1 |
Hi gentoo-user, |
2 |
|
3 |
I'm attempting to configure a mid-range video card: the Radeon R7 370. Running |
4 |
on the Pitcairn chipset and a member of the Southern Islands family, I am |
5 |
surprised at the complexity of setting up the Radeon driver in comparison to its |
6 |
NVIDIA counterpart. |
7 |
|
8 |
I followed [1] carefully. Initially opting to compile AMDGPU into the kernel, I |
9 |
emerged linux-firmware with the following files. All of the relevant files were |
10 |
added to the kernel's CONFIG_EXTRA_FIRMWARE string, using /lib/firmware as the |
11 |
base directory. |
12 |
|
13 |
radeon/pitcairn_smc.bin |
14 |
radeon/pitcairn_ce.bin |
15 |
radeon/pitcairn_mc.bin |
16 |
radeon/pitcairn_me.bin |
17 |
radeon/pitcairn_pfp.bin |
18 |
radeon/pitcairn_k_smc.bin |
19 |
radeon/pitcairn_rlc.bin |
20 |
radeon/TAHITI_uvd.bin |
21 |
radeon/TAHITI_vce.bin |
22 |
|
23 |
Unfortunately, upon booting, the kernel hangs with the following message. This |
24 |
seems to be rather common, with a similar complaint being discussed at [2]. |
25 |
|
26 |
fb0: switching to amdgpudrmfb from EFI VGA |
27 |
|
28 |
As this all occurs pre-OpenRC, I am incapable of creating an S.S.H.\ connection |
29 |
to the machine from my laptop. When booting the kernel with the `nomodesetting` |
30 |
parameter, the X server reports the following after a successful kernel boot |
31 |
(created when executing `startx`): [timestamps omitted] |
32 |
|
33 |
(EE) Failed to load module "fbdev" (module does not exist, 0) |
34 |
(II) LoadModule: "vesa" |
35 |
(WW) Warning, couldn't open module vesa |
36 |
(EE) Failed to load module "vesa" (module does not exist, 0) |
37 |
(II) RADEON: Driver for ATI/AMD Radeon chipsets: |
38 |
ATI Radeon Mobility X600 (M24), ATI FireMV 2400, |
39 |
ATI Radeon Mobility X300 (M24), ATI FireGL M24 GL, |
40 |
|
41 |
[...] |
42 |
|
43 |
ARUBA, TAHITI, PITCAIRN, VERDE, OLAND, HAINAN, BONAIRE, KABINI, |
44 |
MULLINS, KAVERI, HAWAII |
45 |
(II) modesetting: Driver for Modesetting Kernel Drivers: kms |
46 |
(--) using VT number 7 |
47 |
|
48 |
(II) [KMS] drm report modesetting isn't supported. |
49 |
(EE) open /dev/dri/card0: No such file or directory |
50 |
(WW) Falling back to old probe method for modesetting |
51 |
(EE) open /dev/dri/card0: No such file or directory |
52 |
(EE) Screen 0 deleted because of no matching config section. |
53 |
(II) UnloadModule: "radeon" |
54 |
(EE) Screen 0 deleted because of no matching config section. |
55 |
(II) UnloadModule: "modesetting" |
56 |
(EE) Device(s) detected, but none match those in the config file. |
57 |
(EE) |
58 |
Fatal server error: |
59 |
(EE) no screens found(EE) |
60 |
(EE) |
61 |
|
62 |
In a fruitless attempt to fix this, I opted to instead load AMDGPU as a kernel |
63 |
module. That did not go far either, as apparently a bug exists in the Radeon |
64 |
drivers which causes OpenRC to hang when setting the system time according to |
65 |
the internal clock [3]: |
66 |
|
67 |
Setting system clock using hardware clock [Local Time] |
68 |
[hangs forever] |
69 |
|
70 |
Any help with this would be appreciated. I'm moving away from NVIDIA due to the |
71 |
requirement of proprietary drivers to get any decent performance, however now it |
72 |
feels as though the AMD drivers, although open-source, consist of too many bugs |
73 |
(such as hanging the boot-up process for some reason or another) to be of any |
74 |
actual use. Whilst I'm aware that is obviously not the case due to the |
75 |
popularity of their cards, I am bewildered at how difficult this seems. |
76 |
|
77 |
Cheers. |
78 |
|
79 |
[1] https://wiki.gentoo.org/wiki/AMDGPU |
80 |
[2] https://forums.gentoo.org/viewtopic-t-1061750-start-0.html |
81 |
[3] https://www.reddit.com/r/Gentoo/comments/9oax20/ |
82 |
|
83 |
-- |
84 |
|
85 |
Ashley Dixon |
86 |
suugaku.co.uk |
87 |
|
88 |
2A9A 4117 |
89 |
DA96 D18A |
90 |
8A7B B0D2 |
91 |
A30E BF25 |
92 |
F290 A8AA |