Attention is currently required from: jolly, laforge.

View Change

6 comments:

• Commit Message:

  • Patch Set #9, Line 23: to determine when all submissions are completed.

    Sounds like an easy thing to add right? It makes sense to implement this either her eor on a follow-up patch.

• File src/core/osmo_io_internal.h:

  • Patch Set #9, Line 110: bool read_enabled;

    (At some point it may make sense to properly put all these fields into a "read" and "write" substructs in here.)

• File src/core/osmo_io_uring.c:

  • Patch Set #9, Line 126: g_ring = talloc_zero(OTC_GLOBAL, struct osmo_io_uring);

    are you sure OTC_GLOBAL is defined at this point? Or is it actually NULL?
    If it's still NULL better pass NULL explicitly to avoid creating confusion.

  • Patch Set #9, Line 175: g_io_uring_size *= 2;

    Once you move this to an unsigned, you can do "<< 1" here. BTW, I don't see you are making sure we don't go over the "maximum ring size allowed" in there, the one you are checking in previous commit.

  • Patch Set #9, Line 208: /* All subsequent read SQEs must be on the same ring. */

    This should imho go inside iofd_uring_get_sqe(), based on "current_ring" being passed to it.

  • Patch Set #9, Line 282: while (iofd->u.uring.reads_submitted < ((iofd->u.uring.num_read_sqes) ? : g_io_uring_read_sqes)) {

    why is the g_io_uring_read_sqes reference needed here? How can iofd have num_read_sqes = 0, and if it does, why are we overriding it?

To view, visit change 40725. To unsubscribe, or for help writing mail filters, visit settings.