Thank you for replying.

On Mon 11/Nov/2024 01:59:48 +0100 Ludovic Bellière via Dng wrote:
> Hello Alessandro.
>
> Several quick observations
>
> 1. You can attach a file to your email instead of pasting text. That helps with
>    readability.


I didn't want to clog everybody's mailbox with useless data. I just posted the
few lines near the hanging point.


> 2. When there is a failure in the kernel, there is a neat "[ cut here ]" mark.
>    That means that the content under that mark is what is relevant to the
>    problem.
> 3. tomoyo is a LSM (linux security module), from Japan. It's in the same style
>    as AppArmor and SELinux.


After googling a bit, I found it is cooler than AppArmor in that you have to
explicitly enable it if you want it, rather than need to disable it if you don't.


> With that out of the way, about the hang. To me, it seems pretty clear...
> There's an issue with the graphics driver:
>
>> [ 2075.240903] ------------[ cut here ]------------
>> [ 2075.240912] Can't enable IRQ/MSI because no handler is installed
>> [ 2075.240972] WARNING: CPU: 5 PID: 3136 at drivers/gpu/drm/radeon/r100.c:727
>> r100_irq_set+0x93/0xb0 [radeon]

Hm...  The server has an ATI ES1000.  Indeed, the code at line 727 says:
if (!rdev->irq.installed) {
    WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
    WREG32(R_000040_GEN_INT_CNTL, 0);
    return -EINVAL;
}

However, I found no address to report to. As I have no particular problems
with the display, I'll leave it at that.


> Can't make head or tail of whatever else from the partial logs you pasted.


Do you think something went wrong much earlier than the hanging point?


> It hangs, yet you can use your keyboard? Can't you access your ttys? If you have
> access to your ttys, then you can do live troubleshooting. You can look at what
> is running and what should be running. You can try to start/restart the X
> server. All that stuff.


Ctrl-Alt-F<n>, n=1, 2, 3, ... doesn't work. Ctrl-Alt-Delete does.

X doesn't start automatically. I run startx after logging in on the terminal.
I have 2 as initdefault.


> Seems like the kernel loads properly, but the rest of the boot process isn't. A
> point to look into.


Perhaps I should have started in single-user mode and then manually run the
scripts in /etc/rc2.d, in lexicographical order, with some checks between one
and the next. Would that work?


Best
Ale
--