1 |
On 27/08/2013 09:14, Pandu Poluan wrote: |
2 |
>>>> That list is the list of kernels that nVidia supports, which is easy |
3 |
>>>> > >> to find. |
4 |
>>> > > |
5 |
>>> > > Where? AIUI from reading various threads about this, sometimes that |
6 |
>>> > > info can be found in nVidia's developer web forum, but I've never been |
7 |
>>> > > able to find it there. nVidia's READMEs give a minimum kernel version, |
8 |
>>> > > but no max. |
9 |
>> > |
10 |
>> > |
11 |
>> > So ask nVidia to clearly and unambiguously state in an easily found |
12 |
>> > place what kernels *they* support. |
13 |
>> > |
14 |
>> > Look, all issues with building the driver shim are directly the |
15 |
>> > responsibility of nVidia themselves, a result of *their* business |
16 |
>> > decisions. The correct thing to do is to make it nVidia's problem and |
17 |
>> > not force the community to jump through hoops trying to track down what |
18 |
>> > does and does not work today. |
19 |
>> > |
20 |
>> > Or, you could do the heavy lifting yourself. You test all current |
21 |
>> > drivers with all recent kernels and maintain a gentoo wiki page that |
22 |
>> > lists the info you want. |
23 |
> |
24 |
> Hmm... reading this thread makes me understand why Linus gave nVidia 'the bird'. |
25 |
|
26 |
|
27 |
|
28 |
I'm not so sure. I can understand nVidia's position. They make and sell |
29 |
hardware. They want their stuff to run on as many things as possible. |
30 |
|
31 |
They also have a huge codebase driven mostly by the primary OS of their |
32 |
users - Windows. Maintaining that is a large job so they'd like to |
33 |
re-use the bulk of it across all OSes. Business-wise this does make |
34 |
sense. But it does mean that they have to drop support for much built-in |
35 |
goodness on Linux (KMS, shipped OpenGL and more) and provide that bit |
36 |
themselves. No biggy - they have it all already for Windows. |
37 |
|
38 |
The kernel shim module is GPL'ed, but not in mainline, and there's this |
39 |
little thing about the Linux kernel - the famous stable api nonsense. So |
40 |
they are forever playing catch-up, and the kernel DOES rip out huge |
41 |
chunks of the api as and when needed. |
42 |
|
43 |
So what's nVidia to do? By and large their support for X11 is pretty |
44 |
good, and I've seen much worse. Yes, they are behind current kernel |
45 |
releases. No, they are not years behind. For the most part their code |
46 |
keeps up with the major binary distros. |
47 |
|
48 |
To those users who want to run whatever kernel Linux shipped today and |
49 |
expect nVidia to always keep up, I have an answer: get real people. |
50 |
|
51 |
nvidia support Linux, they never promised to keep up with Linus. Why do |
52 |
users think they have a right to demand something from a vendor that the |
53 |
vendor never promised to do? |
54 |
|
55 |
Why do some users think it correct and proper to demand the Gentoo |
56 |
maintainers jump through hoops to provide functionality that nVidia |
57 |
never promised, for versions they do not support *yet*? |
58 |
|
59 |
Seriously, to all the nVidia dumpers (not you Pandu), get a life people |
60 |
and get real. You bought hardware knowing full well what the conditions |
61 |
for drivers were going to be. The vendor does an OK job in the market |
62 |
place and if you don't like that, well that's tough. Buy different |
63 |
hardware. But don't expect entities to do stuff they never agreed to do. |
64 |
|
65 |
To Gentoo users who dump on this matter, when you installed Gentoo you |
66 |
implicitly agreed to be your own packager. There is no PPA or Yum repo |
67 |
and there's no paid staff member building packages. The work the |
68 |
packager does for Ubuntu and RedHat is now you now have to do yourself, |
69 |
and part of that is dealing with the times when shit don't work. |
70 |
|
71 |
|
72 |
|
73 |
-- |
74 |
Alan McKinnon |
75 |
alan.mckinnon@×××××.com |