Merge tag '9p-for-6.12-rc5' of https://github.com/martinetd/linux

Pull more 9p reverts from Dominique Martinet:
 "Revert patches causing inode collision problems.

  The code simplification introduced significant regressions on servers
  that do not remap inode numbers when exporting multiple underlying
  filesystems with colliding inodes. See the top-most revert (commit
  be2ca3825372) for details.

  This problem had been ignored for too long and the reverts will also
  head to stable (6.9+).

  I'm confident this set of patches gets us back to previous behaviour
  (another related patch had already been reverted back in April and
  we're almost back to square 1, and the rest didn't touch inode
  lifecycle)"

* tag '9p-for-6.12-rc5' of https://github.com/martinetd/linux:
  Revert "fs/9p: simplify iget to remove unnecessary paths"
  Revert "fs/9p: fix uaf in in v9fs_stat2inode_dotl"
  Revert "fs/9p: remove redundant pointer v9ses"
  Revert " fs/9p: mitigate inode collisions"

5 files changed, 192 insertions, 87 deletions

diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index 1775fcc7f0e8..698c43dd5dc8 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -179,14 +179,16 @@ extern int v9fs_vfs_rename(struct mnt_idmap *idmap,
 			   struct inode *old_dir, struct dentry *old_dentry,
 			   struct inode *new_dir, struct dentry *new_dentry,
 			   unsigned int flags);
-extern struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid,
-						bool new);
+extern struct inode *v9fs_inode_from_fid(struct v9fs_session_info *v9ses,
+					 struct p9_fid *fid,
+					 struct super_block *sb, int new);
 extern const struct inode_operations v9fs_dir_inode_operations_dotl;
 extern const struct inode_operations v9fs_file_inode_operations_dotl;
 extern const struct inode_operations v9fs_symlink_inode_operations_dotl;
 extern const struct netfs_request_ops v9fs_req_ops;
-extern struct inode *v9fs_fid_iget_dotl(struct super_block *sb,
-						struct p9_fid *fid, bool new);
+extern struct inode *v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses,
+					      struct p9_fid *fid,
+					      struct super_block *sb, int new);
 
 /* other default globals */
 #define V9FS_PORT	564
@@ -225,12 +227,30 @@ static inline int v9fs_proto_dotl(struct v9fs_session_info *v9ses)
  */
 static inline struct inode *
 v9fs_get_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-			struct super_block *sb, bool new)
+			struct super_block *sb)
 {
 	if (v9fs_proto_dotl(v9ses))
-		return v9fs_fid_iget_dotl(sb, fid, new);
+		return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 0);
 	else
-		return v9fs_fid_iget(sb, fid, new);
+		return v9fs_inode_from_fid(v9ses, fid, sb, 0);
+}
+
+/**
+ * v9fs_get_new_inode_from_fid - Helper routine to populate an inode by
+ * issuing a attribute request
+ * @v9ses: session information
+ * @fid: fid to issue attribute request for
+ * @sb: superblock on which to create inode
+ *
+ */
+static inline struct inode *
+v9fs_get_new_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
+			    struct super_block *sb)
+{
+	if (v9fs_proto_dotl(v9ses))
+		return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 1);
+	else
+		return v9fs_inode_from_fid(v9ses, fid, sb, 1);
 }
 
 #endif
diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index 7923c3c347cb..d3aefbec4de6 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -42,7 +42,7 @@ struct inode *v9fs_alloc_inode(struct super_block *sb);
 void v9fs_free_inode(struct inode *inode);
 void v9fs_set_netfs_context(struct inode *inode);
 int v9fs_init_inode(struct v9fs_session_info *v9ses,
-		    struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev);
+		    struct inode *inode, umode_t mode, dev_t rdev);
 void v9fs_evict_inode(struct inode *inode);
 #if (BITS_PER_LONG == 32)
 #define QID2INO(q) ((ino_t) (((q)->path+2) ^ (((q)->path) >> 32)))
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index fd72fc38c8f5..3e68521f4e2f 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -256,12 +256,9 @@ void v9fs_set_netfs_context(struct inode *inode)
 }
 
 int v9fs_init_inode(struct v9fs_session_info *v9ses,
-		    struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev)
+		    struct inode *inode, umode_t mode, dev_t rdev)
 {
 	int err = 0;
-	struct v9fs_inode *v9inode = V9FS_I(inode);
-
-	memcpy(&v9inode->qid, qid, sizeof(struct p9_qid));
 
 	inode_init_owner(&nop_mnt_idmap, inode, NULL, mode);
 	inode->i_blocks = 0;
@@ -365,59 +362,105 @@ void v9fs_evict_inode(struct inode *inode)
 		clear_inode(inode);
 }
 
-struct inode *
-v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid, bool new)
+static int v9fs_test_inode(struct inode *inode, void *data)
+{
+	int umode;
+	dev_t rdev;
+	struct v9fs_inode *v9inode = V9FS_I(inode);
+	struct p9_wstat *st = (struct p9_wstat *)data;
+	struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode);
+
+	umode = p9mode2unixmode(v9ses, st, &rdev);
+	/* don't match inode of different type */
+	if (inode_wrong_type(inode, umode))
+		return 0;
+
+	/* compare qid details */
+	if (memcmp(&v9inode->qid.version,
+		   &st->qid.version, sizeof(v9inode->qid.version)))
+		return 0;
+
+	if (v9inode->qid.type != st->qid.type)
+		return 0;
+
+	if (v9inode->qid.path != st->qid.path)
+		return 0;
+	return 1;
+}
+
+static int v9fs_test_new_inode(struct inode *inode, void *data)
+{
+	return 0;
+}
+
+static int v9fs_set_inode(struct inode *inode,  void *data)
+{
+	struct v9fs_inode *v9inode = V9FS_I(inode);
+	struct p9_wstat *st = (struct p9_wstat *)data;
+
+	memcpy(&v9inode->qid, &st->qid, sizeof(st->qid));
+	return 0;
+}
+
+static struct inode *v9fs_qid_iget(struct super_block *sb,
+				   struct p9_qid *qid,
+				   struct p9_wstat *st,
+				   int new)
 {
 	dev_t rdev;
 	int retval;
 	umode_t umode;
 	struct inode *inode;
-	struct p9_wstat *st;
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
+	int (*test)(struct inode *inode, void *data);
 
-	inode = iget_locked(sb, QID2INO(&fid->qid));
-	if (unlikely(!inode))
-		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW)) {
-		if (!new) {
-			goto done;
-		} else {
-			p9_debug(P9_DEBUG_VFS, "WARNING: Inode collision %ld\n",
-						inode->i_ino);
-			iput(inode);
-			remove_inode_hash(inode);
-			inode = iget_locked(sb, QID2INO(&fid->qid));
-			WARN_ON(!(inode->i_state & I_NEW));
-		}
-	}
+	if (new)
+		test = v9fs_test_new_inode;
+	else
+		test = v9fs_test_inode;
 
+	inode = iget5_locked(sb, QID2INO(qid), test, v9fs_set_inode, st);
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+	if (!(inode->i_state & I_NEW))
+		return inode;
 	/*
 	 * initialize the inode with the stat info
 	 * FIXME!! we may need support for stale inodes
 	 * later.
 	 */
-	st = p9_client_stat(fid);
-	if (IS_ERR(st)) {
-		retval = PTR_ERR(st);
-		goto error;
-	}
-
+	inode->i_ino = QID2INO(qid);
 	umode = p9mode2unixmode(v9ses, st, &rdev);
-	retval = v9fs_init_inode(v9ses, inode, &fid->qid, umode, rdev);
-	v9fs_stat2inode(st, inode, sb, 0);
-	p9stat_free(st);
-	kfree(st);
+	retval = v9fs_init_inode(v9ses, inode, umode, rdev);
 	if (retval)
 		goto error;
 
+	v9fs_stat2inode(st, inode, sb, 0);
 	v9fs_set_netfs_context(inode);
 	v9fs_cache_inode_get_cookie(inode);
 	unlock_new_inode(inode);
-done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
+
+}
+
+struct inode *
+v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
+		    struct super_block *sb, int new)
+{
+	struct p9_wstat *st;
+	struct inode *inode = NULL;
+
+	st = p9_client_stat(fid);
+	if (IS_ERR(st))
+		return ERR_CAST(st);
+
+	inode = v9fs_qid_iget(sb, &st->qid, st, new);
+	p9stat_free(st);
+	kfree(st);
+	return inode;
 }
 
 /**
@@ -449,15 +492,8 @@ static int v9fs_at_to_dotl_flags(int flags)
  */
 static void v9fs_dec_count(struct inode *inode)
 {
-	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2) {
-		if (inode->i_nlink) {
-			drop_nlink(inode);
-		} else {
-			p9_debug(P9_DEBUG_VFS,
-						"WARNING: unexpected i_nlink zero %d inode %ld\n",
-						inode->i_nlink, inode->i_ino);
-		}
-	}
+	if (!S_ISDIR(inode->i_mode) || inode->i_nlink > 2)
+		drop_nlink(inode);
 }
 
 /**
@@ -508,9 +544,6 @@ static int v9fs_remove(struct inode *dir, struct dentry *dentry, int flags)
 		} else
 			v9fs_dec_count(inode);
 
-		if (inode->i_nlink <= 0)	/* no more refs unhash it */
-			remove_inode_hash(inode);
-
 		v9fs_invalidate_inode_attr(inode);
 		v9fs_invalidate_inode_attr(dir);
 
@@ -576,7 +609,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
 		/*
 		 * instantiate inode and assign the unopened fid to the dentry
 		 */
-		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, true);
+		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
 		if (IS_ERR(inode)) {
 			err = PTR_ERR(inode);
 			p9_debug(P9_DEBUG_VFS,
@@ -704,8 +737,10 @@ struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 		inode = NULL;
 	else if (IS_ERR(fid))
 		inode = ERR_CAST(fid);
+	else if (v9ses->cache & (CACHE_META|CACHE_LOOSE))
+		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb);
 	else
-		inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb, false);
+		inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
 	/*
 	 * If we had a rename on the server and a parallel lookup
 	 * for the new name, then make sure we instantiate with
diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c
index c61b97bd13b9..143ac03b7425 100644
--- a/fs/9p/vfs_inode_dotl.c
+++ b/fs/9p/vfs_inode_dotl.c
@@ -52,50 +52,80 @@ static kgid_t v9fs_get_fsgid_for_create(struct inode *dir_inode)
 	return current_fsgid();
 }
 
+static int v9fs_test_inode_dotl(struct inode *inode, void *data)
+{
+	struct v9fs_inode *v9inode = V9FS_I(inode);
+	struct p9_stat_dotl *st = (struct p9_stat_dotl *)data;
 
+	/* don't match inode of different type */
+	if (inode_wrong_type(inode, st->st_mode))
+		return 0;
 
-struct inode *
-v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid, bool new)
+	if (inode->i_generation != st->st_gen)
+		return 0;
+
+	/* compare qid details */
+	if (memcmp(&v9inode->qid.version,
+		   &st->qid.version, sizeof(v9inode->qid.version)))
+		return 0;
+
+	if (v9inode->qid.type != st->qid.type)
+		return 0;
+
+	if (v9inode->qid.path != st->qid.path)
+		return 0;
+	return 1;
+}
+
+/* Always get a new inode */
+static int v9fs_test_new_inode_dotl(struct inode *inode, void *data)
+{
+	return 0;
+}
+
+static int v9fs_set_inode_dotl(struct inode *inode,  void *data)
+{
+	struct v9fs_inode *v9inode = V9FS_I(inode);
+	struct p9_stat_dotl *st = (struct p9_stat_dotl *)data;
+
+	memcpy(&v9inode->qid, &st->qid, sizeof(st->qid));
+	inode->i_generation = st->st_gen;
+	return 0;
+}
+
+static struct inode *v9fs_qid_iget_dotl(struct super_block *sb,
+					struct p9_qid *qid,
+					struct p9_fid *fid,
+					struct p9_stat_dotl *st,
+					int new)
 {
 	int retval;
 	struct inode *inode;
-	struct p9_stat_dotl *st;
 	struct v9fs_session_info *v9ses = sb->s_fs_info;
+	int (*test)(struct inode *inode, void *data);
 
-	inode = iget_locked(sb, QID2INO(&fid->qid));
-	if (unlikely(!inode))
-		return ERR_PTR(-ENOMEM);
-	if (!(inode->i_state & I_NEW)) {
-		if (!new) {
-			goto done;
-		} else { /* deal with race condition in inode number reuse */
-			p9_debug(P9_DEBUG_ERROR, "WARNING: Inode collision %lx\n",
-						inode->i_ino);
-			iput(inode);
-			remove_inode_hash(inode);
-			inode = iget_locked(sb, QID2INO(&fid->qid));
-			WARN_ON(!(inode->i_state & I_NEW));
-		}
-	}
+	if (new)
+		test = v9fs_test_new_inode_dotl;
+	else
+		test = v9fs_test_inode_dotl;
 
+	inode = iget5_locked(sb, QID2INO(qid), test, v9fs_set_inode_dotl, st);
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+	if (!(inode->i_state & I_NEW))
+		return inode;
 	/*
 	 * initialize the inode with the stat info
 	 * FIXME!! we may need support for stale inodes
 	 * later.
 	 */
-	st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN);
-	if (IS_ERR(st)) {
-		retval = PTR_ERR(st);
-		goto error;
-	}
-
-	retval = v9fs_init_inode(v9ses, inode, &fid->qid,
+	inode->i_ino = QID2INO(qid);
+	retval = v9fs_init_inode(v9ses, inode,
 				 st->st_mode, new_decode_dev(st->st_rdev));
-	v9fs_stat2inode_dotl(st, inode, 0);
-	kfree(st);
 	if (retval)
 		goto error;
 
+	v9fs_stat2inode_dotl(st, inode, 0);
 	v9fs_set_netfs_context(inode);
 	v9fs_cache_inode_get_cookie(inode);
 	retval = v9fs_get_acl(inode, fid);
@@ -103,11 +133,27 @@ v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid, bool new)
 		goto error;
 
 	unlock_new_inode(inode);
-done:
 	return inode;
 error:
 	iget_failed(inode);
 	return ERR_PTR(retval);
+
+}
+
+struct inode *
+v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid,
+			 struct super_block *sb, int new)
+{
+	struct p9_stat_dotl *st;
+	struct inode *inode = NULL;
+
+	st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN);
+	if (IS_ERR(st))
+		return ERR_CAST(st);
+
+	inode = v9fs_qid_iget_dotl(sb, &st->qid, fid, st, new);
+	kfree(st);
+	return inode;
 }
 
 struct dotl_openflag_map {
@@ -259,7 +305,7 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry,
 		p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
 		goto out;
 	}
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
+	inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", err);
@@ -309,6 +355,7 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 			       umode_t omode)
 {
 	int err;
+	struct v9fs_session_info *v9ses;
 	struct p9_fid *fid = NULL, *dfid = NULL;
 	kgid_t gid;
 	const unsigned char *name;
@@ -318,6 +365,7 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	struct posix_acl *dacl = NULL, *pacl = NULL;
 
 	p9_debug(P9_DEBUG_VFS, "name %pd\n", dentry);
+	v9ses = v9fs_inode2v9ses(dir);
 
 	omode |= S_IFDIR;
 	if (dir->i_mode & S_ISGID)
@@ -352,7 +400,7 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	}
 
 	/* instantiate inode and assign the unopened fid to the dentry */
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
+	inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
@@ -749,6 +797,7 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 	kgid_t gid;
 	const unsigned char *name;
 	umode_t mode;
+	struct v9fs_session_info *v9ses;
 	struct p9_fid *fid = NULL, *dfid = NULL;
 	struct inode *inode;
 	struct p9_qid qid;
@@ -758,6 +807,7 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 		 dir->i_ino, dentry, omode,
 		 MAJOR(rdev), MINOR(rdev));
 
+	v9ses = v9fs_inode2v9ses(dir);
 	dfid = v9fs_parent_fid(dentry);
 	if (IS_ERR(dfid)) {
 		err = PTR_ERR(dfid);
@@ -788,7 +838,7 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir,
 			 err);
 		goto error;
 	}
-	inode = v9fs_fid_iget_dotl(dir->i_sb, fid, true);
+	inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n",
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index f52fdf42945c..489db161abc9 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -139,7 +139,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags,
 	else
 		sb->s_d_op = &v9fs_dentry_operations;
 
-	inode = v9fs_get_inode_from_fid(v9ses, fid, sb, true);
+	inode = v9fs_get_new_inode_from_fid(v9ses, fid, sb);
 	if (IS_ERR(inode)) {
 		retval = PTR_ERR(inode);
 		goto release_sb;