[PATCH v3 06/13] drm: bridge: samsung-dsim: Add DSI init in bridge pre_enable()

[Dave Stevenson]

Hi Jagan

On Mon, 29 Aug 2022 at 19:31, Jagan Teki <jagan at amarulasolutions.com> wrote:
>
> Hi Dave,
>
> On Fri, Jul 22, 2022 at 9:35 PM Dave Stevenson
> <dave.stevenson at raspberrypi.com> wrote:
> >
> > Hi Jagan and Marek.
> >
> > On Fri, 22 Jul 2022 at 16:35, Marek Szyprowski <m.szyprowski at samsung.com> wrote:
> > >
> > > On 20.07.2022 17:52, Jagan Teki wrote:
> > > > Host transfer() in DSI master will invoke only when the DSI commands
> > > > are sent from DSI devices like DSI Panel or DSI bridges and this
> > > > host transfer wouldn't invoke for I2C-based-DSI bridge drivers.
> > > >
> > > > Handling DSI host initialization in transfer calls misses the
> > > > controller setup for I2C configured DSI bridges.
> > > >
> > > > This patch adds the DSI initialization from transfer to bridge
> > > > pre_enable as the bridge pre_enable API is invoked by core as
> > > > it is common across all classes of DSI device drivers.
> > >
> > > This is still problematic in case of Exynos. Without a workaround like this
> > >
> > > https://github.com/mszyprow/linux/commit/11bbfc61272da9610dd5c574bb8ef838dc150961
> > >
> > > the display on the all real DSI panels on my Exynos based boards is broken.
> >
> > I'd queried on the other thread trying to address DSI operation [1] as
> > to whether the test for STOP_STATE (presumably LP-11) at [2] was
> > actually valid, but had no response.
> > There is no need to check for bus contention at that point, but should
> > it happen the driver doesn't write the registers in lines 862-868
> > having returned -EFAULT at line 853. The controller is therefore only
> > partially initialised.
>
> Can you link me if you have any updated series on this? or the
> existing one is the latest one itself?

I've not updated my patch set as I didn't think there had been any
significant review comments to action -
https://patchwork.freedesktop.org/series/100252/ is still the latest.

Sam had suggested changing upstream to prev/next, but seeing as no one
else has expressed a view on that I didn't see much point in
respinning. If others agree with Sam, then I'll do it.

Checking I do note that the suggested change to drop
drm_bridge_chain_* API has been done by Sam. I don't see that having
been merged, but once it is patch 1 needs to be dropped / reworked.

I don't have any Exynos hardware, so can't really help out on the DSI
issues there other than making suggestions by inspection of the code.

  Dave

> Thanks,
> Jagan.