home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!mintaka.lcs.mit.edu!hal.gnu.ai.mit.edu!mycroft
- From: mycroft@hal.gnu.ai.mit.edu (Charles Hannum)
- Newsgroups: comp.unix.aix
- Subject: Re: AIX 3.2.2 and amd
- Message-ID: <1992Sep1.181754.26655@mintaka.lcs.mit.edu>
- Date: 1 Sep 92 18:17:54 GMT
- References: <5396@unixhub.SLAC.Stanford.EDU>
- Sender: news@mintaka.lcs.mit.edu
- Organization: /etc/organization
- Lines: 105
-
-
- In article <5396@unixhub.SLAC.Stanford.EDU>
- joann@unixhub.slac.stanford.edu writes:
- >
- > We are now trying to bring up 3.2.2, and are having trouble with amd.
- > Is anyone else running it?
-
- I'm not sure whether you're talking about `amd' or really about
- `automount' here. If `amd', then you probably need two patches I
- recently sent to the amd-workers list:
-
-
- From: mycroft@gnu.ai.mit.edu
- Message-Id: <9208261848.AA22775@hal.gnu.ai.mit.edu>
- Subject: Patch to 5.3beta1; memory used after being free()d
- To: amd-workers@acl.lanl.gov
- Date: Wed, 26 Aug 92 14:48:55 EDT
-
-
- The following patch fixes a condition where amd was free()ing memory
- and then using it immediately. This was causing it to dump core on
- some systems with finicky malloc()s.
-
- Also, config/os-aix3.h should #include <sys/time.h> for AIX 3.2.
-
- -----8<-----snip-----8<-----snip-----8<-----snip-----8<-----snip-----8<-----
- *** amd/afs_ops.c.orig Sun May 31 12:36:42 1992
- --- amd/afs_ops.c Wed Aug 26 14:28:59 1992
- ***************
- *** 1537,1544 ****
- return new_mp;
- }
-
- ! if (error && (cp->mp->am_mnt->mf_ops == &efs_ops))
- ! cp->mp->am_error = error;
-
- assign_error_mntfs(new_mp);
-
- --- 1537,1544 ----
- return new_mp;
- }
-
- ! if (error && (new_mp->am_mnt->mf_ops == &efs_ops))
- ! new_mp->am_error = error;
-
- assign_error_mntfs(new_mp);
-
- -----8<-----snip-----8<-----snip-----8<-----snip-----8<-----snip-----8<-----
-
-
- From: mycroft@gnu.ai.mit.edu
- Message-Id: <9208270802.AA55732@hal.gnu.ai.mit.edu>
- Subject: Another bug in 5.3beta1
- To: amd-workers@acl.lanl.gov
- Date: Thu, 27 Aug 92 4:02:35 EDT
-
-
- The old version of str3cat() interacted badly with the mnt_dup() in
- mtab_aix.c. The following patch fixes this by just having str3cat()
- use xmalloc() rather than xrealloc(). I don't see any good reason not
- to have it work this way.
-
- (I don't know of any more bugs at the moment.)
-
- -----8<-----snip-----8<-----snip-----8<-----snip-----8<-----snip-----8<-----
- *** util.c.orig Sat Mar 7 12:51:42 1992
- --- util.c Thu Aug 27 03:52:32 1992
- ***************
- *** 79,92 ****
- char *s2;
- char *s3;
- {
- int l1 = strlen(s1);
- int l2 = strlen(s2);
- int l3 = strlen(s3);
- ! p = (char *) xrealloc(p, l1 + l2 + l3 + 1);
- ! bcopy(s1, p, l1);
- ! bcopy(s2, p + l1, l2);
- ! bcopy(s3, p + l1 + l2, l3 + 1);
- ! return p;
- }
-
- char *strealloc(p, s)
- --- 79,94 ----
- char *s2;
- char *s3;
- {
- + char *new_p;
- int l1 = strlen(s1);
- int l2 = strlen(s2);
- int l3 = strlen(s3);
- ! new_p = (char *) xmalloc(l1 + l2 + l3 + 1);
- ! bcopy(s1, new_p, l1);
- ! bcopy(s2, new_p + l1, l2);
- ! bcopy(s3, new_p + l1 + l2, l3 + 1);
- ! free(p);
- ! return new_p;
- }
-
- char *strealloc(p, s)
- -----8<-----snip-----8<-----snip-----8<-----snip-----8<-----snip-----8<-----
-
- --
- \ Charles Hannum, mycroft@ai.mit.edu
- /\ White heterosexual atheist male (WHAM) pride!
-