RDMA/iw_cxgb4: ensure eps don't get freed while the mutex is held

In rx_data(), with the ep in FPDU_MODE, refcnt=2, if we get unexpected
streaming data, we call c4iw_modify_rc_qp() and move the qp from
RTS -> TERMINATE.  In c4iw_modify_rc_qp(), if rdma_fini() returns
an error, the ep will be dereferenced (refcnt=1).  Then rx_data()
calls c4iw_ep_disconnect() which starts the close operation.
But if send_halfclose() fails in c4iw_ep_disconnect(), we  will call
release_ep_resources() derefing the ep which reduces the refcnt to 0 and
and frees the ep. However we still has the ep mutex at that point, so we
have a touch-after-free bug.  There is a similar issue where
peer_close() calls c4iw_ep_disconnect().

The solution is to add a reference to the ep in c4iw_ep_disconnect()
after acquiring  the mutex, and release it after releasing the mutex.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>

diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index cc9836e468005809f2992b4558e20b309d8ff657..12eac98661c143d59d92212588bd8f2cac764716 100644 (file)
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -3416,6 +3416,12 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
        PDBG("%s ep %p state %s, abrupt %d\n", __func__, ep,
             states[ep->com.state], abrupt);
 
+       /*
+        * Ref the ep here in case we have fatal errors causing the
+        * ep to be released and freed.
+        */
+       c4iw_get_ep(&ep->com);
+
        rdev = &ep->com.dev->rdev;
        if (c4iw_fatal_error(rdev)) {
                fatal = 1;
@@ -3476,6 +3482,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
                }
        }
        mutex_unlock(&ep->com.mutex);
+       c4iw_put_ep(&ep->com);
        if (fatal)
                release_ep_resources(ep);
        return ret;