From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.9) Gecko/20020314 Description of problem: When using the pilot-link tools to access my USB Visor Prism, the tools complete successfully, and then segfault on exit. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Press hotsync button on cradle to allow hotplug to load drivers, etc. 2. Run a pilot-link program (dlpsh or pilot-xfer I've tried so far) 3. Let program finish Actual Results: Program completes successfully, then segmentation faults on exit. Expected Results: Program should complete successfully, and exit cleanly. Additional info: I also tried with a version I had around of pilot-link 0.9.6 from cvs, that did the same thing. Also, there is a log message about an error loading the visor module/driver. It is shown loaded in an lsmod, however. I'll attach /var/log/messages - let me know if there's anything else you need
Created attachment 50024 [details] /var/log/messages - there are a few instances of the segfault near the end
Could you please try with kernel-2.4.9-21 if this bug is fixed? thanks
Where can I get this from? Do you mean 2.4.9-31 from 7.2 updates? I can't find a copy of 2.4.9-21 for i386/i686. For what it's worth, this had been working in 7.2 before I upgraded to the beta, with 2.4.9-31.
i think it should be a bug in kernel.
Just wondering since I havne't heard anything yet - do you still need me to try another kernel? If so, where can I get it?
It would be useful if you installed the current Raw Hide kernel from ftp.redhat.com /pub/redhat/linux/rawhide/i386/RedHat/RPMS. It is far from production, but helps to see if it fixes the problem. The maintainer (Greg K-H) told me that he fixed some similar things around 2.4.12.
BTW, DO NOT REMOVE 2.4.9-31. If Raw Hide bombs on you, you have to have a backup. Have them side by side in /boot/grub/grub.conf.
I tried the Rawhide kernel (2.4.18-0.10), and that produced the same segfault error as the skipjack kernel that was installed by default (2.4.18-0.4). I also installed 2.4.9-31 from 7.2 updates, and with this kernel, there is no segmentation fault. I also tried (while running 2.4.18-0.10) to rebuild the pilot-link package from SRPMS to see if that maybe helped, and that package produced the same error. Let me know what else you want me to try, or if there's any more information I can give you.
This is a known bug, sorry. It's fixed in the 2.5 tree, and there's a quick fix at: http://www.kernel.org/pub/linux/kernel/people/gregkh/usb/2.4/usb-visor_hack-2.4.18.patch But that patch will break all other usb-serial drivers. I'll make a patch that will work for all devices in a short bit.
Real fix for this bug is now at: http://www.kernel.org/pub/linux/kernel/people/gregkh/usb/2.4/usb-serial_visor_fix-2.4.18.patch Pete, feel free to add this to your kernel :)
It is too close for a release, and I do not have the hardware even for a sanity test. I am casting about for help from Meridian people. The fix looks good, but I have a history of botching even trivial fixes, so no bending the procedure anymore for me...
The fix made beta4, /pub/redhat/linux/rawhide/...... kernel-2.4.18-0.13.src.rpm. Anyone is up for testing?
I'll try that kernel out sometime this weekend after I upgrade to skipjack2 and let you know.
OK, this appears to me to be fixed. As long as nobody else is having problems, it's resolved.