kernelnewbies.kernelnewbies.org archive mirror
 help / color / mirror / Atom feed
* confuse about generic_file_read_iter
@ 2014-05-05  3:39 Jimmy Li
  0 siblings, 0 replies; only message in thread
From: Jimmy Li @ 2014-05-05  3:39 UTC (permalink / raw)
  To: kernelnewbies

Hello everyone,

I'm trying to figure out how asyncchronous I/O implemented, but I'm
confusing about the generic_file_read_iter in mm/filemap.c.
major file systems use this function as read_iter for it's
file_operations structure(eg, ext3).

And it's called by syscall io_submit(fs/aio.c:1359):
   sys_io_submit -> do_io_submit -> io_submit_one -> aio_run_iocb

The point I'm confusing is generic_file_read_iter seems do noting about
asyncchronous I/O at all, as I see what generic_file_read_iter does:
   if opened with O_DIRECT
     flush pages;
	 call direct_IO;
   else
     call do_generic_file_read();
Actually, I think it's syncchronous.

It's that means file systems using generic_file_read_iter does not
support aio(such as ext3)?

I also tried `git log fs/aio.c`. I found retry-based AIO is removed at
commit 41003a7bcfed1255032e1e7c7b487e505b22e298(Tue May 7 16:18:25 2013)
Could anyone tell me how does linux AIO implement now?

I'm working on linux-next(20140502)
thanks.

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2014-05-05  3:39 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-05-05  3:39 confuse about generic_file_read_iter Jimmy Li

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).