Return-Path: linux-nfs-owner@vger.kernel.org Received: from fieldses.org ([174.143.236.118]:59857 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754685Ab1JTILx (ORCPT ); Thu, 20 Oct 2011 04:11:53 -0400 Date: Thu, 20 Oct 2011 04:11:47 -0400 From: "J. Bruce Fields" To: "Aneesh Kumar K.V" Cc: agruen@kernel.org, akpm@linux-foundation.org, viro@zeniv.linux.org.uk, dhowells@redhat.com, linux-fsdevel@vger.kernel.org, linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH -V7 09/26] vfs: Add delete child and delete self permission flags Message-ID: <20111020081147.GB5444@fieldses.org> References: <1318951981-5508-1-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1318951981-5508-10-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <20111019220915.GA1874@fieldses.org> <87mxcw5d4x.fsf@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <87mxcw5d4x.fsf@linux.vnet.ibm.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Thu, Oct 20, 2011 at 01:05:26PM +0530, Aneesh Kumar K.V wrote: > On Wed, 19 Oct 2011 18:09:15 -0400, "J. Bruce Fields" wrote: > > Maybe I'm dense, but that big if-else-if is still giving me a headache. > > > > The point is just to delay the ns_capable() check to avoid setting > > PF_SUPERPRIV in cases where we weren't before? > > > > How about putting using a helper function for the richacl check, and > > calling it from check_sticky instead? That makes the above: > > > > error = inode_permission(dir, mask | replace_mask); > > if (error && !richacl_may_delete(dir, inode, replace_mask)) > > return error; > > if (check_sticky(dir, inode, replace_mask)) > > return -EPERM; > > > > (As in the following--totally untested and possibly wrong.) > > > > Also: the comment before may_delete() needs updating. > > > > Thanks for the suggestion. That made the code simpler. Updated patch > below. Looks good to me if it passes your tests, thanks! Feel free to add Reviewed-by or Acked-by for "J. Bruce Fields" . --b. > > commit 3c92363ce2dee22aa174327c21726f8f02cbcd6e > Author: Andreas Gruenbacher > Date: Tue Oct 18 15:17:50 2011 +0530 > > vfs: Add delete child and delete self permission flags > > Normally, deleting a file requires write access to the parent directory. > Some permission models use a different permission on the parent > directory to indicate delete access. In addition, a process can have > per-file delete access even without delete access on the parent > directory. > > Introduce two new inode_permission() mask flags and use them in > may_delete() > > Acked-by: David Howells > Signed-off-by: Andreas Gruenbacher > Signed-off-by: Aneesh Kumar K.V > > diff --git a/fs/namei.c b/fs/namei.c > index f6184b8..044b6d1 100644 > --- a/fs/namei.c > +++ b/fs/namei.c > @@ -337,7 +337,7 @@ static inline int do_inode_permission(struct inode *inode, int mask) > * are used for other things. > * > * When checking for MAY_APPEND, MAY_CREATE_FILE, MAY_CREATE_DIR, > - * MAY_WRITE must also be set in @mask. > + * MAY_DELETE_CHILD, MAY_DELETE_SELF, MAY_WRITE must also be set in @mask. > */ > int inode_permission(struct inode *inode, int mask) > { > @@ -1835,11 +1835,25 @@ static int user_path_parent(int dfd, const char __user *path, > return error; > } > > + > +/* > + * We should have exec permission on directory and MAY_DELETE_SELF > + * on the object being deleted. > + */ > +static int richacl_may_selfdelete(struct inode *dir, > + struct inode *inode, int replace_mask) > +{ > + return (IS_RICHACL(inode) && > + (inode_permission(dir, MAY_EXEC | replace_mask) == 0) && > + (inode_permission(inode, MAY_DELETE_SELF) == 0)); > +} > + > /* > * It's inline, so penalty for filesystems that don't use sticky bit is > * minimal. > */ > -static inline int check_sticky(struct inode *dir, struct inode *inode) > +static inline int check_sticky(struct inode *dir, > + struct inode *inode, int replace_mask) > { > uid_t fsuid = current_fsuid(); > > @@ -1851,7 +1865,8 @@ static inline int check_sticky(struct inode *dir, struct inode *inode) > return 0; > if (dir->i_uid == fsuid) > return 0; > - > + if (richacl_may_selfdelete(dir, inode, replace_mask)) > + return 0; > other_userns: > return !ns_capable(inode_userns(inode), CAP_FOWNER); > } > @@ -1875,30 +1890,38 @@ other_userns: > * 10. We don't allow removal of NFS sillyrenamed files; it's handled by > * nfs_async_unlink(). > */ > -static int may_delete(struct inode *dir,struct dentry *victim,int isdir) > +static int may_delete(struct inode *dir, struct dentry *victim, > + int isdir, int replace) > { > - int error; > + int mask, replace_mask = 0, error; > + struct inode *inode = victim->d_inode; > > - if (!victim->d_inode) > + if (!inode) > return -ENOENT; > > BUG_ON(victim->d_parent->d_inode != dir); > audit_inode_child(victim, dir); > > - error = inode_permission(dir, MAY_WRITE | MAY_EXEC); > + mask = MAY_WRITE | MAY_EXEC | MAY_DELETE_CHILD; > + if (replace) > + replace_mask = S_ISDIR(inode->i_mode) ? > + MAY_CREATE_DIR : MAY_CREATE_FILE; > + error = inode_permission(dir, mask | replace_mask); > + if (error && richacl_may_selfdelete(dir, inode, replace_mask)) > + error = 0; > if (error) > return error; > if (IS_APPEND(dir)) > return -EPERM; > - if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)|| > - IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode)) > + if (check_sticky(dir, inode, replace_mask) || IS_APPEND(inode) || > + IS_IMMUTABLE(inode) || IS_SWAPFILE(inode)) > return -EPERM; > if (isdir) { > - if (!S_ISDIR(victim->d_inode->i_mode)) > + if (!S_ISDIR(inode->i_mode)) > return -ENOTDIR; > if (IS_ROOT(victim)) > return -EBUSY; > - } else if (S_ISDIR(victim->d_inode->i_mode)) > + } else if (S_ISDIR(inode->i_mode)) > return -EISDIR; > if (IS_DEADDIR(dir)) > return -ENOENT; > @@ -2605,7 +2628,7 @@ void dentry_unhash(struct dentry *dentry) > > int vfs_rmdir(struct inode *dir, struct dentry *dentry) > { > - int error = may_delete(dir, dentry, 1); > + int error = may_delete(dir, dentry, 1, 0); > > if (error) > return error; > @@ -2700,7 +2723,7 @@ SYSCALL_DEFINE1(rmdir, const char __user *, pathname) > > int vfs_unlink(struct inode *dir, struct dentry *dentry) > { > - int error = may_delete(dir, dentry, 0); > + int error = may_delete(dir, dentry, 0, 0); > > if (error) > return error; > @@ -3096,14 +3119,14 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry, > if (old_dentry->d_inode == new_dentry->d_inode) > return 0; > > - error = may_delete(old_dir, old_dentry, is_dir); > + error = may_delete(old_dir, old_dentry, is_dir, 0); > if (error) > return error; > > if (!new_dentry->d_inode) > error = may_create(new_dir, new_dentry, is_dir); > else > - error = may_delete(new_dir, new_dentry, is_dir); > + error = may_delete(new_dir, new_dentry, is_dir, 1); > if (error) > return error; > > diff --git a/include/linux/fs.h b/include/linux/fs.h > index 60361c6..ccece40 100644 > --- a/include/linux/fs.h > +++ b/include/linux/fs.h > @@ -69,6 +69,8 @@ struct inodes_stat_t { > #define MAY_NOT_BLOCK 0x00000080 > #define MAY_CREATE_FILE 0x00000100 > #define MAY_CREATE_DIR 0x00000200 > +#define MAY_DELETE_CHILD 0x00000400 > +#define MAY_DELETE_SELF 0x00000800 > > /* > * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond >