search for: child_termin

...is a patch against an older version of OpenSSH with other stuff anyways, BUT, it's so TRIVIAL(*), that you can see how it would apply to newer versions (which I've not tried). Here's the gist: server_loop2() has a race condition with respect to reception of SIGCHLD and checking/setting child_terminated. This patch does two things: wait_until_can_do_something() adds a 1 second timeout to select() IF AND ONLY IF (!channel_still_open) AND, server_loop2() breaks out of its loop when there are no sessions left. Blocking SIGCHLD before select()ing would not fix the problem, nor would that be very...

...children of sshd remain (even as zombies), and it has no other interesting open fds. If you send a SIGCHLD to the hung sshd, it wakes up and exits. As far as I can see, there's a race condition in wait_until_can_do_something(), both in RedHat 2.5.2p2-5 and in the latest CVS sources. It tests child_terminated, and sets a non-zero timeout if so, before calling select(). However, there is a very small window (between checking child_terminated and calling select() in which a SIGCHLD can arrive and set child_terminated. If this happens, and there is no other activity from the client or the child fds, ss...

http://bugzilla.mindrot.org/show_bug.cgi?id=52 dtucker at zip.com.au changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #1075 is|0 |1 obsolete| | Attachment #1098 is|0 |1 obsolete|

...th hangs on exit from ssh and sftp sessions. A quick look at the OpenSSH source code revealed the following: In serverloop.c there is a signal handler defined for SIGCHLD as follows: static void sigchld_handler(int sig) { int save_errno = errno; debug("Received SIGCHLD."); child_terminated = 1; #ifndef _UNICOS mysignal(SIGCHLD, sigchld_handler); #endif notify_parent(); errno = save_errno; } As far as I can tell the primary purpose of this method is to set a value for child_terminated which is referenced by the following method: static void collect_children(vo...

I don't already figure out what is the real impact of this but I think there is a typo in function sigchld_handler() in serverloop.c (l 75). It is written if (WIFEXITED(child_wait_status) || WIFSIGNALED(child_wait_status)) child_terminated = 1; child_has_selected = 0; But I think one actually means: if (WIFEXITED(child_wait_status) || WIFSIGNALED(child_wait_status)) { child_terminated = 1; child_has_selected = 0; } Regards, Philippe

...u -r1.82 serverloop.c --- serverloop.c 10 Oct 2001 22:18:47 -0000 1.82 +++ serverloop.c 11 Oct 2001 18:06:33 -0000 @@ -92,6 +92,45 @@ /* prototypes */ static void server_init_dispatch(void); +/* + * we write to this pipe if a SIGCHLD is caught in order to avoid + * the race between select() and child_terminated + */ +static int notify_pipe[2]; +static void +notify_setup(void) +{ + if (pipe(notify_pipe) < 0) { + error("pipe(notify_pipe) failed %s", strerror(errno)); + notify_pipe[0] = -1; /* read end */ + notify_pipe[1] = -1; /* write end */ + } else { + set_nonblock(notify_pipe[0]); +...

...a diff of what I did (basically just resetting it as it was for 2.3.0p1): *** serverloop.c.orig Thu Jan 18 08:41:13 EST 2001 --- serverloop.c Wed Jan 17 16:01:41 EST 2001 *************** *** 109,115 **** int save_errno = errno; debug("Received SIGCHLD."); child_terminated = 1; - signal(SIGCHLD, sigchld_handler2); errno = save_errno; } --- 109,114 ---- *************** *** 667,672 **** --- 666,672 ---- if (child_terminated) { while ((pid = waitpid(-1, &status, WNOHANG)) > 0)...

...ren() not handling ECHILD: { pid_t pid; sigset_t oset, nset; int status; /* block SIGCHLD while we check for dead children */ sigemptyset(&nset); sigaddset(&nset, SIGCHLD); sigprocmask(SIG_BLOCK, &nset, &oset); if (child_terminated) { while ((pid = waitpid(-1, &status, WNOHANG)) > 0 || (pid < 0 && errno == EINTR)) if (pid > 0) session_close_by_pid(pid, status); child_termi...

...static int fderr_eof = 0; /* EOF encountered readung from fderr. */ static int fdin_is_tty = 0; /* fdin points to a tty. */ @@ -107,6 +108,9 @@ wait_pid, child_pid); if (WIFEXITED(child_wait_status) || - WIFSIGNALED(child_wait_status)) + WIFSIGNALED(child_wait_status)) { child_terminated = 1; + if (fdout_maybe_eof) + fdout_eof = 1; + } } signal(SIGCHLD, sigchld_handler); @@ -338,10 +342,14 @@ /* Read and buffer any available stdout data from the program. */ - if (!fdout_eof && FD_ISSET(fdout, readset)) { + if (!fdout_eof && (child_terminated || FD_...

http://bugzilla.mindrot.org/show_bug.cgi?id=52 dtucker at zip.com.au changed: What |Removed |Added ---------------------------------------------------------------------------- OtherBugsDependingO| |1274 nThis| | ------- You are receiving this mail because: ------- You are the assignee for

...r2() when exiting a protocol 2 session apparently because we don't call wait before reinstalling the handler. Any thoughts on this issue or how to address it? serverloop.c from latest snapshots: void sigchld_handler2(int sig) { int save_errno = errno; debug("Received SIGCHLD."); child_terminated = 1; signal(SIGCHLD, sigchld_handler2); errno = save_errno; } from signal(5) on HP-UX 11 (SIGCLD is the same as SIGCHLD): If one of the signal interface routines is used to set the action for SIGCLD to be caught (that is, a...

...ic void collect_children(void) { pid_t pid; sigset_t oset, nset; int status; /* block SIGCHLD while we check for dead children */ sigemptyset(&nset); sigaddset(&nset, SIGCHLD); sigprocmask(SIG_BLOCK, &nset, &oset); if (child_terminated) { while ((pid = waitpid(-1, &status, WNOHANG)) > 0 || (pid < 0 && errno == EINTR)) if (pid > 0) session_close_by_pid(pid, status); child_terminated = 0; }...

...--------------- I have tried my hand at gdb, and this is the output and backtrace before the session hangs. (gdb) wait_until_can_do_something (readsetp=0x7ffff8e4, writesetp=0x7ffff8e0, maxfdp=0x7ffff8dc, nallocp=0x7ffff8d8, max_time_milliseconds=0) at serverloop.c:313 313 if (child_terminated && packet_not_very_much_data_to_write()) (gdb) 317 if (max_time_milliseconds == 0) (gdb) 318 tvp = NULL; (gdb) 326 ret = select((*maxfdp)+1, *readsetp, *writesetp, NULL, tvp); (gdb) p connection_closed $1 = 0 (gdb) bt #0 wait_until_can_do_som...

The 4.6p1 sshd is logging this error during remote commands or file transfers: error: channel 0: chan_read_failed for istate 3 Platform is Solaris 8, 4.6p1 + OpenSSL 0.9.8d. The commands and transfers work correctly, so the error message appears to be spurious. The error message does not appear when processing logins. Otherwise 4.6p1 is running without any apparent problems. This error

...mp; !defined(DISABLE_LASTLOG) && !defined(WITH_AIXAUTHENTICATE) */ } /* Records that the user has logged out. */ --- serverloop.c.orig Wed May 10 14:34:00 2000 +++ serverloop.c Thu May 11 08:17:17 2000 @@ -85,7 +85,6 @@ int save_errno = errno; debug("Received SIGCHLD."); child_terminated = 1; - signal(SIGCHLD, sigchld_handler2); errno = save_errno; } @@ -640,6 +639,7 @@ while ((pid = waitpid(-1, &status, WNOHANG)) > 0) session_close_by_pid(pid, status); child_terminated = 0; + signal(SIGCHLD, sigchld_handler2); } channel_after_select(&readse...

Hello, I've been trying to get openssh working at our site recently, but have been running into these problems. In using the "release" version (openssh-2.3.0p1) we kept getting these broken pipe errors: zcat: stdout: Broken pipe Damien suggested we try out the snapshot versions, so I've been trying out the daily versions since last week. With the snapshots, the pipe problem

Hi OpenSSH developers, I am using openSSH (now 2.9.9p2, but prob occurs in 2.9p2 also) to execute commands on a remote machine which outputs data to stdout then pipes it to another invocation of ssh which connects back to the first machine in the same way, where it starts a program to read and store the output from the command on the second machine. I am using the "command" option in

Hi guys, and another feature request for sshd which I would classify as really useful. And I think this behaviour is currently not available (If yes, sorry, I must have missed it): > I believe that the sshd should log which RSA key was used to connect to > an account. When there are a number of keys in the authorized_keys file > it is often useful to know which one was used for each

This patch is against Cray patch against openssh-SNAP-20010710. Here a few notes about them: 1) rijndael does not work on cray due to the fact it is rooted in 32 bits. I looking for a fix, it may come form Wendy Palam. For now the cray default to the following cihpers for ssh version 2 ssh are: 3des-cbc,blowfish-cbc,cast128-cbc,arcfour 2) Crays don't have setitimer so I

...len==0 from the read(), and does: verbose("Connection closed by remote host."); fatal_cleanup(); Comparing the v1 server_loop and the v2 server_loop2, the v1 loop appears to have significantly more checks for breaking the loop: if (((fdout_eof && fderr_eof) || (child_terminated && child_has_selected)) && !packet_have_data_to_write() && (buffer_len(&stdout_buffer) == 0) && (buffer_len(&stderr_buffer) == 0)) { if (!channel_still_open()) break; as well as other useful logging...