On Mon, Aug 20, 2012 at 06:37:47PM -0400, bfields wrote:
> From: "J. Bruce Fields" <bfields@xxxxxxxxxx>
>
> The rpc server tries to ensure that there will be room to send a reply
> before it receives a request.
>
> It does this by tracking, in xpt_reserved, an upper bound on the total
> size of the replies that is has already committed to for the socket.
>
> Currently it is adding in the estimate for a new reply *before* it
> checks whether there is space available. If it finds that there is not
> space, it then subtracts the estimate back out.
>
> This may lead the subsequent svc_xprt_enqueue to decide that there is
> space after all.
>
> The results is a svc_recv() that will repeatedly return -EAGAIN, causing
> server threads to loop without doing any actual work.
>
> Cc: stable@xxxxxxxxxxxxxxx
> Reported-by: Michael Tokarev <mjt@xxxxxxxxxx>
> Tested-by: Michael Tokarev <mjt@xxxxxxxxxx>
> Signed-off-by: J. Bruce Fields <bfields@xxxxxxxxxx>
> ---
> net/sunrpc/svc_xprt.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> Queuing up for 3.6 absent any objections.--b.
By the way, one thing I'm still curious about is how this got
introduced. mjt bisected it to f03d78db65085609938fdb686238867e65003181
"net: refine {udp|tcp|sctp}_mem limits", which looks like it just made
the problem a little more likely.
The last substantive change to has_wspace logic was Trond's
47fcb03fefee2501e79176932a4184fc24d6f8ec, but I have a tough time
figuring out whether that would have affected it one way or the other.
As far as I can tell we've always added to xpt_reserved in this way, so
that svc_recv and svc_xprt_enqueue are comparing different things, and
surely this was always wrong even if the problem must have been harder
to trigger before.
But some of the wspace logic I don't understand, so cc'ing Neil and
Trond in case they see any other problem I missed.
--b.
>
> diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
> index 0d693a8..bac973a 100644
> --- a/net/sunrpc/svc_xprt.c
> +++ b/net/sunrpc/svc_xprt.c
> @@ -316,7 +316,6 @@ static bool svc_xprt_has_something_to_do(struct svc_xprt *xprt)
> */
> void svc_xprt_enqueue(struct svc_xprt *xprt)
> {
> - struct svc_serv *serv = xprt->xpt_server;
> struct svc_pool *pool;
> struct svc_rqst *rqstp;
> int cpu;
> @@ -362,8 +361,6 @@ void svc_xprt_enqueue(struct svc_xprt *xprt)
> rqstp, rqstp->rq_xprt);
> rqstp->rq_xprt = xprt;
> svc_xprt_get(xprt);
> - rqstp->rq_reserved = serv->sv_max_mesg;
> - atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
> pool->sp_stats.threads_woken++;
> wake_up(&rqstp->rq_wait);
> } else {
> @@ -640,8 +637,6 @@ int svc_recv(struct svc_rqst *rqstp, long timeout)
> if (xprt) {
> rqstp->rq_xprt = xprt;
> svc_xprt_get(xprt);
> - rqstp->rq_reserved = serv->sv_max_mesg;
> - atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
>
> /* As there is a shortage of threads and this request
> * had to be queued, don't allow the thread to wait so
> @@ -738,6 +733,8 @@ int svc_recv(struct svc_rqst *rqstp, long timeout)
> else
> len = xprt->xpt_ops->xpo_recvfrom(rqstp);
> dprintk("svc: got len=%d\n", len);
> + rqstp->rq_reserved = serv->sv_max_mesg;
> + atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
> }
> svc_xprt_received(xprt);
>
> --
> 1.7.9.5
>
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
