It's looking like IO_uring zero-copy receive support should be ready for the Linux 6.15 kernel cycle this spring.

IO_uring lead developer Jens Axboe [1]shared on X today that the networking side of the IO_uring receive zero-copy support was finally merged to the net-next code ahead of the Linux 6.15 kernel.

The IO_uring side patches still need to be queued up by Axboe, but with the networking subsystem side work now queued and one month to go until the Linux 6.15 cycle, it looks like it could cross the finish line for the next kernel version.

As explained of this IO_uring zero-copy receive support in the networking [2]merge :

"This patchset contains net/ patches needed by a new io_uring request implementing zero copy rx into userspace pages, eliminating a kernel to user copy.

We configure a page pool that a driver uses to fill a hw rx queue to hand out user pages instead of kernel pages. Any data that ends up hitting this hw rx queue will thus be dma'd into userspace memory directly, without needing to be bounced through kernel memory. 'Reading' data out of a socket instead becomes a _notification_ mechanism, where the kernel tells userspace where the data is. The overall approach is similar to the devmem TCP proposal."

The benchmark results shown in that merge are extremely positive:

Looking great and exciting to finally see this approaching mainline after thirteen rounds of patch review!



[1] https://x.com/axboe/status/1887868202295660654?s=09

[2] https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=71f0dd5a3293d75d26d405ffbaedfdda4836af32