On 8/14/2012 5:55 AM, Neil Horman wrote:
On Mon, Aug 13, 2012 at 07:43:27PM -0700, John Fastabend wrote:
quoted
A socket fd passed in a SCM_RIGHTS datagram was not getting
updated with the new tasks cgrp prioidx. This leaves IO on
the socket tagged with the old tasks priority.
To fix this add a check in the scm recvmsg path to update the
sock cgrp prioidx with the new tasks value.
Thanks to Al Viro for catching this.
CC: Neil Horman <nhorman@tuxdriver.com>
Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: John Fastabend <redacted>
---
[...]
quoted
@@ -281,6 +283,9 @@ void scm_detach_fds(struct msghdr *msg, struct scm_cookie *scm)
}
/* Bump the usage count and install the file. */
get_file(fp[i]);
+ sock = sock_from_file(fp[i], &err);
+ if (sock)
+ sock->sk->sk_cgrp_prioidx = prioidx;
nit: You can replace the prioidx variable above and this set with a call to
sock_update_netprioidx
Neil
OK but then I should also make sock_update_netprioidx inline and drop
the in_interrupt() call. I'll send a v3 with this change and also a
third patch to fix a race between write_priomap and cgrp_create (also
spotted by Al Viro).