More AMD Radeon DRM Changes For Linux 3.14

[89c51]

I get to choose my kernel in grub, and a second or so into the boot process my screen goes black and the monitor claims there's no signal.

See the following for more info. It's not my bug, but it matches my symptoms:
https://bugs.freedesktop.org/show_bug.cgi?id=73053

Honestly i never had that happen. If i solve the colour thing my 6850 will be perfect (period).

[TemplarGR]

I have an ATI3870(r670) gpu and dpm shows some weird behaviour. It sometimes leads to black screen on boot, it sometimes works fine. I just throw a coin every time and cross fingers...

PS: Kernel 3.12.7, Mesa 10.0.2, latest Xorg, ArchLinux.

[Veerappan]

What kind of trouble do you get with the 6850?? I have the same and haven't noticed anything weird.

One thing i get lately is inverted colour after wakeup from blanking but i have no idea where this comes from since it doesn't happen all the time.

I get to choose my kernel in grub, and a second or so into the boot process my screen goes black and the monitor claims there's no signal.

See the following for more info. It's not my bug, but it matches my symptoms:
https://bugs.freedesktop.org/show_bug.cgi?id=73053

[89c51]

Yeah, DPM works just fine on 3 of my 4 radeon-based systems (Llano A6, Radeon 7850, Radeon 5400, Radeon 6850). It's just the 6850 that has issues. It's not my primary system anyway, so it's not a huge issue for me to lose DPM, but I'm sure that other people feel differently.

What kind of trouble do you get with the 6850?? I have the same and haven't noticed anything weird.

One thing i get lately is inverted colour after wakeup from blanking but i have no idea where this comes from since it doesn't happen all the time.

[Krejzi]

I have a Laptop with Hybrid Graphics, Intel HD 3000 and Radeon HD 6470M (CAICOS), and I haven't yet noticed any problems. Even the new "on demand" power-on/off feature seems to work flawlessly as well as UVD (through DRI_PRIME). But well, there's still radeon.dpm=1 ...

Although, there was an anoying bug with pci hotplug module which broke dpm in every possible way but that's fixed in latest kernels :/

[whitecat]

You can ignore that. It's just a memory allocation failure. I think the test code needs to be updated to take into account some additional allocations the driver makes before the test is run.

Ok ! Thx

[darkbasic]

Hopefully dpm will stop switching power states like mad while watching videos using UVD on HD 7950

[Veerappan]

Works fine here (few teething troubles at first before it because stable) On my turks card

Yeah, DPM works just fine on 3 of my 4 radeon-based systems (Llano A6, Radeon 7850, Radeon 5400, Radeon 6850). It's just the 6850 that has issues. It's not my primary system anyway, so it's not a huge issue for me to lose DPM, but I'm sure that other people feel differently.

[agd5f]

@agd5f

I have a RV770 card that fails the radeon.test=1 mode at boot. What does that mean? My card is down or a bug ?

The error is:
[drm:radeon_do_test_moves] *ERROR* Failed to create GTT object 0
Error while testing BO move.

You can ignore that. It's just a memory allocation failure. I think the test code needs to be updated to take into account some additional allocations the driver makes before the test is run.

[Vim_User]

It's working fine on all the BTC parts I have as well, but there are quite a few reports of stability problems on certain boards. You can still manually enable it by adding radeon.dpm=1 to the kernel command line in grub until we sort out what the problem is on those boards and re-enable it by default.

Thanks, I have radeon.dpm=1 set as default option in my lilo.conf anyways, so I don't even have to change something.