Re: [Libvir] PATCH: Prevent zombie ssh tunnels
Daniel Veillard wrote:
On Tue, Sep 11, 2007 at 11:35:46AM +0200, Gerd Hoffmann wrote:
> Daniel Veillard wrote:
>> - the ssh process dies
>> - libvirt based application takes some time to notice it
>> - the OS span a new process with the same PID after a PID rollabck
> Can not happen as long as libvirt hasn't asked for the exist status via
> waitpid() because the pid is still in use by the zombie ssh process.
Hum, which is precisely why we need the patch. Still I would feel a bit
better if we could check that priv->pid is a child of the current process
something like (getppid(priv->pid) == getpid()) test before any kill would
do this easilly I think.
I think Gerd's point was that as long as we haven't waited for the PID
within this process before, the PID cannot be reused.
That doesn't mean the situation cannot arise -- for example the main
program might be using other libraries as well as libvirt, and those
other libraries might blindly wait(2) for children.
Rich.
--
Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/
Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod
Street, Windsor, Berkshire, SL4 1TE, United Kingdom. Registered in
England and Wales under Company Registration No. 03798903
Attachments:
- smime.p7s (application/x-pkcs7-signature — 3.2 KB)