> I noticed that you filed a bug [1] on K3b using a sarge version. Have
> you also tried to reproduce the problem using the version of K3b that is
> in unstable?
Yes, I just reproduced it using a fresh sid chroot.
While k3b started from one user was writing, k3b started from another user
hanged at 'Scanning for devices' stage for a long time (several minutes).
Although it unblocked before writing was complete, several minutes of
freeze does not seem to be dummy-user-friendly :).
Most likely, this is caused by the fact that k3b uses some device ioclt
while scanning, which in-kernel implementation tries to gain an exclusive
lock to the device. Which in turn blocks entire process until another
in-kernel usage unblocks the device.
Proper fix will probably need some cooperation between k3b and kernel
people, to make it possible to read device characteristics (or whatever
k3b needs while "scanning devices") without blocking.
To make situation a bit better involving k3b only, k3b probably should
split scanning ioctls into a separate thread (to avoid GUI freeze) and
inform user what's going on (e.g. "looks like device is locked by another
user, waiting for lock to be released"). This will not solve the problem -
user will still be unable to start writing on other writer - but anyway
it's better than hanging.
Nikita
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
