From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; X11; Linux i686) Opera 7.54 [en] Description of problem: With at least kernel-2.6.10-1.770_FC3 and kernel-2.6.11-1.14_FC3, when WordPerfect 8 xwp is started, the WordPerfect print server wpexc will crash with a SIGSEGV when started by xwp. With the 2.6.9 and earlier kernels, there has never been this kind of problem with WordPerfect 8. WordPerfect 8 uses the libc5 compatability libraries. I have many documents and templates in WP format, so trashing WordPerfect is not an option. Version-Release number of selected component (if applicable): kernel-2.6.11-1.14_FC3 How reproducible: Always Steps to Reproduce: 1.xwp 2. 3. Actual Results: xwp opens an Error window with the message "Cannot create a new process". wpexc has crashed. xwp will restart wpexc, but the usual wpexc files in /tmp/wpc-<hostname>/ are not created, and xwp cannot connect to the print server. Additional info:
Can you attach the output of 'dmesg' after its crashed ? if you do.. echo 0 > /proc/sys/kernel/exec-shield as root, and then start xwp, does it work ? Another thing to try.. echo 0 > /proc/sys/kernel/randomize_va_space or maybe.. echo 1 > /proc/sys/vm/legacy_va_layout If one of those makes it work, it'll narrow things down a little where the problem is.
No joy so far... Nothing is added to dmesg by the wpexc crash or the wpexc restart. Changing the flags makes no difference. Hopefully, even a null result says something useful. I notice that when wpexc starts, either manually or from xwp, it writes some files successfully to /tmp/wpc-<hostname>/, but the pipe "excmsg8" is not created. Is there anything in the new kernel that might prevent wpexc from creating a pipe in the tmp directory?
the pipe code did get quite a workout circa 2.6.10 iirc, so that could certainly explain things. Can you strace -f the process and find out what it was doing ? hopefully we can construct a reproducer from that.
Frustratingly, wpexc attempts to block SIGTRAP, and this doesn't play well with trace. All I get is 9231 sigaction(SIGTRAP, {0x8050090, [], SA_INTERRUPT|SA_NOMASK|SA_ONESHOT}, {0xbffff51c, ~[HUP INT ILL ABRT BUS USR1 USR2 SYS], 0}, 0) = ? ERESTARTNOINTR (To be restarted) looping until a control-C. If you have any ideas about getting around this, let me know. Is there any other tracing tool to try?
I am seeing what appears to be the same problem currrently under RH ES4. I am using the 2.6.9-5.EL kernel. I am attempting to run the text version of WP and get the error "Could not create a new process -- wpexc" when trying to startup WP. If I execute wpexc by itself it seems to start, at least it stays running in the process table. We are using WP extensively for some maritime reporting features so, unfortunately, we can't abondon it yet. If this sheds any light on anything or if I can help to test anything out or provide more information please let me know.
An update has been released for Fedora Core 3 (kernel-2.6.12-1.1372_FC3) which may contain a fix for your problem. Please update to this new kernel, and report whether or not it fixes your problem. If you have updated to Fedora Core 4 since this bug was opened, and the problem still occurs with the latest updates for that release, please change the version field of this bug to 'fc4'. Thank you.
Thanks. I've since upgraded to FC4 and kernel-2.6.12-1.1390_FC4, and I see that this bug is gone. I also checked kernel-2.6.12-1.1372_FC3 on another machine and also, this bug is gone. Great!