Sorry, I forgot I was on email... make that *attach* your dmesg output ;) -----Original Message----- From: Bridgman, John Sent: Saturday, January 30, 2010 1:57 PM To: 'Attila Kinali'; radeonhd@opensuse.org Subject: RE: [radeonhd] 0x9505:0x1002:0x9505: Saphire HD3850 - No XVideo adaptors, seemingly no EXA If you look further down the log there are error messages when the drm interface is being set up, resulting in "DRI initialization failed" and "acceleration disabled". Looks like the system fails to allocate a 10MB depth buffer, which I don't remember seeing before. Can you pastebin your dmesg output ? There might be clues there. -----Original Message----- From: Attila Kinali [mailto:attila@kinali.ch] Sent: Friday, January 29, 2010 6:03 PM To: radeonhd@opensuse.org Subject: [radeonhd] 0x9505:0x1002:0x9505: Saphire HD3850 - No XVideo adaptors, seemingly no EXA Hi, I have here a Radeon HD 3850 with two CRTs connected. xorg 7.5, radeonhd 1.3.0 (debian/testing), kernel 2.6.32.2 (vanilla, self compiled) According to the Xorg.0.log, EXA has been enabled, but the movement of windows (even Eterm) is so sluggish, that i have to assume that it isnt enabled. Also, I dont get any XV adaptors at all: --- $ xvinfo X-Video Extension version 2.2 screen #0 no adaptors present --- There i have no clue what it could be, as the log does not give any indication that XV doesnt work, although the XVideo extension has been loaded. I tried google, but nothing that came up seemed to be related to any of my problems. Could someone be so kind to give me some hints how to get XV and EXA correctly working, or at least what i should look for in the log files to debug this problem? Thanks in advance Attila Kinali -- Xiphias's First Law of Bookshelves: There is no such thing as too many books, only not enough bookcases. Xiphias's Second Law of Bookshelves: There is no such thing as enough bookcases, only insufficient books. -- Ian Osmond -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

I'm not sure how well the stack works with multiple adapters from different vendors right now. Maybe someone else on the list can comment. One easy workaround would be to choose ShadowFB rather than EXA as the AccelMethod. You won't get hardwrae acceleration but the software rendering will draw into system memory rather than video memory, which should be a *lot* faster. -----Original Message----- From: Attila Kinali [mailto:attila@kinali.ch] Sent: Saturday, January 30, 2010 2:32 PM To: Bridgman, John Cc: radeonhd@opensuse.org Subject: Re: [radeonhd] 0x9505:0x1002:0x9505: Saphire HD3850 - No XVideo adaptors, seemingly no EXA On Sat, 30 Jan 2010 13:57:13 -0500 "Bridgman, John" wrote:

If you look further down the log there are error messages when the drm

interface is being set up, resulting in "DRI initialization failed" and "acceleration disabled".

Looks like the system fails to allocate a 10MB depth buffer, which I don't remember seeing before.

Oops. I seem to have missed those error messages. Thanks for opening my eyes :-)

Can you pastebin your dmesg output ? There might be clues there.

Here the relevant part about DRM --- [drm] Initialized drm 1.1.0 20060810 [drm] radeon defaulting to userspace modesetting. pci 0000:02:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16 [drm] Initialized radeon 1.31.0 20080528 for 0000:02:00.0 on minor 0 pci 0000:00:0d.0: enabling device (0000 -> 0002) pci 0000:00:0d.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18 [drm] Initialized mga 3.2.1 20051102 for 0000:00:0d.0 on minor 1 pci 0000:01:00.0: enabling device (0000 -> 0002) pci 0000:01:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 [drm] Initialized mga 3.2.1 20051102 for 0000:01:00.0 on minor 2 pci 0000:01:04.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 [drm] Initialized mga 3.2.1 20051102 for 0000:01:04.0 on minor 3 pci 0000:01:08.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 [drm] Initialized mga 3.2.1 20051102 for 0000:01:08.0 on minor 4 pci 0000:01:0c.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 [drm] Initialized mga 3.2.1 20051102 for 0000:01:0c.0 on minor 5 --- (Yes, there are 5 Matrox G200 installed :-) Other than this, i haven't seen anything related to DRM/DRI. According to /sys/class/drm/card0/device, the correct dri device has been opened. Unless i'm deceived somehow. If it would be the wrong dri device, it would explain why that allocation fails, as the G200 have only 8MB of RAM installed. After i wrote the mail yesterday, i also tried radeonhd from git, with exactly the same effect. Attila Kinali -- Xiphias's First Law of Bookshelves: There is no such thing as too many books, only not enough bookcases. Xiphias's Second Law of Bookshelves: There is no such thing as enough bookcases, only insufficient books. -- Ian Osmond -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

On Jan 30, 10 22:09:21 +0100, Attila Kinali wrote:

On Sat, 30 Jan 2010 14:35:14 -0500 "Bridgman, John" wrote:

...

I'm not sure how well the stack works with multiple adapters from different vendors right now. Maybe someone else on the list can comment.

Well, now that this works. here the next problem: When switching to console (non-fb, plain VGA text mode) everything works fine. When going back to X, the machine freezes completely. Checked with both, radeonhd 1.3.0 release, and current git.

You would have to make sure that you're using the latest Xserver as the very first thing. Multicard support was broken in 1.6.x, and I think it still is partially broken in 1.7.x. Also, you might need some patches to sort out driver symbol clashes (if there are any between matrox and radeonhd - dunno). We have one in our SuSE packages (for radeon vs. radeonhd) - but I don't remember at hand where it is, sorry. Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org