similar to: openssh-2.9.9p2 assumes pid_t, uid_t, etc. are not 'long'

Is anyone bothered by the compiler warnings that indicate that the format strings don't match the associated variables? I was, so I cast most of the objectionable args (pids, uids, gids) to "long", and added an "l" (el) to the format string. A single item was cast to an int. Here's the patch. If you haven't applied my UseLogin patch, the line numbers in

In several places, openssh-2.9.9p2 passes a 'char' value to a ctype macro like 'isdigit'. This has undefined behavior on hosts with signed characters, if the character value happens to be negative. For example, isdigit('\200') expands to an array access that is a subscript error on hosts with signed characters where '\200' == -128. This leads to incorrect results,

I found this problem on 32-bit Solaris 8 sparc. session.c uses two void functions without declaring them first; this violates the C standard, which requires declarations for such functions. Here's a patch. 2001-09-28 Paul Eggert <eggert at twinsun.com> * session.c (do_pre_login, record_utmp_only): Add missing declarations.

I'm using OpenSSH 2.3.0p1. When my users use ssh for the first time, using rhosts authentication, entropy.c drops the privs in prng_write_seedfile() at the setuid(original_uid) line (line 550, approx): void prng_write_seedfile(void) { int fd; char seed[1024]; char filename[1024]; struct passwd *pw; /* Don't bother if we have already saved a seed */ if (prng_seed_saved) return;

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greetings, I am not sure if this is the correct place to ask these question, if I am at the wrong place please advise. I am currently working on some modifications to openssh which record the users rsa/dsa identity comment file to a log file when the user logs in (password authentication is disabled). The ssh1 portion of the modification works

Hello, all- Apparently, under Solaris (I can personally confirm SunOS 5.7 and 5.8), pam_open_session will generate a segfault if PAM_TTY is not set. The obvious symptom of this is that OpenSSH 2.9.9p2 will segfault on any operation that does not request a tty (do_exec_no_pty). Based on a quick google search, this seems to have been encountered by others, though the specific symptoms seem to

Hi, Solaris (tested with 2.6) needs a username in the logout record in the wtmpx file. Currently openssh (version 2.3.0p1) leaves the username (utmpx.ut_user) empty in logout records, which leads to conflicting results from the last command. Example: # last -5 siegert siegert pts/186 stikine.ucs.sfu. Mon Jan 15 14:26 still logged in siegert pts/105 stikine.ucs.sfu. Mon Jan 15

http://bugzilla.mindrot.org/show_bug.cgi?id=419 Summary: HP-UX PAM problems with 3.5p1 Product: Portable OpenSSH Version: -current Platform: HPPA OS/Version: HP-UX Status: NEW Severity: normal Priority: P2 Component: sshd AssignedTo: openssh-unix-dev at mindrot.org ReportedBy:

Without this patch "ssh-agent -d > ~/ssh-agent.sh" will produce a zero byte file. Obviously a corner case, but for what I'm doing it's a show-stopper, and it _seems_ like an obvious improvement to correctness, rather than relying on implicit newline flushing with TTYs and flush-on-exit with the forking mode. Not subscribed, so please CC me on any replies. ---

I'm fairly new to the list and new to submitting patches. Can someone please verify the attached patch for running a HP-UX Trusted System with PAM and OpenSSH 3.4p1? The problem seemed to be that pam couldn't verify the user via __pamh after the call to permanently_set_uid in session.c. So I called do_pam_session prior to the call and added a function do_pam_set_tty in order to set the

Hi. I have a box (a net5501 with a Perle serial octal RS-232 card in it) that I want to use as a console server for a bunch of headless computers. I was wondering if there's a trivial app that I can run as the shell that ssh (or telnet) would run, example: #!/bin/bash port=$(echo "$SSH_CONNECT" | cut -d' ' -f4) let -i tty=$port-2000 ttyname=$(printf

After upgrading to openssh-2.5.2p2, my users were unable to login using ssh Protocol 1. Entries like this were showing up in syslog: Apr 5 19:29:45 maple sshd[16726]: Accepted password for anthonyu from ::ffff:192.168.0.2 port 1019 Apr 5 19:29:45 maple sshd[16726]: fatal: stat(/dev/pts/1 19:29:45 sshd[16726]: Accepted password for anthonyu) failed: No such file or directory Apr 5 19:29:45

In case it is any help, here is the patch against openssh-3.0.1p1 that corrects the problem with last reporting on Solaris that I sent to the list a week or so ago against 3.0p1. There was no conversation about this aside from Rip Loomis' comment about including it to support BSM auditing - does this present a problem for other OSes to include the ut_name field in the utmpx entry? Should this

Hello, The error message in the subject line occurs with the new 2.3.0 openssh version and appeared in the previous snapshots on our Solaris systems. As far as I remember it was reported, but have not seen any more about this. I have looked into it a little bit. First, the file session.c (line 1849 onwards): debug("session_exit_message: release channel %d", s->chanid);

On Fri, Feb 12, 2016 at 9:25 AM, Randall S. Becker <rsbecker at nexbridge.com> wrote: > On February 10, 2016 10:23 AM, I wrote: >> On February 9, 2016 9:30 PM, Darren Tucker wrote: >> [...] >> > This one looks odd. The ssh session itself looks OK: it authenticates >> > then sends a printf shell command (basically, just a way of >> > guaranteeing a

Hi All, I've been poking around various parts of the auth code for a while. Some platforms support failed login counters and it occurs to me that there's as few too many instances of: #ifdef [PLATFORM] if (authenticated == 0 && strcmp(method, "password") == 0) some_login_failure_func(); #endif The attached patch creates a record_failed_login() function in

http://bugzilla.mindrot.org/show_bug.cgi?id=84 Summary: last command provides incorrect information on Solaris 8 Product: Portable OpenSSH Version: -current Platform: UltraSparc OS/Version: Solaris Status: NEW Severity: normal Priority: P2 Component: ssh AssignedTo: openssh-unix-dev at mindrot.org

Hi All. First the questions: Is there anything objectionable in this patch? Is AUDIT_FAIL_AUTH appropriate for the "Reason" field? Now the details: attached is a patch that changes some of the #includes for AIX. It moves the AIX-specific includes to port-aix.h and adds includes that contain the prototypes for many of the authentication functions. The idea isto fix some warnings.

Hi, News from the "AIX is different than the rest of the world" department... AIX has something similar to setluid() on SCO, just that it uses text strings (similar to setenv()) and calls it "usrinfo". I've appended the man page below. Under normal conditions, well-behaved applications use ttyname(), logname() and getuid() get the relevant informations, but today

I have put together a simple set of diffs that corrects the problem described by Steven Fishback <sfishback at interpath.net> on 10-30 on this list regarding incorrect information reported by last on Solaris. The patches merely pass along the username in the utmpx record for a logout. Is there any reason why this would be a problem with other OSes? If not, maybe this could be rolled into the