2021-08-25 21:10:48

by Trond Myklebust

Subject: [PATCH] SUNRPC: Fix XPT_BUSY flag leakage in svc_handle_xprt()...

From: Trond Myklebust <[email protected]>

If the attempt to reserve a slot fails, we currently leak the XPT_BUSY
flag on the socket. Among other things, this make it impossible to close
the socket.

Fixes: 82011c80b3ec ("SUNRPC: Move svc_xprt_received() call sites")
Signed-off-by: Trond Myklebust <[email protected]>
---
net/sunrpc/svc_xprt.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
index 5f0d33ca4bdb..b3cff4077899 100644
--- a/net/sunrpc/svc_xprt.c
+++ b/net/sunrpc/svc_xprt.c
@@ -975,7 +975,8 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt)
rqstp->rq_stime = ktime_get();
rqstp->rq_reserved = serv->sv_max_mesg;
atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
- }
+ } else
+ svc_xprt_received(xprt);
out:
trace_svc_handle_xprt(xprt, len);
return len;
--
2.31.1

2021-08-25 21:10:48

by Chuck Lever

Subject: Re: [PATCH] SUNRPC: Fix XPT_BUSY flag leakage in svc_handle_xprt()...

> On Aug 25, 2021, at 3:33 PM, [email protected] wrote:
>
> From: Trond Myklebust <[email protected]>
>
> If the attempt to reserve a slot fails, we currently leak the XPT_BUSY
> flag on the socket. Among other things, this make it impossible to close
> the socket.
>
> Fixes: 82011c80b3ec ("SUNRPC: Move svc_xprt_received() call sites")
> Signed-off-by: Trond Myklebust <[email protected]>
> ---
> net/sunrpc/svc_xprt.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
> index 5f0d33ca4bdb..b3cff4077899 100644
> --- a/net/sunrpc/svc_xprt.c
> +++ b/net/sunrpc/svc_xprt.c
> @@ -975,7 +975,8 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt)
> rqstp->rq_stime = ktime_get();
> rqstp->rq_reserved = serv->sv_max_mesg;
> atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
> - }
> + } else
> + svc_xprt_received(xprt);
> out:
> trace_svc_handle_xprt(xprt, len);
> return len;
> --
> 2.31.1

Looks correct. Bruce, perhaps you should pull this for 5.14-rc.
Linus says he expects to release 5.14-final this Sunday, fyi.

--
Chuck Lever

2021-08-25 21:12:37

by J. Bruce Fields

Subject: Re: [PATCH] SUNRPC: Fix XPT_BUSY flag leakage in svc_handle_xprt()...

I wonder if this would explain
https://bugzilla.kernel.org/show_bug.cgi?id=213887.

On Wed, Aug 25, 2021 at 3:51 PM Chuck Lever III <[email protected]> wrote:
>
>
>
> > On Aug 25, 2021, at 3:33 PM, [email protected] wrote:
> >
> > From: Trond Myklebust <[email protected]>
> >
> > If the attempt to reserve a slot fails, we currently leak the XPT_BUSY
> > flag on the socket. Among other things, this make it impossible to close
> > the socket.
> >
> > Fixes: 82011c80b3ec ("SUNRPC: Move svc_xprt_received() call sites")
> > Signed-off-by: Trond Myklebust <[email protected]>
> > ---
> > net/sunrpc/svc_xprt.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
> > index 5f0d33ca4bdb..b3cff4077899 100644
> > --- a/net/sunrpc/svc_xprt.c
> > +++ b/net/sunrpc/svc_xprt.c
> > @@ -975,7 +975,8 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt)
> > rqstp->rq_stime = ktime_get();
> > rqstp->rq_reserved = serv->sv_max_mesg;
> > atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved);
> > - }
> > + } else
> > + svc_xprt_received(xprt);
> > out:
> > trace_svc_handle_xprt(xprt, len);
> > return len;
> > --
> > 2.31.1
>
> Looks correct. Bruce, perhaps you should pull this for 5.14-rc.
> Linus says he expects to release 5.14-final this Sunday, fyi.
>
> --
> Chuck Lever
>
>
>