io_uring: remove bogus RLIMIT_NOFILE check in file registration
authorJens Axboe <axboe@kernel.dk>
Fri, 3 Apr 2020 19:54:26 +0000 (13:54 -0600)
committerJens Axboe <axboe@kernel.dk>
Fri, 3 Apr 2020 19:56:44 +0000 (13:56 -0600)
We already checked this limit when the file was opened, and we keep it
open in the file table. Hence when we added unit_inflight to the count
we want to register, we're doubly accounting these files. This results
in -EMFILE for file registration, if we're at half the limit.

Cc: stable@vger.kernel.org # v5.1+
Signed-off-by: Jens Axboe <axboe@kernel.dk>
fs/io_uring.c

index 2460c3333f700999dd394407c13e8db7d424d25e..ce76157c2f959721889e1191264194da5947aa15 100644 (file)
@@ -6216,13 +6216,6 @@ static int __io_sqe_files_scm(struct io_ring_ctx *ctx, int nr, int offset)
        struct sk_buff *skb;
        int i, nr_files;
 
-       if (!capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN)) {
-               unsigned long inflight = ctx->user->unix_inflight + nr;
-
-               if (inflight > task_rlimit(current, RLIMIT_NOFILE))
-                       return -EMFILE;
-       }
-
        fpl = kzalloc(sizeof(*fpl), GFP_KERNEL);
        if (!fpl)
                return -ENOMEM;