AMD A6-3400M not working on Linux kernel 3.0

[bridgman]

Just asking for clarification, does that include llano systems using the open-source driver?

agd5f *only* talks about the open source driver

What kind of display output are you using - VGA, DVI-D, DVI-I/A, HDCP, DP etc... ?

[letharion]

agd5f *only* talks about the open source driver

What kind of display output are you using - VGA, DVI-D, DVI-I/A, HDCP, DP etc... ?

I understand that now
I wish to point out I meant no offence to agd5f. What little knowledge I have on the topic comes from reading phoronix articles, and the last one I remember, it didn't work, but that could either just not be true, I could remember wrong, or not having read the latest articles on the topic.

My cable is the same one as the this example image from wikipedia, https://upload.wikimedia.org/wikiped...-Dvi-cable.jpg
3 by 3, by 2 + 1. That's a DVI-D single link, if I read the wikipedia page right.
I tried upgrading my software to: kernel: 3.1-rc7, mesa: git, libdrm: git, xf86-video-ati: git, xorg-server: 1.11. The result is the same as with the stable versions in Gentoo. I see a message about SUMO code loading when the kernel is starting, image freezes for maybe 15 seconds, then I get lots of nice colors that make no sense at all. Can't switch to a text tty, but system responds to and restarts on ctrl+alt+delete.

It's very hopeful to learn that a lot of hardware is expected to work with the foss-drivers. Is there anything a normal user like me can do? Is there a project that accepts bug reports? I realize it could be near impossible to understand a driver issue when you can't reproduce it.

[nepwk]

No idea, never really tested such things. I mostly use it with the AC adapter plugged in. On battery seems to last 2.5 - 3 hours when not doing anything strenuous.

You should check your settings, I get 5+ hours of light use. Ubuntu 11.04 and Catalyst 11.8 actually get better battery life than Windows 7.

If I use 11.10 and the FOSS driver (attached to a HDMI monitor, since the laptop display won't work), then I get something like an hour or an hour and a half, which doesn't make me miss the FOSS driver I had been using since Ubuntu 10.04.

[bridgman]

and the last one I remember, it didn't work, but that could either just not be true, I could remember wrong, or not having read the latest articles on the topic.

Unfortunately the web has been around long enough now that there is probably more obsolete info than current info, and very few people go back and update/delete the obsolete stuff so it just keeps growing. The impact of that on open source drivers (where most of the development happens in public including the transition from "works on one board" to "works on pretty much everything") you will find more "doesn't work" posts than "does work" posts for maybe a year after it "does work".

You pretty much have to ignore everything that doesn't have a date, then put everything that does have a date in date sequence and stare at it until a pattern emerges. Only then is it safe to take any individual post seriously

My cable is the same one as the this example image from wikipedia, https://upload.wikimedia.org/wikiped...-Dvi-cable.jpg 3 by 3, by 2 + 1. That's a DVI-D single link, if I read the wikipedia page right.

Yep, that looks like DVI-D to me as well. I think that is in the "works frequently" category right now - agd5f would know better but he is away this week.

I see a message about SUMO code loading when the kernel is starting,

image freezes for maybe 15 seconds, then I get lots of nice colors that make no sense at all. Can't switch to a text tty, but system responds to and restarts on ctrl+alt+delete.

Does the message say that the code loaded successfully ? I don't remember seeing that 15 second freeze except when ucode loads but I haven't had time to reboot anything for a few months ;(

Is there anything a normal user like me can do? Is there a project that accepts bug reports? I realize it could be near impossible to understand a driver issue when you can't reproduce it. [/QUOTE]

Go to the bottom of : http://www.x.org/wiki/RadeonFeature and you'll find info on where to file bugs for each component. You want the "KMS DRM" in that list. Please check to see if a very similar bug exists already and if so add info to that rather than starting a new one. The xorg log and dmesg output are always useful; dmesg output is the most useful for the kernel driver.

[AlbertP]

The failed systems mentioned in this topic are both BeaverCreek which are the Radeon HD 6500G/D series plus 6620G and 6480G. They are used in Fusion A6/A8 and the mobile A4.

Perhaps it's just a BeaverCreek regression with WinterPark/Sumo2 working fine?

Letharion, do yo know what Radeon / Fusion type you're using? Also make sure the SUMO firmware is properly installed. How to get SUMO firmware is described earlier in this topic. You can add radeon.modeset=0 to the kernel command-line to boot with Vesa.

[letharion]

The failed systems mentioned in this topic are both BeaverCreek which are the Radeon HD 6500G/D series plus 6620G and 6480G. They are used in Fusion A6/A8 and the mobile A4.

Perhaps it's just a BeaverCreek regression with WinterPark/Sumo2 working fine?

Letharion, do yo know what Radeon / Fusion type you're using? Also make sure the SUMO firmware is properly installed. How to get SUMO firmware is described earlier in this topic. You can add radeon.modeset=0 to the kernel command-line to boot with Vesa.

I have a A8-3850, which I believe is a BeaverCreek? I naively believed that the message "Loading SUMO firmware" meant that the firmware was actually available. I fixed the firmware, and now my X actually starts! Can't get a higher resolution than 1024x768, but it's great progress!

Thanks a lot

[letharion]

Actually, 1280x1024, which is my max resolution, works just fine. Not sure what changed.

[agd5f]

It's not the APU itself so much as the combinations of monitors in use and connectors on the motherboard. Certain combinations are problematic.

[Ex-Cyber]

Which board are you using? I have an Asus F1A75-V PRO, and I get a max resolution of 1024x768 initially. What seems to be happening is that the driver detects a (nonexistent) connection on the VGA port, fails to get EDID (since there's no monitor there), uses a builtin list of "safe" modes that maxes out at 1024x768, and mirrors by default (so the main display is also limited to 1024x768). Once I'm in GNOME, I can disable mirroring and choose the max resolution of my actual display.

[letharion]

Which board are you using? I have an Asus F1A75-V PRO, and I get a max resolution of 1024x768 initially. What seems to be happening is that the driver detects a (nonexistent) connection on the VGA port, fails to get EDID (since there's no monitor there), uses a builtin list of "safe" modes that maxes out at 1024x768, and mirrors by default (so the main display is also limited to 1024x768). Once I'm in GNOME, I can disable mirroring and choose the max resolution of my actual display.

I'm using a Gigabyte a75-ud4h, and what you say could also apply to me. Every time I boot, I get a 1024x768 resolution, but I can easily switch to full with the krandrtray application. The program does list the VGA connection as 'available' at a max resolution of 1024x768, despite there being nothing connected to it. Perhaps I should find a vga cable somewhere and see what happens if I connect my monitor to it instead.