Gentoo Archives: gentoo-user

From: Alan McKinnon <alan.mckinnon@×××××.com>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] ati-drivers:legacy fail to build with kernel 3.12
Date: Thu, 03 Oct 2013 09:35:25
Message-Id: 524D393E.7090909@gmail.com
In Reply to: Re: [gentoo-user] ati-drivers:legacy fail to build with kernel 3.12 by Helmut Jarausch
1 On 02/10/2013 14:19, Helmut Jarausch wrote:
2 > On 10/02/2013 09:06:42 AM, Alan McKinnon wrote:
3 >> On 02/10/2013 09:02, Helmut Jarausch wrote:
4 >> > On 10/01/2013 08:16:23 PM, Volker Armin Hemmann wrote:
5 >> >> Am 01.10.2013 10:00, schrieb Helmut Jarausch:
6 >> >> > Hi,
7 >> >> >
8 >> >> > in "good" tradition the new 3.12 Linux kernel breaks ati-drivers
9 >> >> > again (as always in the past).
10 >> >> > Does anybody know about a patch to make
11 >> >> > x11-drivers/ati-drivers:legacy
12 >> >> > compile with linux-3.12-rc3 ?
13 >> >> >
14 >> >> > Trying to emerge ati-drivers-13.1_pre897 (currently the only legacy
15 >> >> > driver)
16 >> >> > with 3.12-rc3 gives
17 >> >> >
18 >> >> > MODPOST 1 modules
19 >> >> > FATAL: modpost: GPL-incompatible module fglrx.ko uses GPL-only
20 >> symbol
21 >> >> > 'acpi_bus_get_device'
22 >> >> >
23 >> >> > Many thanks for a hint,
24 >> >> > Helmut
25 >> >> >
26 >> >>
27 >> >> so you are using a driver meant for stable (old) systems, with a pre
28 >> >> release kernel - and you don't even know what do do with that message?
29 >> >
30 >> > I'm quite sure there are a lot of things which you don't known either.
31 >> >
32 >> >> Does that do no seem a bit silly? hm?
33 >> >
34 >> > Thanks for this compliment!
35 >> >
36 >> >> acpi_bus_get_device is gpl only. You can undo that. If you want.
37 >> >>
38 >> >> But seriously, what is wrong with using stable releases like
39 >> 3.10.x? If
40 >> >> you use ati-drivers, you don't need the amd driver improvements in
41 >> 3.11
42 >> >> or 3.12.
43 >> >
44 >> > There are two points here:
45 >> > - first I consider it as my contribution to GenToo to help testing.
46 >> > The problem with the 3.12 prerelease will definitely continue with
47 >> the
48 >> > main release
49 >> > coming in about 4 weeks.
50 >> >
51 >> > - This 3.12 prerelease has a lot of BTRFS pataches. Since I'm using
52 >> > BTRFS for most of
53 >> > my file systems, I like to have the most recent fixes.
54 >>
55 >>
56 >> Calm down boys.
57 >>
58 >>
59 >> Helmut, the problem is that you are trying to use ATI's proprietary code
60 >> with GPL code in the kernel.
61 >>
62 >> The kernel devs do not want you to do that and that is their right.
63 >>
64 >> What you should have done in your mail is indicate that you read the
65 >> error message and understand it, and why you want to continue
66 >> regradless. Then you should have mentioned what Google told you about
67 >> the problem, or something to show what you already did to help yourself.
68 >>
69 >> As it stands you post is really just asking someone else to do your
70 >> homework for you and that you are too lazy to do it yourself. I'm sure
71 >> that's not want you are really doing, it just looks that way.
72 >>
73 >
74 > Hi Alan,
75 >
76 > the only "fix" I have found with Google is a patch to the kernel itself.
77 > This works but I don't like it.
78 > Up to now, all similar problems in the past have been fixed by patches
79 > to the ati-drivers.
80 > But I haven't found such a patch, perhaps it's impossible to fix it by
81 > only patching
82 > the ati-drivers, this time. Linus makes our life a bit harder.
83
84
85 I missed this one earlier.
86
87 ATI will need to modify their driver code. The non-GPL portion is trying
88 to access GPL-only kernel code, so that portion needs to move into ATI's
89 shim.
90
91 Nothing is broken here, this is the expected result occasionally. You
92 implicitly agreed to have to put up with this occasionally when you
93 decided to use fglrx, and also when you decided to use the Linux kernel.
94 See:
95
96 stable_api_nonsense.txt
97
98 or you could just use radeon and all these troubles go away.
99
100
101 --
102 Alan McKinnon
103 alan.mckinnon@×××××.com