On 05/29/2012 09:51 PM, Wen Congyang wrote:
At 05/30/2012 09:20 AM, Eric Blake Wrote:
> KAMEZAWA Hiroyuki reported a nasty double-free bug when virCommand
> is used to convert a string into input to a child command. The
> problem is that the poll() loop of virCommandProcessIO would close()
> the write end of the pipe in order to let the child see EOF, then
> the caller virCommandRun() would also close the same fd number, with
> the second close possibly nuking an fd opened by some other thread
> in the meantime. This in turn can have all sorts of bad effects.
>
> This is based on his first attempt at a patch, at
>
https://bugzilla.redhat.com/show_bug.cgi?id=823716
close fd more twice is the cause of this bug. But there are some
other codes that have the same problem. I am searching all such
codes recent days.
Thanks for helping on that front.
> + if (inoff == inlen &&
VIR_CLOSE(*inpipe) < 0)
> + VIR_DEBUG("ignoring failed close on fd %d",
infd);
> + infd = -1;
if inoff != inlen, we should not set infd to -1.
Oh, good catch. I'll post a v2.
--
Eric Blake eblake(a)redhat.com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org