[Bug 1656] GSoC AIO patch

bugzilla-daemon at rtems.org bugzilla-daemon at rtems.org
Mon Aug 16 11:58:30 CDT 2010


--- Comment #25 from Alin Rus <alin.codejunkie at gmail.com> 2010-08-16 11:58:29 CDT ---
(In reply to comment #19)
> aio_init() appears to use a posix thread, mutex, and condition variable.
> These should be accounted for in confdefs.h and in doc/user/conf.t.
> A configuration variable like CONFIGURE_ENABLE_POSIX_AIO and then propagate
> the requirements into the various objects. The GNAT/Ada configuration should
> be a good guide.

Ok will take a look at it.

> > As before, no actual testing yet.
> I just did a coverage run on erc32.  Both psxaio01 and psxaio02 failed.
> Obviously both should pass as a minimum.  But the aio code needs to have
> 100% code coverage so getting these to run is a minimum.  With the addition
> of AIO code and these tests failing, we went from 72 bytes (18 instructions)
> uncovered to this much just in the aio code (bytes).
> 1772 aio_misc.c
> 28 aio_suspend.c
> 28 aio_fsync.c
> 12 aio_write.c
> 12 aio_read.c
> So submit a patch against the cvs head to get these tests running and then
> we can see what is not exercised by your tests.

Will do.

> I noticed that psxenosys broke with this patch.  It broke because some
> of the aio_ routines are implemented but there remain 3 that are not:

> $ grep ENOSYS aio*
> aio_cancel.c:  rtems_set_errno_and_return_minus_one( ENOSYS );
> aio_fsync.c:  rtems_set_errno_and_return_minus_one( ENOSYS );
> aio_suspend.c:  rtems_set_errno_and_return_minus_one( ENOSYS );

aio_cancel is being worked on, it's finished but not tested.

aio_fsync() and aio_suspend are missing because of this
http://www.rtems.com/wiki/index.php/POSIX_Asynchronous_IO . I see you're the
last editor of the page.

Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.

More information about the rtems-bugs mailing list