http://qa.mandrakesoft.com/show_bug.cgi?id=4709
[EMAIL PROTECTED] changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
------- Additional Comments From [EMAIL PROTECTED] 2003-30-08 16:57 -------
I haven't seen any hangs since my upgrade to kernel 2.4.22-0.6mdk.
The bug is apparently fixed with some version after 2.4.22-0.3mdk which was the
last one with deadlocks that I've tested.
--
Configure bugmail: http://qa.mandrakesoft.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
------- Reminder: -------
assigned_to: [EMAIL PROTECTED]
status: RESOLVED
creation_date:
description:
I originally had this problem on Mandrake 9.1, but thought it could be caused by
specific configuration of that system - I've installed lots of software on it,
and some from Cooker.
Recently I've clean-installed Mandrake 9.2 beta2 (formatting the '/' partition,
and leaving only '/home' intact), and the problem still exists.
The symptoms are: after some time the system is running, it gets impossible t
start new shell processes - bash,csh, any.
All applications that try to launch a shell for some task hang at that stage too
- so non-interactive shells cannot start too.
The shells that are already started are working fine - it's only impossible to
start new ones.
I've run a screen session with multiple windows and waited for the problem to
occur, and then did strace on various programs that launch shells.
They look similarly - they susually hang in waitpit() or read() function.
For example this strace end comes from bash:
<snip>
stat64("/bin/bash", {st_mode=S_IFREG|0755, st_size=641868, ...}) = 0
getpgrp() = 5960
rt_sigaction(SIGCHLD, {0x80776a0, [], SA_RESTORER, 0x40051c68}, {SIG_DFL}, 8) = 0
rt_sigprocmask(SIG_BLOCK, NULL, [RTMIN], 8) = 0
fcntl64(0, F_GETFL) = 0x2 (flags O_RDWR)
fstat64(0, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 5), ...}) = 0
_llseek(0, 0, 0xbfffefe8, SEEK_CUR) = -1 ESPIPE (Illegal seek)
rt_sigprocmask(SIG_BLOCK, NULL, [RTMIN], 8) = 0
read(0, <unfinished ...>
<snip>
and this strace end comes from csh:
<snip>
rt_sigprocmask(SIG_SETMASK, [INT RTMIN], [INT RTMIN], 8) = 0
close(0) = -1 EBADF (Bad file descriptor)
dup(19) = 0
fcntl64(0, F_SETFD, 0) = 0
close(1) = -1 EBADF (Bad file descriptor)
dup(17) = 1
fcntl64(1, F_SETFD, 0) = 0
close(2) = -1 EBADF (Bad file descriptor)
dup(18) = 2
fcntl64(2, F_SETFD, 0) = 0
pipe([5, 6]) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [INT RTMIN], 8) = 0
fork() = 5991
gettimeofday({1060688761, 435762}, NULL) = 0
rt_sigprocmask(SIG_SETMASK, [INT RTMIN], [INT CHLD RTMIN], 8) = 0
close(6) = 0
read(5, <unfinished ...>
<snip>
and this strace end comes from xterm:
<snip>
close(4) = 0
close(4) = -1 EBADF (Bad file descriptor)
rt_sigprocmask(SIG_BLOCK, [INT CHLD], [CHLD RTMIN], 8) = 0
fork() = 5967
rt_sigprocmask(SIG_SETMASK, [CHLD RTMIN], NULL, 8) = 0
close(3) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [CHLD RTMIN], 8) = 0
rt_sigprocmask(SIG_SETMASK, [CHLD RTMIN], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [CHLD RTMIN], 8) = 0
rt_sigaction(SIGINT, {0x8076710, [], SA_RESTORER, 0x40051c68}, {SIG_DFL}, 8) = 0
waitpid(-1, [WIFEXITED(s) && WEXITSTATUS(s) == 0], 0) = 5966
waitpid(-1, <unfinished ...>
<snip>
The launch operation for programs like xterm, csh can be interrupted with
CTRL-C, but when I run bash - I get the following errors in strace and the
process won't terminate when I press CTRL-C:
<snip>
close(3) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [CHLD RTMIN], 8) = 0
rt_sigprocmask(SIG_SETMASK, [CHLD RTMIN], NULL, 8) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [CHLD RTMIN], 8) = 0
rt_sigaction(SIGINT, {0x8076710, [], SA_RESTORER, 0x40051c68}, {0x8087030, [],
SA_RESTORER, 0x40051c68}, 8) = 0
waitpid(-1, [WIFEXITED(s) && WEXITSTATUS(s) == 0], 0) = 5944
waitpid(-1, 0xbfffe358, 0) = ? ERESTARTSYS (To be restarted)
--- SIGINT (Interrupt) @ 0 (0) ---
sigreturn() = ? (mask now [CHLD RTMIN])
waitpid(-1, 0xbfffe358, 0) = ? ERESTARTSYS (To be restarted)
--- SIGINT (Interrupt) @ 0 (0) ---
sigreturn() = ? (mask now [CHLD RTMIN])
waitpid(-1,
<snip>
I'm attaching full strace logs.
