More syscall messaging commits, and some testing code as well.
dillon at apollo.backplane.com
Wed Aug 13 10:40:17 PDT 2003
:> That doesn't mean every system call needs to be made interruptable. It
:> just means that long-lived system calls need to be interruptable: like
:> NFS mounts are if you turn on intr (not sure why anyone doesn't turn
:> that on).
:because a lot of (stupid?) apps don't really handle the case of their backing
:store going away very efficiently. if its a temporary thing you really should
:have your apps block, but if I send an app a kill -9, even if its just as root,
:I'd prefer them to die.
:*sigh* we can all wish..
:Adrian Chadd <angryskul> learning is bad
I don't think it's all that difficult to make kill -9 work even if a
program is stuck in a hard mounted NFS operation. I would agree
with such a change if it were submitted, as long as (like with 'intr'
mounts) the signal does not abort the operation until NFS has gone
through a few retries. Maybe we could call it the 'medium well done'
option. Not too hard, not too soft :-).
<dillon at xxxxxxxxxxxxx>
More information about the Kernel