Vladimir Davydov wrote: > >> This is achieved by keeping the position where a process stopped as a > >> pointer to mount entry and resuming reading from the position. If a mount entry > >> is removed, all processes that stopped on the entry are advanced i.e. their > >> position is moved to the next entry. To achieve this, all processes reading > >> /proc/mounts are organized in a linked list. > >> > >> An example of /proc/mounts inconsistency is here: > >> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=593516 > > > > In the specific case of schroot I'm not convinced that you shouldn't > > just increase the user space buffer size if you don't read everything. > > Or to simply use mount namespaces to make unmounting unnecessary. > > I agree, but there may be a lot of programs that already read > /proc/mounts and don't care about its consistency. I have cared about its consistency, but didn't realise it didn't provide it. Oops! I will use a bigger buffer in future, thanks for pointing it out ;-) -- Jamie -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
