| 1 |
On 06/02/2014 18:08, James wrote: |
| 2 |
>> [working theory: the kernel throws permission denied errors when it's |
| 3 |
>> > asked in weird ways to load wrong versioned modules. Pure speculation, |
| 4 |
>> > I've never done this at all and don't know what the error is] |
| 5 |
> The permission are all consistent now (/lib/modules/*). I'm not sure how |
| 6 |
> they got wacked, as I have not done anything with modules yet. Nore |
| 7 |
> anything messing with those perms...... |
| 8 |
> |
| 9 |
> Obviously, from the strings command, the kernel(s) need fixing up a bit. |
| 10 |
> I only got them to a point, to get the openbox stuff setup. The audio |
| 11 |
> and usb automounting are all that is left to fix.... The points is the |
| 12 |
> kernels should be good enough to work with? I suspect grub2, as this |
| 13 |
> is my first forray into a bootable system with grub2........ |
| 14 |
> |
| 15 |
> |
| 16 |
> What is stumping me is why all three kernels boot, but the modules |
| 17 |
> only point to 3.10.25, even when boot either the second |
| 18 |
> (kernel-3.13.0-gentoo-r1) kernel or the third (config-3.13.1-gentoo) |
| 19 |
> kernel. |
| 20 |
> |
| 21 |
> I'm going to work on this and scratch a bit..... So any other suggestions |
| 22 |
> are welcome, although it'll be a few days until I post back. Got |
| 23 |
> any strings options/scripts to only filter out the english readable |
| 24 |
> parts of mostly binary files? Manual parsing is a drag...... |
| 25 |
> |
| 26 |
> What/where could the system be corrupted to only attempt to use those |
| 27 |
> modules from 3.10.25? |
| 28 |
> |
| 29 |
|
| 30 |
|
| 31 |
Your kernels look fine actually. All kernel images have that large |
| 32 |
collection of attention-grabbing strings, they seem to be regular |
| 33 |
compiled in error messages. And the version numbers are fine. |
| 34 |
|
| 35 |
I've just built 3.13.1-gentoo here and my perms come out right, so it's |
| 36 |
not a bug that e.g. snuck into that one version's Makefile. And I know |
| 37 |
of no way for a bootloader to influence what a kernel image thinks it's |
| 38 |
version is or how to get to it's modules (it would be a huge attack |
| 39 |
vector if a bootloader could do that) |
| 40 |
|
| 41 |
I think I'm all out of ideas now, you might have to consult with some |
| 42 |
kernel guys. Per my understanding, what you describe cannot happen, so |
| 43 |
it's an odd one indeed. |
| 44 |
|
| 45 |
|
| 46 |
|
| 47 |
-- |
| 48 |
Alan McKinnon |
| 49 |
alan.mckinnon@×××××.com |
Archives