public inbox for [email protected]
 help / color / mirror / Atom feed
* IOSQE_IO_LINK vs. short send of SOCK_STREAM
@ 2023-01-11 15:26 Ming Lei
  2023-01-11 15:49 ` Jens Axboe
  2023-01-11 16:32 ` Stefan Metzmacher
  0 siblings, 2 replies; 14+ messages in thread
From: Ming Lei @ 2023-01-11 15:26 UTC (permalink / raw)
  To: io-uring, Pavel Begunkov, Jens Axboe
  Cc: ming.lei, Stefan Metzmacher, David Ahern

Hello Guy,

Per my understanding, a short send on SOCK_STREAM should terminate the
remainder of the SQE chain built by IOSQE_IO_LINK.

But from my observation, this point isn't true when using io_sendmsg or
io_sendmsg_zc on TCP socket, and the other remainder of the chain still
can be completed after one short send is found. MSG_WAITALL is off.

For SOCK_STREAM, IOSQE_IO_LINK probably is the only way of io_uring for
sending data correctly in batch. However, it depends on the assumption
of chain termination by short send.

Appreciate any comment.

Thanks,
Ming


^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2023-01-14  3:47 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-01-11 15:26 IOSQE_IO_LINK vs. short send of SOCK_STREAM Ming Lei
2023-01-11 15:49 ` Jens Axboe
2023-01-11 16:32 ` Stefan Metzmacher
2023-01-11 16:36   ` Jens Axboe
2023-01-12  3:27   ` Ming Lei
2023-01-12  3:40     ` Jens Axboe
2023-01-12  7:35       ` Stefan Metzmacher
2023-01-13 10:12         ` Ming Lei
2023-01-13 17:51           ` Jens Axboe
2023-01-13 18:01             ` Jens Axboe
2023-01-14  0:27               ` Ming Lei
2023-01-14  1:39                 ` Ming Lei
2023-01-14  2:12                 ` Ming Lei
2023-01-14  3:47                   ` Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox