[PATCH RESEND] fs/epoll: fix the edge-triggered mode for nested epoll

[hev]

From: Heiher <r@xxxxxx>

The structure of event pools:
 efd[1]: { efd[2] (EPOLLIN) }        efd[0]: { efd[2] (EPOLLIN | EPOLLET) }
               |                                   |
               +-----------------+-----------------+
                                 |
                                 v
                             efd[2]: { sfd[0] (EPOLLIN) }

When sfd[0] to be readable:
 * the epoll_wait(efd[0], ..., 0) should return efd[2]'s events on first call,
   and returns 0 on next calls, because efd[2] is added in edge-triggered mode.
 * the epoll_wait(efd[1], ..., 0) should returns efd[2]'s events on every calls
   until efd[2] is not readable (epoll_wait(efd[2], ...) => 0), because efd[1]
   is added in level-triggered mode.
 * the epoll_wait(efd[2], ..., 0) should returns sfd[0]'s events on every calls
   until sfd[0] is not readable (read(sfd[0], ...) => EAGAIN), because sfd[0]
   is added in level-triggered mode.

Test code:
 #include <stdio.h>
 #include <unistd.h>
 #include <sys/epoll.h>
 #include <sys/socket.h>

 int main(int argc, char *argv[])
 {
 	int sfd[2];
 	int efd[3];
 	int nfds;
 	struct epoll_event e;

 	if (socketpair(AF_UNIX, SOCK_STREAM, 0, sfd) < 0)
 		goto out;

 	efd[0] = epoll_create(1);
 	if (efd[0] < 0)
 		goto out;

 	efd[1] = epoll_create(1);
 	if (efd[1] < 0)
 		goto out;

 	efd[2] = epoll_create(1);
 	if (efd[2] < 0)
 		goto out;

 	e.events = EPOLLIN;
 	if (epoll_ctl(efd[2], EPOLL_CTL_ADD, sfd[0], &e) < 0)
 		goto out;

 	e.events = EPOLLIN;
 	if (epoll_ctl(efd[1], EPOLL_CTL_ADD, efd[2], &e) < 0)
 		goto out;

 	e.events = EPOLLIN | EPOLLET;
 	if (epoll_ctl(efd[0], EPOLL_CTL_ADD, efd[2], &e) < 0)
 		goto out;

 	if (write(sfd[1], "w", 1) != 1)
 		goto out;

 	nfds = epoll_wait(efd[0], &e, 1, 0);
 	if (nfds != 1)
 		goto out;

 	nfds = epoll_wait(efd[0], &e, 1, 0);
 	if (nfds != 0)
 		goto out;

 	nfds = epoll_wait(efd[1], &e, 1, 0);
 	if (nfds != 1)
 		goto out;

 	nfds = epoll_wait(efd[1], &e, 1, 0);
 	if (nfds != 1)
 		goto out;

 	nfds = epoll_wait(efd[2], &e, 1, 0);
 	if (nfds != 1)
 		goto out;

 	nfds = epoll_wait(efd[2], &e, 1, 0);
 	if (nfds != 1)
 		goto out;

 	close(efd[2]);
 	close(efd[1]);
 	close(efd[0]);
 	close(sfd[0]);
 	close(sfd[1]);

 	printf("PASS\n");
 	return 0;

 out:
 	printf("FAIL\n");
 	return -1;
 }

Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Davide Libenzi <davidel@xxxxxxxxxxxxxxx>
Cc: Davidlohr Bueso <dave@xxxxxxxxxxxx>
Cc: Dominik Brodowski <linux@xxxxxxxxxxxxxxxxxxxx>
Cc: Eric Wong <e@xxxxxxxxx>
Cc: Jason Baron <jbaron@xxxxxxxxxx>
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Roman Penyaev <rpenyaev@xxxxxxx>
Cc: Sridhar Samudrala <sridhar.samudrala@xxxxxxxxx>
Cc: linux-kernel@xxxxxxxxxxxxxxx
Cc: linux-fsdevel@xxxxxxxxxxxxxxx
Signed-off-by: hev <r@xxxxxx>
---
 fs/eventpoll.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index d7f1f5011fac..a44cb27c636c 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -672,6 +672,7 @@ static __poll_t ep_scan_ready_list(struct eventpoll *ep,
 {
 	__poll_t res;
 	int pwake = 0;
+	int nwake = 0;
 	struct epitem *epi, *nepi;
 	LIST_HEAD(txlist);
 
@@ -685,6 +686,9 @@ static __poll_t ep_scan_ready_list(struct eventpoll *ep,
 	if (!ep_locked)
 		mutex_lock_nested(&ep->mtx, depth);
 
+	if (!depth || list_empty(&ep->rdllist))
+		nwake = 1;
+
 	/*
 	 * Steal the ready list, and re-init the original one to the
 	 * empty list. Also, set ep->ovflist to NULL so that events
@@ -739,7 +743,7 @@ static __poll_t ep_scan_ready_list(struct eventpoll *ep,
 	list_splice(&txlist, &ep->rdllist);
 	__pm_relax(ep->ws);
 
-	if (!list_empty(&ep->rdllist)) {
+	if (nwake && !list_empty(&ep->rdllist)) {
 		/*
 		 * Wake up (if active) both the eventpoll wait list and
 		 * the ->poll() wait list (delayed after we release the lock).
-- 
2.23.0