similar to: [Bug 374] New: uidswap.c doesn't compile on SCO 3.2v4.2

Ok, for those running NeXT and other platforms with broken/missing _POSIX_SAVED_ID please try this patch, and anyone that has spent any amount of time dealing with this problem. I believe it's right. BTW, this patch is no where near as big as it looks. The patch was done against an earily version of the tree which had an issue with white space. - Ben --- ../openssh/uidswap.c Sun Apr 22

Hello, I was trying to build a chrooted sftp account when I faced a problem. The chroot is done with the patch present in the contrib subdirectory in the portable version (I'm under linux slackware current). My problem is that verifying access rights on directories and files are too tight and then I couldn't have the following things : The user sftp, with primary group sftp, is chrooted

Hi, the below patch drops another test for uid 0 on Cygwin. It's embarassing that I never found it. Actually temporarily_use_uid never worked on Cygwin due to that. So far that had no influence, but now that we have activated another feature which makes Cygwin more POSIX-like, somebody on the Cygwin list found that agent forwarding didn't work anymore. The reason is that due to the

openssh-2.9.9p2 assumes that pid_t, uid_t, gid_t, and mode_t are no wider than int. GCC complains about this assumption on 32-bit Solaris 8 sparc, where these types are 'long', not 'int'. This isn't an actual problem at runtime on this host, as long and int are the same width, but it is a problem on other hosts where pid_t is wider than int. E.g., I've heard that 64-bit

Hello Porters, I've finally (thanks to Wendy Palm of Cray) ported OpenSSH to Dynix v4.4.4. I had to make sure that "UseLogin" was set to "no" in the sshd_config file. Also, here are the old-style contextual diffs (obtained with 'diff -c' on the Dynix box) of the two files I had to change: *** configure Sat Jun 16 17:09:50 2001 --- configure.new Mon Oct 8

Hello Porters, I am attempting to compile OpenSSH 2.9.9p2 on a Dynix V4.4.4 host. I have set USE_PIPES and BROKEN_SAVED_UIDS (the latter because there are no functions for set{eu,eg}id() that I can find). I configured with "./configure '--with-libs=-lnsl -lsec'". Each time I attempt to login, I get this error: No utmp entry. You must exec "login" from

Here's a unified diff, proposed by Christos Zoulas (with a little reworking to get the configure.in part a bit cleaner). The main target of the patches are NetBSD compatability, although there are some changes making if (pointer X) into if (X != NULL) as well. Christos also proposed a change to Makefile.in, but I'll send that as a seperate mail. Thanks, David Index: configure.in

http://bugzilla.mindrot.org/show_bug.cgi?id=374 mouring at eviladmin.org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED ------- Additional Comments From mouring at eviladmin.org

Hi, the following patches are in the NetBSD packages collection to get OpenSSH 1.2.2 going. Changes: * /dev/urandom may be there but not in the kernel - make sure * Pull in some more headers needed by configure * Add proper ld-flags for ELF platforms * Some code cleanup * Install example files always to a different dir, and use out build system to DTRT WRT getting them to /etc.

http://bugzilla.mindrot.org/show_bug.cgi?id=136 Summary: setgid() deemed to fail for non-suid ssh client on linux if using other than primary group Product: Portable OpenSSH Version: 3.0.2p1 Platform: ix86 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: ssh

Linux has just raised the NGROUPS_MAX limit from 32 to 64k. In doing an audit of various tools, openssh turned up as having incorrect groups handling. Almost no user-space apps really care about NGROUPS_MAX. A proposed patch (untested, since the CVS build won't compile on my RH box.. :-/) : What think? Index: uidswap.c ===================================================================

Below I've includes a patch which helps build OpenSSH outside from a read-only source tree, find OpenSSL on Mac OS X, and fix a typo. This applies to OpenSSH 2.1.1p4. You should already have gotten a note from Melissa O'Neil about a conflict with the crc32() symbol in zlib, which was causing a crash on Darwin. I've noticed another bug. If ssh is setuid, I get a permission

All, Could someone explain the purpose of the uidswap functions with respect to ssh ( the client ). From what I gathered , ssh installs as setuid root and swaps ids when reading potential key files that may be read only by root. Also , I think when binding to a privileged port ssh swaps id. Is that so? What are the consequnences if you do not install ssh setuid root? ( As far I as know no uid

I am enclosing a context diff of the changes that I made to get openssh working on UWIN. UWIN is a UNIX operating system layer that runs on Win32 systems. For more information on UWIN go to http://www.research.att.com/sw/tools/uwin/. I also ran configure using -with-cppflags=-D_BSDCOMP=2. I don't know where that information would go with the source code. Let me know if you need more

A change was recently introduced into uidswap.c to cover the case where the user is root. The change is "&& pw->pw_uid != 0 &&". /* Try restoration of GID if changed (test clearing of saved gid) */ if (old_gid != pw->pw_gid && pw->pw_uid != 0 && (setgid(old_gid) != -1 || setegid(old_gid) != -1))

Hi, I'm terribly sorry that I missed this before 3.7p1 was out. The permanently_set_uid() function fails on Cygwin since the test to revert to the saved uid unfortunately works on Cygwin though it shouldn't. The reason is that a Windows NT process always can revert to its previous privileges. There's no such concept of giving up rights in a process permanently. This is only

Fixes a potential (but probably rather unlikely) use after free bug in function temporarily_use_uid(), file uidswap.c. --- a/uidswap.c +++ b/uidswap.c @@ -113,8 +113,9 @@ temporarily_use_uid(struct passwd *pw) } } /* Set the effective uid to the given (unprivileged) uid. */ - if (setgroups(user_groupslen, user_groups) < 0) -

http://bugzilla.mindrot.org/show_bug.cgi?id=787 Summary: Minor security problem due to use of deprecated NGROUPS_MAX in uidswap.c (sshd) Product: Portable OpenSSH Version: 3.7.1p2 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: sshd AssignedTo:

[resending with uidswap.c instead of uidwrap.c] Once I got past the missing inet_ntoa.h weirdness, I ran into an sshd that died a lot. It appears that IRIX doesn't like some of the extra checks added between 1.23 and 1.24 of uidswap.c. Not sure if that constitutes an IRIX bug or not, but helpfully this helps someone. -- Mail: mjo at dojo.mi.org WWW: http://dojo.mi.org/~mjo/ Phone: +1

EHLO, Somehow I don't think it makes any sense to test whether the gid/egid can be changed, if the original uid happened to be root. Root can always change the gid/egid anyhow. So, I would like to propose the following change to 3.9p1... --- uidswap.c.orig Sun Aug 29 15:43:57 2004 +++ uidswap.c Sun Aug 29 15:44:05 2004 @@ -201,7 +201,7 @@ #endif /* Try restoration of GID if