From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Neil Brown <neilb@suse.de>,
linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: [PATCH 08/24] namei.c: separate the parts of follow_link() that find the link body
Date: Mon, 20 Apr 2015 19:12:52 +0100 [thread overview]
Message-ID: <1429553588-24764-8-git-send-email-viro@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20150420181222.GK889@ZenIV.linux.org.uk>
From: Al Viro <viro@zeniv.linux.org.uk>
Calling conventions for ->follow_link() are rather unfortunate. It
would be better to have it return ERR_PTR(error) on error, NULL on
jumps and link body for normal symlinks. What we currently return
(opaque pointer used by ->put_link() once we are done) should've
been given to analogue (and replacement) of nd_set_link().
For now let's just split a piece of fs/namei.c:follow_link() that
does obtaining the link body into a separate function. When we
get around to changing ->follow_link() API, the changes will be
contained in that sucker. follow_link() itself converted to
calling get_link() and then doing the body traversal (if any).
The next step will expand follow_link() call in link_path_walk()
and this helps to keep the size down...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
fs/namei.c | 79 +++++++++++++++++++++++++++++++++++---------------------------
1 file changed, 45 insertions(+), 34 deletions(-)
diff --git a/fs/namei.c b/fs/namei.c
index a1f6271..e07bf5c 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -871,21 +871,23 @@ static int may_linkat(struct path *link)
return -EPERM;
}
-static __always_inline int
-follow_link(struct path *link, struct nameidata *nd, void **p)
+static __always_inline char *
+get_link(struct path *link, struct nameidata *nd, void **p)
{
struct dentry *dentry = link->dentry;
+ struct inode *inode = dentry->d_inode;
+ void *cookie;
int error;
- char *s;
+ char *res;
BUG_ON(nd->flags & LOOKUP_RCU);
if (link->mnt == nd->path.mnt)
mntget(link->mnt);
- error = -ELOOP;
+ res = ERR_PTR(-ELOOP);
if (unlikely(nd->total_link_count >= 40))
- goto out_put_nd_path;
+ goto out;
cond_resched();
nd->total_link_count++;
@@ -893,44 +895,53 @@ follow_link(struct path *link, struct nameidata *nd, void **p)
touch_atime(link);
nd_set_link(NULL);
- error = security_inode_follow_link(link->dentry);
+ error = security_inode_follow_link(dentry);
+ res = ERR_PTR(error);
if (error)
- goto out_put_nd_path;
+ goto out;
nd->last_type = LAST_BIND;
- *p = dentry->d_inode->i_op->follow_link(dentry);
- error = PTR_ERR(*p);
- if (IS_ERR(*p))
- goto out_put_nd_path;
+ res = cookie = inode->i_op->follow_link(dentry);
+ if (IS_ERR(cookie))
+ goto out;
- error = 0;
- s = nd_get_link(nd);
- if (s) {
- if (unlikely(IS_ERR(s))) {
- path_put(&nd->path);
- put_link(nd, link, *p);
- return PTR_ERR(s);
- }
- if (*s == '/') {
- if (!nd->root.mnt)
- set_root(nd);
- path_put(&nd->path);
- nd->path = nd->root;
- path_get(&nd->root);
- nd->flags |= LOOKUP_JUMPED;
- }
- nd->inode = nd->path.dentry->d_inode;
- error = link_path_walk(s, nd);
- if (unlikely(error))
- put_link(nd, link, *p);
+ res = nd_get_link(nd);
+ if (!IS_ERR(res)) {
+ *p = cookie;
+ return res;
}
- return error;
-
-out_put_nd_path:
+ if (inode->i_op->put_link)
+ inode->i_op->put_link(dentry, res, cookie);
+out:
*p = NULL;
path_put(&nd->path);
path_put(link);
+ return res;
+}
+
+static __always_inline int
+follow_link(struct path *link, struct nameidata *nd, void **p)
+{
+ char *s = get_link(link, nd, p);
+ int error;
+
+ if (unlikely(IS_ERR(s)))
+ return PTR_ERR(s);
+ if (unlikely(!s))
+ return 0;
+ if (*s == '/') {
+ if (!nd->root.mnt)
+ set_root(nd);
+ path_put(&nd->path);
+ nd->path = nd->root;
+ path_get(&nd->root);
+ nd->flags |= LOOKUP_JUMPED;
+ }
+ nd->inode = nd->path.dentry->d_inode;
+ error = link_path_walk(s, nd);
+ if (unlikely(error))
+ put_link(nd, link, *p);
return error;
}
--
2.1.4
next prev parent reply other threads:[~2015-04-20 18:12 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-04-20 18:12 [RFC][PATCHSET] non-recursive link_path_walk() and reducing stack footprint Al Viro
2015-04-20 18:12 ` [PATCH 01/24] lustre: rip the private symlink nesting limit out Al Viro
2015-04-20 19:08 ` Andreas Dilger
2015-04-20 19:22 ` Al Viro
2015-04-20 20:35 ` Al Viro
2015-04-20 18:12 ` [PATCH 02/24] VFS: replace {, total_}link_count in task_struct with pointer to nameidata Al Viro
2015-04-20 18:12 ` [PATCH 03/24] ovl: rearrange ovl_follow_link to it doesn't need to call ->put_link Al Viro
2015-04-20 18:12 ` [PATCH 04/24] VFS: replace nameidata arg to ->put_link with a char* Al Viro
2015-04-20 18:12 ` [PATCH 05/24] SECURITY: remove nameidata arg from inode_follow_link Al Viro
2015-04-20 18:12 ` [PATCH 06/24] VFS: remove nameidata args from ->follow_link Al Viro
2015-04-20 18:12 ` [PATCH 07/24] namei: expand nested_symlink() in its only caller Al Viro
2015-04-20 18:12 ` Al Viro [this message]
2015-04-20 18:12 ` [PATCH 09/24] namei: fold follow_link() into link_path_walk() Al Viro
2015-04-20 18:12 ` [PATCH 10/24] link_path_walk: handle get_link() returning ERR_PTR() immediately Al Viro
2015-04-20 18:12 ` [PATCH 11/24] link_path_walk: don't bother with walk_component() after jumping link Al Viro
2015-04-20 18:12 ` [PATCH 12/24] link_path_walk: turn inner loop into explicit goto Al Viro
2015-04-20 18:12 ` [PATCH 13/24] link_path_walk: massage a bit more Al Viro
2015-04-20 18:12 ` [PATCH 14/24] link_path_walk: get rid of duplication Al Viro
2015-04-20 18:12 ` [PATCH 15/24] link_path_walk: final preparations to killing recursion Al Viro
2015-04-20 18:13 ` [PATCH 16/24] link_path_walk: kill the recursion Al Viro
2015-04-20 21:04 ` Linus Torvalds
2015-04-20 21:32 ` Al Viro
2015-04-20 21:39 ` Linus Torvalds
2015-04-20 21:51 ` Al Viro
2015-04-20 21:41 ` Linus Torvalds
2015-04-20 21:42 ` Linus Torvalds
2015-04-20 21:59 ` Al Viro
2015-04-20 21:52 ` Al Viro
2015-04-20 18:13 ` [PATCH 17/24] link_path_walk: split "return from recursive call" path Al Viro
2015-04-20 18:13 ` [PATCH 18/24] link_path_walk: cleanup - turn goto start; into continue; Al Viro
2015-04-20 18:13 ` [PATCH 19/24] namei: fold may_follow_link() into follow_link() Al Viro
2015-04-20 18:13 ` [PATCH 20/24] namei: introduce nameidata->stack Al Viro
2015-04-20 18:13 ` [PATCH 21/24] namei: regularize use of put_link() and follow_link(), trim arguments Al Viro
2015-04-20 18:13 ` [PATCH 22/24] namei: trim the arguments of get_link() Al Viro
2015-04-20 18:13 ` [PATCH 23/24] new ->follow_link() and ->put_link() calling conventions Al Viro
2015-04-20 18:13 ` [PATCH 24/24] uninline walk_component() Al Viro
2015-04-21 14:49 ` [RFC][PATCHSET] non-recursive link_path_walk() and reducing stack footprint Al Viro
2015-04-21 15:04 ` Christoph Hellwig
2015-04-21 15:12 ` Richard Weinberger
2015-04-21 15:45 ` Al Viro
2015-04-21 16:46 ` Boaz Harrosh
2015-04-21 21:20 ` Al Viro
2015-04-22 18:07 ` Al Viro
2015-04-22 20:12 ` Al Viro
2015-04-22 21:05 ` Al Viro
2015-04-23 7:45 ` NeilBrown
2015-04-23 18:07 ` Al Viro
2015-04-24 6:35 ` NeilBrown
2015-04-24 13:42 ` Al Viro
2015-05-04 5:11 ` Al Viro
2015-05-04 7:30 ` NeilBrown
2015-04-23 5:01 ` Al Viro
2015-04-21 14:51 ` [PATCH] logfs: fix a pagecache leak for symlinks Al Viro
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1429553588-24764-8-git-send-email-viro@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=neilb@suse.de \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).