Re: making GNU ls -i (--inode) work around the linux readdir bug

[Tony Finch]

On Mon, 7 Jul 2008, Ian Jackson wrote:
>
> When files are only in a single filesystem the inum is sufficient to
> uniquely identify a file.  But when we consider more than one
> filesystem, the inum and device are needed together because inums on
> different filesystems are unrelated and may be (often are) the same.
>
> Thus any program which uses _only_ inums to tell files apart is broken
> if it works near a mountpoint.  Either the documentation for or
> filesystem layout used by that program must ensure that mountpoints
> are not relevant, or the program must take extra special care somehow
> itself.  ls -i  prints only inums.  So by using  ls -i  a program
> promises that mountpoints are not relevant.

Thinking along the same lines... It is correct to pair the d_ino values
from readdir() with the st_dev value of the directory to identify unique
files within a single directory. This is because, according to POSIX,

(1) Mount points have to be directories and directories can't be hard-
linked [*], so a sub-directory's d_ino is unique whether or not it is a
mount point. Therefore only non-directories can have duplicate d_ino
numbers.

(2) All non-mount-points have the same st_dev as their parent, since the
definition of mount points is that only they have a different st_dev
number from their parent directory. Therefore it isn't necessary to
lstat() each non-directory to find out its st_dev number.

If the d_ino of a mount point comes from the mounted file system instead
of the underlying one then this logic breaks, because the d_ino of a
sub-directory is no longer guaranteed to be unique. Such a change would
make the d_ino value entirely useless because you would have to lstat()
every directory entry to get a meaningful dev+ino pair.

[*] The logic also breaks on crufty old systems that permit root to cause
exciting filesystem breakage by hard-linking directories.

Tony.
-- 
f.anthony.n.finch  <address@hidden>  http://dotat.at/
CROMARTY FORTH: VARIABLE, MAINLY NORTH 3 OR 4, OCCASIONALLY 5 LATER. MODERATE
OR ROUGH. OCCASIONAL RAIN. MODERATE OR GOOD, OCCASIONALLY POOR.