[RFC] corner cases of open() on procfs symlinks

From: Al Viro
Date: Wed Jun 05 2013 - 21:20:20 EST


I'm not sure whether to treat that as a bug or as a weird misfeature
enshrined in userland ABI:
open("/tmp", O_CREAT, 0) => -EISDIR // LAST_NORM case
open("/", O_CREAT, 0) => -EISDIR // LAST_ROOT
open(".", O_CREAT, 0) => -EISDIR // LAST_DOT
open("..", O_CREAT, 0) => -EISDIR // LAST_DOTDOT
open("/proc/self/cwd", O_CREAT, 0) => success // LAST_BIND
open("/proc/self/cwd/", O_CREAT, 0) => -EISDIR // trailing slashes
At the very least, it's inconsistent. OTOH, it's exposed to userland.
OTTH, SuS says that O_CREAT without O_RDWR or O_WRONLY is undefined, and
either of those two would suffice for -EISDIR in all cases (may_open() takes
care of that).

Another unpleasantness:
open("/proc/self/exe", LOOKUP_DIRECTORY, 0) -> success
That one is clearly a bug. Moreover, getting rid of both of those
bugs actually simplifies the mess in do_last(). I would obviously
like to do that - do_last() is far too convoluted as it is; the only
question is whether we can change the first weirdness... Comments?

FWIW, the simplification of do_last() would look like that:

diff --git a/fs/namei.c b/fs/namei.c
index 85e40d1..617599c 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -2690,28 +2690,10 @@ static int do_last(struct nameidata *nd, struct path *path,
nd->flags &= ~LOOKUP_PARENT;
nd->flags |= op->intent;

- switch (nd->last_type) {
- case LAST_DOTDOT:
- case LAST_DOT:
+ if (nd->last_type != LAST_NORM) {
error = handle_dots(nd, nd->last_type);
if (error)
return error;
- /* fallthrough */
- case LAST_ROOT:
- error = complete_walk(nd);
- if (error)
- return error;
- audit_inode(name, nd->path.dentry, 0);
- if (open_flag & O_CREAT) {
- error = -EISDIR;
- goto out;
- }
- goto finish_open;
- case LAST_BIND:
- error = complete_walk(nd);
- if (error)
- return error;
- audit_inode(name, dir, 0);
goto finish_open;
}

@@ -2841,6 +2823,7 @@ finish_lookup:
}
nd->inode = inode;
/* Why this, you ask? _Now_ we might have grown LOOKUP_JUMPED... */
+finish_open:
error = complete_walk(nd);
if (error) {
path_put(&save_parent);
@@ -2853,7 +2836,6 @@ finish_lookup:
if ((nd->flags & LOOKUP_DIRECTORY) && !nd->inode->i_op->lookup)
goto out;
audit_inode(name, nd->path.dentry, 0);
-finish_open:
if (!S_ISREG(nd->inode->i_mode))
will_truncate = false;

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/