On 4/18/22 2:50 PM, Jens Axboe wrote: > On 4/18/22 1:51 PM, Pavel Begunkov wrote: >> All io_assign_file() callers do error handling themselves, >> req_set_fail() in the io_assign_file()'s fail path needlessly bloats the >> kernel and is not the best abstraction to have. Simplify the error path. >> >> Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx> >> --- >> fs/io_uring.c | 6 +----- >> 1 file changed, 1 insertion(+), 5 deletions(-) >> >> diff --git a/fs/io_uring.c b/fs/io_uring.c >> index 423427e2203f..9626bc1cb0a0 100644 >> --- a/fs/io_uring.c >> +++ b/fs/io_uring.c >> @@ -7117,12 +7117,8 @@ static bool io_assign_file(struct io_kiocb *req, unsigned int issue_flags) >> req->file = io_file_get_fixed(req, req->cqe.fd, issue_flags); >> else >> req->file = io_file_get_normal(req, req->cqe.fd); >> - if (req->file) >> - return true; >> >> - req_set_fail(req); >> - req->cqe.res = -EBADF; >> - return false; >> + return !!req->file; > > Wouldn't it be cleaner to just do: As soon as I sent that, realize we're missing the file assignment in that case. Stupid. Looks fine as-is. -- Jens Axboe
