[PATCH] Don't bother checking for O_ASYNC, glibc's had it since '98.

Alan Coopersmith Alan.Coopersmith at Sun.COM
Fri Mar 12 07:59:30 PST 2010

Matt Turner wrote:
> On Fri, Mar 12, 2010 at 2:33 AM, Alan Coopersmith
> <Alan.Coopersmith at sun.com> wrote:
>> But other kernels/libc's still don't have it in 2010.   Yes, seriously, 2010.
>> (Hence the #ifdef I_SETSIG code in there for Solaris & other SysV flavors.)
>> Delete the Linux libc5 comment, and the first hunk to define O_ASYNC to FSYNC
>> if you want, but please don't delete the later #ifdefs.
>>        -alan-
> Ugh. OK.
> I noticed that there aren't code paths for ifndef O_ASYNC. Is this OK?
> It operates cleanly without it?

If you don't have O_ASYNC, I_SETSIG or some other equivalent, you don't get
SIGIO support, and simply fall back to reading the input events when select()
reports there's new data to read from the fd, via the same code in os/*.c
that's used for all the non-xfree86 ddxen.

	-Alan Coopersmith-           alan.coopersmith at sun.com
	 Oracle Solaris Platform Engineering: X Window System

More information about the xorg-devel mailing list