Re: [RFC PATCHv2 1/1] io_uring: add support for IORING_SETUP_SQE_MIXED

[Keith Busch <kbusch@kernel.org>]

On Wed, Sep 10, 2025 at 10:44:10AM -0700, Caleb Sander Mateos wrote:
> On Thu, Sep 4, 2025 at 12:27 PM Keith Busch <kbusch@meta.com> wrote:
> > diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
> > index 04ebff33d0e62..9cef9085f52ee 100644
> > --- a/include/uapi/linux/io_uring.h
> > +++ b/include/uapi/linux/io_uring.h
> > @@ -146,6 +146,7 @@ enum io_uring_sqe_flags_bit {
> >         IOSQE_ASYNC_BIT,
> >         IOSQE_BUFFER_SELECT_BIT,
> >         IOSQE_CQE_SKIP_SUCCESS_BIT,
> > +       IOSQE_SQE_128B_BIT,
> 
> Have you given any thought to how we would handle the likely scenario
> that we want to define more SQE flags in the future? Are there
> existing unused bytes of the SQE where the new flags could go? If not,
> we may need to repurpose some existing but rarely used field. And then
> we'd likely want to reserve this last flag bit to specify whether the
> SQE is using this "extended flags" field.

Yeah, I mentioned in the cover letter it may not okay to take this bit
for the cause. Using it this way is just a simple way forward for the
proof-of-concept to iron out handling mixed SQEs everywhere else. I
wouldn't remove the "RFC" prefix until we have agreement on how to flag
a big SQE command on a mixed SQ. One option, for example, might take the
highest opcode bit since we're a ways off off from needing it for more
ops.