Hello list,
I have a smbfs mounted samba-share that got lost due to a network error.
Every process accessing the mount point falls into uniterruptable sleep - till the reboot. Even a lsof is frozen.
Is this a bug or a known issue?

It's known, and it's a bug. The network part of smbfs has a few problems.

I did some patches that used poll to check if they could send things and timeout after 30 seconds. Perhaps combined with checking the tcp state would work for this case.

The overly long delay (and most people would assume it is frozen) is one or more network timeouts (tcp level?) that needs to pass before the process stops.

That the mount fails to recover could be related to smbmount not managing to stay alive.

For 2.5 there is work going on that should fix this at the same time as it allows oplock support and stuff. The process should always be interruptible.

/Urban