| Message ID | 20250107162724.GA18926@redhat.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | poll_wait: add mb() to fix theoretical race between waitqueue_active() and .poll() | expand |
diff --git a/include/linux/poll.h b/include/linux/poll.h index fc641b50f129..57b6d1ccd8bf 100644 --- a/include/linux/poll.h +++ b/include/linux/poll.h @@ -41,7 +41,7 @@ typedef struct poll_table_struct { static inline void poll_wait(struct file * filp, wait_queue_head_t * wait_address, poll_table *p) { - if (p && p->_qproc && wait_address) { + if (p && p->_qproc) { p->_qproc(filp, wait_address, p); /* * This memory barrier is paired in the wq_has_sleeper().
This check is historical and no longer needed, wait_address is never NULL. These days we rely on the poll_table->_qproc check. NULL if select/poll is not going to sleep, or it already has a data to report, or all waiters have already been registered after the 1st iteration. However, poll_table *p can be NULL, see p9_fd_poll() for example, so we can't remove the "p != NULL" check. Link: https://lore.kernel.org/all/20250106180325.GF7233@redhat.com/ Signed-off-by: Oleg Nesterov <oleg@redhat.com> --- include/linux/poll.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)