Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <m3sg1uq6xu.fsf@t19.piap.pl> <dbb99d7b-18eb-317c-911a-b982486848fa@xs4all.nl>
 <m3eeddhora.fsf@t19.piap.pl>
In-Reply-To: <m3eeddhora.fsf@t19.piap.pl>
From:   Tim Harvey <tharvey@gateworks.com>
Date:   Mon, 7 Jun 2021 08:48:19 -0700
Message-ID: <CAJ+vNU0E_0pB-1T+VpdmjJNVirAwCUNjKVbEV4wEbqHOzURj_A@mail.gmail.com>
Subject: Re: [PATCH] TDA1997x: enable EDID support
To:     =?UTF-8?Q?Krzysztof_Ha=C5=82asa?= <khalasa@piap.pl>
Cc:     Hans Verkuil <hverkuil@xs4all.nl>,
        Mauro Carvalho Chehab <mchehab@kernel.org>,
        linux-media <linux-media@vger.kernel.org>,
        lkml <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Jun 7, 2021 at 4:56 AM Krzysztof Ha=C5=82asa <khalasa@piap.pl> wrot=
e:
>
> Hi Hans,
>
> Hans Verkuil <hverkuil@xs4all.nl> writes:
>
> >> Without this patch, the TDA19971 chip's EDID is inactive.
> >
> > Was this wrong from the very beginning? How can this ever have been tes=
ted
> > without an EDID?
>
> It seems so. I suspect it might have worked in tests because this
> register isn't cleared on reboot. I.e., setting it once after power up
> makes it work to the next power up.
> Or, maybe, the HDMI signal source didn't need EDID.
>

Krzysztof,

Most likely it was that the HDMI signal source I tested with didn't
need EDID. I primarily used a V-SG4K HMDI signal generator in my
testing and development of the driver
(http://www.marshall-usa.com/monitors/model/V-SG4K-HDI.php) which
definitely doesn't need it. Other devices I tested with were another
Gateworks board with HDMI out (which also didn't need EDID) and
occasionally a 1st gen Google Chromecast and Amazon Fire stick (which
I'm not sure about).

> I'm looking at the previous version of this driver from Gateworks and it
> contains:
>
>      /* Configure EDID
>       *
>       * EDID_ENABLE bits:
>       *  7 - nack_off
>       *  6 - edid_only
>       *  1 - edid_b_en
>       *  0 - edid_a_en
>       */
>      reg =3D io_read(REG_EDID_ENABLE);
>      if (!tda1997x->internal_edid)
>          reg &=3D ~0x83; /* EDID Nack ON */
>      else
>          reg |=3D 0x83;  /* EDID Nack OFF */
>      io_write(REG_EDID_ENABLE, reg);
>
> Not sure what the "non-internal" EDID could be - a separate I2C EEPROM
> chip? I'm using this on Gateworks' GW54xx boards and I can't see any
> such EEPROM in the vicinity of the TDA19971, but I don't know how it is
> wired - perhaps Tim has some idea?

Not sure where the source above is from (this was all so long ago) but
my guess is that 'internal_edid' meant an EDID had been provided via
software and the else case meant there was no EDID available. There is
no support on that chip for an external EEPROM.

Tim