linux-kernel - Debugfs : inotify, multiple calls to debugfs_create

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20061120181838.GB7328@Krystal>
Date:	Mon, 20 Nov 2006 13:18:38 -0500
From:	Mathieu Desnoyers <compudj@...stal.dyndns.org>
To:	Greg Kroah-Hartman <greg@...ah.com>
Cc:	ltt-dev@...fik.org, linux-kernel@...r.kernel.org
Subject: Debugfs : inotify, multiple calls to debugfs_create_file, remove

Hi Greg,

I just had to add inotify support to my LTTng consumer so I could inform it
of the presence of new CPUs (for CPU hotplug). I noticed that no
notification event was being sent when a debugfs file is created from within
the kernel through debugfs_create. There are probably other notifications
missing, but here is the patch adding the one I care about. Should it be added
in libfs or in debugfs ?

A second problem I noticed is when a caller calls debugfs_create_file more than
once : the result is that the debugfs_remove will fail. I guess the second call
to debugfs_create_file increments the reference counts (there is not fix for
this issue in my patch).

Third problem : a failing call to debugfs_remove keeps the filesystem pinned.
(fixed by calling simple_release_fs in the error path).

The third problem : When a process is in a directory, the call to simple_rmdir
will fail. Debugfs does not use its return value. I noticed that calling
simple_unlink on a directory when simple_rmdir fails removes the directory that
would otherwise be left there. I am not sure if this approach is correct
through.

This patch is against Linux 2.6.18.

Regards,

Mathieu

diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
index e8ae304..d88203f 100644
--- a/fs/debugfs/inode.c
+++ b/fs/debugfs/inode.c
@@ -23,6 +23,7 @@ #include <linux/pagemap.h>
 #include <linux/init.h>
 #include <linux/namei.h>
 #include <linux/debugfs.h>
+#include <linux/fsnotify.h>

 #define DEBUGFS_MAGIC	0x64626720

@@ -156,7 +157,8 @@ static int debugfs_create_by_name(const 
 	} else
 		error = PTR_ERR(dentry);
 	mutex_unlock(&parent->d_inode->i_mutex);
-
+	if (!error)
+		fsnotify_create(parent->d_inode, *dentry);
 	return error;
 }

@@ -204,7 +206,10 @@ struct dentry *debugfs_create_file(const

 	error = debugfs_create_by_name(name, mode, parent, &dentry);
 	if (error) {
+		/* FIXME : a failing call to debugfs_create_file leaves a
+		 * stalled entry after a remove. */
 		dentry = NULL;
+		simple_release_fs(&debugfs_mount, &debugfs_mount_count);
 		goto exit;
 	}

@@ -265,6 +270,7 @@ EXPORT_SYMBOL_GPL(debugfs_create_dir);
 void debugfs_remove(struct dentry *dentry)
 {
 	struct dentry *parent;
+	int ret = 0;

 	if (!dentry)
 		return;
@@ -277,8 +283,12 @@ void debugfs_remove(struct dentry *dentr
 	if (debugfs_positive(dentry)) {
 		if (dentry->d_inode) {
 			if (S_ISDIR(dentry->d_inode->i_mode))
-				simple_rmdir(parent->d_inode, dentry);
+				ret = simple_rmdir(parent->d_inode, dentry);
 			else
+				ret = simple_unlink(parent->d_inode, dentry);
+			/* simple_rmdir failed because the directory
+			 * is used. Force deletion. */
+			if (ret)
 				simple_unlink(parent->d_inode, dentry);
 		dput(dentry);
 		}

OpenPGP public key:              http://krystal.dyndns.org:8080/key/compudj.gpg
Key fingerprint:     8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/