Mailinglist Archive: radeonhd (408 mails)

[Alex Deucher <alexdeucher@xxxxxxxxx>]

On Tue, Jun 16, 2009 at 12:49 AM, Matthias Hopf<mhopf@xxxxxxx> wrote:
> On May 07, 09 23:38:19 -0400, Roman Stanchak wrote:
> > I'm trying to get the radeonhd driver working with the IGP on this
> > motherboard (it has an RS740 chipset).  I'm using radeonhd and drm
> > from the git master 05-07-2009 on an 2.6.28 ubuntu kernel.  Basically,
> > the behavior I'm getting when I start-up X is that the display powers
> > down and doesn't come back up unless I log in remotely and re-run X
> > with the VESA driver.
>
> Roman, given that my patch was dead wrong, and yours just worked by
> accident :-] I'd like to investigate this a bit more.
>
> Your patch basically changed reading the MC registers from PCI space to
> memory mapped register space - but the register addresses stayed the
> same, so you basically read out the value of a completely different
> register. While still writing into the correct register.
>
> I assume you read out 0 or something similar, and that's why it worked.
> I'd rather hardcode this value than read something that is completely
> non-related ;-)
>
> So can you compile the driver with your patch once more, this time with
> "-D RHD_DEBUG=1" ? The -logverbose 7 logfile should then contain the
> value the driver is reading from the "wrong" register. It would be best
> to have two logfiles, one directly after a hard reset, and the second of
> a subsequent Xserver start. Assuming it still works :-]
>
> I'll also go and ask John about this register, maybe something springs
> into his mind. Alex is on vacation right now.

I don't have the specs in front of me, but I would check what radeon
does as I believe it works correctly for these chips.

Alex
--
To unsubscribe, e-mail: radeonhd+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: radeonhd+help@xxxxxxxxxxxx