"Ask ATI" dev thread

[RealNC]

But he wouldn't be allowed to tell us. You see, that is vital information that would bring down AMD. "Hear ye, hear ye. Catalyst 9.6 is gonna support kernel 2.6.29." After that, AMD's stock market values go down by 80% and 3 weeks later they declare bankruptcy.

So better keep stuff like that top secret. Who knows who might be listening.

[curaga]

Why don't Vesa modes work on some later series AMD cards?
I've personally seen a desktop HD4850 fail there. Is this an attempt to save space by not including the VESA bios?

edit: this is using either vesafb or Xvesa.

[energyman]

ask the manufacturer of your card why their card's bios sucks.

[Kano]

Well i had a similar issue with a VIA based motherboard. PCI-E cards "forgot" in that board after some time that they support vesa modes like 791. Temporary replacing helped, well thats hard to do with a laptop

[bridgman]

I'm not aware of any change to VESA support, so this sounds like a bug in either the VBIOS or the VESA driver. I'm not sure where best to file a bug report - it will probably be the same bugzilla we use for fglrx, but let me check.

[Mazur]

You guys know what I don't like in AMD? Few posts up we are asking something strange about drivers and nobody will repeat whereas few posts down somebody will post a question about something else and he will get anser from Dev which will looks like that: "I don't know ... I will check .." and really he don't even care.

[bridgman]

Mazur, my role at AMD is primarily on the open source side, but I try to help with fglrx and other issues where I can. Which earlier issue are you referring to ?

[glock24]

Why don't Vesa modes work on some later series AMD cards?
I've personally seen a desktop HD4850 fail there. Is this an attempt to save space by not including the VESA bios?

edit: this is using either vesafb or Xvesa.

I have a mobility HD3650, I cannot get the vesa console to work, and the vesa X driver doesn't work either. This card comes in an Asus laptop.

[Mazur]

#bridgman
Anyway I am drunk today so I am writing idiotic things but... is card HD3200 going to be supported early in open source drivers? I meant mainly KMS and 3d, I wanted to try mesa-git, radeonhd-git and kernel26-git to check but it takes too long. From that what I know radeon-rewrite was merged to mesa so it should give support of 3d for use, isn't it?

[bridgman]

Anyway I am drunk today so I am writing idiotic things but...

It's Saturday, everyone is either drunk or hung over

is card HD3200 going to be supported early in open source drivers? I meant mainly KMS and 3d, I wanted to try mesa-git, radeonhd-git and kernel26-git to check but it takes too long.

We're doing most of the 3D development on discrete cards but I expect 3200 support will come at pretty much the same time. The 3200 3D engine is pretty much the same as RV610 (HD2400) so 3D should come pretty soon.

KMS is actually mostly done AFAIK but is waiting for memory management - apparently we need interrupt support so we'll need to write some sample code and release some docs before that can happen.

From that what I know radeon-rewrite was merged to mesa so it should give support of 3d for use, isn't it?

The radeon-rewrite code only covered GPUs up to 5xx, so nothing there for you. That said, merging radeon-rewrite is important because it means we can now start merging our 6xx code to master as well. Alex synced up our 6xx tree with the master+radeon_rewrite code yesterday, so no obstacles there.

It's probably a good time to start trying to build agd5f's 6xx/7xx drm and the 6xx-rewrite branch of mesa. It's not at the point where it's worth testing yet but I think we're getting close.