On Thu, 2 May 2013 21:29:36 -0400, Dennis New wrote:
I'm currently running 3.9.0, and the bootup process hangs/locks my machine before the init stage is reached, unless I specify radeon.modeset=0. But modesetting /does/ work if I load the radeon module with modeset=1 after the kernel loads. So what could be going wrong during the kernel loading?
The last things printed on the screen are either radeon 0000:01:00.0: radeon: using MSI. or [drm] Driver supports precise vblank timestamp query. (if I disable msi)
I think the next thing that /should/ get printed is: [drm] radeon: irq initialized. (?)
Here's what dmesg says when I reload the radeon module with KMS: http://dennisn.dyndns.org/guest/pubstuff/dmesg-radeon-kms-module.log
As Александр (Alex M) pointed out a few months ago on The Internet, undoing this patch fixes this bug!
2099810f903caa1920f3ef6014fb7f36e4786490 drm/radeon: enable pci bus mastering after card is initialised (v2)
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/...
(On an unrelated note, when coming back from a suspend-to-ram with this video card (and probably many others), I get a blank screen or random line patterns, but I am able to fix things with radeontool, specifically with: radeontool light off radeonreg regset 0x00000284 0x1bd34208 radeonreg regset 0x000002d0 0x003c00a1 radeonreg regset CL:03 0x001c0007 radeonreg regset CL:04 0x0002004a radeonreg regset CL:05 0x0002004a radeonreg regset CL:06 0x0002004a radeontool light on Is this supposed to be something that userland is meant to handle?)