Attention is currently required from: jolly, pespin.

laforge has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40490?usp=email )

Change subject: Submit all SQEs to kernel if they don't fit in the io_uring ......................................................................

Patch Set 3: Code-Review-1

(1 comment)

File src/core/osmo_io_uring.c:

https://gerrit.osmocom.org/c/libosmocore/+/40490/comment/5026b1ac_d27929cb?u... : PS3, Line 139: sqe = io_uring_get_sqe(&g_ring.ring);

I bet this can still probably return NULL since between call to io_uring_submit() and this line the […]

I doubt this patch is useful in general. The *submit* will not release a sqe. It just notifies the kernel that there are new sqe in the shared memory region. The kernel can then start to process them. Let's assume they're all reads, which will only ever complete if some remote node has sent us something. We cannot assume that this will be true between the two lines above. Even if some of our seq are write to a datagram socket, we have no guarantee that the kernel will have completed those between submit and get_sqe.