On 8/12/24 15:46, Mark Harmstone wrote:
On 12/8/24 12:26, Christoph Hellwig wrote:
What is the point if this doesn't actually do anything but returning
-EIOCBQUEUED?
It returns EIOCBQUEUED to say that io_uring has queued the request, and
adds the task to io_uring's thread pool for it to be completed.
No, it doesn't. With your patch it'll be executed by the
task submitting the request and seemingly get blocked,
which is not how an async interface can work.
I'll comment on the main patch.
Note that that the internals of the btrfs encoded read is built
around kiocbs anyway, so you might as well turn things upside down,
implement a real async io_uring cmd and just wait for it to complete
to implement the existing synchronous ioctl.
I'd have to look into it, but that sounds like it could be an
interesting future refactor.
Mark
--
Pavel Begunkov
