[PATCH glint] Fix timing errors, as per documentation

[Matt Turner]

On Thu, Sep 16, 2010 at 3:02 AM, Mark Kettenis <mark.kettenis at xs4all.nl> wrote:
>> From: Matt Turner <mattst88 at gmail.com>
>> Date: Wed, 15 Sep 2010 23:37:55 -0400
>>
>> pm2_dac.c:
>> -     VsEnd and VsStart should have 1 subtracted, say the docs. This also
>>       matches with the PM3 and PM4 docs, and is what the pm2fb Linux
>>       kernel driver does.
>> -     Move VsTotal subtract-1 to instantiation, since it doesn't need to
>>       be shifted before the subtraction, unlike horizontal timings
>>       (HTotal, for instance).
>> -     Remove subtract-1 from HsStart, as the docs don't say this is
>>       correct. This also matches with the PM3 and PM4 docs, and is what
>>       the pm2fb Linux kernel driver does.
>>
>> pm3_dac.c:
>> -     Subtract 1 from HTotal after the Shiftbpp, as the docs say. This
>>       error has existed since xf86-video-glint was imported into git and
>>       has been propagated to the pm3fb Linux kernel driver.
>>
>> Signed-off-by: Matt Turner <mattst88 at gmail.com>
>> ---
>> I know there isn't much interest in this hardware, but I'd really appreciate
>> a couple of people looking over this change to make sure I didn't totally
>> hose something. Acked-by's would be excellent.
>
> I do care; Sun sold a fairly large number of PGX32 framebuffers with
> their UltraSPARC hardware.  I have one, and I'll try to test this
> change on that card in the next couple of days.

Great! I had not heard of this card before, but it looks to be a
Permedia2 (or 2V?) with OpenFirmware. Which reminds me that I probably
should make similar changes to pm_dac.c and pm2v_dac.c.

I think I can consolidate these code paths quite easily for my KMS driver.

Matt