On Sun, 4 Jun 2000, Alan Cox wrote:
> > I Asynchronous I/O (aio_*)
>
> In glibc
I'll say yes and no to this one. Yes, the aio_* calls should be
implemented in glibc -- POSIX AIO just isn't meant to be in the kernel.
But we do need a mechanism for exporting async io to userland, and
read/write from clones just isn't it.
Btw, I have cooked up an async io mechanism which I'm just about ready for
comments on. Briefly, it's /dev/aio which you submit requests to, and
read() (or mmap a ring buffer) to get the results of completed requests
from. Requests are grouped, and you only get a completion notice when the
entire group is done. Internally, everything is done via a fully
asyncronous rw_kiovec call. At present, it's working on raw devices, with
the generic page cache routines still to come. As soon as it's tidied up,
I'll post it for feedback.
-ben
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:18 EST