From: "MAEDA Naoaki" > > Sorry to interrupt the discussion. I found a different rcfs > mount/umount problem on f0. > > I tried to run the rapid mount/unmount script on f0, and then > stopped the script by SIGINT. After that, I've sometimes observed > that further rcfs mount request is failed by EPERM permanently. > > I guess the reason is that rcfs_mount_count is wrongly > incremented in rcfs_fill_super() on error return case. > > Following patch fix the problem.
I have also observed the EPERM behaviour on f0 after some further testing. Then I applied the patch, and following make/reboot, some *VERY* strange things started happening. I ran the "rapid" script, and interrupted it; "df -a" showed 2 (yes *two*) entries for rcfs (both showing all zeroes). I issued a umount, and the "df -a" showed just the one entry, but with usage figures that exactly matched those for / (root) - not zeroes as I would have expected. A subsequent umount for rcfs generated the following (all 4) messages ; umount: none: not found umount: /rcfs: not mounted umount: none: not found umount: /rcfs: not mounted I was subsequently able to replicate this to some extent after another reboot; ran the script and interrupted it. Only one entry in the "df -a", but it showed usage numbers as per /, and the 4 messges above issued after a umount. The "echo" trace from the loop showed a umount happened to be the last command issued, by there was still that entry in the "df". Food for thought maybe. As a side-note, I now have two (different) systems that consistently display the umount hang at e19. Shane ... ------------------------------------------------------- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for free - -and be entered to win a 42" plasma tv or your very own Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech
