[RFC][PATCH 0/6][v3] Container-init signal semantics
Eric W. Biederman
ebiederm at xmission.com
Mon Dec 22 16:27:32 PST 2008
Sukadev Bhattiprolu <sukadev at linux.vnet.ibm.com> writes:
> Eric W. Biederman [ebiederm at xmission.com] wrote:
> | - container-init is responsible for setting the rest of the signals
> | to SIG_IGN.
> Oleg pointed out that we could drop SIG_DFL signals to global init early
> to ensure wait_for_completion_killable/lock_page_killable don't incorrectly
> believe that a fatal signal is pending. (patch 2/6).
> If that patch is valid regardless of containers, it would be a minor
> extension to get container-inits to drop SIG_DFL signals too, right ?
> So the bigger problem/unknown for me is the sig_from_user() in patch 4/6
> (i.e determining if it safe to deref the pid-ns of sender). We went from
> !in_interrupt() to the SIG_FROM_USER flag to this.
> If that is correct, I am hoping it would come down to opitmizing the code
> if possible (eg: can/should we avoid passing same_ns into sig_ignored()
> There is probably some ugliness :-) but do you see any other correctness
> issues ?
I haven't dug in too deep but right now my concern are user space semantics,
I don't want to wind up with something ugly there because we can not change
So if we can write a description of what happens to signals to cinit
that is right 100% of the time. Something we can write a test case
for that tests all of the corner cases and it always get the same
results. I am happy.
I don't mind dropping signals early as an optimization, but if it
is just an optimization we can't count on it in cinit.
So I would rather deliver less and make user space deal with it,
then deliver more cause problems for user space.
More information about the Containers