[arch-general] Kernel freeze on boot with kernel 3.10.x

[Martti Kühne]

On Sun, Sep 1, 2013 at 1:48 PM, Nils Becker <nilsc.becker at gmail.com> wrote:
> Hello everyone,
>
> since kernel version >3.9.x (e.g. 3.10.3, 3.10.10) my system freezes
> during the boot. I was hoping for a fix to appear in later versions of
> 3.10. but as this was not the case, I am in dire need of help to debug
> this problem as currently I am not even getting any logs from the crash.
>
> Problem: After grub is loaded the system loads the initial ram disk and
> I am prompted to enter the password for my root partition. After that
> the system starts normally (systemd messages) and freezes at some
> point. The last message printed differs a little bit between boots, but
> it is roughly about "Mounted Config File System".
> The freeze is apparently a complete crash: I started a systemd rescue
> console on tty9 (which startup is printed out) but I cannot switch in
> that console after the freeze (no response on key events).
> Apparently the logs are not yet written to the harddisk as I can not
> find any entries on that boot if I startup another kernel.
>
> I would be grateful for any pointers on this issue and/or some
> suggestions how I can at least get a useful log from the crash to open
> a bug report.
>
> I am using the standard arch linux packages (at least where the bootup
> is involved). My system is using a raid1 with encrypted root partition.
> And yes, it works flawlessly if I downgrade the linux package to
> <3.10.x.
>
> My hardware configuration (hwinfo output) can be found under:
> www.ncbecker.de/pc/hwinfo
> I ran hwinfo with the lts-kernel.

Dunno what to do with that, I can't even tell if you're not a HP
pavillon as in [2].

As I made sure to fully be clear about with my last kernel problem,
could you tell us which 3.9.x exactly was the first to break for you?
I made absolutely sure it was 3.10.6 which was the first that broke
for me and I found the fix in the arch forums [0], I just believe
there's quite an importance in knowing this - as it also would shorten
down bisecting kernel-git quite a bit. Look, similar issues can be
found in [1], [2] and [3] at least. To be very unfair, seriously, you
took your time to speak up...

cheers!
mar77i

[0] https://bbs.archlinux.org/viewtopic.php?pid=1319586
[1] https://bbs.archlinux.org/viewtopic.php?pid=1299482
[2] https://bugs.archlinux.org/task/34358
[3] https://bugs.archlinux.org/task/35188