home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!vnet.ibm.com
- From: Mickey_Coggins@vnet.ibm.com
- Message-ID: <19930123.071307.356@almaden.ibm.com>
- Date: Sat, 23 Jan 93 15:43:25 CET
- Cc: steves@ssecmail.ssec.wisc.edu
- Newsgroups: comp.unix.aix
- Subject: Re: Problem with AIX3.2 paging subsystem?
- Reply-To: Mickey_Coggins@vnet.ibm.com
- Organization: IBM
- Disclaimer: This posting represents the poster's views, not those of IBM
- News-Software: UReply 3.1
- References: <1993Jan21.230535.4139@cs.wisc.edu>
- Lines: 52
-
- In <1993Jan21.230535.4139@cs.wisc.edu> Steve Shirshac writes:
- >I think I may have found a problem in the AIX 3.2 paging subsystem. If
- >you have any experience with developing device drivers for AIX, I would
- >appreciate any insights you might have into this problem or how to
- >isolate it further. A rather detailed description follows.
- > ....
- > privileged user process ("fred") allocates a large shared mem
- > segment (16Mb), "fred" pins shared mem seg, "fred" opens our
- > device and passes a pointer to the shared mem seg to my device
- > driver.
- > ....
- > upon privileged user process completion:
- >
- > "svmon" indicates that the pages used to back the shared mem
- > segment have been unpinned.
-
- svmon shows the segment to be unpinned, but is it still there?
- Run a nasty script like the following to see what svmon thinks
- the paging space usage is:
- svmon -Sa 10000 | cut -c45-53 | awk '/0-9Y/{print $1"+f"}' | dc | tail -1
-
- Compare this with pstat -s, vmstat 2 2, lsps -a, (they will all return
- slightly different values, by the way).
-
- If you are getting wildly different values, this could be a bug in AIX.
- Take a system dump, using the sysdumpstart command, and send it into
- the AIX defect support people. You'll have to create some kind of
- "simple" test case.
-
- I'd keep checking your driver for possible problems.
-
- Another thing to try is to map a 16MB file (create and touch all pages
- before you use it) to shared memory using the shmat() system call,
- (see info for the appropriate flags). You might want to create,
- bzero, unlink, map, run, exit and see if the file space is still
- taken up. Could provide a clue.
-
- > "ipcs" indicates that the shared memory segments which were
- > being used are no longer there.
-
- >Steve Shirshac (steves@ssecmail.ssec.wisc.edu)
- >
-
- Viel spass,
- --
- Mickey Coggins Mickey_Coggins@vnet.ibm.com
- +41-1-436-7274 / +41-1-436-8525 (fax)
- Hohlstrasse 560, CH-8048 Zurich, Switzerland
- _______________
- "All who have meditated on the art of governing mankind
- have been convinced that the fate of empires depends on
- the education of youth." - Aristotle
-