1 |
On Sun, Feb 04, 2007 at 01:43:22PM -0600, Dan Farrell wrote: |
2 |
> On Sun, 4 Feb 2007 12:39:07 -0500 |
3 |
> Jerry McBride <mcbrides9@×××××××.net> wrote: |
4 |
> |
5 |
> > On Sunday 04 February 2007 08:51:44 am Etaoin Shrdlu wrote: |
6 |
> > > On Sunday 4 February 2007 13:45, Jorge Almeida wrote: |
7 |
> > > > I was under the impression that it is necessary to install Nvidia |
8 |
> > > > drivers. Is this correct? |
9 |
> > > |
10 |
> > > If the link is not in place, nvidia-drivers will not build. |
11 |
> > > |
12 |
> > > > And what should be done with the System.map file? Copy it to /boot |
13 |
> > > > under which name? (I mean, when booting with a particular kernel, |
14 |
> > > > how does the kernel know the path to the correct System.map?) |
15 |
> > > |
16 |
> > > There used to be a good system.map explanation here: |
17 |
> > > |
18 |
> > > http://dirac.org/linux/system.map/ |
19 |
> > > |
20 |
> > > however, it seems to have some problem at the moment. The google |
21 |
> > > cached copy works (just do a search for linux system.map, it's the |
22 |
> > > first hit). |
23 |
> > > |
24 |
> > > Sorry, cannot comment on whether the symlink should or should not be |
25 |
> > > created, and it's been a lot since I built my last LFS. I have |
26 |
> > > almost always used it, and never had any problem. IIRC the |
27 |
> > > handbook's advice is to create the link. But, as always, YMMV. |
28 |
> > |
29 |
> > The server ar dirac.org need a kick, I emailed the admin there. As |
30 |
> > for the symlink with LFS... LFS is a totally different animal than |
31 |
> > most distributions. For us, the link is required in order to emerge |
32 |
> > anything that touches the kernel sources. |
33 |
> > |
34 |
> I stopped copying System.map very often and it doesn't seem very |
35 |
> useful.. I think copying it to /boot is mostly for recovery purposes, |
36 |
> like .config being copied into /boot. that way you can rebuild the |
37 |
> kernel if you need to. |
38 |
|
39 |
AFAIK about System.map is it is not for recovery, but for |
40 |
debugging. If kernel crashes, it tries get one to tell where it crashed, |
41 |
give you backtrace and like that. It's easier to find the bug with it, |
42 |
so unless you do not have a kernel that will ever crash, or you do not |
43 |
want to send bug reports, you do not need one. |
44 |
|
45 |
-- |
46 |
BOFH Excuse #452: |
47 |
|
48 |
Somebody ran the operating system through a spelling checker. |
49 |
|
50 |
Michal 'vorner' Vaner |