Hi... 2009/1/5 Paraneetharan Chandrasekaran <paraneetharanc@xxxxxxxxx>: > Hi Mulyadi, > These are the sequence of commands i used.. > > dd if=/dev/zero of=./fsimage.img bs=1k count=10000 > /sbin/losetup /dev/loop0 ./fsimage.img > /sbin/mke2fs /de/loop0 10000 > mkdir mntpoint > mount /dev/loop0 ./mntpoint > cd mntpoint > vim newfile > /* entered some contents, saved and exited*/ > cd .. > rm fsimage.img > cd mntpoint > /* I am now able to change to this directory ! */ > cat newfile > /* the contents are here without any change ! */ I think, by directly doing "rm <image file>" without first detaching the loop device from the backing file, it still makes /dev/loop associated with the disk blocks. In other word, backing file inode is not needed in this context, because that's replaced by /dev/loop metadata (or inode). That's why you can still access the directories and the files and their contents. These are all my humble interpretation on how VFS works...CMIIW people regards, Mulyadi. -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
