Openchrome 0.3.1 XV problem with CX700

Help with booting MiniMyth

Moderator: Pablo

Post Reply
mattj
Member
Posts: 18
Joined: Sat Dec 04, 2010 3:23 pm

Openchrome 0.3.1 XV problem with CX700

Post by mattj » Sat Dec 15, 2012 12:22 pm

Hi Pablo

There's a bug in the openchrome 0.3.1 release which causes XV to fail on CX700 chipsets ( and maybe others ), upgrading to latest git master solves the problem. The symptom of this is no video on playback, although audio will be playing, and in the log you will see :

(EE) CHROME(0): [Xv] Port 0: Bandwidth is insufficient. Check bios memory settings.

There are also some other gotchas which it may be helpful for others to know :

1. If you have a single monitor section with specific refresh and sync settings and you're using DVI or TV out, it's possible the driver might now assign that section to the VGA output, and autodetect the settings for the other output, which can be incorrect. In my case I needed a specific refresh setting which the driver wasn't switching to on boot. You can see in the logs what the driver thinks the outputs are, and which output it's assigning the monitor section to. To fix this, specify which output the monitor section is for.

In my case, it's for the DVI output ( DP-1 ), so I in my device section I now have

Option "monitor-DP-1" "dvi"

and in my monitor section I have

Identifier "dvi"

2. You no longer need to force VBEModes to get DVI output working on CX700, you also don't need to force AGPDMA or SWcursor.

Matt

Pablo
Site Admin
Posts: 4182
Joined: Tue Dec 14, 2004 2:13 am
Location: La Jolla
Contact:

Re: Openchrome 0.3.1 XV problem with CX700

Post by Pablo » Sat Dec 15, 2012 10:51 pm

mattj wrote:Hi Pablo

There's a bug in the openchrome 0.3.1 release which causes XV to fail on CX700 chipsets ( and maybe others ), upgrading to latest git master solves the problem. The symptom of this is no video on playback, although audio will be playing, and in the log you will see :

(EE) CHROME(0): [Xv] Port 0: Bandwidth is insufficient. Check bios memory settings.
I have updated the driver to the latest git version. It will be in the next build.
There are also some other gotchas which it may be helpful for others to know :

1. If you have a single monitor section with specific refresh and sync settings and you're using DVI or TV out, it's possible the driver might now assign that section to the VGA output, and autodetect the settings for the other output, which can be incorrect. In my case I needed a specific refresh setting which the driver wasn't switching to on boot. You can see in the logs what the driver thinks the outputs are, and which output it's assigning the monitor section to. To fix this, specify which output the monitor section is for.

In my case, it's for the DVI output ( DP-1 ), so I in my device section I now have

Option "monitor-DP-1" "dvi"

and in my monitor section I have

Identifier "dvi"
I have seen a same thing with the Intel driver. However, as long my device section has
Option "Monitor-outputname" "monitor1"
then the driver uses monitor1 for both the first discovered output (usually VGA) and for the output outputname (DP-1 in your case, HDMI1 in my case). Do you see the same behavior.
MiniMyth running on an Acer ApireRevo 3610 and a Zotac ZBOX-ID80-U. Find out more at my MythTV page.

mattj
Member
Posts: 18
Joined: Sat Dec 04, 2010 3:23 pm

Re: Openchrome 0.3.1 XV problem with CX700

Post by mattj » Sun Dec 16, 2012 11:00 am

Yep, that's exactly the same behaviour :

[2649004.186] (II) CHROME(0): Output VGA-1 using monitor section dvi
[2649004.186] (II) CHROME(0): ViaVT162xDetect
[2649004.186] (II) CHROME(0): I2C device "I2C bus 2:VT162x" registered at address 0x40.
[2649004.187] (--) CHROME(0): Detected VIA Technologies VT1625 TV Encoder
[2649004.187] (II) CHROME(0): ViaVT162xInit
[2649004.187] (II) CHROME(0): Output TV-1 has no monitor section
[2649004.187] (II) CHROME(0): VT162xSave
[2649004.278] (II) CHROME(0): Output DP-1 using monitor section dvi

So it looks like the VGA output will always use the single monitor section, but you need to specify in the device section if you want that to also apply to another output.

mattj
Member
Posts: 18
Joined: Sat Dec 04, 2010 3:23 pm

Re: Openchrome 0.3.1 XV problem with CX700

Post by mattj » Sun Dec 16, 2012 11:02 am

I'm still trying to debug the issue with CLE266, the latest git master segfaults on this platform. I've posted the segfault to openchrome-users, and hopefully one of the openchrome guys might be able to shed some light on it.

Pablo
Site Admin
Posts: 4182
Joined: Tue Dec 14, 2004 2:13 am
Location: La Jolla
Contact:

Re: Openchrome 0.3.1 XV problem with CX700

Post by Pablo » Sat Dec 22, 2012 3:53 am

The openchrome updates is in the latest test build.
mattj wrote:I'm still trying to debug the issue with CLE266, the latest git master segfaults on this platform. I've posted the segfault to openchrome-users, and hopefully one of the openchrome guys might be able to shed some light on it.
Given that MythTV has dropped support for XvMC and XvMC-VLD what is the reason for trying to get newer MiniMyth versions running with VIA CLE266?
MiniMyth running on an Acer ApireRevo 3610 and a Zotac ZBOX-ID80-U. Find out more at my MythTV page.

mattj
Member
Posts: 18
Joined: Sat Dec 04, 2010 3:23 pm

Re: Openchrome 0.3.1 XV problem with CX700

Post by mattj » Sun Jan 06, 2013 12:48 pm

I've never had to use XvMC for standard resolution video using Epia M, and I still have frontends which have Epia M's in them so it's worth the effort from my perspective to try and debug the issue. I don't mind rolling my own as I appreciate I may be the only openchrome user of minimyth left !

Post Reply