buck@siswat.UUCP (A. Lester Buck) (11/29/90)
I don't have first hand experience with SVR4 yet, though I have access to all the manuals that Prentice-Hall sells. I have heard various rumblings that asynchronous I/O was in the SVR4 design documents distributed some time ago. I can't find anything in the manuals about any type of asynchronous I/O support. But today a friend told me about all sorts of asynchronous I/O functions mentioned in standard header files in SVR4, in particular: /usr/include/sys/asyncio.h /usr/include/sys/asyncsys.h hints of an async(2) system call, an aiodaemon, an aioinit function lots of defines for AIOxxx symbols mention that this is Version 2, and Version 1 was a prototype Does anyone know the story? It would be best if AT&T stepped forward and spilled the beans, but, in the best traditions of Usenet, rumors and innuendo are welcome, too. :-) Even something from AT&T along the lines of "Sure it exists but if we told you about it, we would have to kill you." would be okay. Maybe only large application vendors get the details. :-( -- A. Lester Buck buck@siswat.lonestar.org ...!uhnix1!lobster!siswat!buck
bob@MorningStar.Com (Bob Sutterfield) (12/01/90)
In article <579@siswat.UUCP> buck@siswat.UUCP (A. Lester Buck) writes:
Does anyone know the story? It would be best if AT&T stepped
forward and spilled the beans... Even something from AT&T along the
lines of "Sure it exists but if we told you about it, we would have
to kill you." would be okay. Maybe only large application vendors
get the details. :-(
How about everyone with a Sun? SunOS 4.1 (billed as Sun's transition
to SVR4) includes all of <sys/asynch.h>, aiocancel(3), aioread(3),
aiowait(3), and aiowrite(3).
guy@auspex.auspex.com (Guy Harris) (12/07/90)
>How about everyone with a Sun? SunOS 4.1 (billed as Sun's transition >to SVR4) Yes, that's one thing that it is. It's also Sun's release that adds an asynchronous I/O mechanism, but the mechanism doesn't come from S5R4, as far as I know....
davidsen@sixhub.UUCP (Wm E. Davidsen Jr) (12/09/90)
In article <4719@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes: | >How about everyone with a Sun? SunOS 4.1 (billed as Sun's transition | >to SVR4) | | Yes, that's one thing that it is. It's also Sun's release that adds an | asynchronous I/O mechanism, but the mechanism doesn't come from S5R4, as | far as I know.... The header files are in Dell V.4, but I don't see aiowait or asiowait in the masterliblist file, which shows all the routines in every library. Don't know if they only got half, or if they took some out, or if the async stuff has those names redefined in the header files. I will look, but not this morning. -- bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) sysop *IX BBS and Public Access UNIX moderator of comp.binaries.ibm.pc and 80386 mailing list "Stupidity, like virtue, is its own reward" -me
buck@siswat.UUCP (A. Lester Buck) (12/10/90)
In article <2541@sixhub.UUCP>, davidsen@sixhub.UUCP (Wm E. Davidsen Jr) writes: < In article <4719@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes: < | >How about everyone with a Sun? SunOS 4.1 (billed as Sun's transition < | >to SVR4) < | < | Yes, that's one thing that it is. It's also Sun's release that adds an < | asynchronous I/O mechanism, but the mechanism doesn't come from S5R4, as < | far as I know.... < < The header files are in Dell V.4, but I don't see aiowait or asiowait < in the masterliblist file, which shows all the routines in every < library. Don't know if they only got half, or if they took some out, or < if the async stuff has those names redefined in the header files. I will < look, but not this morning. < -- < bill davidsen - davidsen@sixhub.uucp (uunet!crdgw1!sixhub!davidsen) And to answer my own original question... :-) Someone from UNIX System Laboratories returned my call and explained the situation with SVR4. No, SVR4 does not have any form of asynchronous I/O. A pre-release version did exist, but it was removed from the distributed code. Some of the STREAMS functionality depends on definitions in the asynch header files, so they had to stay. -- A. Lester Buck buck@siswat.lonestar.org ...!uhnix1!lobster!siswat!buck
shri@ncst.ernet.in (H.Shrikumar) (12/17/90)
In article <4719@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes: >>How about everyone with a Sun? SunOS 4.1 (billed as Sun's transition >>to SVR4) > >Yes, that's one thing that it is. It's also Sun's release that adds an >asynchronous I/O mechanism, but the mechanism doesn't come from S5R4, as >far as I know.... More info on the aynch I/O requested ... or at least do you have any pointers to it ? Is it done as a posted system call, with an early return, and if so does it invent a new SIG exception to handle failures ? And if (that too) so, how are pre-emptive quotas etc. implemented ? EMWTK. Thanx in advance. -- shrikumar ( shri@ncst.in )