1 |
> If /proc/config.gz exists when you boot from that image you could |
2 |
>> compare that to the .config file that you're using to see what the |
3 |
>> differences are. |
4 |
>> |
5 |
>> |
6 |
> I tried it. I spend much time on it. |
7 |
|
8 |
|
9 |
For me the only way is to use a GUI diff program like Meld, otherwise i find |
10 |
it too difficult to notice the differences. :) |
11 |
|
12 |
|
13 |
> I also tried to use the config of the allround boot disks (in |
14 |
> /proc/config.gz of the running livecd system). I tried to figure out |
15 |
> differences and also tried to copy it to the Xen-Kernel and then started |
16 |
> 'make menuconfig' for the Xen kernel to fix dependencies cleared in 'make |
17 |
> menuconfig'. It leads to kernels that |
18 |
> * cannot be build due to compile errors, |
19 |
> * or kernels that run perfectly on G3 (generation 3 HP hardware) but not on |
20 |
> G4 or G6. |
21 |
> |
22 |
> Maybe this is an driver version issue for the 'Compaq SMART2 support' |
23 |
> and the 'Compaq Smart Array 5xxx support'. |
24 |
> But fixing a kernel, delivered by portage, is completly new to me. So i |
25 |
> tried to ship around this... |
26 |
> |
27 |
> Would you suggest to try this way? Should I try to fix the Xen-Kernel from |
28 |
> Gentoo with some HP driver source code? Or could it be a completly other |
29 |
> issue? |
30 |
> |
31 |
> I'd be very confident that the driver code will be identical in xen kernel |
32 |
to regular kernels of the same version. You could verify this by running |
33 |
md5sum against the modules source codes. If they are the same there's no |
34 |
need to worry about the driver code. |
35 |
|
36 |
Put your xen .config and the config from the boot disk that works up |
37 |
somewhere public - maybe we can spot the problem? |