[PATCH] viafb: don't setup TMDS if it doesn't exist

[PATCH] viafb: don't setup TMDS if it doesn't exist

[Daniel Drake]

The OLPC XO-1.5 does not have this particular output interface.
However, init_tmds_chip_info() always assumes it to be present,
ignoring failure codes from viafb_tmds_trasmitter_identify().

This fixes lack of output on the XO-1.5 laptop display.
Thanks to Jon Nettleton for helping identify the problematic code.

Signed-off-by: Daniel Drake <dsd@laptop.org>
---
 drivers/video/via/hw.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/video/via/hw.c b/drivers/video/via/hw.c
index 7dcb4d5..a24d03f 100644
--- a/drivers/video/via/hw.c
+++ b/drivers/video/via/hw.c
@@ -2125,7 +2125,8 @@ static void init_gfx_chip_info(int chip_type)
 
 static void init_tmds_chip_info(void)
 {
-	viafb_tmds_trasmitter_identify();
+	if (viafb_tmds_trasmitter_identify() != OK)
+		return;
 
 	if (INTERFACE_NONE = viaparinfo->chip_info->tmds_chip_info.
 		output_interface) {
-- 
1.7.2.2

Re: [PATCH] viafb: don't setup TMDS if it doesn't exist

[Florian Tobias Schandinat]

Hi,

Daniel Drake schrieb:
> The OLPC XO-1.5 does not have this particular output interface.
> However, init_tmds_chip_info() always assumes it to be present,
> ignoring failure codes from viafb_tmds_trasmitter_identify().

This is true. I noticed it too while working on the identify logic. I will carry 
it on in my tree after I adjusted it (because I replaced the OK/FAIL madness by 
bool, it will change to "if (!viafb_tmds_trasmitter_identify())")

> This fixes lack of output on the XO-1.5 laptop display.

I wonder about that. Current linux-next with an unpatched viafb seems to work 
well on my XO 1.5. I guess I should notice a missing output?

> Thanks to Jon Nettleton for helping identify the problematic code.
> 
> Signed-off-by: Daniel Drake <dsd@laptop.org>

Thanks,

Florian Tobias Schandiant

> ---
>  drivers/video/via/hw.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/video/via/hw.c b/drivers/video/via/hw.c
> index 7dcb4d5..a24d03f 100644
> --- a/drivers/video/via/hw.c
> +++ b/drivers/video/via/hw.c
> @@ -2125,7 +2125,8 @@ static void init_gfx_chip_info(int chip_type)
>  
>  static void init_tmds_chip_info(void)
>  {
> -	viafb_tmds_trasmitter_identify();
> +	if (viafb_tmds_trasmitter_identify() != OK)
> +		return;
>  
>  	if (INTERFACE_NONE = viaparinfo->chip_info->tmds_chip_info.
>  		output_interface) {

Re: [PATCH] viafb: don't setup TMDS if it doesn't exist

[Daniel Drake]

On 25 September 2010 21:20, Florian Tobias Schandinat
<FlorianSchandinat@gmx.de> wrote:
> I wonder about that. Current linux-next with an unpatched viafb seems to
> work well on my XO 1.5. I guess I should notice a missing output?

Strange. I was testing with linus master. Nothing comes up on the
screen at all, and the display controller dies a scary death (screen
fades in and out as if the world is ending). Do you know which version
of the XO-1.5 you are working with?

Daniel

Re: [PATCH] viafb: don't setup TMDS if it doesn't exist

[Florian Tobias Schandinat]

Daniel Drake schrieb:
> On 25 September 2010 21:20, Florian Tobias Schandinat
> <FlorianSchandinat@gmx.de> wrote:
>> I wonder about that. Current linux-next with an unpatched viafb seems to
>> work well on my XO 1.5. I guess I should notice a missing output?
> 
> Strange. I was testing with linus master. Nothing comes up on the
> screen at all, and the display controller dies a scary death (screen
> fades in and out as if the world is ending). Do you know which version
> of the XO-1.5 you are working with?

It's probably just that you tested
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=summary
while I'm talking about
http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=summary
which is pretty much what linus tree will look like after the next merge window 
and includes
http://github.com/schandinat/linux-2.6/tree/viafb-next

I think the most relevant patch in it concerning the issue you mentioned is
http://github.com/schandinat/linux-2.6/commit/e029ab0d02b94d633d8e82a46dfdb7fd94a62216
at least that was the point in time where I got rid of the old OLPC-specific 
patch which disabled tmds and lvds identification.
(there still is a big diff between linus tree and the XO 1.5 tree I run but at 
least viafb should work well after the next merge window)


Thanks,

Florian Tobias Schandinat

Re: [PATCH] viafb: don't setup TMDS if it doesn't exist

[Daniel Drake]

On 25 September 2010 22:06, Florian Tobias Schandinat
<FlorianSchandinat@gmx.de> wrote:
>> Strange. I was testing with linus master. Nothing comes up on the
>> screen at all, and the display controller dies a scary death (screen
>> fades in and out as if the world is ending). Do you know which version
>> of the XO-1.5 you are working with?
>
> It's probably just that you tested
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=summary
> while I'm talking about
> http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=summary
> which is pretty much what linus tree will look like after the next merge

You're right, linux-next works. It looks like you've fixed this
particular problem somehow for 2.6.37. Thanks!

Daniel