[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Kernel 2.6.30 blocks on udev initialisation



On 20.01.2010 10:39, Rémi Moyen wrote:
> Hi,
> 
> Until recently, I was still on a (quite old) 2.6.26 kernel (stock one
> from testing). I also had an old version of udev (< 150). A few days
> ago, I made a full update of my system. It first tried to upgrade udev
> to 150, but when doing this complained that my current kernel had the
> CONFIG_SYSFS_DEPRECATED option activated and that it was no more
> supported and likely not to work at all. Although it was also trying
> to install the new kernel at the same time (in which I believe this
> option is no more activated), I guess since I was not the current
> active kernel, this caused the error.
> 
> So I first installed the new kernel (2.6.30), then wanted to reboot on
> this kernel (so still with the old udev) in order to complete the udev
> installation. When rebooting, the new kernel loads, one of the very
> first thing it says is "udev: waiting for /dev to be fully populated",
> but then nothing happens. After 1 min (and every minute after), I get
> a message about the CPU being stuck since 60 secs, but even after 4-5
> minutes, nothing more happens.
> 
> After rebooting on the old kernel (2.6.26), I installed the 2.6.32
> kernel from unstable. This one worked fine with the old udev. I
> finished the update of udev (150) under this kernel, and rebooted
> again: the 2.6.32 kernel works fine, the 2.6.30 has still exactly the
> same problem. The 2.6.26 kernel, although still with the non-supported
> kernel config option on, also works correctly. On all working kernels,
> I get a bunch of warning about deprecated udev options that have to be
> changed in various files, but apart from this, I don't see any
> problems and all my hardware is working fine.
> 
> Has anyone an idea as to what happens with the 2.6.30 kernel? I didn't
> find anything in the bug reports nor on the web about this.
> 
> Thanks!

I think it was the 2.6.31 kernel version where the
CONFIG_SYSFS_DEPRECATED bug was fixed. The kernel option had the sense
inverted.
I hope i remember correctly... didn't lookup the change-log again.

best regards

Mart


Reply to: