->props.fb_blank != FB_BLANK_UNBLANK) brightness = 0; return adp8870_bl_set(bl, brightness); } Much nicer! The old code reads the brightness property direct. I prefer the small helper so we do not hardcode too much of the internals in the drivers. Also the above is simpler than trying to maintain the correct value in props.brightness all the time. And can be introduced gradually. I will rework the series to use this helper. I will also fix so I can use a const backlight_device *. Thanks for the suggestion. Sam > It effectively makes 13/13 and the above mentioned follow-up series obsolete. > > That said, series look spot on. With the documentation fixed (pointer > by Daniel) patches 1-12 are: > Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com> > > -Emil > _______________________________________________ > dri-devel mailing list > dri-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/dri-devel