6.14.4 driver hanging w/ Radeon HD 7660G GPU

[Alex Deucher]

On Fri, Jul 13, 2012 at 10:35 AM, criego duncan <gunkilatur at gmail.com> wrote:
> On 07/13/2012 02:49 AM, Michel Dänzer wrote:
>>
>> On Don, 2012-07-12 at 18:02 -0700, criego duncan wrote:
>>>
>>> I just got a new HP Pavilion dv6 laptop with an AMD A10-4600M APU (which
>>> contains the Radeon HD 7660G GPU).
>>>
>>> I read that the 3.4 kernel now supports Trinity, so i tried it
>>> (installed the 3.4-trunk-amd64 from debian) but the first problem was
>>> that the Xorg.log told me that kernel modesetting was required.  I have
>>> "/etc/modprobe.d/radeon-kms.conf" which contains: options radeon
>>> modeset=1
>>>
>>> This didn't turn it on, though, so i added it to grub and now X just
>>> hangs in an uninterruptable sleep state.
>>
>> One problem is that the firmware-linux-nonfree package doesn't contain
>> the microcode for Trinity yet. Get the ARUBA_*.bin files from
>> http://people.freedesktop.org/~agd5f/radeon_ucode/ , put them
>> in /lib/firmware/radeon/ and re-generate the initrd. Then see what if
>> any problems remain.
>>
>>
>
> I'd already found some extra firmware i needed... but not those.  Should
> have noticed the: ni_cp: Failed to load firmware "radeon/ARUBA_pfp.bin" msg
> in the kernel log.
> However...
>
> modesetting now works (i took it out of grub).  I'm still seeing this in the
> kernel log:
>
> Jul 13 00:16:12 hplap kernel: [    6.740486] [drm:radeon_dp_link_train_ce]
> *ERROR* channel eq failed: 5 tries
> Jul 13 00:16:12 hplap kernel: [    6.740489] [drm:radeon_dp_link_train_ce]
> *ERROR* channel eq failed
>
> X now gets much farther but then spits out (in the terminal i started it
> from) :
>
> (II) [KMS] Kernel modesetting enabled.
> radeon: Invalid PCI ID.
>
> In the Xorg log file i see:
>
> [   200.194] (II) RADEON(0): Unknown vendor-specific block 2
> [   200.194] (II) RADEON(0): EDID (in hex):
> [   200.194] (II) RADEON(0):    00ffffffffffff0030e44a0300000000
> [   200.194] (II) RADEON(0):    00150104952313780a51759659569028
> [   200.194] (II) RADEON(0):    1e505400000001010101010101010101
> [   200.195] (II) RADEON(0):    010101010101ce1d56f4500016303020
> [   200.195] (II) RADEON(0):    350059c21000001bdf1356f450001630
> [   200.195] (II) RADEON(0):    3020350059c21000001b000000000000
> [   200.195] (II) RADEON(0):    00000000000000000000000000000002
> [   200.195] (II) RADEON(0):    000c4cff0a3c641214226400000000c7
> [   200.195] (II) RADEON(0): EDID vendor "LGD", prod id 842
>
> Then it goes on initializing extensions and adding all sorts of devices
> until it ends with the following:
>
> [   219.144] (II) RADEON(0): EDID vendor "LGD", prod id 842
> [   219.144] (II) RADEON(0): Printing DDC gathered Modelines:
> [   219.144] (II) RADEON(0): Modeline "1366x768"x0.0   76.30  1366 1414 1446
> 1610  768 771 776 790 +hsync -vsync (47.4 kHz eP)
> [   219.144] (II) RADEON(0): Modeline "1366x768"x0.0   50.87  1366 1414 1446
> 1610  768 771 776 790 +hsync -vsync (31.6 kHz e)
>
> and there it hangs.

Probably best to file a bug (https://bugs.freedesktop.org component:
DRI, Product: DRM/Radeon) and attach your full dmesg output and xorg
log.

Alex