1 |
Bill Longman wrote: |
2 |
> On 06/21/2010 12:01 PM, Dale wrote: |
3 |
> |
4 |
>> Alex Schuster wrote: |
5 |
>> |
6 |
>>> Dale writes: |
7 |
>>> |
8 |
>>> |
9 |
>>> |
10 |
>>>> I don't use genkernel anymore. I just roll my own. That way, I know |
11 |
>>>> what is in there and what is not. Then if something doesn't work, I |
12 |
>>>> know if it is the kernel or something else. With genkernel, you won't |
13 |
>>>> have a clue what it is since you don't know much if anything about the |
14 |
>>>> kernel and how it is configured. |
15 |
>>>> |
16 |
>>>> |
17 |
>>> That's not necessarily true. When I create a new kernel, I copy |
18 |
>>> /usr/src/linux/.config into the new kernel directory, make oldconfig and |
19 |
>>> menuconfig just as I like my kernel to be, and recreate the linux symlink |
20 |
>>> to the new kernel directory. Then I do a genkernel --install --lvm --luks |
21 |
>>> all&& emerge -a @module-rebuild, and am done. |
22 |
>>> I never noticed genkernel changing anything in my configuration, .config, |
23 |
>>> /proc/config.gz and the stuff in /etc/kernels/ are identical. Until not |
24 |
>>> long ago, I did not even know that genkernel was intended to create a |
25 |
>>> working kernel from scratch. |
26 |
>>> |
27 |
>>> Wonko |
28 |
>>> |
29 |
>>> |
30 |
>>> |
31 |
>> I always do mine this way. I copy the .config from the old kernel to |
32 |
>> the new kernel, run make oldconfig then afterwards make all&& make |
33 |
>> modules_install and then copy the kernel to /boot with my own numbering |
34 |
>> system. That way I know which version and series the kernel is. After |
35 |
>> that, edit grub with the new kernel and I'm done. I have only had that |
36 |
>> fail once in the past six years or so and the kernel made some serious |
37 |
>> changes and I had to start from scratch that one time. They moved |
38 |
>> things around and oldconfig couldn't reorganize things on the new kernel. |
39 |
>> |
40 |
>> Point being, genkernal causes issues for people and they don't know how |
41 |
>> to fix it because they expect genkernel to do everything. Problem with |
42 |
>> that is that usually when someone has a kernel problem, they use |
43 |
>> genkernel. If they do their own, it just works. Now someone new to |
44 |
>> building a kernel may need some help but apparently genkernel needs some |
45 |
>> help anyway. May as well learn how to roll your own. This is Gentoo |
46 |
>> after all. |
47 |
>> |
48 |
> The only thing that genkernel would add is your initrd. The kernel is |
49 |
> exactly the same, whether you compile it with "make" or through |
50 |
> "genkernel". Do a test and you'll see. (I'm assuming we're both talking |
51 |
> about gentoo-sources, not vanilla-sources. Either way, they'd be the |
52 |
> same.) Some might be confused about what happens in the steps if they |
53 |
> haven't been down the "kernel compilation trail" more than once or |
54 |
> twice, but for folks who just want to compile their kernel and plop it |
55 |
> into place, along with a hands-off initrd, it's rather handy. |
56 |
> |
57 |
> |
58 |
|
59 |
But only if it works. When I compile my kernel, I KNOW for sure what is |
60 |
in there. When genkernel does one, especially on a new install, I have |
61 |
no idea what is in it or what is not. If something goes wrong, I don't |
62 |
know where to start. Is it a kernel problem or is it something else? |
63 |
Who knows. Then you have to go back and start from the bottom, usually |
64 |
the kernel, and work your way back up to find out what is broken. |
65 |
|
66 |
Genkernel may work for you but that doesn't mean it does for everyone |
67 |
else. Should I mention hal here? When someone comes for help, your |
68 |
looking for the failure not the successes. If it was sucessful, they |
69 |
wouldn't need help. |
70 |
|
71 |
Dale |
72 |
|
73 |
:-) :-) |