On 04.05.2012 16:48, Ian Kent wrote: > On Thu, 2012-05-03 at 09:32 +0400, Michael Tokarev wrote: >> Ping? Any news on this? Unfortunately I >> don't have much understanding in kernel >> internals to look at this... > > I haven't looked at it yet. Oh. No problem, I just wanted to be sure if my email actually made to somewhere.. We all have our time constraints... :) Especially now when the main issue with autofs has been fixed (in a very interesting way!), this refcount issue becomes much less urgent. It was very annoying for me to have to reboot each time I wanted to try a new variant of autofs4.ko with another version of the fix. > I don't think it is the module, most likely something elsewhere. > How have you established that the count is negative? $ lsmod | grep auto autofs4 22422 2147483647 This is on my system with last fix from Linus, I wanted to check behavour of some other program with pre-fixed version but wasn't able to and had no time to reboot, and later was distracted by something else. That happens on normal /etc/init.d/autofs stop. When I kill the daemon manually with -9 and umount the filesystem, the usage count becomes 0 wich is how it should be. Note that 2147483647 is not really -1, it is 0x7fffffff, so it might not be negative after all. But this definitely prevents the module from being unloaded. Thank you! /mjt -- To unsubscribe from this list: send the line "unsubscribe autofs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
