On 9/14/20 2:05 AM, Stefano Garzarella wrote: > On Fri, Sep 11, 2020 at 09:36:02AM -0600, Jens Axboe wrote: >> On 9/11/20 7:34 AM, Stefano Garzarella wrote: >>> These new errors are added with the restriction series recently >>> merged in io_uring (Linux 5.10). >>> >>> Signed-off-by: Stefano Garzarella <sgarzare@xxxxxxxxxx> >>> --- >>> man/io_uring_enter.2 | 18 ++++++++++++++++++ >>> 1 file changed, 18 insertions(+) >>> >>> diff --git a/man/io_uring_enter.2 b/man/io_uring_enter.2 >>> index 5443d5f..4773dfd 100644 >>> --- a/man/io_uring_enter.2 >>> +++ b/man/io_uring_enter.2 >>> @@ -842,6 +842,16 @@ is set appropriately. >>> .PP >>> .SH ERRORS >>> .TP >>> +.B EACCES >>> +The >>> +.I flags >>> +field or >>> +.I opcode >>> +in a submission queue entry is not allowed due to registered restrictions. >>> +See >>> +.BR io_uring_register (2) >>> +for details on how restrictions work. >>> +.TP >>> .B EAGAIN >>> The kernel was unable to allocate memory for the request, or otherwise ran out >>> of resources to handle it. The application should wait for some completions and >>> @@ -861,6 +871,14 @@ field in the submission queue entry is invalid, or the >>> flag was set in the submission queue entry, but no files were registered >>> with the io_uring instance. >>> .TP >>> +.B EBADFD >>> +The >>> +.I fd >>> +field in the submission queue entry is valid, but the io_uring ring is not >>> +in the right state (enabled). See >>> +.BR io_uring_register (2) >>> +for details on how to enable the ring. >>> +.TP >> >> I actually think some of this needs general updating. io_uring_enter() >> will not return an error on behalf of an sqe, it'll only return an error >> if one happened outside the context of a specific sqe. Any error >> specific to an sqe will generate a cqe with the result. > > Mmm, right. > > For example in this case, EACCES is returned by a cqe and EBADFD is > returned by io_uring_enter(). > > Should we create 2 error sections? Yep, I think we should. One that describes that io_uring_enter() would return in terms of errors, and one that describes cqe->res returns. Are you up for this? Would be a great change, making it a lot more accurate. -- Jens Axboe
