9P: Fix race in p9_read_work()
Race scenario between p9_read_work() and p9_poll_mux()
Data arrive, Rworksched is set, p9_read_work() is called.
thread A thread B
p9_read_work()
.
reads data
.
checks if new data ready. No.
.
gets preempted
.
More data arrive, p9_poll_mux() is called. .
.
.
p9_poll_mux() .
.
if (!test_and_set_bit(Rworksched, .
&m->wsched)) { .
schedule_work(&m->rq); .
} .
.
-> does not schedule work because .
Rworksched is set .
.
clear_bit(Rworksched, &m->wsched);
return;
No work has been scheduled, and yet data are waiting.
Currently p9_read_work() checks if there is data to read,
and if not, it clears Rworksched.
I think it should clear Rworksched first, and then check if there is data to read.
Signed-off-by: Simon Derr <simon.derr@bull.net>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c
index 6449bae..de1bbad 100644
--- a/net/9p/trans_fd.c
+++ b/net/9p/trans_fd.c
@@ -316,8 +316,7 @@
m->rsize - m->rpos);
p9_debug(P9_DEBUG_TRANS, "mux %p got %d bytes\n", m, err);
if (err == -EAGAIN) {
- clear_bit(Rworksched, &m->wsched);
- return;
+ goto end_clear;
}
if (err <= 0)
@@ -379,19 +378,20 @@
m->req = NULL;
}
+end_clear:
+ clear_bit(Rworksched, &m->wsched);
+
if (!list_empty(&m->req_list)) {
if (test_and_clear_bit(Rpending, &m->wsched))
n = POLLIN;
else
n = p9_fd_poll(m->client, NULL);
- if (n & POLLIN) {
+ if ((n & POLLIN) && !test_and_set_bit(Rworksched, &m->wsched)) {
p9_debug(P9_DEBUG_TRANS, "sched read work %p\n", m);
schedule_work(&m->rq);
- } else
- clear_bit(Rworksched, &m->wsched);
- } else
- clear_bit(Rworksched, &m->wsched);
+ }
+ }
return;
error: