Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh...
Linus Torvalds [Tue, 11 Jan 2011 22:36:08 +0000 (14:36 -0800)]
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
  fs/9p: Don't set dentry->d_op in create routines
  fs/9p: fix spelling typo
  fs/9p: TREADLINK bugfix
  net/9p: Use proper data types
  fs/9p: Simplify the .L create operation
  fs/9p: Move dotl inode operations into a seperate file
  fs/9p: fix menu presentation
  fs/9p: Fix the return error on default acl removal
  fs/9p: Remove unnecessary semicolons

fs/9p/Kconfig
fs/9p/Makefile
fs/9p/acl.c
fs/9p/v9fs.h
fs/9p/vfs_inode.c
fs/9p/vfs_inode_dotl.c [new file with mode: 0644]
fs/9p/xattr.c
net/9p/protocol.c

index 7e05114..814ac4e 100644 (file)
@@ -9,6 +9,8 @@ config 9P_FS
 
          If unsure, say N.
 
+if 9P_FS
+
 config 9P_FSCACHE
        bool "Enable 9P client caching support (EXPERIMENTAL)"
        depends on EXPERIMENTAL
@@ -20,7 +22,6 @@ config 9P_FSCACHE
 
 config 9P_FS_POSIX_ACL
        bool "9P POSIX Access Control Lists"
-       depends on 9P_FS
        select FS_POSIX_ACL
        help
          POSIX Access Control Lists (ACLs) support permissions for users and
@@ -30,3 +31,5 @@ config 9P_FS_POSIX_ACL
          Linux website <http://acl.bestbits.at/>.
 
          If you don't know what Access Control Lists are, say N
+
+endif
index f8ba37e..ab8c127 100644 (file)
@@ -3,6 +3,7 @@ obj-$(CONFIG_9P_FS) := 9p.o
 9p-objs := \
        vfs_super.o \
        vfs_inode.o \
+       vfs_inode_dotl.o \
        vfs_addr.o \
        vfs_file.o \
        vfs_dir.o \
index 6e58c4c..02a2cf6 100644 (file)
@@ -28,7 +28,7 @@ static struct posix_acl *__v9fs_get_acl(struct p9_fid *fid, char *name)
 {
        ssize_t size;
        void *value = NULL;
-       struct posix_acl *acl = NULL;;
+       struct posix_acl *acl = NULL;
 
        size = v9fs_fid_xattr_get(fid, name, NULL, 0);
        if (size > 0) {
@@ -365,7 +365,7 @@ static int v9fs_xattr_set_acl(struct dentry *dentry, const char *name,
        case ACL_TYPE_DEFAULT:
                name = POSIX_ACL_XATTR_DEFAULT;
                if (!S_ISDIR(inode->i_mode)) {
-                       retval = -EINVAL;
+                       retval = acl ? -EINVAL : 0;
                        goto err_out;
                }
                break;
index cb63968..c4b5d88 100644 (file)
@@ -113,9 +113,27 @@ struct v9fs_session_info {
 
 struct p9_fid *v9fs_session_init(struct v9fs_session_info *, const char *,
                                                                        char *);
-void v9fs_session_close(struct v9fs_session_info *v9ses);
-void v9fs_session_cancel(struct v9fs_session_info *v9ses);
-void v9fs_session_begin_cancel(struct v9fs_session_info *v9ses);
+extern void v9fs_session_close(struct v9fs_session_info *v9ses);
+extern void v9fs_session_cancel(struct v9fs_session_info *v9ses);
+extern void v9fs_session_begin_cancel(struct v9fs_session_info *v9ses);
+extern struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
+                       struct nameidata *nameidata);
+extern int v9fs_vfs_unlink(struct inode *i, struct dentry *d);
+extern int v9fs_vfs_rmdir(struct inode *i, struct dentry *d);
+extern int v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
+                       struct inode *new_dir, struct dentry *new_dentry);
+extern void v9fs_vfs_put_link(struct dentry *dentry, struct nameidata *nd,
+                       void *p);
+extern struct inode *v9fs_inode(struct v9fs_session_info *v9ses,
+                       struct p9_fid *fid,
+                       struct super_block *sb);
+
+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 struct inode *v9fs_inode_dotl(struct v9fs_session_info *v9ses,
+                       struct p9_fid *fid,
+                       struct super_block *sb);
 
 /* other default globals */
 #define V9FS_PORT      564
@@ -138,3 +156,21 @@ static inline int v9fs_proto_dotl(struct v9fs_session_info *v9ses)
 {
        return v9ses->flags & V9FS_PROTO_2000L;
 }
+
+/**
+ * v9fs_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_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
+                               struct super_block *sb)
+{
+       if (v9fs_proto_dotl(v9ses))
+               return v9fs_inode_dotl(v9ses, fid, sb);
+       else
+               return v9fs_inode(v9ses, fid, sb);
+}
index 5978298..5076eeb 100644 (file)
 
 static const struct inode_operations v9fs_dir_inode_operations;
 static const struct inode_operations v9fs_dir_inode_operations_dotu;
-static const struct inode_operations v9fs_dir_inode_operations_dotl;
 static const struct inode_operations v9fs_file_inode_operations;
-static const struct inode_operations v9fs_file_inode_operations_dotl;
 static const struct inode_operations v9fs_symlink_inode_operations;
-static const struct inode_operations v9fs_symlink_inode_operations_dotl;
-
-static int
-v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode,
-                   dev_t rdev);
 
 /**
  * unixmode2p9mode - convert unix mode bits to plan 9
@@ -251,41 +244,6 @@ void v9fs_destroy_inode(struct inode *inode)
 #endif
 
 /**
- * v9fs_get_fsgid_for_create - Helper function to get the gid for creating a
- * new file system object. This checks the S_ISGID to determine the owning
- * group of the new file system object.
- */
-
-static gid_t v9fs_get_fsgid_for_create(struct inode *dir_inode)
-{
-       BUG_ON(dir_inode == NULL);
-
-       if (dir_inode->i_mode & S_ISGID) {
-               /* set_gid bit is set.*/
-               return dir_inode->i_gid;
-       }
-       return current_fsgid();
-}
-
-/**
- * v9fs_dentry_from_dir_inode - helper function to get the dentry from
- * dir inode.
- *
- */
-
-static struct dentry *v9fs_dentry_from_dir_inode(struct inode *inode)
-{
-       struct dentry *dentry;
-
-       spin_lock(&inode->i_lock);
-       /* Directory should have only one entry. */
-       BUG_ON(S_ISDIR(inode->i_mode) && !list_is_singular(&inode->i_dentry));
-       dentry = list_entry(inode->i_dentry.next, struct dentry, d_alias);
-       spin_unlock(&inode->i_lock);
-       return dentry;
-}
-
-/**
  * v9fs_get_inode - helper function to setup an inode
  * @sb: superblock
  * @mode: mode to setup inode with
@@ -454,7 +412,7 @@ void v9fs_evict_inode(struct inode *inode)
 #endif
 }
 
-static struct inode *
+struct inode *
 v9fs_inode(struct v9fs_session_info *v9ses, struct p9_fid *fid,
        struct super_block *sb)
 {
@@ -489,60 +447,6 @@ error:
        return ERR_PTR(err);
 }
 
-static struct inode *
-v9fs_inode_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-       struct super_block *sb)
-{
-       struct inode *ret = NULL;
-       int err;
-       struct p9_stat_dotl *st;
-
-       st = p9_client_getattr_dotl(fid, P9_STATS_BASIC);
-       if (IS_ERR(st))
-               return ERR_CAST(st);
-
-       ret = v9fs_get_inode(sb, st->st_mode);
-       if (IS_ERR(ret)) {
-               err = PTR_ERR(ret);
-               goto error;
-       }
-
-       v9fs_stat2inode_dotl(st, ret);
-       ret->i_ino = v9fs_qid2ino(&st->qid);
-#ifdef CONFIG_9P_FSCACHE
-       v9fs_vcookie_set_qid(ret, &st->qid);
-       v9fs_cache_inode_get_cookie(ret);
-#endif
-       err = v9fs_get_acl(ret, fid);
-       if (err) {
-               iput(ret);
-               goto error;
-       }
-       kfree(st);
-       return ret;
-error:
-       kfree(st);
-       return ERR_PTR(err);
-}
-
-/**
- * v9fs_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_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid,
-                       struct super_block *sb)
-{
-       if (v9fs_proto_dotl(v9ses))
-               return v9fs_inode_dotl(v9ses, fid, sb);
-       else
-               return v9fs_inode(v9ses, fid, sb);
-}
-
 /**
  * v9fs_remove - helper function to remove files and directories
  * @dir: directory inode that is being deleted
@@ -633,12 +537,6 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir,
                P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n", err);
                goto error;
        }
-
-       if (v9ses->cache)
-               d_set_d_op(dentry, &v9fs_cached_dentry_operations);
-       else
-               d_set_d_op(dentry, &v9fs_dentry_operations);
-
        d_instantiate(dentry, inode);
        err = v9fs_fid_add(dentry, fid);
        if (err < 0)
@@ -657,144 +555,6 @@ error:
 }
 
 /**
- * v9fs_vfs_create_dotl - VFS hook to create files for 9P2000.L protocol.
- * @dir: directory inode that is being created
- * @dentry:  dentry that is being deleted
- * @mode: create permissions
- * @nd: path information
- *
- */
-
-static int
-v9fs_vfs_create_dotl(struct inode *dir, struct dentry *dentry, int omode,
-               struct nameidata *nd)
-{
-       int err = 0;
-       char *name = NULL;
-       gid_t gid;
-       int flags;
-       mode_t mode;
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *fid = NULL;
-       struct p9_fid *dfid, *ofid;
-       struct file *filp;
-       struct p9_qid qid;
-       struct inode *inode;
-       struct posix_acl *pacl = NULL, *dacl = NULL;
-
-       v9ses = v9fs_inode2v9ses(dir);
-       if (nd && nd->flags & LOOKUP_OPEN)
-               flags = nd->intent.open.flags - 1;
-       else {
-               /*
-                * create call without LOOKUP_OPEN is due
-                * to mknod of regular files. So use mknod
-                * operation.
-                */
-               return v9fs_vfs_mknod_dotl(dir, dentry, omode, 0);
-       }
-
-       name = (char *) dentry->d_name.name;
-       P9_DPRINTK(P9_DEBUG_VFS, "v9fs_vfs_create_dotl: name:%s flags:0x%x "
-                       "mode:0x%x\n", name, flags, omode);
-
-       dfid = v9fs_fid_lookup(dentry->d_parent);
-       if (IS_ERR(dfid)) {
-               err = PTR_ERR(dfid);
-               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
-               return err;
-       }
-
-       /* clone a fid to use for creation */
-       ofid = p9_client_walk(dfid, 0, NULL, 1);
-       if (IS_ERR(ofid)) {
-               err = PTR_ERR(ofid);
-               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
-               return err;
-       }
-
-       gid = v9fs_get_fsgid_for_create(dir);
-
-       mode = omode;
-       /* Update mode based on ACL value */
-       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
-       if (err) {
-               P9_DPRINTK(P9_DEBUG_VFS,
-                          "Failed to get acl values in creat %d\n", err);
-               goto error;
-       }
-       err = p9_client_create_dotl(ofid, name, flags, mode, gid, &qid);
-       if (err < 0) {
-               P9_DPRINTK(P9_DEBUG_VFS,
-                               "p9_client_open_dotl failed in creat %d\n",
-                               err);
-               goto error;
-       }
-       /* instantiate inode and assign the unopened fid to the dentry */
-       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE ||
-           (nd && nd->flags & LOOKUP_OPEN)) {
-               fid = p9_client_walk(dfid, 1, &name, 1);
-               if (IS_ERR(fid)) {
-                       err = PTR_ERR(fid);
-                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
-                               err);
-                       fid = NULL;
-                       goto error;
-               }
-
-               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
-                               err);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_cached_dentry_operations);
-               d_instantiate(dentry, inode);
-               err = v9fs_fid_add(dentry, fid);
-               if (err < 0)
-                       goto error;
-               /* The fid would get clunked via a dput */
-               fid = NULL;
-       } else {
-               /*
-                * Not in cached mode. No need to populate
-                * inode with stat. We need to get an inode
-                * so that we can set the acl with dentry
-                */
-               inode = v9fs_get_inode(dir->i_sb, mode);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_dentry_operations);
-               d_instantiate(dentry, inode);
-       }
-       /* Now set the ACL based on the default value */
-       v9fs_set_create_acl(dentry, dacl, pacl);
-
-       /* if we are opening a file, assign the open fid to the file */
-       if (nd && nd->flags & LOOKUP_OPEN) {
-               filp = lookup_instantiate_filp(nd, dentry, generic_file_open);
-               if (IS_ERR(filp)) {
-                       p9_client_clunk(ofid);
-                       return PTR_ERR(filp);
-               }
-               filp->private_data = ofid;
-       } else
-               p9_client_clunk(ofid);
-
-       return 0;
-
-error:
-       if (ofid)
-               p9_client_clunk(ofid);
-       if (fid)
-               p9_client_clunk(fid);
-       return err;
-}
-
-/**
  * v9fs_vfs_create - VFS hook to create files
  * @dir: directory inode that is being created
  * @dentry:  dentry that is being deleted
@@ -884,107 +644,6 @@ static int v9fs_vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
        return err;
 }
 
-
-/**
- * v9fs_vfs_mkdir_dotl - VFS mkdir hook to create a directory
- * @dir:  inode that is being unlinked
- * @dentry: dentry that is being unlinked
- * @mode: mode for new directory
- *
- */
-
-static int v9fs_vfs_mkdir_dotl(struct inode *dir,
-                              struct dentry *dentry, int omode)
-{
-       int err;
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *fid = NULL, *dfid = NULL;
-       gid_t gid;
-       char *name;
-       mode_t mode;
-       struct inode *inode;
-       struct p9_qid qid;
-       struct dentry *dir_dentry;
-       struct posix_acl *dacl = NULL, *pacl = NULL;
-
-       P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", dentry->d_name.name);
-       err = 0;
-       v9ses = v9fs_inode2v9ses(dir);
-
-       omode |= S_IFDIR;
-       if (dir->i_mode & S_ISGID)
-               omode |= S_ISGID;
-
-       dir_dentry = v9fs_dentry_from_dir_inode(dir);
-       dfid = v9fs_fid_lookup(dir_dentry);
-       if (IS_ERR(dfid)) {
-               err = PTR_ERR(dfid);
-               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
-               dfid = NULL;
-               goto error;
-       }
-
-       gid = v9fs_get_fsgid_for_create(dir);
-       mode = omode;
-       /* Update mode based on ACL value */
-       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
-       if (err) {
-               P9_DPRINTK(P9_DEBUG_VFS,
-                          "Failed to get acl values in mkdir %d\n", err);
-               goto error;
-       }
-       name = (char *) dentry->d_name.name;
-       err = p9_client_mkdir_dotl(dfid, name, mode, gid, &qid);
-       if (err < 0)
-               goto error;
-
-       /* instantiate inode and assign the unopened fid to the dentry */
-       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
-               fid = p9_client_walk(dfid, 1, &name, 1);
-               if (IS_ERR(fid)) {
-                       err = PTR_ERR(fid);
-                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
-                               err);
-                       fid = NULL;
-                       goto error;
-               }
-
-               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
-                               err);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_cached_dentry_operations);
-               d_instantiate(dentry, inode);
-               err = v9fs_fid_add(dentry, fid);
-               if (err < 0)
-                       goto error;
-               fid = NULL;
-       } else {
-               /*
-                * Not in cached mode. No need to populate
-                * inode with stat. We need to get an inode
-                * so that we can set the acl with dentry
-                */
-               inode = v9fs_get_inode(dir->i_sb, mode);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_dentry_operations);
-               d_instantiate(dentry, inode);
-       }
-       /* Now set the ACL based on the default value */
-       v9fs_set_create_acl(dentry, dacl, pacl);
-
-error:
-       if (fid)
-               p9_client_clunk(fid);
-       return err;
-}
-
 /**
  * v9fs_vfs_lookup - VFS lookup hook to "walk" to a new inode
  * @dir:  inode that is being walked from
@@ -993,7 +652,7 @@ error:
  *
  */
 
-static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
+struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
                                      struct nameidata *nameidata)
 {
        struct super_block *sb;
@@ -1063,7 +722,7 @@ error:
  *
  */
 
-static int v9fs_vfs_unlink(struct inode *i, struct dentry *d)
+int v9fs_vfs_unlink(struct inode *i, struct dentry *d)
 {
        return v9fs_remove(i, d, 0);
 }
@@ -1075,7 +734,7 @@ static int v9fs_vfs_unlink(struct inode *i, struct dentry *d)
  *
  */
 
-static int v9fs_vfs_rmdir(struct inode *i, struct dentry *d)
+int v9fs_vfs_rmdir(struct inode *i, struct dentry *d)
 {
        return v9fs_remove(i, d, 1);
 }
@@ -1089,7 +748,7 @@ static int v9fs_vfs_rmdir(struct inode *i, struct dentry *d)
  *
  */
 
-static int
+int
 v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
                struct inode *new_dir, struct dentry *new_dentry)
 {
@@ -1196,42 +855,6 @@ v9fs_vfs_getattr(struct vfsmount *mnt, struct dentry *dentry,
        return 0;
 }
 
-static int
-v9fs_vfs_getattr_dotl(struct vfsmount *mnt, struct dentry *dentry,
-                struct kstat *stat)
-{
-       int err;
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *fid;
-       struct p9_stat_dotl *st;
-
-       P9_DPRINTK(P9_DEBUG_VFS, "dentry: %p\n", dentry);
-       err = -EPERM;
-       v9ses = v9fs_inode2v9ses(dentry->d_inode);
-       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE)
-               return simple_getattr(mnt, dentry, stat);
-
-       fid = v9fs_fid_lookup(dentry);
-       if (IS_ERR(fid))
-               return PTR_ERR(fid);
-
-       /* Ask for all the fields in stat structure. Server will return
-        * whatever it supports
-        */
-
-       st = p9_client_getattr_dotl(fid, P9_STATS_ALL);
-       if (IS_ERR(st))
-               return PTR_ERR(st);
-
-       v9fs_stat2inode_dotl(st, dentry->d_inode);
-       generic_fillattr(dentry->d_inode, stat);
-       /* Change block size to what the server returned */
-       stat->blksize = st->st_blksize;
-
-       kfree(st);
-       return 0;
-}
-
 /**
  * v9fs_vfs_setattr - set file metadata
  * @dentry: file whose metadata to set
@@ -1291,64 +914,6 @@ static int v9fs_vfs_setattr(struct dentry *dentry, struct iattr *iattr)
 }
 
 /**
- * v9fs_vfs_setattr_dotl - set file metadata
- * @dentry: file whose metadata to set
- * @iattr: metadata assignment structure
- *
- */
-
-int v9fs_vfs_setattr_dotl(struct dentry *dentry, struct iattr *iattr)
-{
-       int retval;
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *fid;
-       struct p9_iattr_dotl p9attr;
-
-       P9_DPRINTK(P9_DEBUG_VFS, "\n");
-
-       retval = inode_change_ok(dentry->d_inode, iattr);
-       if (retval)
-               return retval;
-
-       p9attr.valid = iattr->ia_valid;
-       p9attr.mode = iattr->ia_mode;
-       p9attr.uid = iattr->ia_uid;
-       p9attr.gid = iattr->ia_gid;
-       p9attr.size = iattr->ia_size;
-       p9attr.atime_sec = iattr->ia_atime.tv_sec;
-       p9attr.atime_nsec = iattr->ia_atime.tv_nsec;
-       p9attr.mtime_sec = iattr->ia_mtime.tv_sec;
-       p9attr.mtime_nsec = iattr->ia_mtime.tv_nsec;
-
-       retval = -EPERM;
-       v9ses = v9fs_inode2v9ses(dentry->d_inode);
-       fid = v9fs_fid_lookup(dentry);
-       if (IS_ERR(fid))
-               return PTR_ERR(fid);
-
-       retval = p9_client_setattr(fid, &p9attr);
-       if (retval < 0)
-               return retval;
-
-       if ((iattr->ia_valid & ATTR_SIZE) &&
-           iattr->ia_size != i_size_read(dentry->d_inode)) {
-               retval = vmtruncate(dentry->d_inode, iattr->ia_size);
-               if (retval)
-                       return retval;
-       }
-
-       setattr_copy(dentry->d_inode, iattr);
-       mark_inode_dirty(dentry->d_inode);
-       if (iattr->ia_valid & ATTR_MODE) {
-               /* We also want to update ACL when we update mode bits */
-               retval = v9fs_acl_chmod(dentry);
-               if (retval < 0)
-                       return retval;
-       }
-       return 0;
-}
-
-/**
  * v9fs_stat2inode - populate an inode structure with mistat info
  * @stat: Plan 9 metadata (mistat) structure
  * @inode: inode to populate
@@ -1426,77 +991,6 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode,
 }
 
 /**
- * v9fs_stat2inode_dotl - populate an inode structure with stat info
- * @stat: stat structure
- * @inode: inode to populate
- * @sb: superblock of filesystem
- *
- */
-
-void
-v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode)
-{
-
-       if ((stat->st_result_mask & P9_STATS_BASIC) == P9_STATS_BASIC) {
-               inode->i_atime.tv_sec = stat->st_atime_sec;
-               inode->i_atime.tv_nsec = stat->st_atime_nsec;
-               inode->i_mtime.tv_sec = stat->st_mtime_sec;
-               inode->i_mtime.tv_nsec = stat->st_mtime_nsec;
-               inode->i_ctime.tv_sec = stat->st_ctime_sec;
-               inode->i_ctime.tv_nsec = stat->st_ctime_nsec;
-               inode->i_uid = stat->st_uid;
-               inode->i_gid = stat->st_gid;
-               inode->i_nlink = stat->st_nlink;
-               inode->i_mode = stat->st_mode;
-               inode->i_rdev = new_decode_dev(stat->st_rdev);
-
-               if ((S_ISBLK(inode->i_mode)) || (S_ISCHR(inode->i_mode)))
-                       init_special_inode(inode, inode->i_mode, inode->i_rdev);
-
-               i_size_write(inode, stat->st_size);
-               inode->i_blocks = stat->st_blocks;
-       } else {
-               if (stat->st_result_mask & P9_STATS_ATIME) {
-                       inode->i_atime.tv_sec = stat->st_atime_sec;
-                       inode->i_atime.tv_nsec = stat->st_atime_nsec;
-               }
-               if (stat->st_result_mask & P9_STATS_MTIME) {
-                       inode->i_mtime.tv_sec = stat->st_mtime_sec;
-                       inode->i_mtime.tv_nsec = stat->st_mtime_nsec;
-               }
-               if (stat->st_result_mask & P9_STATS_CTIME) {
-                       inode->i_ctime.tv_sec = stat->st_ctime_sec;
-                       inode->i_ctime.tv_nsec = stat->st_ctime_nsec;
-               }
-               if (stat->st_result_mask & P9_STATS_UID)
-                       inode->i_uid = stat->st_uid;
-               if (stat->st_result_mask & P9_STATS_GID)
-                       inode->i_gid = stat->st_gid;
-               if (stat->st_result_mask & P9_STATS_NLINK)
-                       inode->i_nlink = stat->st_nlink;
-               if (stat->st_result_mask & P9_STATS_MODE) {
-                       inode->i_mode = stat->st_mode;
-                       if ((S_ISBLK(inode->i_mode)) ||
-                                               (S_ISCHR(inode->i_mode)))
-                               init_special_inode(inode, inode->i_mode,
-                                                               inode->i_rdev);
-               }
-               if (stat->st_result_mask & P9_STATS_RDEV)
-                       inode->i_rdev = new_decode_dev(stat->st_rdev);
-               if (stat->st_result_mask & P9_STATS_SIZE)
-                       i_size_write(inode, stat->st_size);
-               if (stat->st_result_mask & P9_STATS_BLOCKS)
-                       inode->i_blocks = stat->st_blocks;
-       }
-       if (stat->st_result_mask & P9_STATS_GEN)
-                       inode->i_generation = stat->st_gen;
-
-       /* Currently we don't support P9_STATS_BTIME and P9_STATS_DATA_VERSION
-        * because the inode structure does not have fields for them.
-        */
-}
-
-/**
  * v9fs_qid2ino - convert qid into inode number
  * @qid: qid to hash
  *
@@ -1602,7 +1096,7 @@ static void *v9fs_vfs_follow_link(struct dentry *dentry, struct nameidata *nd)
  *
  */
 
-static void
+void
 v9fs_vfs_put_link(struct dentry *dentry, struct nameidata *nd, void *p)
 {
        char *s = nd_get_link(nd);
@@ -1646,94 +1140,6 @@ static int v9fs_vfs_mkspecial(struct inode *dir, struct dentry *dentry,
 }
 
 /**
- * v9fs_vfs_symlink_dotl - helper function to create symlinks
- * @dir: directory inode containing symlink
- * @dentry: dentry for symlink
- * @symname: symlink data
- *
- * See Also: 9P2000.L RFC for more information
- *
- */
-
-static int
-v9fs_vfs_symlink_dotl(struct inode *dir, struct dentry *dentry,
-               const char *symname)
-{
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *dfid;
-       struct p9_fid *fid = NULL;
-       struct inode *inode;
-       struct p9_qid qid;
-       char *name;
-       int err;
-       gid_t gid;
-
-       name = (char *) dentry->d_name.name;
-       P9_DPRINTK(P9_DEBUG_VFS, "v9fs_vfs_symlink_dotl : %lu,%s,%s\n",
-                       dir->i_ino, name, symname);
-       v9ses = v9fs_inode2v9ses(dir);
-
-       dfid = v9fs_fid_lookup(dentry->d_parent);
-       if (IS_ERR(dfid)) {
-               err = PTR_ERR(dfid);
-               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
-               return err;
-       }
-
-       gid = v9fs_get_fsgid_for_create(dir);
-
-       /* Server doesn't alter fid on TSYMLINK. Hence no need to clone it. */
-       err = p9_client_symlink(dfid, name, (char *)symname, gid, &qid);
-
-       if (err < 0) {
-               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_symlink failed %d\n", err);
-               goto error;
-       }
-
-       if (v9ses->cache) {
-               /* Now walk from the parent so we can get an unopened fid. */
-               fid = p9_client_walk(dfid, 1, &name, 1);
-               if (IS_ERR(fid)) {
-                       err = PTR_ERR(fid);
-                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
-                                       err);
-                       fid = NULL;
-                       goto error;
-               }
-
-               /* instantiate inode and assign the unopened fid to dentry */
-               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
-                                       err);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_cached_dentry_operations);
-               d_instantiate(dentry, inode);
-               err = v9fs_fid_add(dentry, fid);
-               if (err < 0)
-                       goto error;
-               fid = NULL;
-       } else {
-               /* Not in cached mode. No need to populate inode with stat */
-               inode = v9fs_get_inode(dir->i_sb, S_IFLNK);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_dentry_operations);
-               d_instantiate(dentry, inode);
-       }
-
-error:
-       if (fid)
-               p9_client_clunk(fid);
-
-       return err;
-}
-
-/**
  * v9fs_vfs_symlink - helper function to create symlinks
  * @dir: directory inode containing symlink
  * @dentry: dentry for symlink
@@ -1792,77 +1198,6 @@ clunk_fid:
 }
 
 /**
- * v9fs_vfs_link_dotl - create a hardlink for dotl
- * @old_dentry: dentry for file to link to
- * @dir: inode destination for new link
- * @dentry: dentry for link
- *
- */
-
-static int
-v9fs_vfs_link_dotl(struct dentry *old_dentry, struct inode *dir,
-               struct dentry *dentry)
-{
-       int err;
-       struct p9_fid *dfid, *oldfid;
-       char *name;
-       struct v9fs_session_info *v9ses;
-       struct dentry *dir_dentry;
-
-       P9_DPRINTK(P9_DEBUG_VFS, "dir ino: %lu, old_name: %s, new_name: %s\n",
-                       dir->i_ino, old_dentry->d_name.name,
-                       dentry->d_name.name);
-
-       v9ses = v9fs_inode2v9ses(dir);
-       dir_dentry = v9fs_dentry_from_dir_inode(dir);
-       dfid = v9fs_fid_lookup(dir_dentry);
-       if (IS_ERR(dfid))
-               return PTR_ERR(dfid);
-
-       oldfid = v9fs_fid_lookup(old_dentry);
-       if (IS_ERR(oldfid))
-               return PTR_ERR(oldfid);
-
-       name = (char *) dentry->d_name.name;
-
-       err = p9_client_link(dfid, oldfid, (char *)dentry->d_name.name);
-
-       if (err < 0) {
-               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_link failed %d\n", err);
-               return err;
-       }
-
-       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
-               /* Get the latest stat info from server. */
-               struct p9_fid *fid;
-               struct p9_stat_dotl *st;
-
-               fid = v9fs_fid_lookup(old_dentry);
-               if (IS_ERR(fid))
-                       return PTR_ERR(fid);
-
-               st = p9_client_getattr_dotl(fid, P9_STATS_BASIC);
-               if (IS_ERR(st))
-                       return PTR_ERR(st);
-
-               v9fs_stat2inode_dotl(st, old_dentry->d_inode);
-
-               kfree(st);
-       } else {
-               /* Caching disabled. No need to get upto date stat info.
-                * This dentry will be released immediately. So, just hold the
-                * inode
-                */
-               ihold(old_dentry->d_inode);
-       }
-
-       d_set_d_op(dentry, old_dentry->d_op);
-       d_instantiate(dentry, old_dentry->d_inode);
-
-       return err;
-}
-
-/**
  * v9fs_vfs_mknod - create a special file
  * @dir: inode destination for new link
  * @dentry: dentry for file
@@ -1907,160 +1242,6 @@ v9fs_vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
        return retval;
 }
 
-/**
- * v9fs_vfs_mknod_dotl - create a special file
- * @dir: inode destination for new link
- * @dentry: dentry for file
- * @mode: mode for creation
- * @rdev: device associated with special file
- *
- */
-static int
-v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode,
-               dev_t rdev)
-{
-       int err;
-       char *name;
-       mode_t mode;
-       struct v9fs_session_info *v9ses;
-       struct p9_fid *fid = NULL, *dfid = NULL;
-       struct inode *inode;
-       gid_t gid;
-       struct p9_qid qid;
-       struct dentry *dir_dentry;
-       struct posix_acl *dacl = NULL, *pacl = NULL;
-
-       P9_DPRINTK(P9_DEBUG_VFS,
-               " %lu,%s mode: %x MAJOR: %u MINOR: %u\n", dir->i_ino,
-               dentry->d_name.name, omode, MAJOR(rdev), MINOR(rdev));
-
-       if (!new_valid_dev(rdev))
-               return -EINVAL;
-
-       v9ses = v9fs_inode2v9ses(dir);
-       dir_dentry = v9fs_dentry_from_dir_inode(dir);
-       dfid = v9fs_fid_lookup(dir_dentry);
-       if (IS_ERR(dfid)) {
-               err = PTR_ERR(dfid);
-               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
-               dfid = NULL;
-               goto error;
-       }
-
-       gid = v9fs_get_fsgid_for_create(dir);
-       mode = omode;
-       /* Update mode based on ACL value */
-       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
-       if (err) {
-               P9_DPRINTK(P9_DEBUG_VFS,
-                          "Failed to get acl values in mknod %d\n", err);
-               goto error;
-       }
-       name = (char *) dentry->d_name.name;
-
-       err = p9_client_mknod_dotl(dfid, name, mode, rdev, gid, &qid);
-       if (err < 0)
-               goto error;
-
-       /* instantiate inode and assign the unopened fid to the dentry */
-       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
-               fid = p9_client_walk(dfid, 1, &name, 1);
-               if (IS_ERR(fid)) {
-                       err = PTR_ERR(fid);
-                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
-                               err);
-                       fid = NULL;
-                       goto error;
-               }
-
-               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
-                               err);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_cached_dentry_operations);
-               d_instantiate(dentry, inode);
-               err = v9fs_fid_add(dentry, fid);
-               if (err < 0)
-                       goto error;
-               fid = NULL;
-       } else {
-               /*
-                * Not in cached mode. No need to populate inode with stat.
-                * socket syscall returns a fd, so we need instantiate
-                */
-               inode = v9fs_get_inode(dir->i_sb, mode);
-               if (IS_ERR(inode)) {
-                       err = PTR_ERR(inode);
-                       goto error;
-               }
-               d_set_d_op(dentry, &v9fs_dentry_operations);
-               d_instantiate(dentry, inode);
-       }
-       /* Now set the ACL based on the default value */
-       v9fs_set_create_acl(dentry, dacl, pacl);
-error:
-       if (fid)
-               p9_client_clunk(fid);
-       return err;
-}
-
-static int
-v9fs_vfs_readlink_dotl(struct dentry *dentry, char *buffer, int buflen)
-{
-       int retval;
-       struct p9_fid *fid;
-       char *target = NULL;
-
-       P9_DPRINTK(P9_DEBUG_VFS, " %s\n", dentry->d_name.name);
-       retval = -EPERM;
-       fid = v9fs_fid_lookup(dentry);
-       if (IS_ERR(fid))
-               return PTR_ERR(fid);
-
-       retval = p9_client_readlink(fid, &target);
-       if (retval < 0)
-               return retval;
-
-       strncpy(buffer, target, buflen);
-       P9_DPRINTK(P9_DEBUG_VFS, "%s -> %s\n", dentry->d_name.name, buffer);
-
-       retval = strnlen(buffer, buflen);
-       return retval;
-}
-
-/**
- * v9fs_vfs_follow_link_dotl - follow a symlink path
- * @dentry: dentry for symlink
- * @nd: nameidata
- *
- */
-
-static void *
-v9fs_vfs_follow_link_dotl(struct dentry *dentry, struct nameidata *nd)
-{
-       int len = 0;
-       char *link = __getname();
-
-       P9_DPRINTK(P9_DEBUG_VFS, "%s n", dentry->d_name.name);
-
-       if (!link)
-               link = ERR_PTR(-ENOMEM);
-       else {
-               len = v9fs_vfs_readlink_dotl(dentry, link, PATH_MAX);
-               if (len < 0) {
-                       __putname(link);
-                       link = ERR_PTR(len);
-               } else
-                       link[min(len, PATH_MAX-1)] = 0;
-       }
-       nd_set_link(nd, link);
-
-       return NULL;
-}
-
 static const struct inode_operations v9fs_dir_inode_operations_dotu = {
        .create = v9fs_vfs_create,
        .lookup = v9fs_vfs_lookup,
@@ -2075,25 +1256,6 @@ static const struct inode_operations v9fs_dir_inode_operations_dotu = {
        .setattr = v9fs_vfs_setattr,
 };
 
-static const struct inode_operations v9fs_dir_inode_operations_dotl = {
-       .create = v9fs_vfs_create_dotl,
-       .lookup = v9fs_vfs_lookup,
-       .link = v9fs_vfs_link_dotl,
-       .symlink = v9fs_vfs_symlink_dotl,
-       .unlink = v9fs_vfs_unlink,
-       .mkdir = v9fs_vfs_mkdir_dotl,
-       .rmdir = v9fs_vfs_rmdir,
-       .mknod = v9fs_vfs_mknod_dotl,
-       .rename = v9fs_vfs_rename,
-       .getattr = v9fs_vfs_getattr_dotl,
-       .setattr = v9fs_vfs_setattr_dotl,
-       .setxattr = generic_setxattr,
-       .getxattr = generic_getxattr,
-       .removexattr = generic_removexattr,
-       .listxattr = v9fs_listxattr,
-       .check_acl = v9fs_check_acl,
-};
-
 static const struct inode_operations v9fs_dir_inode_operations = {
        .create = v9fs_vfs_create,
        .lookup = v9fs_vfs_lookup,
@@ -2111,16 +1273,6 @@ static const struct inode_operations v9fs_file_inode_operations = {
        .setattr = v9fs_vfs_setattr,
 };
 
-static const struct inode_operations v9fs_file_inode_operations_dotl = {
-       .getattr = v9fs_vfs_getattr_dotl,
-       .setattr = v9fs_vfs_setattr_dotl,
-       .setxattr = generic_setxattr,
-       .getxattr = generic_getxattr,
-       .removexattr = generic_removexattr,
-       .listxattr = v9fs_listxattr,
-       .check_acl = v9fs_check_acl,
-};
-
 static const struct inode_operations v9fs_symlink_inode_operations = {
        .readlink = generic_readlink,
        .follow_link = v9fs_vfs_follow_link,
@@ -2129,14 +1281,3 @@ static const struct inode_operations v9fs_symlink_inode_operations = {
        .setattr = v9fs_vfs_setattr,
 };
 
-static const struct inode_operations v9fs_symlink_inode_operations_dotl = {
-       .readlink = v9fs_vfs_readlink_dotl,
-       .follow_link = v9fs_vfs_follow_link_dotl,
-       .put_link = v9fs_vfs_put_link,
-       .getattr = v9fs_vfs_getattr_dotl,
-       .setattr = v9fs_vfs_setattr_dotl,
-       .setxattr = generic_setxattr,
-       .getxattr = generic_getxattr,
-       .removexattr = generic_removexattr,
-       .listxattr = v9fs_listxattr,
-};
diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c
new file mode 100644 (file)
index 0000000..fe3ffa9
--- /dev/null
@@ -0,0 +1,824 @@
+/*
+ *  linux/fs/9p/vfs_inode_dotl.c
+ *
+ * This file contains vfs inode ops for the 9P2000.L protocol.
+ *
+ *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
+ *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/pagemap.h>
+#include <linux/stat.h>
+#include <linux/string.h>
+#include <linux/inet.h>
+#include <linux/namei.h>
+#include <linux/idr.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/xattr.h>
+#include <linux/posix_acl.h>
+#include <net/9p/9p.h>
+#include <net/9p/client.h>
+
+#include "v9fs.h"
+#include "v9fs_vfs.h"
+#include "fid.h"
+#include "cache.h"
+#include "xattr.h"
+#include "acl.h"
+
+static int
+v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode,
+                   dev_t rdev);
+
+/**
+ * v9fs_get_fsgid_for_create - Helper function to get the gid for creating a
+ * new file system object. This checks the S_ISGID to determine the owning
+ * group of the new file system object.
+ */
+
+static gid_t v9fs_get_fsgid_for_create(struct inode *dir_inode)
+{
+       BUG_ON(dir_inode == NULL);
+
+       if (dir_inode->i_mode & S_ISGID) {
+               /* set_gid bit is set.*/
+               return dir_inode->i_gid;
+       }
+       return current_fsgid();
+}
+
+/**
+ * v9fs_dentry_from_dir_inode - helper function to get the dentry from
+ * dir inode.
+ *
+ */
+
+static struct dentry *v9fs_dentry_from_dir_inode(struct inode *inode)
+{
+       struct dentry *dentry;
+
+       spin_lock(&inode->i_lock);
+       /* Directory should have only one entry. */
+       BUG_ON(S_ISDIR(inode->i_mode) && !list_is_singular(&inode->i_dentry));
+       dentry = list_entry(inode->i_dentry.next, struct dentry, d_alias);
+       spin_unlock(&inode->i_lock);
+       return dentry;
+}
+
+struct inode *
+v9fs_inode_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid,
+       struct super_block *sb)
+{
+       struct inode *ret = NULL;
+       int err;
+       struct p9_stat_dotl *st;
+
+       st = p9_client_getattr_dotl(fid, P9_STATS_BASIC);
+       if (IS_ERR(st))
+               return ERR_CAST(st);
+
+       ret = v9fs_get_inode(sb, st->st_mode);
+       if (IS_ERR(ret)) {
+               err = PTR_ERR(ret);
+               goto error;
+       }
+
+       v9fs_stat2inode_dotl(st, ret);
+       ret->i_ino = v9fs_qid2ino(&st->qid);
+#ifdef CONFIG_9P_FSCACHE
+       v9fs_vcookie_set_qid(ret, &st->qid);
+       v9fs_cache_inode_get_cookie(ret);
+#endif
+       err = v9fs_get_acl(ret, fid);
+       if (err) {
+               iput(ret);
+               goto error;
+       }
+       kfree(st);
+       return ret;
+error:
+       kfree(st);
+       return ERR_PTR(err);
+}
+
+/**
+ * v9fs_vfs_create_dotl - VFS hook to create files for 9P2000.L protocol.
+ * @dir: directory inode that is being created
+ * @dentry:  dentry that is being deleted
+ * @mode: create permissions
+ * @nd: path information
+ *
+ */
+
+static int
+v9fs_vfs_create_dotl(struct inode *dir, struct dentry *dentry, int omode,
+               struct nameidata *nd)
+{
+       int err = 0;
+       char *name = NULL;
+       gid_t gid;
+       int flags;
+       mode_t mode;
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *fid = NULL;
+       struct p9_fid *dfid, *ofid;
+       struct file *filp;
+       struct p9_qid qid;
+       struct inode *inode;
+       struct posix_acl *pacl = NULL, *dacl = NULL;
+
+       v9ses = v9fs_inode2v9ses(dir);
+       if (nd && nd->flags & LOOKUP_OPEN)
+               flags = nd->intent.open.flags - 1;
+       else {
+               /*
+                * create call without LOOKUP_OPEN is due
+                * to mknod of regular files. So use mknod
+                * operation.
+                */
+               return v9fs_vfs_mknod_dotl(dir, dentry, omode, 0);
+       }
+
+       name = (char *) dentry->d_name.name;
+       P9_DPRINTK(P9_DEBUG_VFS, "v9fs_vfs_create_dotl: name:%s flags:0x%x "
+                       "mode:0x%x\n", name, flags, omode);
+
+       dfid = v9fs_fid_lookup(dentry->d_parent);
+       if (IS_ERR(dfid)) {
+               err = PTR_ERR(dfid);
+               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
+               return err;
+       }
+
+       /* clone a fid to use for creation */
+       ofid = p9_client_walk(dfid, 0, NULL, 1);
+       if (IS_ERR(ofid)) {
+               err = PTR_ERR(ofid);
+               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
+               return err;
+       }
+
+       gid = v9fs_get_fsgid_for_create(dir);
+
+       mode = omode;
+       /* Update mode based on ACL value */
+       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
+       if (err) {
+               P9_DPRINTK(P9_DEBUG_VFS,
+                          "Failed to get acl values in creat %d\n", err);
+               goto error;
+       }
+       err = p9_client_create_dotl(ofid, name, flags, mode, gid, &qid);
+       if (err < 0) {
+               P9_DPRINTK(P9_DEBUG_VFS,
+                               "p9_client_open_dotl failed in creat %d\n",
+                               err);
+               goto error;
+       }
+
+       /* instantiate inode and assign the unopened fid to the dentry */
+       fid = p9_client_walk(dfid, 1, &name, 1);
+       if (IS_ERR(fid)) {
+               err = PTR_ERR(fid);
+               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err);
+               fid = NULL;
+               goto error;
+       }
+       inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
+       if (IS_ERR(inode)) {
+               err = PTR_ERR(inode);
+               P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n", err);
+               goto error;
+       }
+       d_instantiate(dentry, inode);
+       err = v9fs_fid_add(dentry, fid);
+       if (err < 0)
+               goto error;
+
+       /* Now set the ACL based on the default value */
+       v9fs_set_create_acl(dentry, dacl, pacl);
+
+       /* Since we are opening a file, assign the open fid to the file */
+       filp = lookup_instantiate_filp(nd, dentry, generic_file_open);
+       if (IS_ERR(filp)) {
+               p9_client_clunk(ofid);
+               return PTR_ERR(filp);
+       }
+       filp->private_data = ofid;
+       return 0;
+
+error:
+       if (ofid)
+               p9_client_clunk(ofid);
+       if (fid)
+               p9_client_clunk(fid);
+       return err;
+}
+
+/**
+ * v9fs_vfs_mkdir_dotl - VFS mkdir hook to create a directory
+ * @dir:  inode that is being unlinked
+ * @dentry: dentry that is being unlinked
+ * @mode: mode for new directory
+ *
+ */
+
+static int v9fs_vfs_mkdir_dotl(struct inode *dir,
+                              struct dentry *dentry, int omode)
+{
+       int err;
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *fid = NULL, *dfid = NULL;
+       gid_t gid;
+       char *name;
+       mode_t mode;
+       struct inode *inode;
+       struct p9_qid qid;
+       struct dentry *dir_dentry;
+       struct posix_acl *dacl = NULL, *pacl = NULL;
+
+       P9_DPRINTK(P9_DEBUG_VFS, "name %s\n", dentry->d_name.name);
+       err = 0;
+       v9ses = v9fs_inode2v9ses(dir);
+
+       omode |= S_IFDIR;
+       if (dir->i_mode & S_ISGID)
+               omode |= S_ISGID;
+
+       dir_dentry = v9fs_dentry_from_dir_inode(dir);
+       dfid = v9fs_fid_lookup(dir_dentry);
+       if (IS_ERR(dfid)) {
+               err = PTR_ERR(dfid);
+               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
+               dfid = NULL;
+               goto error;
+       }
+
+       gid = v9fs_get_fsgid_for_create(dir);
+       mode = omode;
+       /* Update mode based on ACL value */
+       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
+       if (err) {
+               P9_DPRINTK(P9_DEBUG_VFS,
+                          "Failed to get acl values in mkdir %d\n", err);
+               goto error;
+       }
+       name = (char *) dentry->d_name.name;
+       err = p9_client_mkdir_dotl(dfid, name, mode, gid, &qid);
+       if (err < 0)
+               goto error;
+
+       /* instantiate inode and assign the unopened fid to the dentry */
+       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
+               fid = p9_client_walk(dfid, 1, &name, 1);
+               if (IS_ERR(fid)) {
+                       err = PTR_ERR(fid);
+                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
+                               err);
+                       fid = NULL;
+                       goto error;
+               }
+
+               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
+                               err);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+               err = v9fs_fid_add(dentry, fid);
+               if (err < 0)
+                       goto error;
+               fid = NULL;
+       } else {
+               /*
+                * Not in cached mode. No need to populate
+                * inode with stat. We need to get an inode
+                * so that we can set the acl with dentry
+                */
+               inode = v9fs_get_inode(dir->i_sb, mode);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+       }
+       /* Now set the ACL based on the default value */
+       v9fs_set_create_acl(dentry, dacl, pacl);
+
+error:
+       if (fid)
+               p9_client_clunk(fid);
+       return err;
+}
+
+static int
+v9fs_vfs_getattr_dotl(struct vfsmount *mnt, struct dentry *dentry,
+                struct kstat *stat)
+{
+       int err;
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *fid;
+       struct p9_stat_dotl *st;
+
+       P9_DPRINTK(P9_DEBUG_VFS, "dentry: %p\n", dentry);
+       err = -EPERM;
+       v9ses = v9fs_inode2v9ses(dentry->d_inode);
+       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE)
+               return simple_getattr(mnt, dentry, stat);
+
+       fid = v9fs_fid_lookup(dentry);
+       if (IS_ERR(fid))
+               return PTR_ERR(fid);
+
+       /* Ask for all the fields in stat structure. Server will return
+        * whatever it supports
+        */
+
+       st = p9_client_getattr_dotl(fid, P9_STATS_ALL);
+       if (IS_ERR(st))
+               return PTR_ERR(st);
+
+       v9fs_stat2inode_dotl(st, dentry->d_inode);
+       generic_fillattr(dentry->d_inode, stat);
+       /* Change block size to what the server returned */
+       stat->blksize = st->st_blksize;
+
+       kfree(st);
+       return 0;
+}
+
+/**
+ * v9fs_vfs_setattr_dotl - set file metadata
+ * @dentry: file whose metadata to set
+ * @iattr: metadata assignment structure
+ *
+ */
+
+int v9fs_vfs_setattr_dotl(struct dentry *dentry, struct iattr *iattr)
+{
+       int retval;
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *fid;
+       struct p9_iattr_dotl p9attr;
+
+       P9_DPRINTK(P9_DEBUG_VFS, "\n");
+
+       retval = inode_change_ok(dentry->d_inode, iattr);
+       if (retval)
+               return retval;
+
+       p9attr.valid = iattr->ia_valid;
+       p9attr.mode = iattr->ia_mode;
+       p9attr.uid = iattr->ia_uid;
+       p9attr.gid = iattr->ia_gid;
+       p9attr.size = iattr->ia_size;
+       p9attr.atime_sec = iattr->ia_atime.tv_sec;
+       p9attr.atime_nsec = iattr->ia_atime.tv_nsec;
+       p9attr.mtime_sec = iattr->ia_mtime.tv_sec;
+       p9attr.mtime_nsec = iattr->ia_mtime.tv_nsec;
+
+       retval = -EPERM;
+       v9ses = v9fs_inode2v9ses(dentry->d_inode);
+       fid = v9fs_fid_lookup(dentry);
+       if (IS_ERR(fid))
+               return PTR_ERR(fid);
+
+       retval = p9_client_setattr(fid, &p9attr);
+       if (retval < 0)
+               return retval;
+
+       if ((iattr->ia_valid & ATTR_SIZE) &&
+           iattr->ia_size != i_size_read(dentry->d_inode)) {
+               retval = vmtruncate(dentry->d_inode, iattr->ia_size);
+               if (retval)
+                       return retval;
+       }
+
+       setattr_copy(dentry->d_inode, iattr);
+       mark_inode_dirty(dentry->d_inode);
+       if (iattr->ia_valid & ATTR_MODE) {
+               /* We also want to update ACL when we update mode bits */
+               retval = v9fs_acl_chmod(dentry);
+               if (retval < 0)
+                       return retval;
+       }
+       return 0;
+}
+
+/**
+ * v9fs_stat2inode_dotl - populate an inode structure with stat info
+ * @stat: stat structure
+ * @inode: inode to populate
+ * @sb: superblock of filesystem
+ *
+ */
+
+void
+v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode)
+{
+
+       if ((stat->st_result_mask & P9_STATS_BASIC) == P9_STATS_BASIC) {
+               inode->i_atime.tv_sec = stat->st_atime_sec;
+               inode->i_atime.tv_nsec = stat->st_atime_nsec;
+               inode->i_mtime.tv_sec = stat->st_mtime_sec;
+               inode->i_mtime.tv_nsec = stat->st_mtime_nsec;
+               inode->i_ctime.tv_sec = stat->st_ctime_sec;
+               inode->i_ctime.tv_nsec = stat->st_ctime_nsec;
+               inode->i_uid = stat->st_uid;
+               inode->i_gid = stat->st_gid;
+               inode->i_nlink = stat->st_nlink;
+               inode->i_mode = stat->st_mode;
+               inode->i_rdev = new_decode_dev(stat->st_rdev);
+
+               if ((S_ISBLK(inode->i_mode)) || (S_ISCHR(inode->i_mode)))
+                       init_special_inode(inode, inode->i_mode, inode->i_rdev);
+
+               i_size_write(inode, stat->st_size);
+               inode->i_blocks = stat->st_blocks;
+       } else {
+               if (stat->st_result_mask & P9_STATS_ATIME) {
+                       inode->i_atime.tv_sec = stat->st_atime_sec;
+                       inode->i_atime.tv_nsec = stat->st_atime_nsec;
+               }
+               if (stat->st_result_mask & P9_STATS_MTIME) {
+                       inode->i_mtime.tv_sec = stat->st_mtime_sec;
+                       inode->i_mtime.tv_nsec = stat->st_mtime_nsec;
+               }
+               if (stat->st_result_mask & P9_STATS_CTIME) {
+                       inode->i_ctime.tv_sec = stat->st_ctime_sec;
+                       inode->i_ctime.tv_nsec = stat->st_ctime_nsec;
+               }
+               if (stat->st_result_mask & P9_STATS_UID)
+                       inode->i_uid = stat->st_uid;
+               if (stat->st_result_mask & P9_STATS_GID)
+                       inode->i_gid = stat->st_gid;
+               if (stat->st_result_mask & P9_STATS_NLINK)
+                       inode->i_nlink = stat->st_nlink;
+               if (stat->st_result_mask & P9_STATS_MODE) {
+                       inode->i_mode = stat->st_mode;
+                       if ((S_ISBLK(inode->i_mode)) ||
+                                               (S_ISCHR(inode->i_mode)))
+                               init_special_inode(inode, inode->i_mode,
+                                                               inode->i_rdev);
+               }
+               if (stat->st_result_mask & P9_STATS_RDEV)
+                       inode->i_rdev = new_decode_dev(stat->st_rdev);
+               if (stat->st_result_mask & P9_STATS_SIZE)
+                       i_size_write(inode, stat->st_size);
+               if (stat->st_result_mask & P9_STATS_BLOCKS)
+                       inode->i_blocks = stat->st_blocks;
+       }
+       if (stat->st_result_mask & P9_STATS_GEN)
+                       inode->i_generation = stat->st_gen;
+
+       /* Currently we don't support P9_STATS_BTIME and P9_STATS_DATA_VERSION
+        * because the inode structure does not have fields for them.
+        */
+}
+
+static int
+v9fs_vfs_symlink_dotl(struct inode *dir, struct dentry *dentry,
+               const char *symname)
+{
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *dfid;
+       struct p9_fid *fid = NULL;
+       struct inode *inode;
+       struct p9_qid qid;
+       char *name;
+       int err;
+       gid_t gid;
+
+       name = (char *) dentry->d_name.name;
+       P9_DPRINTK(P9_DEBUG_VFS, "v9fs_vfs_symlink_dotl : %lu,%s,%s\n",
+                       dir->i_ino, name, symname);
+       v9ses = v9fs_inode2v9ses(dir);
+
+       dfid = v9fs_fid_lookup(dentry->d_parent);
+       if (IS_ERR(dfid)) {
+               err = PTR_ERR(dfid);
+               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
+               return err;
+       }
+
+       gid = v9fs_get_fsgid_for_create(dir);
+
+       /* Server doesn't alter fid on TSYMLINK. Hence no need to clone it. */
+       err = p9_client_symlink(dfid, name, (char *)symname, gid, &qid);
+
+       if (err < 0) {
+               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_symlink failed %d\n", err);
+               goto error;
+       }
+
+       if (v9ses->cache) {
+               /* Now walk from the parent so we can get an unopened fid. */
+               fid = p9_client_walk(dfid, 1, &name, 1);
+               if (IS_ERR(fid)) {
+                       err = PTR_ERR(fid);
+                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
+                                       err);
+                       fid = NULL;
+                       goto error;
+               }
+
+               /* instantiate inode and assign the unopened fid to dentry */
+               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
+                                       err);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+               err = v9fs_fid_add(dentry, fid);
+               if (err < 0)
+                       goto error;
+               fid = NULL;
+       } else {
+               /* Not in cached mode. No need to populate inode with stat */
+               inode = v9fs_get_inode(dir->i_sb, S_IFLNK);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+       }
+
+error:
+       if (fid)
+               p9_client_clunk(fid);
+
+       return err;
+}
+
+/**
+ * v9fs_vfs_link_dotl - create a hardlink for dotl
+ * @old_dentry: dentry for file to link to
+ * @dir: inode destination for new link
+ * @dentry: dentry for link
+ *
+ */
+
+static int
+v9fs_vfs_link_dotl(struct dentry *old_dentry, struct inode *dir,
+               struct dentry *dentry)
+{
+       int err;
+       struct p9_fid *dfid, *oldfid;
+       char *name;
+       struct v9fs_session_info *v9ses;
+       struct dentry *dir_dentry;
+
+       P9_DPRINTK(P9_DEBUG_VFS, "dir ino: %lu, old_name: %s, new_name: %s\n",
+                       dir->i_ino, old_dentry->d_name.name,
+                       dentry->d_name.name);
+
+       v9ses = v9fs_inode2v9ses(dir);
+       dir_dentry = v9fs_dentry_from_dir_inode(dir);
+       dfid = v9fs_fid_lookup(dir_dentry);
+       if (IS_ERR(dfid))
+               return PTR_ERR(dfid);
+
+       oldfid = v9fs_fid_lookup(old_dentry);
+       if (IS_ERR(oldfid))
+               return PTR_ERR(oldfid);
+
+       name = (char *) dentry->d_name.name;
+
+       err = p9_client_link(dfid, oldfid, (char *)dentry->d_name.name);
+
+       if (err < 0) {
+               P9_DPRINTK(P9_DEBUG_VFS, "p9_client_link failed %d\n", err);
+               return err;
+       }
+
+       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
+               /* Get the latest stat info from server. */
+               struct p9_fid *fid;
+               struct p9_stat_dotl *st;
+
+               fid = v9fs_fid_lookup(old_dentry);
+               if (IS_ERR(fid))
+                       return PTR_ERR(fid);
+
+               st = p9_client_getattr_dotl(fid, P9_STATS_BASIC);
+               if (IS_ERR(st))
+                       return PTR_ERR(st);
+
+               v9fs_stat2inode_dotl(st, old_dentry->d_inode);
+
+               kfree(st);
+       } else {
+               /* Caching disabled. No need to get upto date stat info.
+                * This dentry will be released immediately. So, just hold the
+                * inode
+                */
+               ihold(old_dentry->d_inode);
+       }
+       d_instantiate(dentry, old_dentry->d_inode);
+
+       return err;
+}
+
+/**
+ * v9fs_vfs_mknod_dotl - create a special file
+ * @dir: inode destination for new link
+ * @dentry: dentry for file
+ * @mode: mode for creation
+ * @rdev: device associated with special file
+ *
+ */
+static int
+v9fs_vfs_mknod_dotl(struct inode *dir, struct dentry *dentry, int omode,
+               dev_t rdev)
+{
+       int err;
+       char *name;
+       mode_t mode;
+       struct v9fs_session_info *v9ses;
+       struct p9_fid *fid = NULL, *dfid = NULL;
+       struct inode *inode;
+       gid_t gid;
+       struct p9_qid qid;
+       struct dentry *dir_dentry;
+       struct posix_acl *dacl = NULL, *pacl = NULL;
+
+       P9_DPRINTK(P9_DEBUG_VFS,
+               " %lu,%s mode: %x MAJOR: %u MINOR: %u\n", dir->i_ino,
+               dentry->d_name.name, omode, MAJOR(rdev), MINOR(rdev));
+
+       if (!new_valid_dev(rdev))
+               return -EINVAL;
+
+       v9ses = v9fs_inode2v9ses(dir);
+       dir_dentry = v9fs_dentry_from_dir_inode(dir);
+       dfid = v9fs_fid_lookup(dir_dentry);
+       if (IS_ERR(dfid)) {
+               err = PTR_ERR(dfid);
+               P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err);
+               dfid = NULL;
+               goto error;
+       }
+
+       gid = v9fs_get_fsgid_for_create(dir);
+       mode = omode;
+       /* Update mode based on ACL value */
+       err = v9fs_acl_mode(dir, &mode, &dacl, &pacl);
+       if (err) {
+               P9_DPRINTK(P9_DEBUG_VFS,
+                          "Failed to get acl values in mknod %d\n", err);
+               goto error;
+       }
+       name = (char *) dentry->d_name.name;
+
+       err = p9_client_mknod_dotl(dfid, name, mode, rdev, gid, &qid);
+       if (err < 0)
+               goto error;
+
+       /* instantiate inode and assign the unopened fid to the dentry */
+       if (v9ses->cache == CACHE_LOOSE || v9ses->cache == CACHE_FSCACHE) {
+               fid = p9_client_walk(dfid, 1, &name, 1);
+               if (IS_ERR(fid)) {
+                       err = PTR_ERR(fid);
+                       P9_DPRINTK(P9_DEBUG_VFS, "p9_client_walk failed %d\n",
+                               err);
+                       fid = NULL;
+                       goto error;
+               }
+
+               inode = v9fs_inode_from_fid(v9ses, fid, dir->i_sb);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       P9_DPRINTK(P9_DEBUG_VFS, "inode creation failed %d\n",
+                               err);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+               err = v9fs_fid_add(dentry, fid);
+               if (err < 0)
+                       goto error;
+               fid = NULL;
+       } else {
+               /*
+                * Not in cached mode. No need to populate inode with stat.
+                * socket syscall returns a fd, so we need instantiate
+                */
+               inode = v9fs_get_inode(dir->i_sb, mode);
+               if (IS_ERR(inode)) {
+                       err = PTR_ERR(inode);
+                       goto error;
+               }
+               d_instantiate(dentry, inode);
+       }
+       /* Now set the ACL based on the default value */
+       v9fs_set_create_acl(dentry, dacl, pacl);
+error:
+       if (fid)
+               p9_client_clunk(fid);
+       return err;
+}
+
+/**
+ * v9fs_vfs_follow_link_dotl - follow a symlink path
+ * @dentry: dentry for symlink
+ * @nd: nameidata
+ *
+ */
+
+static void *
+v9fs_vfs_follow_link_dotl(struct dentry *dentry, struct nameidata *nd)
+{
+       int retval;
+       struct p9_fid *fid;
+       char *link = __getname();
+       char *target;
+
+       P9_DPRINTK(P9_DEBUG_VFS, "%s\n", dentry->d_name.name);
+
+       if (!link) {
+               link = ERR_PTR(-ENOMEM);
+               goto ndset;
+       }
+       fid = v9fs_fid_lookup(dentry);
+       if (IS_ERR(fid)) {
+               __putname(link);
+               link = ERR_PTR(PTR_ERR(fid));
+               goto ndset;
+       }
+       retval = p9_client_readlink(fid, &target);
+       if (!retval) {
+               strcpy(link, target);
+               kfree(target);
+               goto ndset;
+       }
+       __putname(link);
+       link = ERR_PTR(retval);
+ndset:
+       nd_set_link(nd, link);
+       return NULL;
+}
+
+const struct inode_operations v9fs_dir_inode_operations_dotl = {
+       .create = v9fs_vfs_create_dotl,
+       .lookup = v9fs_vfs_lookup,
+       .link = v9fs_vfs_link_dotl,
+       .symlink = v9fs_vfs_symlink_dotl,
+       .unlink = v9fs_vfs_unlink,
+       .mkdir = v9fs_vfs_mkdir_dotl,
+       .rmdir = v9fs_vfs_rmdir,
+       .mknod = v9fs_vfs_mknod_dotl,
+       .rename = v9fs_vfs_rename,
+       .getattr = v9fs_vfs_getattr_dotl,
+       .setattr = v9fs_vfs_setattr_dotl,
+       .setxattr = generic_setxattr,
+       .getxattr = generic_getxattr,
+       .removexattr = generic_removexattr,
+       .listxattr = v9fs_listxattr,
+       .check_acl = v9fs_check_acl,
+};
+
+const struct inode_operations v9fs_file_inode_operations_dotl = {
+       .getattr = v9fs_vfs_getattr_dotl,
+       .setattr = v9fs_vfs_setattr_dotl,
+       .setxattr = generic_setxattr,
+       .getxattr = generic_getxattr,
+       .removexattr = generic_removexattr,
+       .listxattr = v9fs_listxattr,
+       .check_acl = v9fs_check_acl,
+};
+
+const struct inode_operations v9fs_symlink_inode_operations_dotl = {
+       .readlink = generic_readlink,
+       .follow_link = v9fs_vfs_follow_link_dotl,
+       .put_link = v9fs_vfs_put_link,
+       .getattr = v9fs_vfs_getattr_dotl,
+       .setattr = v9fs_vfs_setattr_dotl,
+       .setxattr = generic_setxattr,
+       .getxattr = generic_getxattr,
+       .removexattr = generic_removexattr,
+       .listxattr = v9fs_listxattr,
+};
index 43ec7df..d288773 100644 (file)
@@ -133,7 +133,7 @@ int v9fs_xattr_set(struct dentry *dentry, const char *name,
                        "p9_client_xattrcreate failed %d\n", retval);
                goto error;
        }
-       msize = fid->clnt->msize;;
+       msize = fid->clnt->msize;
        while (value_len) {
                if (value_len > (msize - P9_IOHDRSZ))
                        write_count = msize - P9_IOHDRSZ;
index 798beac..1e308f2 100644 (file)
@@ -178,27 +178,24 @@ p9pdu_vreadf(struct p9_fcall *pdu, int proto_version, const char *fmt,
                        break;
                case 's':{
                                char **sptr = va_arg(ap, char **);
-                               int16_t len;
-                               int size;
+                               uint16_t len;
 
                                errcode = p9pdu_readf(pdu, proto_version,
                                                                "w", &len);
                                if (errcode)
                                        break;
 
-                               size = max_t(int16_t, len, 0);
-
-                               *sptr = kmalloc(size + 1, GFP_KERNEL);
+                               *sptr = kmalloc(len + 1, GFP_KERNEL);
                                if (*sptr == NULL) {
                                        errcode = -EFAULT;
                                        break;
                                }
-                               if (pdu_read(pdu, *sptr, size)) {
+                               if (pdu_read(pdu, *sptr, len)) {
                                        errcode = -EFAULT;
                                        kfree(*sptr);
                                        *sptr = NULL;
                                } else
-                                       (*sptr)[size] = 0;
+                                       (*sptr)[len] = 0;
                        }
                        break;
                case 'Q':{
@@ -234,14 +231,14 @@ p9pdu_vreadf(struct p9_fcall *pdu, int proto_version, const char *fmt,
                        }
                        break;
                case 'D':{
-                               int32_t *count = va_arg(ap, int32_t *);
+                               uint32_t *count = va_arg(ap, uint32_t *);
                                void **data = va_arg(ap, void **);
 
                                errcode =
                                    p9pdu_readf(pdu, proto_version, "d", count);
                                if (!errcode) {
                                        *count =
-                                           min_t(int32_t, *count,
+                                           min_t(uint32_t, *count,
                                                  pdu->size - pdu->offset);
                                        *data = &pdu->sdata[pdu->offset];
                                }
@@ -404,9 +401,10 @@ p9pdu_vwritef(struct p9_fcall *pdu, int proto_version, const char *fmt,
                        break;
                case 's':{
                                const char *sptr = va_arg(ap, const char *);
-                               int16_t len = 0;
+                               uint16_t len = 0;
                                if (sptr)
-                                       len = min_t(int16_t, strlen(sptr), USHRT_MAX);
+                                       len = min_t(uint16_t, strlen(sptr),
+                                                               USHRT_MAX);
 
                                errcode = p9pdu_writef(pdu, proto_version,
                                                                "w", len);
@@ -438,7 +436,7 @@ p9pdu_vwritef(struct p9_fcall *pdu, int proto_version, const char *fmt,
                                                 stbuf->n_gid, stbuf->n_muid);
                        } break;
                case 'D':{
-                               int32_t count = va_arg(ap, int32_t);
+                               uint32_t count = va_arg(ap, uint32_t);
                                const void *data = va_arg(ap, const void *);
 
                                errcode = p9pdu_writef(pdu, proto_version, "d",