6.5-rc1 breakage in samsung-dsim

[Tim Harvey]

On Wed, Jul 12, 2023 at 2:43 PM Marek Vasut <marex at denx.de> wrote:
>
> On 7/12/23 20:52, Tim Harvey wrote:
> > Greetings,
>
> Tim,
>
> > I've noticed a regression in 6.5-rc1 that I'm having trouble bisecting
> > between 6.4 with regards to imx8mm MIPI DSI.
> >
> > I'm testing on an imx8mm-venice-gw72xx-0x with the following display:
> >   - Powertip PH800480T013-IDF02 compatible panel
> >   - Toshiba TC358762 compatible DSI to DBI bridge
> >   - ATTINY based regulator used for backlight controller and panel enable
>
> You mean RPi 7" display wired to non-RPi hardware like many people do ? ;-)

yup.... good source of readily available cheap expansion hardware :)

>
> > I'm using a dt overlay to support this [1] which works on 6.4 but on
> > 6.5-rc1 I get the following error:
> > [    6.110585] samsung-dsim 32e60000.dsi: xfer timed out: 29 06 00 00
> > 64 01 05 00 00 00
> > [    6.326588] tc358762 32e60000.dsi.0: error initializing bridge (-110)
> >
> > I'm trying to bisect this for some reason. Does anyone have any idea
> > what may be causing this or how I can debug it?
>
> Try and do something like ...
>
> git revert --no-edit v6.4..v6.5-rc1 -- drivers/gpu/drm/bridge/tc358762.c
> drivers/gpu/drm/bridge/samsung-dsim.c drivers/regulator/*attiny*.c

$ git revert --no-edit v6.4...v6.5-rc1 --
drivers/gpu/drm/bridge/tc358762.c
drivers/gpu/drm/bridge/samsung-dsim.c drivers/regulator/*attiny*.c
fatal: bad revision 'drivers/gpu/drm/bridge/tc358762.c'

your filenames look correct to me however:
$ ls drivers/gpu/drm/bridge/tc358762.c
drivers/gpu/drm/bridge/samsung-dsim.c drivers/regulator/*attiny*.c
drivers/gpu/drm/bridge/samsung-dsim.c
drivers/gpu/drm/bridge/tc358762.c
drivers/regulator/rpi-panel-attiny-regulator.c

is that intended to revert all patches between v6.4 and v6.5-rc1 in
files matching that file pattern? That would be very useful indeed if
I can figure out the syntax!

I did by hand revert the patches to samsung-dsim by hand and found the
breakage to occur with commit 0c14d3130654 ("drm: bridge:
samsung-dsim: Fix i.MX8M enable flow to meet spec"). Reverting that
works for me.

I'll respond to that thread directly but I would love to know how to
revert a patches to a series of files like you suggested if you know
what's wrong with the syntax.

thanks,

Tim



>
> (I might have the filenames wrong)
>
> Does that start working afterward ?
>
> If so, you can reverse-bisect on the reverts.
>
> I wouldn't be surprised if this was somehow related to the non-burst
> mode delay calculation again, sigh.
>
> [...]