Re: [PATCH 1/2] ARM: tegra: Add Tegra20 host1x support

[Thierry Reding <thierry.reding-RM9K5IK7kjKj5M59NBduVrNAH6kLmebB@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 1359 bytes --]

On Wed, Nov 14, 2012 at 05:29:33PM +0200, Terje Bergström wrote:
> On 14.11.2012 17:01, Thierry Reding wrote:
> > Funny. I just tested with this line removed and I also get the freeze.
> > With the line I don't get the freeze. Does the freeze only occur with
> > additional patches on top? If so I think we should keep the line in for
> > now because it is what most people have tested against and which has
> > proven to work. We can fix any remaining issues with host1x specific
> > things when actual patches emerge.
> 
> I was running with 2D acceleration support added on top of your
> gitorious tree. It was the nvhost driver, which probes host1x, and
> initializes it. In middle of initialization (writing zeros to sync point
> registers), host1x stops responding.
> 
> I agree, if removing the line causes regression, keep it and let's debug
> this issue later.
> 
> It might be that we have a difference in bootloader. Does your
> bootloader enable display? U-Boot and fastboot do, and it might be that
> has an effect. They would need to initialize host1x clocks, and it might
> be that kernel initialization somehow clashes with bootloader's.

I use an upstream U-Boot version that performs no display initialization
on the board that I use. Perhaps that indeed causes your setup to work
properly but not mine.

Thierry

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]