Daniel P. Berrange wrote:
On Tue, Jul 03, 2007 at 04:12:56PM +0100, Richard W.M. Jones wrote:
> Here we just fix the remote driver so that if the server goes down, the
> process doesn't die on SIGPIPE.
This isn't thread safe. Changing signals with sigaction affects the
entire process - so if you've got multiple threads running with
multiple remote connections open, you can get overlapping calls.
Either we should just block sigpipe in virInitialize, or mandate
that the application needs to block sigpipe itself. I'm not sure
really what best course is - I'd probably lean towards saying it is
the application's responsibility to deal with sig handlers.
Yes, you're right. I'll change this so that instead it adds some
supporting documentation for app writers.
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