io_uring: fix return value when removing provided buffers

When a request to remove buffers is submitted, and the given number to be
removed is larger than available in the specified buffer group, the
resulting CQE result will be the number of removed buffers + 1, which is
1 more than it should be.

Previously, the head was part of the list and it got removed after the
loop, so the increment was needed. Now, the head is not an element of
the list, so the increment shouldn't be there anymore.

Fixes: dbc7d45 ("io_uring: manage provided buffers strictly ordered")
Signed-off-by: Wojciech Lukowicz <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jens Axboe <[email protected]>

Author: wlukowicz

io_uring/kbuf.c

@@ -228,7 +228,6 @@ static int __io_remove_buffers(struct io_ring_ctx *ctx,
 		return i;
 	}
 
-	/* the head kbuf is the list itself */
 	while (!list_empty(&bl->buf_list)) {
 		struct io_buffer *nxt;
 
@@ -238,7 +237,6 @@ static int __io_remove_buffers(struct io_ring_ctx *ctx,
 			return i;
 		cond_resched();
 	}
-	i++;
 
 	return i;
 }