[v5,05/12] drm/modes: Rewrite the command line parser

[Thomas Graichen]

hi maxim,

On Tue, Aug 20, 2019 at 5:00 PM Maxime Ripard <maxime.ripard at bootlin.com> wrote:
>
> Hi,
>
> On Mon, Aug 19, 2019 at 09:20:00PM +0200, Thomas Graichen wrote:
> > On Mon, Aug 19, 2019 at 8:54 PM Jernej Škrabec <jernej.skrabec at gmail.com> wrote:
> > >
> > > +CC: Thomas Graichen
> > >
> > > Dne ponedeljek, 17. junij 2019 ob 16:51:32 CEST je Maxime Ripard napisal(a):
> > > > From: Maxime Ripard <maxime.ripard at free-electrons.com>
> > > >
> > > > Rewrite the command line parser in order to get away from the state machine
> > > > parsing the video mode lines.
> > > >
> > > > Hopefully, this will allow to extend it more easily to support named modes
> > > > and / or properties set directly on the command line.
> > > >
> > > > Reviewed-by: Noralf Trønnes <noralf at tronnes.org>
> > > > Signed-off-by: Maxime Ripard <maxime.ripard at free-electrons.com>
> > >
> > > Thomas reported to me that this patch breaks "video=CONNECTOR:e" kernel
> > > parameter which he currently uses as a workaround for H6 HDMI monitor
> > > detection issue on one STB.
> > >
> > > I suppose this is the same issue that Dmitry noticed.
> > >
> > > Thomas Graichen (in CC) can provide more information if needed.
> >
> > as jernej already mentioned i am currently having to use the kernel
> > cmdline option video=HDMI-A-1:e to get a working hdmi output on an
> > eachlink h6 mini tv box and was wondering that i did not get any hdmi
> > output even with this option when switching from the
> > https://github.com/megous/linux oprange-pi-5.2 to the orange-pi-5.3
> > branch which seems to contain this patch.
>
> Which kernel version is that based on?

5.3-rc3

> > as i had no idea what might have caused the breakage of the hdmi
> > output and did a full bisect of the kernel between those two
> > versions, which ended reliably at exactly this patch - so i guess
> > there is a regression at least with the video=CONNECTOR:e option
> > (maybe others too?) with this patches code which makes it not
> > working anymore.
>
> I'm not sure I'll have the time to look into it this week (or the
> next, unfortunately). However, the e parameter is supposed to be
> parsed by drm_mode_parse_cmdline_extra, which in turn is supposed to
> be called there:
> https://elixir.bootlin.com/linux/v5.3-rc5/source/drivers/gpu/drm/drm_modes.c#L1810
>
> If you can test that, having an idea of if that function is called,
> which return code it returns, and if it isn't if why would be super
> helpful.

i just added a printk and it looks like it is not getting called:

diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index b0369e690f36..4c58fdb1d7be 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -1813,6 +1813,7 @@ bool
drm_mode_parse_command_line_for_connector(const char *mode_option,

                ret = drm_mode_parse_cmdline_extra(extra_ptr, len,
                                                   connector, mode);
+               printk(KERN_WARNING "DEBUG -
drm_mode_parse_cmdline_extra %d", ret);
                if (ret)
                        return false;
        }

no output from it in dmesg (my loglevel=8 and on the kernel cmdline
and in /proc/cmdline i have "video=HDMI-A-1:e") - so looks like it
really gets lost somewhere along the way ...

best wishes - thomas