Jason Andryuk writes ("Re: [PATCH 1/2] tools/helpers: Introduce
cmp-fd-file-inode utility"):
> I'd be happy to use stat if it works. The comment in locking.sh above
> the usage is:
> # We can't just stat /dev/stdin or /proc/self/fd/$_lockfd or
> # use bash's test -ef because those all go through what is
> # actually a synthetic symlink in /proc and we aren't
> # guaranteed that our stat(2) won't lose the race with an
> # rm(1) between reading the synthetic link and traversing the
> # file system to find the inum. Perl is very fast so use that.
>
> ...which I thought ruled out stat.
Well read.
I have done some more testing and in my tests (on Debian stretch)
stat -L - <some-file
does this
fstat64(0, {st_mode=S_IFREG|0664, st_size=117844, ...}) = 0
(according to strace) which is precisely what is needed.
Oddly, it also does this
fstat64(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 171), ...}) = 0
but it doesn't seem to do anything with the results, so I think
that's harmless.
I wrote that comment in 2012. Presumably `stat -L -' has appeared in
the meantime.
The synthetic symlink may be a red herring anyway; nowadays at least,
I am told by someone who read the Linux kernel source that
the name comes from the `readlink' method on the link inode, but a
different method entirely -- `get_link' -- is used by `namei' to
actually resolve the link to a destination inode.
But using `-' is clearly fine, like this I think:
mariner:~> stat -c%D.%i -L - t <t
fe04.844307
fe04.844307
mariner:~>
Sorry to muddy the waters.
Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
