Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.5 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 86B08C43381 for ; Sun, 3 Mar 2019 22:21:35 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4920B20815 for ; Sun, 3 Mar 2019 22:21:35 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1725983AbfCCWSt (ORCPT ); Sun, 3 Mar 2019 17:18:49 -0500 Received: from ipmail06.adl2.internode.on.net ([150.101.137.129]:29529 "EHLO ipmail06.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725933AbfCCWSt (ORCPT ); Sun, 3 Mar 2019 17:18:49 -0500 Received: from ppp59-167-129-252.static.internode.on.net (HELO dastard) ([59.167.129.252]) by ipmail06.adl2.internode.on.net with ESMTP; 04 Mar 2019 08:48:44 +1030 Received: from dave by dastard with local (Exim 4.80) (envelope-from ) id 1h0ZRU-0005WF-2k; Mon, 04 Mar 2019 09:18:44 +1100 Date: Mon, 4 Mar 2019 09:18:44 +1100 From: Dave Chinner To: Wang Shilong Cc: linux-fsdevel@vger.kernel.org, linux-ext4@vger.kernel.org, linux-xfs@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, lixi@ddn.com, adilger@dilger.ca, Wang Shilong Subject: Re: [PATCH 4/8] xfs: support project ID in xfs_setattr() Message-ID: <20190303221844.GR23020@dastard> References: <1551449141-7884-1-git-send-email-wshilong1991@gmail.com> <1551449141-7884-5-git-send-email-wshilong1991@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1551449141-7884-5-git-send-email-wshilong1991@gmail.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-ext4-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-ext4@vger.kernel.org On Fri, Mar 01, 2019 at 11:05:37PM +0900, Wang Shilong wrote: > From: Wang Shilong > > From: Wang Shilong > > Signed-off-by: Wang Shilong > --- > fs/xfs/xfs_iops.c | 51 +++++++++++++++++++++++++++++++++++++--------- > fs/xfs/xfs_linux.h | 10 +++++++++ > 2 files changed, 51 insertions(+), 10 deletions(-) > > diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c > index f48ffd7a8d3e..c10466fe6ed4 100644 > --- a/fs/xfs/xfs_iops.c > +++ b/fs/xfs/xfs_iops.c > @@ -589,7 +589,8 @@ xfs_vn_change_ok( > struct dentry *dentry, > struct iattr *iattr) > { > - struct xfs_mount *mp = XFS_I(d_inode(dentry))->i_mount; > + struct xfs_inode *ip = XFS_I(d_inode(dentry)); > + struct xfs_mount *mp = ip->i_mount; > > if (mp->m_flags & XFS_MOUNT_RDONLY) > return -EROFS; > @@ -597,6 +598,13 @@ xfs_vn_change_ok( > if (XFS_FORCED_SHUTDOWN(mp)) > return -EIO; > > + if ((iattr->ia_valid & ATTR_PROJID) && > + current_user_ns() != &init_user_ns) { > + if (!projid_eq(xfs_projid_to_kprojid(xfs_get_projid(ip)), > + iattr->ia_projid)) > + return -EPERM; > + } See previous comments about this. > + > return setattr_prepare(dentry, iattr); > } > > @@ -619,8 +627,10 @@ xfs_setattr_nonsize( > int error; > kuid_t uid = GLOBAL_ROOT_UID, iuid = GLOBAL_ROOT_UID; > kgid_t gid = GLOBAL_ROOT_GID, igid = GLOBAL_ROOT_GID; > - struct xfs_dquot *udqp = NULL, *gdqp = NULL; > + kprojid_t projid, iprojid; So, uninitialised, unlink the uid/gids. These shoul dbe GLOBAL_ROOT_PROJID, which probably should be: #define GLOBAL_ROOT_PROJID KPROJIDT_INIT(0) to match these: #define XFS_PROJID_DEFAULT 0 #define F2FS_DEF_PROJID 0 /* default project ID */ #define EXT4_DEF_PROJID 0 > + struct xfs_dquot *udqp = NULL, *gdqp = NULL, *pdqp = NULL; > struct xfs_dquot *olddquot1 = NULL, *olddquot2 = NULL; > + struct xfs_dquot *olddquot3 = NULL; > > ASSERT((mask & ATTR_SIZE) == 0); > > @@ -632,7 +642,7 @@ xfs_setattr_nonsize( > * If the IDs do change before we take the ilock, we're covered > * because the i_*dquot fields will get updated anyway. > */ > - if (XFS_IS_QUOTA_ON(mp) && (mask & (ATTR_UID|ATTR_GID))) { > + if (XFS_IS_QUOTA_ON(mp) && (mask & (ATTR_UID|ATTR_GID|ATTR_PROJID))) { > uint qflags = 0; > > if ((mask & ATTR_UID) && XFS_IS_UQUOTA_ON(mp)) { > @@ -647,18 +657,25 @@ xfs_setattr_nonsize( > } else { > gid = inode->i_gid; > } > + if ((mask & ATTR_PROJID) && XFS_IS_PQUOTA_ON(mp)) { > + projid = iattr->ia_projid; > + qflags |= XFS_QMOPT_PQUOTA; > + } else { > + projid = xfs_projid_to_kprojid(xfs_get_projid(ip)); > + } Hmmm. why would we convert the XFS on-disk project ID to a kernel representation, only to immediately: > > /* > - * We take a reference when we initialize udqp and gdqp, > + * We take a reference when we initialize udqp,gdqp and pdqp, > * so it is important that we never blindly double trip on > * the same variable. See xfs_create() for an example. > */ > ASSERT(udqp == NULL); > ASSERT(gdqp == NULL); > + ASSERT(pdqp == NULL); > error = xfs_qm_vop_dqalloc(ip, xfs_kuid_to_uid(uid), > xfs_kgid_to_gid(gid), > - xfs_get_projid(ip), > - qflags, &udqp, &gdqp, NULL); > + xfs_kprojid_to_projid(projid), > + qflags, &udqp, &gdqp, &pdqp); Convert it back to the XFS on disk representation? Perhaps: if ((mask & ATTR_PROJID) && XFS_IS_PQUOTA_ON(mp)) { projid = xfs_kprojid_to_projid(iattr->ia_projid); qflags |= XFS_QMOPT_PQUOTA; } else { projid = xfs_get_projid(ip); } Unless, of course, we promote the the project ID into the struct inode so it matches uid and gid. This code is really telling me that we should be promoting it before we make thse changes... > @@ -673,7 +690,7 @@ xfs_setattr_nonsize( > /* > * Change file ownership. Must be the owner or privileged. > */ > - if (mask & (ATTR_UID|ATTR_GID)) { > + if (mask & (ATTR_UID|ATTR_GID|ATTR_PROJID)) { > /* > * These IDs could have changed since we last looked at them. > * But, we're assured that if the ownership did change > @@ -682,8 +699,10 @@ xfs_setattr_nonsize( > */ > iuid = inode->i_uid; > igid = inode->i_gid; > + iprojid = xfs_projid_to_kprojid(xfs_get_projid(ip)); > gid = (mask & ATTR_GID) ? iattr->ia_gid : igid; > uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid; > + projid = (mask & ATTR_PROJID) ? iattr->ia_projid : iprojid; > > /* > * Do a quota reservation only if uid/gid is actually > @@ -691,10 +710,11 @@ xfs_setattr_nonsize( > */ > if (XFS_IS_QUOTA_RUNNING(mp) && > ((XFS_IS_UQUOTA_ON(mp) && !uid_eq(iuid, uid)) || > - (XFS_IS_GQUOTA_ON(mp) && !gid_eq(igid, gid)))) { > + (XFS_IS_GQUOTA_ON(mp) && !gid_eq(igid, gid)) || > + (XFS_IS_PQUOTA_ON(mp) && !projid_eq(iprojid, projid)))) { > ASSERT(tp); > error = xfs_qm_vop_chown_reserve(tp, ip, udqp, gdqp, > - NULL, capable(CAP_FOWNER) ? > + pdqp, capable(CAP_FOWNER) ? > XFS_QMOPT_FORCE_RES : 0); > if (error) /* out of quota */ > goto out_cancel; > @@ -704,7 +724,7 @@ xfs_setattr_nonsize( > /* > * Change file ownership. Must be the owner or privileged. > */ > - if (mask & (ATTR_UID|ATTR_GID)) { > + if (mask & (ATTR_UID|ATTR_GID|ATTR_PROJID)) { > /* > * CAP_FSETID overrides the following restrictions: > * > @@ -741,6 +761,15 @@ xfs_setattr_nonsize( > ip->i_d.di_gid = xfs_kgid_to_gid(gid); > inode->i_gid = gid; > } > + if (!projid_eq(iprojid, projid)) { > + if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_PQUOTA_ON(mp)) { > + ASSERT(mask & ATTR_PROJID); > + ASSERT(pdqp); > + olddquot3 = xfs_qm_vop_chown(tp, ip, > + &ip->i_pdquot, pdqp); > + } > + xfs_set_projid(ip, xfs_kprojid_to_projid(projid)); > + } > } Ok, this adds another set of boilerplate code here. This is starting to need some factoring work. Not needed for this patchset, though. Cheers, Dave. -- Dave Chinner david@fromorbit.com