xfs stable candidate patches for 5.4.y (from v5.11 & v5.12)
cover xfs stable candidate patches for 5.4.y (from v5.11 & v5.12)
Hi Greg,
This 5.4.y backport series contains XFS fixes from v5.11 & v5.12. The patchset has been acked by Darrick.
As a side note, where applicable, patches have been cherry picked from 5.10.y rather than from v5.11/v5.12.
Brian Foster (2):
xfs: consider shutdown in bmapbt cursor delete assert
xfs: don't reuse busy extents on extent trim
Christoph Hellwig (10):
xfs: merge the projid fields in struct xfs_icdinode
xfs: ensure that the inode uid/gid match values match the icdinode
ones
xfs: remove the icdinode di_uid/di_gid members
xfs: remove the kuid/kgid conversion wrappers
xfs: add a new xfs_sb_version_has_v3inode helper
xfs: only check the superblock version for dinode size calculation
xfs: simplify di_flags2 inheritance in xfs_ialloc
xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
xfs: remove the di_version field from struct icdinode
xfs: fix up non-directory creation in SGID directories
Darrick J. Wong (3):
xfs: report corruption only as a regular error
xfs: shut down the filesystem if we screw up quota reservation
xfs: force log and push AIL to clear pinned inodes when aborting mount
Jeffrey Mitchell (1):
xfs: set inode size after creating symlink
Kaixu Xia (1):
xfs: show the proper user quota options
fs/xfs/libxfs/xfs_attr_leaf.c | 5 +- fs/xfs/libxfs/xfs_bmap.c | 10 ++-- fs/xfs/libxfs/xfs_btree.c | 30 +++++------- fs/xfs/libxfs/xfs_format.h | 33 ++++++++++--- fs/xfs/libxfs/xfs_ialloc.c | 6 +-- fs/xfs/libxfs/xfs_inode_buf.c | 54 +++++++------------- fs/xfs/libxfs/xfs_inode_buf.h | 8 +-- fs/xfs/libxfs/xfs_inode_fork.c | 2 +- fs/xfs/libxfs/xfs_inode_fork.h | 9 +--- fs/xfs/libxfs/xfs_log_format.h | 10 ++-- fs/xfs/libxfs/xfs_trans_resv.c | 2 +- fs/xfs/xfs_acl.c | 12 +++-- fs/xfs/xfs_bmap_util.c | 16 +++--- fs/xfs/xfs_buf_item.c | 2 +- fs/xfs/xfs_dquot.c | 6 +-- fs/xfs/xfs_error.c | 2 +- fs/xfs/xfs_extent_busy.c | 14 ------ fs/xfs/xfs_icache.c | 8 ++- fs/xfs/xfs_inode.c | 61 ++++++++--------------- fs/xfs/xfs_inode.h | 21 +------- fs/xfs/xfs_inode_item.c | 20 ++++---- fs/xfs/xfs_ioctl.c | 22 ++++----- fs/xfs/xfs_iops.c | 11 +---- fs/xfs/xfs_itable.c | 8 +-- fs/xfs/xfs_linux.h | 32 +++--------- fs/xfs/xfs_log_recover.c | 6 +-- fs/xfs/xfs_mount.c | 90 +++++++++++++++++----------------- fs/xfs/xfs_qm.c | 43 +++++++++------- fs/xfs/xfs_qm_bhv.c | 2 +- fs/xfs/xfs_quota.h | 4 +- fs/xfs/xfs_super.c | 10 ++-- fs/xfs/xfs_symlink.c | 7 ++- fs/xfs/xfs_trans_dquot.c | 16 ++++-- 33 files changed, 248 insertions(+), 334 deletions(-)
Signature · 1 line
2.39.1
PATCH 1/17 xfs: show the proper user quota options
From: Kaixu Xia <[email protected]>
commit 237d7887ae723af7d978e8b9a385fdff416f357b upstream.
The quota option 'usrquota' should be shown if both the XFS_UQUOTA_ACCT and XFS_UQUOTA_ENFD flags are set. The option 'uqnoenforce' should be shown when only the XFS_UQUOTA_ACCT flag is set. The current code logic seems wrong, Fix it and show proper options.
--- fs/xfs/xfs_super.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
fs/xfs/xfs_super.c +6 -4
@@ -490,10 +490,12 @@ xfs_showargs(seq_printf (m,",swidth=%d" , (int )XFS_FSB_TO_BB (mp, mp->m_swidth)); -if (mp->m_qflags & (XFS_UQUOTA_ACCT |XFS_UQUOTA_ENFD )) -seq_puts (m,",usrquota" ); -else if (mp->m_qflags &XFS_UQUOTA_ACCT ) -seq_puts (m,",uqnoenforce" ); +if (mp->m_qflags &XFS_UQUOTA_ACCT ) { +if (mp->m_qflags &XFS_UQUOTA_ENFD ) +seq_puts (m,",usrquota" ); +else +seq_puts (m,",uqnoenforce" ); + }if (mp->m_qflags &XFS_PQUOTA_ACCT ) {if (mp->m_qflags &XFS_PQUOTA_ENFD )
Signature · 1 line
2.39.1
PATCH 2/17 xfs: merge the projid fields in struct xfs_icdinode
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit de7a866fd41b227b0aa6e9cbeb0dae221c12f542 upstream.
There is no point in splitting the fields like this in an purely in-memory structure.
--- fs/xfs/libxfs/xfs_inode_buf.c | 11 +++++------ fs/xfs/libxfs/xfs_inode_buf.h | 3 +-- fs/xfs/xfs_dquot.c | 2 +- fs/xfs/xfs_icache.c | 4 ++-- fs/xfs/xfs_inode.c | 6 +++--- fs/xfs/xfs_inode.h | 21 +-------------------- fs/xfs/xfs_inode_item.c | 4 ++-- fs/xfs/xfs_ioctl.c | 8 ++++---- fs/xfs/xfs_iops.c | 2 +- fs/xfs/xfs_itable.c | 2 +- fs/xfs/xfs_qm.c | 8 ++++---- fs/xfs/xfs_qm_bhv.c | 2 +- 12 files changed, 26 insertions(+), 47 deletions(-)
fs/xfs/libxfs/xfs_inode_buf.c +5 -6
@@ -213,13 +213,12 @@ xfs_inode_from_disk( to->di_version = from->di_version;if (to->di_version ==1 ) {set_nlink (inode,be16_to_cpu (from->di_onlink)); - to->di_projid_lo =0 ; - to->di_projid_hi =0 ; + to->di_projid =0 ; to->di_version =2 ; }else {set_nlink (inode,be32_to_cpu (from->di_nlink)); - to->di_projid_lo =be16_to_cpu (from->di_projid_lo); - to->di_projid_hi =be16_to_cpu (from->di_projid_hi); + to->di_projid = (prid_t )be16_to_cpu (from->di_projid_hi) <<16 | +be16_to_cpu (from->di_projid_lo); } to->di_format = from->di_format;
@@ -279,8 +278,8 @@ xfs_inode_to_disk( to->di_format = from->di_format; to->di_uid =cpu_to_be32 (from->di_uid); to->di_gid =cpu_to_be32 (from->di_gid); - to->di_projid_lo =cpu_to_be16 (from->di_projid_lo); - to->di_projid_hi =cpu_to_be16 (from->di_projid_hi); + to->di_projid_lo =cpu_to_be16 (from->di_projid &0xffff ); + to->di_projid_hi =cpu_to_be16 (from->di_projid >>16 );memset (to->di_pad,0 ,sizeof (to->di_pad)); to->di_atime.t_sec =cpu_to_be32 (inode->i_atime.tv_sec);
fs/xfs/libxfs/xfs_inode_buf.h +1 -2
@@ -21,8 +21,7 @@ struct xfs_icdinode {uint16_t di_flushiter;/* incremented on flush */ uint32_t di_uid;/* owner's user id */ uint32_t di_gid;/* owner's group id */ -uint16_t di_projid_lo;/* lower part of owner's project id */ -uint16_t di_projid_hi;/* higher part of owner's project id */ +uint32_t di_projid;/* owner's project id */ xfs_fsize_t di_size;/* number of bytes in file */ xfs_rfsblock_t di_nblocks;/* # of direct & btree blocks used */ xfs_extlen_t di_extsize;/* basic/minimum extent size for file */
fs/xfs/xfs_dquot.c +1 -1
@@ -863,7 +863,7 @@ xfs_qm_id_for_quotatype(case XFS_DQ_GROUP :return ip->i_d.di_gid;case XFS_DQ_PROJ : -return xfs_get_projid (ip); +return ip->i_d.di_projid; }ASSERT (0 );return 0 ;
fs/xfs/xfs_icache.c +2 -2
@@ -1430,7 +1430,7 @@ xfs_inode_match_id(return 0 ;if ((eofb->eof_flags &XFS_EOF_FLAGS_PRID ) && -xfs_get_projid (ip) != eofb->eof_prid) + ip->i_d.di_projid != eofb->eof_prid)return 0 ;return 1 ;
@@ -1454,7 +1454,7 @@ xfs_inode_match_id_union(return 1 ;if ((eofb->eof_flags &XFS_EOF_FLAGS_PRID ) && -xfs_get_projid (ip) == eofb->eof_prid) + ip->i_d.di_projid == eofb->eof_prid)return 1 ;return 0 ;
fs/xfs/xfs_inode.c +3 -3
@@ -809,7 +809,7 @@ xfs_ialloc( ip->i_d.di_uid =xfs_kuid_to_uid (current_fsuid ()); ip->i_d.di_gid =xfs_kgid_to_gid (current_fsgid ()); inode->i_rdev = rdev; -xfs_set_projid (ip, prid); + ip->i_d.di_projid = prid;if (pip &&XFS_INHERIT_GID (pip)) { ip->i_d.di_gid = pip->i_d.di_gid;
@@ -1418,7 +1418,7 @@ xfs_link(* the tree quota mechanism could be circumvented. */ if (unlikely ((tdp->i_d.di_flags &XFS_DIFLAG_PROJINHERIT ) && - (xfs_get_projid (tdp) !=xfs_get_projid (sip)))) { + tdp->i_d.di_projid != sip->i_d.di_projid)) { error = -EXDEV ; goto error_return; }
@@ -3299,7 +3299,7 @@ xfs_rename(* tree quota mechanism would be circumvented. */ if (unlikely ((target_dp->i_d.di_flags &XFS_DIFLAG_PROJINHERIT ) && - (xfs_get_projid (target_dp) !=xfs_get_projid (src_ip)))) { + target_dp->i_d.di_projid != src_ip->i_d.di_projid)) { error = -EXDEV ; goto out_trans_cancel; }
fs/xfs/xfs_inode.h +1 -20
@@ -177,30 +177,11 @@ xfs_iflags_test_and_set(xfs_inode_t *ip, unsigned short flags)return ret; } -/* -* Project quota id helpers (previously projid was 16bit only -* and using two 16bit values to hold new 32bit projid was chosen -* to retain compatibility with "old" filesystems). -*/ -static inline prid_t -xfs_get_projid (struct xfs_inode *ip) -{ -return (prid_t )ip->i_d.di_projid_hi <<16 | ip->i_d.di_projid_lo; -} - -static inline void -xfs_set_projid (struct xfs_inode *ip, -prid_t projid) -{ - ip->i_d.di_projid_hi = (uint16_t ) (projid >>16 ); - ip->i_d.di_projid_lo = (uint16_t ) (projid &0xffff ); -} -static inline prid_t xfs_get_initial_prid (struct xfs_inode *dp) {if (dp->i_d.di_flags &XFS_DIFLAG_PROJINHERIT ) -return xfs_get_projid (dp); +return dp->i_d.di_projid;return XFS_PROJID_DEFAULT ; }
fs/xfs/xfs_inode_item.c +2 -2
@@ -310,8 +310,8 @@ xfs_inode_to_log_dinode( to->di_format = from->di_format; to->di_uid = from->di_uid; to->di_gid = from->di_gid; - to->di_projid_lo = from->di_projid_lo; - to->di_projid_hi = from->di_projid_hi; + to->di_projid_lo = from->di_projid &0xffff ; + to->di_projid_hi = from->di_projid >>16 ;memset (to->di_pad,0 ,sizeof (to->di_pad));memset (to->di_pad3,0 ,sizeof (to->di_pad3));
fs/xfs/xfs_ioctl.c +4 -4
@@ -1144,7 +1144,7 @@ xfs_fill_fsxattr( fa->fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog; fa->fsx_cowextsize = ip->i_d.di_cowextsize << ip->i_mount->m_sb.sb_blocklog; - fa->fsx_projid =xfs_get_projid (ip); + fa->fsx_projid = ip->i_d.di_projid;if (attr) {if (ip->i_afp) {
@@ -1597,7 +1597,7 @@ xfs_ioctl_setattr( }if (XFS_IS_QUOTA_RUNNING (mp) &&XFS_IS_PQUOTA_ON (mp) && -xfs_get_projid (ip) != fa->fsx_projid) { + ip->i_d.di_projid != fa->fsx_projid) { code =xfs_qm_vop_chown_reserve (tp, ip, udqp,NULL , pdqp,capable (CAP_FOWNER ) ?XFS_QMOPT_FORCE_RES :0 );if (code)/* out of quota */
@@ -1634,13 +1634,13 @@ xfs_ioctl_setattr(VFS_I (ip)->i_mode &= ~(S_ISUID |S_ISGID );/* Change the ownerships and register project quota modifications */ -if (xfs_get_projid (ip) != fa->fsx_projid) { +if (ip->i_d.di_projid != fa->fsx_projid) {if (XFS_IS_QUOTA_RUNNING (mp) &&XFS_IS_PQUOTA_ON (mp)) { olddquot =xfs_qm_vop_chown (tp, ip, &ip->i_pdquot, pdqp); }ASSERT (ip->i_d.di_version >1 ); -xfs_set_projid (ip, fa->fsx_projid); + ip->i_d.di_projid = fa->fsx_projid; }/*
fs/xfs/xfs_iops.c +1 -1
@@ -668,7 +668,7 @@ xfs_setattr_nonsize(ASSERT (gdqp ==NULL ); error =xfs_qm_vop_dqalloc (ip,xfs_kuid_to_uid (uid),xfs_kgid_to_gid (gid), -xfs_get_projid (ip), + ip->i_d.di_projid, qflags, &udqp, &gdqp,NULL );if (error)return error;
fs/xfs/xfs_itable.c +1 -1
@@ -84,7 +84,7 @@ xfs_bulkstat_one_int(/* xfs_iget returns the following without needing * further change. */ - buf->bs_projectid =xfs_get_projid (ip); + buf->bs_projectid = ip->i_d.di_projid; buf->bs_ino = ino; buf->bs_uid = dic->di_uid; buf->bs_gid = dic->di_gid;
fs/xfs/xfs_qm.c +4 -4
@@ -347,7 +347,7 @@ xfs_qm_dqattach_locked( }if (XFS_IS_PQUOTA_ON (mp) && !ip->i_pdquot) { - error =xfs_qm_dqattach_one (ip,xfs_get_projid (ip),XFS_DQ_PROJ , + error =xfs_qm_dqattach_one (ip, ip->i_d.di_projid,XFS_DQ_PROJ , doalloc, &ip->i_pdquot);if (error) goto done;
@@ -1715,7 +1715,7 @@ xfs_qm_vop_dqalloc( } }if ((flags &XFS_QMOPT_PQUOTA ) &&XFS_IS_PQUOTA_ON (mp)) { -if (xfs_get_projid (ip) != prid) { +if (ip->i_d.di_projid != prid) {xfs_iunlock (ip, lockflags); error =xfs_qm_dqget (mp, (xfs_dqid_t )prid,XFS_DQ_PROJ , true, &pq);
@@ -1849,7 +1849,7 @@ xfs_qm_vop_chown_reserve( }if (XFS_IS_PQUOTA_ON (ip->i_mount) && pdqp && -xfs_get_projid (ip) !=be32_to_cpu (pdqp->q_core.d_id)) { + ip->i_d.di_projid !=be32_to_cpu (pdqp->q_core.d_id)) { prjflags =XFS_QMOPT_ENOSPC ; pdq_delblks = pdqp;if (delblks) {
@@ -1950,7 +1950,7 @@ xfs_qm_vop_create_dqattach( }if (pdqp &&XFS_IS_PQUOTA_ON (mp)) {ASSERT (ip->i_pdquot ==NULL ); -ASSERT (xfs_get_projid (ip) ==be32_to_cpu (pdqp->q_core.d_id)); +ASSERT (ip->i_d.di_projid ==be32_to_cpu (pdqp->q_core.d_id)); ip->i_pdquot =xfs_qm_dqhold (pdqp);xfs_trans_mod_dquot (tp, pdqp,XFS_TRANS_DQ_ICOUNT ,1 );
fs/xfs/xfs_qm_bhv.c +1 -1
@@ -60,7 +60,7 @@ xfs_qm_statvfs(struct xfs_mount *mp = ip->i_mount;struct xfs_dquot *dqp; -if (!xfs_qm_dqget (mp,xfs_get_projid (ip),XFS_DQ_PROJ , false, &dqp)) { +if (!xfs_qm_dqget (mp, ip->i_d.di_projid,XFS_DQ_PROJ , false, &dqp)) {xfs_fill_statvfs_from_dquot (statp, dqp);xfs_qm_dqput (dqp); }
Signature · 1 line
2.39.1
PATCH 3/17 xfs: ensure that the inode uid/gid match values match the icdinode ones
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 3d8f2821502d0b60bac2789d0bea951fda61de0c upstream.
Instead of only synchronizing the uid/gid values in xfs_setup_inode, ensure that they always match to prepare for removing the icdinode fields.
--- fs/xfs/libxfs/xfs_inode_buf.c | 2 ++ fs/xfs/xfs_icache.c | 4 ++++ fs/xfs/xfs_inode.c | 8 ++++++-- fs/xfs/xfs_iops.c | 3 --- 4 files changed, 12 insertions(+), 5 deletions(-)
fs/xfs/libxfs/xfs_inode_buf.c +2 -0
@@ -223,7 +223,9 @@ xfs_inode_from_disk( to->di_format = from->di_format; to->di_uid =be32_to_cpu (from->di_uid); + inode->i_uid =xfs_uid_to_kuid (to->di_uid); to->di_gid =be32_to_cpu (from->di_gid); + inode->i_gid =xfs_gid_to_kgid (to->di_gid); to->di_flushiter =be16_to_cpu (from->di_flushiter);/*
fs/xfs/xfs_icache.c +4 -0
@@ -289,6 +289,8 @@ xfs_reinit_inode(uint64_t version =inode_peek_iversion (inode);umode_t mode = inode->i_mode;dev_t dev = inode->i_rdev; +kuid_t uid = inode->i_uid; +kgid_t gid = inode->i_gid; error =inode_init_always (mp->m_super, inode);
@@ -297,6 +299,8 @@ xfs_reinit_inode(inode_set_iversion_queried (inode, version); inode->i_mode = mode; inode->i_rdev = dev; + inode->i_uid = uid; + inode->i_gid = gid;return error; }
fs/xfs/xfs_inode.c +6 -2
@@ -806,15 +806,19 @@ xfs_ialloc( inode->i_mode = mode;set_nlink (inode, nlink); - ip->i_d.di_uid =xfs_kuid_to_uid (current_fsuid ()); - ip->i_d.di_gid =xfs_kgid_to_gid (current_fsgid ()); + inode->i_uid =current_fsuid (); + ip->i_d.di_uid =xfs_kuid_to_uid (inode->i_uid); inode->i_rdev = rdev; ip->i_d.di_projid = prid;if (pip &&XFS_INHERIT_GID (pip)) { + inode->i_gid =VFS_I (pip)->i_gid; ip->i_d.di_gid = pip->i_d.di_gid;if ((VFS_I (pip)->i_mode &S_ISGID ) &&S_ISDIR (mode)) inode->i_mode |=S_ISGID ; + }else { + inode->i_gid =current_fsgid (); + ip->i_d.di_gid =xfs_kgid_to_gid (inode->i_gid); }/*
fs/xfs/xfs_iops.c +0 -3
@@ -1288,9 +1288,6 @@ xfs_setup_inode(/* make the inode look hashed for the writeback code */ inode_fake_hash (inode); - inode->i_uid =xfs_uid_to_kuid (ip->i_d.di_uid); - inode->i_gid =xfs_gid_to_kgid (ip->i_d.di_gid); -i_size_write (inode, ip->i_d.di_size);xfs_diflags_to_iflags (inode, ip);
Signature · 1 line
2.39.1
PATCH 4/17 xfs: remove the icdinode di_uid/di_gid members
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 542951592c99ff7b15c050954c051dd6dd6c0f97 upstream.
Use the Linux inode i_uid/i_gid members everywhere and just convert from/to the scalar value when reading or writing the on-disk inode.
--- fs/xfs/libxfs/xfs_inode_buf.c | 10 ++++----- fs/xfs/libxfs/xfs_inode_buf.h | 2 -- fs/xfs/xfs_dquot.c | 4 ++-- fs/xfs/xfs_inode.c | 14 ++++-------- fs/xfs/xfs_inode_item.c | 4 ++-- fs/xfs/xfs_ioctl.c | 6 +++--- fs/xfs/xfs_iops.c | 6 +----- fs/xfs/xfs_itable.c | 4 ++-- fs/xfs/xfs_qm.c | 40 ++++++++++++++++++++++------------- fs/xfs/xfs_quota.h | 4 ++-- fs/xfs/xfs_symlink.c | 4 +--- 11 files changed, 46 insertions(+), 52 deletions(-)
fs/xfs/libxfs/xfs_inode_buf.c +4 -6
@@ -222,10 +222,8 @@ xfs_inode_from_disk( } to->di_format = from->di_format; - to->di_uid =be32_to_cpu (from->di_uid); - inode->i_uid =xfs_uid_to_kuid (to->di_uid); - to->di_gid =be32_to_cpu (from->di_gid); - inode->i_gid =xfs_gid_to_kgid (to->di_gid); + inode->i_uid =xfs_uid_to_kuid (be32_to_cpu (from->di_uid)); + inode->i_gid =xfs_gid_to_kgid (be32_to_cpu (from->di_gid)); to->di_flushiter =be16_to_cpu (from->di_flushiter);/*
@@ -278,8 +276,8 @@ xfs_inode_to_disk( to->di_version = from->di_version; to->di_format = from->di_format; - to->di_uid =cpu_to_be32 (from->di_uid); - to->di_gid =cpu_to_be32 (from->di_gid); + to->di_uid =cpu_to_be32 (xfs_kuid_to_uid (inode->i_uid)); + to->di_gid =cpu_to_be32 (xfs_kgid_to_gid (inode->i_gid)); to->di_projid_lo =cpu_to_be16 (from->di_projid &0xffff ); to->di_projid_hi =cpu_to_be16 (from->di_projid >>16 );
fs/xfs/libxfs/xfs_inode_buf.h +0 -2
@@ -19,8 +19,6 @@ struct xfs_icdinode {int8_t di_version;/* inode version */ int8_t di_format;/* format of di_c data */ uint16_t di_flushiter;/* incremented on flush */ -uint32_t di_uid;/* owner's user id */ -uint32_t di_gid;/* owner's group id */ uint32_t di_projid;/* owner's project id */ xfs_fsize_t di_size;/* number of bytes in file */ xfs_rfsblock_t di_nblocks;/* # of direct & btree blocks used */
fs/xfs/xfs_dquot.c +2 -2
@@ -859,9 +859,9 @@ xfs_qm_id_for_quotatype( {switch (type) {case XFS_DQ_USER : -return ip->i_d.di_uid; +return xfs_kuid_to_uid (VFS_I (ip)->i_uid);case XFS_DQ_GROUP : -return ip->i_d.di_gid; +return xfs_kgid_to_gid (VFS_I (ip)->i_gid);case XFS_DQ_PROJ :return ip->i_d.di_projid; }
fs/xfs/xfs_inode.c +4 -10
@@ -807,18 +807,15 @@ xfs_ialloc( inode->i_mode = mode;set_nlink (inode, nlink); inode->i_uid =current_fsuid (); - ip->i_d.di_uid =xfs_kuid_to_uid (inode->i_uid); inode->i_rdev = rdev; ip->i_d.di_projid = prid;if (pip &&XFS_INHERIT_GID (pip)) { inode->i_gid =VFS_I (pip)->i_gid; - ip->i_d.di_gid = pip->i_d.di_gid;if ((VFS_I (pip)->i_mode &S_ISGID ) &&S_ISDIR (mode)) inode->i_mode |=S_ISGID ; }else { inode->i_gid =current_fsgid (); - ip->i_d.di_gid =xfs_kgid_to_gid (inode->i_gid); }/*
@@ -826,9 +823,8 @@ xfs_ialloc(* ID or one of the supplementary group IDs, the S_ISGID bit is cleared * (and only if the irix_sgid_inherit compatibility variable is set). */ -if ((irix_sgid_inherit) && - (inode->i_mode &S_ISGID ) && - (!in_group_p (xfs_gid_to_kgid (ip->i_d.di_gid)))) +if (irix_sgid_inherit && + (inode->i_mode &S_ISGID ) && !in_group_p (inode->i_gid)) inode->i_mode &= ~S_ISGID ; ip->i_d.di_size =0 ;
@@ -1157,8 +1153,7 @@ xfs_create(/* * Make sure that we have allocated dquot(s) on disk. */ - error =xfs_qm_vop_dqalloc (dp,xfs_kuid_to_uid (current_fsuid ()), -xfs_kgid_to_gid (current_fsgid ()), prid, + error =xfs_qm_vop_dqalloc (dp,current_fsuid (),current_fsgid (), prid,XFS_QMOPT_QUOTALL |XFS_QMOPT_INHERIT , &udqp, &gdqp, &pdqp);if (error)
@@ -1308,8 +1303,7 @@ xfs_create_tmpfile(/* * Make sure that we have allocated dquot(s) on disk. */ - error =xfs_qm_vop_dqalloc (dp,xfs_kuid_to_uid (current_fsuid ()), -xfs_kgid_to_gid (current_fsgid ()), prid, + error =xfs_qm_vop_dqalloc (dp,current_fsuid (),current_fsgid (), prid,XFS_QMOPT_QUOTALL |XFS_QMOPT_INHERIT , &udqp, &gdqp, &pdqp);if (error)
fs/xfs/xfs_inode_item.c +2 -2
@@ -308,8 +308,8 @@ xfs_inode_to_log_dinode( to->di_version = from->di_version; to->di_format = from->di_format; - to->di_uid = from->di_uid; - to->di_gid = from->di_gid; + to->di_uid =xfs_kuid_to_uid (inode->i_uid); + to->di_gid =xfs_kgid_to_gid (inode->i_gid); to->di_projid_lo = from->di_projid &0xffff ; to->di_projid_hi = from->di_projid >>16 ;
fs/xfs/xfs_ioctl.c +3 -3
@@ -1572,9 +1572,9 @@ xfs_ioctl_setattr(* because the i_*dquot fields will get updated anyway. */ if (XFS_IS_QUOTA_ON (mp)) { - code =xfs_qm_vop_dqalloc (ip, ip->i_d.di_uid, - ip->i_d.di_gid, fa->fsx_projid, -XFS_QMOPT_PQUOTA , &udqp,NULL , &pdqp); + code =xfs_qm_vop_dqalloc (ip,VFS_I (ip)->i_uid, +VFS_I (ip)->i_gid, fa->fsx_projid, +XFS_QMOPT_PQUOTA , &udqp,NULL , &pdqp);if (code)return code; }
fs/xfs/xfs_iops.c +1 -5
@@ -666,9 +666,7 @@ xfs_setattr_nonsize(*/ ASSERT (udqp ==NULL );ASSERT (gdqp ==NULL ); - error =xfs_qm_vop_dqalloc (ip,xfs_kuid_to_uid (uid), -xfs_kgid_to_gid (gid), - ip->i_d.di_projid, + error =xfs_qm_vop_dqalloc (ip, uid, gid, ip->i_d.di_projid, qflags, &udqp, &gdqp,NULL );if (error)return error;
@@ -737,7 +735,6 @@ xfs_setattr_nonsize( olddquot1 =xfs_qm_vop_chown (tp, ip, &ip->i_udquot, udqp); } - ip->i_d.di_uid =xfs_kuid_to_uid (uid); inode->i_uid = uid; }if (!gid_eq (igid, gid)) {
@@ -749,7 +746,6 @@ xfs_setattr_nonsize( olddquot2 =xfs_qm_vop_chown (tp, ip, &ip->i_gdquot, gdqp); } - ip->i_d.di_gid =xfs_kgid_to_gid (gid); inode->i_gid = gid; } }
fs/xfs/xfs_itable.c +2 -2
@@ -86,8 +86,8 @@ xfs_bulkstat_one_int(*/ buf->bs_projectid = ip->i_d.di_projid; buf->bs_ino = ino; - buf->bs_uid = dic->di_uid; - buf->bs_gid = dic->di_gid; + buf->bs_uid =xfs_kuid_to_uid (inode->i_uid); + buf->bs_gid =xfs_kgid_to_gid (inode->i_gid); buf->bs_size = dic->di_size; buf->bs_nlink = inode->i_nlink;
fs/xfs/xfs_qm.c +25 -15
@@ -331,16 +331,18 @@ xfs_qm_dqattach_locked(ASSERT (xfs_isilocked (ip,XFS_ILOCK_EXCL ));if (XFS_IS_UQUOTA_ON (mp) && !ip->i_udquot) { - error =xfs_qm_dqattach_one (ip, ip->i_d.di_uid,XFS_DQ_USER , - doalloc, &ip->i_udquot); + error =xfs_qm_dqattach_one (ip, +xfs_kuid_to_uid (VFS_I (ip)->i_uid), +XFS_DQ_USER , doalloc, &ip->i_udquot);if (error) goto done;ASSERT (ip->i_udquot); }if (XFS_IS_GQUOTA_ON (mp) && !ip->i_gdquot) { - error =xfs_qm_dqattach_one (ip, ip->i_d.di_gid,XFS_DQ_GROUP , - doalloc, &ip->i_gdquot); + error =xfs_qm_dqattach_one (ip, +xfs_kgid_to_gid (VFS_I (ip)->i_gid), +XFS_DQ_GROUP , doalloc, &ip->i_gdquot);if (error) goto done;ASSERT (ip->i_gdquot);
@@ -1630,8 +1632,8 @@ xfs_qm_dqfree_one( int xfs_qm_vop_dqalloc( struct xfs_inode *ip, - xfs_dqid_t uid, - xfs_dqid_t gid, + kuid_t uid, + kgid_t gid, prid_t prid, uint flags, struct xfs_dquot **O_udqpp,
@@ -1639,6 +1641,7 @@ xfs_qm_vop_dqalloc( struct xfs_dquot **O_pdqpp) {struct xfs_mount *mp = ip->i_mount; +struct inode *inode =VFS_I (ip);struct xfs_dquot *uq =NULL ;struct xfs_dquot *gq =NULL ;struct xfs_dquot *pq =NULL ;
@@ -1652,7 +1655,7 @@ xfs_qm_vop_dqalloc(xfs_ilock (ip, lockflags);if ((flags &XFS_QMOPT_INHERIT ) &&XFS_INHERIT_GID (ip)) - gid = ip->i_d.di_gid; + gid = inode->i_gid;/* * Attach the dquot(s) to this inode, doing a dquot allocation
@@ -1667,7 +1670,7 @@ xfs_qm_vop_dqalloc( }if ((flags &XFS_QMOPT_UQUOTA ) &&XFS_IS_UQUOTA_ON (mp)) { -if (ip->i_d.di_uid != uid) { +if (!uid_eq (inode->i_uid, uid)) {/* * What we need is the dquot that has this uid, and * if we send the inode to dqget, the uid of the inode
@@ -1678,7 +1681,8 @@ xfs_qm_vop_dqalloc(* holding ilock. */ xfs_iunlock (ip, lockflags); - error =xfs_qm_dqget (mp, uid,XFS_DQ_USER , true, &uq); + error =xfs_qm_dqget (mp,xfs_kuid_to_uid (uid), +XFS_DQ_USER , true, &uq);if (error) {ASSERT (error != -ENOENT );return error;
@@ -1699,9 +1703,10 @@ xfs_qm_vop_dqalloc( } }if ((flags &XFS_QMOPT_GQUOTA ) &&XFS_IS_GQUOTA_ON (mp)) { -if (ip->i_d.di_gid != gid) { +if (!gid_eq (inode->i_gid, gid)) {xfs_iunlock (ip, lockflags); - error =xfs_qm_dqget (mp, gid,XFS_DQ_GROUP , true, &gq); + error =xfs_qm_dqget (mp,xfs_kgid_to_gid (gid), +XFS_DQ_GROUP , true, &gq);if (error) {ASSERT (error != -ENOENT ); goto error_rele;
@@ -1827,7 +1832,8 @@ xfs_qm_vop_chown_reserve( XFS_QMOPT_RES_RTBLKS :XFS_QMOPT_RES_REGBLKS ;if (XFS_IS_UQUOTA_ON (mp) && udqp && - ip->i_d.di_uid !=be32_to_cpu (udqp->q_core.d_id)) { +xfs_kuid_to_uid (VFS_I (ip)->i_uid) != +be32_to_cpu (udqp->q_core.d_id)) { udq_delblks = udqp;/* * If there are delayed allocation blocks, then we have to
@@ -1840,7 +1846,8 @@ xfs_qm_vop_chown_reserve( } }if (XFS_IS_GQUOTA_ON (ip->i_mount) && gdqp && - ip->i_d.di_gid !=be32_to_cpu (gdqp->q_core.d_id)) { +xfs_kgid_to_gid (VFS_I (ip)->i_gid) != +be32_to_cpu (gdqp->q_core.d_id)) { gdq_delblks = gdqp;if (delblks) {ASSERT (ip->i_gdquot);
@@ -1937,14 +1944,17 @@ xfs_qm_vop_create_dqattach(if (udqp &&XFS_IS_UQUOTA_ON (mp)) {ASSERT (ip->i_udquot ==NULL ); -ASSERT (ip->i_d.di_uid ==be32_to_cpu (udqp->q_core.d_id)); +ASSERT (xfs_kuid_to_uid (VFS_I (ip)->i_uid) == +be32_to_cpu (udqp->q_core.d_id)); ip->i_udquot =xfs_qm_dqhold (udqp);xfs_trans_mod_dquot (tp, udqp,XFS_TRANS_DQ_ICOUNT ,1 ); }if (gdqp &&XFS_IS_GQUOTA_ON (mp)) {ASSERT (ip->i_gdquot ==NULL ); -ASSERT (ip->i_d.di_gid ==be32_to_cpu (gdqp->q_core.d_id)); +ASSERT (xfs_kgid_to_gid (VFS_I (ip)->i_gid) == +be32_to_cpu (gdqp->q_core.d_id)); + ip->i_gdquot =xfs_qm_dqhold (gdqp);xfs_trans_mod_dquot (tp, gdqp,XFS_TRANS_DQ_ICOUNT ,1 ); }
fs/xfs/xfs_quota.h +2 -2
@@ -86,7 +86,7 @@ extern int xfs_trans_reserve_quota_bydquots(struct xfs_trans *, struct xfs_mount *, struct xfs_dquot *, struct xfs_dquot *, struct xfs_dquot *, int64_t, long, uint); -extern int xfs_qm_vop_dqalloc (struct xfs_inode *,xfs_dqid_t ,xfs_dqid_t , +extern int xfs_qm_vop_dqalloc (struct xfs_inode *,kuid_t ,kgid_t ,prid_t ,uint ,struct xfs_dquot **,struct xfs_dquot **,struct xfs_dquot **); extern void xfs_qm_vop_create_dqattach(struct xfs_trans *, struct xfs_inode *,
@@ -109,7 +109,7 @@ extern void xfs_qm_unmount_quotas(struct xfs_mount *);#else static inline int -xfs_qm_vop_dqalloc (struct xfs_inode *ip,xfs_dqid_t uid,xfs_dqid_t gid, +xfs_qm_vop_dqalloc (struct xfs_inode *ip,kuid_t kuid,kgid_t kgid,prid_t prid,uint flags,struct xfs_dquot **udqp,struct xfs_dquot **gdqp,struct xfs_dquot **pdqp) {
fs/xfs/xfs_symlink.c +1 -3
@@ -191,9 +191,7 @@ xfs_symlink(/* * Make sure that we have allocated dquot(s) on disk. */ - error =xfs_qm_vop_dqalloc (dp, -xfs_kuid_to_uid (current_fsuid ()), -xfs_kgid_to_gid (current_fsgid ()), prid, + error =xfs_qm_vop_dqalloc (dp,current_fsuid (),current_fsgid (), prid,XFS_QMOPT_QUOTALL |XFS_QMOPT_INHERIT , &udqp, &gdqp, &pdqp);if (error)
Signature · 1 line
2.39.1
PATCH 5/17 xfs: remove the kuid/kgid conversion wrappers
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit ba8adad5d036733d240fa8a8f4d055f3d4490562 upstream.
Remove the XFS wrappers for converting from and to the kuid/kgid types. Mostly this means switching to VFS i_{u,g}id_{read,write} helpers, but in a few spots the calls to the conversion functions is open coded. To match the use of sb->s_user_ns in the helpers and other file systems, sb->s_user_ns is also used in the quota code. The ACL code already does the conversion in a grotty layering violation in the VFS xattr code, so it keeps using init_user_ns for the identity mapping.
--- fs/xfs/libxfs/xfs_inode_buf.c | 8 ++++---- fs/xfs/xfs_acl.c | 12 ++++++++---- fs/xfs/xfs_dquot.c | 4 ++-- fs/xfs/xfs_inode_item.c | 4 ++-- fs/xfs/xfs_itable.c | 4 ++-- fs/xfs/xfs_linux.h | 26 -------------------------- fs/xfs/xfs_qm.c | 23 +++++++++-------------- 7 files changed, 27 insertions(+), 54 deletions(-)
fs/xfs/libxfs/xfs_inode_buf.c +4 -4
@@ -222,8 +222,8 @@ xfs_inode_from_disk( } to->di_format = from->di_format; - inode->i_uid =xfs_uid_to_kuid (be32_to_cpu (from->di_uid)); - inode->i_gid =xfs_gid_to_kgid (be32_to_cpu (from->di_gid)); +i_uid_write (inode,be32_to_cpu (from->di_uid)); +i_gid_write (inode,be32_to_cpu (from->di_gid)); to->di_flushiter =be16_to_cpu (from->di_flushiter);/*
@@ -276,8 +276,8 @@ xfs_inode_to_disk( to->di_version = from->di_version; to->di_format = from->di_format; - to->di_uid =cpu_to_be32 (xfs_kuid_to_uid (inode->i_uid)); - to->di_gid =cpu_to_be32 (xfs_kgid_to_gid (inode->i_gid)); + to->di_uid =cpu_to_be32 (i_uid_read (inode)); + to->di_gid =cpu_to_be32 (i_gid_read (inode)); to->di_projid_lo =cpu_to_be16 (from->di_projid &0xffff ); to->di_projid_hi =cpu_to_be16 (from->di_projid >>16 );
fs/xfs/xfs_acl.c +8 -4
@@ -66,10 +66,12 @@ xfs_acl_from_disk(switch (acl_e->e_tag) {case ACL_USER : - acl_e->e_uid =xfs_uid_to_kuid (be32_to_cpu (ace->ae_id)); + acl_e->e_uid =make_kuid (&init_user_ns, +be32_to_cpu (ace->ae_id));break ;case ACL_GROUP : - acl_e->e_gid =xfs_gid_to_kgid (be32_to_cpu (ace->ae_id)); + acl_e->e_gid =make_kgid (&init_user_ns, +be32_to_cpu (ace->ae_id));break ;case ACL_USER_OBJ :case ACL_GROUP_OBJ :
@@ -102,10 +104,12 @@ xfs_acl_to_disk(struct xfs_acl *aclp, const struct posix_acl *acl) ace->ae_tag =cpu_to_be32 (acl_e->e_tag);switch (acl_e->e_tag) {case ACL_USER : - ace->ae_id =cpu_to_be32 (xfs_kuid_to_uid (acl_e->e_uid)); + ace->ae_id =cpu_to_be32 ( +from_kuid (&init_user_ns, acl_e->e_uid));break ;case ACL_GROUP : - ace->ae_id =cpu_to_be32 (xfs_kgid_to_gid (acl_e->e_gid)); + ace->ae_id =cpu_to_be32 ( +from_kgid (&init_user_ns, acl_e->e_gid));break ;default : ace->ae_id =cpu_to_be32 (ACL_UNDEFINED_ID );
fs/xfs/xfs_dquot.c +2 -2
@@ -859,9 +859,9 @@ xfs_qm_id_for_quotatype( {switch (type) {case XFS_DQ_USER : -return xfs_kuid_to_uid (VFS_I (ip)->i_uid); +return i_uid_read (VFS_I (ip));case XFS_DQ_GROUP : -return xfs_kgid_to_gid (VFS_I (ip)->i_gid); +return i_gid_read (VFS_I (ip));case XFS_DQ_PROJ :return ip->i_d.di_projid; }
fs/xfs/xfs_inode_item.c +2 -2
@@ -308,8 +308,8 @@ xfs_inode_to_log_dinode( to->di_version = from->di_version; to->di_format = from->di_format; - to->di_uid =xfs_kuid_to_uid (inode->i_uid); - to->di_gid =xfs_kgid_to_gid (inode->i_gid); + to->di_uid =i_uid_read (inode); + to->di_gid =i_gid_read (inode); to->di_projid_lo = from->di_projid &0xffff ; to->di_projid_hi = from->di_projid >>16 ;
fs/xfs/xfs_itable.c +2 -2
@@ -86,8 +86,8 @@ xfs_bulkstat_one_int(*/ buf->bs_projectid = ip->i_d.di_projid; buf->bs_ino = ino; - buf->bs_uid =xfs_kuid_to_uid (inode->i_uid); - buf->bs_gid =xfs_kgid_to_gid (inode->i_gid); + buf->bs_uid =i_uid_read (inode); + buf->bs_gid =i_gid_read (inode); buf->bs_size = dic->di_size; buf->bs_nlink = inode->i_nlink;
fs/xfs/xfs_linux.h +0 -26
@@ -163,32 +163,6 @@ struct xstats {extern struct xstats xfsstats; -/* Kernel uid/gid conversion. These are used to convert to/from the on disk -* uid_t/gid_t types to the kuid_t/kgid_t types that the kernel uses internally. -* The conversion here is type only, the value will remain the same since we -* are converting to the init_user_ns. The uid is later mapped to a particular -* user namespace value when crossing the kernel/user boundary. -*/ -static inline uint32_t xfs_kuid_to_uid (kuid_t uid) -{ -return from_kuid (&init_user_ns, uid); -} - -static inline kuid_t xfs_uid_to_kuid (uint32_t uid) -{ -return make_kuid (&init_user_ns, uid); -} - -static inline uint32_t xfs_kgid_to_gid (kgid_t gid) -{ -return from_kgid (&init_user_ns, gid); -} - -static inline kgid_t xfs_gid_to_kgid (uint32_t gid) -{ -return make_kgid (&init_user_ns, gid); -} -static inline dev_t xfs_to_linux_dev_t (xfs_dev_t dev) {return MKDEV (sysv_major (dev) &0x1ff ,sysv_minor (dev));
fs/xfs/xfs_qm.c +9 -14
@@ -331,8 +331,7 @@ xfs_qm_dqattach_locked(ASSERT (xfs_isilocked (ip,XFS_ILOCK_EXCL ));if (XFS_IS_UQUOTA_ON (mp) && !ip->i_udquot) { - error =xfs_qm_dqattach_one (ip, -xfs_kuid_to_uid (VFS_I (ip)->i_uid), + error =xfs_qm_dqattach_one (ip,i_uid_read (VFS_I (ip)),XFS_DQ_USER , doalloc, &ip->i_udquot);if (error) goto done;
@@ -340,8 +339,7 @@ xfs_qm_dqattach_locked( }if (XFS_IS_GQUOTA_ON (mp) && !ip->i_gdquot) { - error =xfs_qm_dqattach_one (ip, -xfs_kgid_to_gid (VFS_I (ip)->i_gid), + error =xfs_qm_dqattach_one (ip,i_gid_read (VFS_I (ip)),XFS_DQ_GROUP , doalloc, &ip->i_gdquot);if (error) goto done;
@@ -1642,6 +1640,7 @@ xfs_qm_vop_dqalloc( {struct xfs_mount *mp = ip->i_mount;struct inode *inode =VFS_I (ip); +struct user_namespace *user_ns = inode->i_sb->s_user_ns;struct xfs_dquot *uq =NULL ;struct xfs_dquot *gq =NULL ;struct xfs_dquot *pq =NULL ;
@@ -1681,7 +1680,7 @@ xfs_qm_vop_dqalloc(* holding ilock. */ xfs_iunlock (ip, lockflags); - error =xfs_qm_dqget (mp,xfs_kuid_to_uid (uid), + error =xfs_qm_dqget (mp,from_kuid (user_ns, uid),XFS_DQ_USER , true, &uq);if (error) {ASSERT (error != -ENOENT );
@@ -1705,7 +1704,7 @@ xfs_qm_vop_dqalloc(if ((flags &XFS_QMOPT_GQUOTA ) &&XFS_IS_GQUOTA_ON (mp)) {if (!gid_eq (inode->i_gid, gid)) {xfs_iunlock (ip, lockflags); - error =xfs_qm_dqget (mp,xfs_kgid_to_gid (gid), + error =xfs_qm_dqget (mp,from_kgid (user_ns, gid),XFS_DQ_GROUP , true, &gq);if (error) {ASSERT (error != -ENOENT );
@@ -1832,8 +1831,7 @@ xfs_qm_vop_chown_reserve( XFS_QMOPT_RES_RTBLKS :XFS_QMOPT_RES_REGBLKS ;if (XFS_IS_UQUOTA_ON (mp) && udqp && -xfs_kuid_to_uid (VFS_I (ip)->i_uid) != -be32_to_cpu (udqp->q_core.d_id)) { +i_uid_read (VFS_I (ip)) !=be32_to_cpu (udqp->q_core.d_id)) { udq_delblks = udqp;/* * If there are delayed allocation blocks, then we have to
@@ -1846,8 +1844,7 @@ xfs_qm_vop_chown_reserve( } }if (XFS_IS_GQUOTA_ON (ip->i_mount) && gdqp && -xfs_kgid_to_gid (VFS_I (ip)->i_gid) != -be32_to_cpu (gdqp->q_core.d_id)) { +i_gid_read (VFS_I (ip)) !=be32_to_cpu (gdqp->q_core.d_id)) { gdq_delblks = gdqp;if (delblks) {ASSERT (ip->i_gdquot);
@@ -1944,16 +1941,14 @@ xfs_qm_vop_create_dqattach(if (udqp &&XFS_IS_UQUOTA_ON (mp)) {ASSERT (ip->i_udquot ==NULL ); -ASSERT (xfs_kuid_to_uid (VFS_I (ip)->i_uid) == -be32_to_cpu (udqp->q_core.d_id)); +ASSERT (i_uid_read (VFS_I (ip)) ==be32_to_cpu (udqp->q_core.d_id)); ip->i_udquot =xfs_qm_dqhold (udqp);xfs_trans_mod_dquot (tp, udqp,XFS_TRANS_DQ_ICOUNT ,1 ); }if (gdqp &&XFS_IS_GQUOTA_ON (mp)) {ASSERT (ip->i_gdquot ==NULL ); -ASSERT (xfs_kgid_to_gid (VFS_I (ip)->i_gid) == -be32_to_cpu (gdqp->q_core.d_id)); +ASSERT (i_gid_read (VFS_I (ip)) ==be32_to_cpu (gdqp->q_core.d_id)); ip->i_gdquot =xfs_qm_dqhold (gdqp);xfs_trans_mod_dquot (tp, gdqp,XFS_TRANS_DQ_ICOUNT ,1 );
Signature · 1 line
2.39.1
PATCH 6/17 xfs: add a new xfs_sb_version_has_v3inode helper
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit b81b79f4eda2ea98ae5695c0b6eb384c8d90b74d upstream.
Add a new wrapper to check if a file system supports the v3 inode format with a larger dinode core. Previously we used xfs_sb_version_hascrc for that, which is technically correct but a little confusing to read.
Also move xfs_dinode_good_version next to xfs_sb_version_has_v3inode so that we have one place that documents the superblock version to inode version relationship.
--- fs/xfs/libxfs/xfs_format.h | 17 +++++++++++++++++ fs/xfs/libxfs/xfs_ialloc.c | 4 ++-- fs/xfs/libxfs/xfs_inode_buf.c | 17 +++-------------- fs/xfs/libxfs/xfs_inode_buf.h | 2 -- fs/xfs/libxfs/xfs_trans_resv.c | 2 +- fs/xfs/xfs_buf_item.c | 2 +- fs/xfs/xfs_log_recover.c | 2 +- 7 files changed, 25 insertions(+), 21 deletions(-)
fs/xfs/libxfs/xfs_format.h +17 -0
@@ -497,6 +497,23 @@ static inline bool xfs_sb_version_hascrc(struct xfs_sb *sbp)return XFS_SB_VERSION_NUM (sbp) ==XFS_SB_VERSION_5 ; } +/* +* v5 file systems support V3 inodes only, earlier file systems support +* v2 and v1 inodes. +*/ +static inline bool xfs_sb_version_has_v3inode (struct xfs_sb *sbp) +{ +return XFS_SB_VERSION_NUM (sbp) ==XFS_SB_VERSION_5 ; +} + +static inline bool xfs_dinode_good_version (struct xfs_sb *sbp, +uint8_t version) +{ +if (xfs_sb_version_has_v3inode (sbp)) +return version ==3 ; +return version ==1 || version ==2 ; +} +static inline bool xfs_sb_version_has_pquotino (struct xfs_sb *sbp) {return XFS_SB_VERSION_NUM (sbp) ==XFS_SB_VERSION_5 ;
fs/xfs/libxfs/xfs_ialloc.c +2 -2
@@ -303,7 +303,7 @@ xfs_ialloc_inode_init(* That means for v3 inode we log the entire buffer rather than just the * inode cores. */ -if (xfs_sb_version_hascrc (&mp->m_sb)) { +if (xfs_sb_version_has_v3inode (&mp->m_sb)) { version =3 ; ino =XFS_AGINO_TO_INO (mp, agno,XFS_AGB_TO_AGINO (mp, agbno));
@@ -2818,7 +2818,7 @@ xfs_ialloc_setup_geometry(* cannot change the behavior. */ igeo->inode_cluster_size_raw =XFS_INODE_BIG_CLUSTER_SIZE ; -if (xfs_sb_version_hascrc (&mp->m_sb)) { +if (xfs_sb_version_has_v3inode (&mp->m_sb)) {int new_size = igeo->inode_cluster_size_raw; new_size *= mp->m_sb.sb_inodesize /XFS_DINODE_MIN_SIZE ;
fs/xfs/libxfs/xfs_inode_buf.c +3 -14
@@ -44,17 +44,6 @@ xfs_inobp_check( }#endif -bool -xfs_dinode_good_version ( -struct xfs_mount *mp, -__u8 version) -{ -if (xfs_sb_version_hascrc (&mp->m_sb)) -return version ==3 ; - -return version ==1 || version ==2 ; -} -/* * If we are doing readahead on an inode buffer, we might be in log recovery * reading an inode allocation buffer that hasn't yet been replayed, and hence
@@ -93,7 +82,7 @@ xfs_inode_buf_verify( dip =xfs_buf_offset (bp, (i << mp->m_sb.sb_inodelog)); unlinked_ino =be32_to_cpu (dip->di_next_unlinked); di_ok =xfs_verify_magic16 (bp, dip->di_magic) && -xfs_dinode_good_version (mp, dip->di_version) && +xfs_dinode_good_version (&mp->m_sb, dip->di_version) &&xfs_verify_agino_or_null (mp, agno, unlinked_ino);if (unlikely (XFS_TEST_ERROR (!di_ok, mp,XFS_ERRTAG_ITOBP_INOTOBP ))) {
@@ -454,7 +443,7 @@ xfs_dinode_verify(/* Verify v3 integrity information first */ if (dip->di_version >=3 ) { -if (!xfs_sb_version_hascrc (&mp->m_sb)) +if (!xfs_sb_version_has_v3inode (&mp->m_sb))return __this_address;if (!xfs_verify_cksum ((char *)dip, mp->m_sb.sb_inodesize,XFS_DINODE_CRC_OFF ))
@@ -629,7 +618,7 @@ xfs_iread(/* shortcut IO on inode allocation if possible */ if ((iget_flags &XFS_IGET_CREATE ) && -xfs_sb_version_hascrc (&mp->m_sb) && +xfs_sb_version_has_v3inode (&mp->m_sb) && !(mp->m_flags &XFS_MOUNT_IKEEP )) {/* initialise the on-disk inode core */ memset (&ip->i_d,0 ,sizeof (ip->i_d));
fs/xfs/libxfs/xfs_inode_buf.h +0 -2
@@ -59,8 +59,6 @@ void xfs_inode_from_disk(struct xfs_inode *ip, struct xfs_dinode *from);void xfs_log_dinode_to_disk (struct xfs_log_dinode *from,struct xfs_dinode *to); -bool xfs_dinode_good_version (struct xfs_mount *mp,__u8 version); -#if defined(DEBUG )void xfs_inobp_check (struct xfs_mount *,struct xfs_buf *);#else
fs/xfs/libxfs/xfs_trans_resv.c +1 -1
@@ -187,7 +187,7 @@ xfs_calc_inode_chunk_res( XFS_FSB_TO_B(mp, 1));if (alloc) {/* icreate tx uses ordered buffers */ -if (xfs_sb_version_hascrc (&mp->m_sb)) +if (xfs_sb_version_has_v3inode (&mp->m_sb))return res; size =XFS_FSB_TO_B (mp,1 ); }
fs/xfs/xfs_buf_item.c +1 -1
@@ -328,7 +328,7 @@ xfs_buf_item_format(* occurs during recovery. */ if (bip->bli_flags &XFS_BLI_INODE_BUF ) { -if (xfs_sb_version_hascrc (&lip->li_mountp->m_sb) || +if (xfs_sb_version_has_v3inode (&lip->li_mountp->m_sb) || !((bip->bli_flags &XFS_BLI_INODE_ALLOC_BUF ) &&xfs_log_item_in_current_chkpt (lip))) bip->__bli_format.blf_flags |=XFS_BLF_INODE_BUF ;
fs/xfs/xfs_log_recover.c +1 -1
@@ -3018,7 +3018,7 @@ xlog_recover_inode_pass2(* superblock flag to determine whether we need to look at di_flushiter * to skip replay when the on disk inode is newer than the log one */ -if (!xfs_sb_version_hascrc (&mp->m_sb) && +if (!xfs_sb_version_has_v3inode (&mp->m_sb) && ldip->di_flushiter <be16_to_cpu (dip->di_flushiter)) {/* * Deal with the wrap case, DI_MAX_FLUSH is less
Signature · 1 line
2.39.1
PATCH 7/17 xfs: only check the superblock version for dinode size calculation
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit e9e2eae89ddb658ea332295153fdca78c12c1e0d upstream.
The size of the dinode structure is only dependent on the file system version, so instead of checking the individual inode version just use the newly added xfs_sb_version_has_large_dinode helper, and simplify various calling conventions.
--- fs/xfs/libxfs/xfs_attr_leaf.c | 5 ++--- fs/xfs/libxfs/xfs_bmap.c | 10 ++++------ fs/xfs/libxfs/xfs_format.h | 16 ++++++++-------- fs/xfs/libxfs/xfs_ialloc.c | 2 +- fs/xfs/libxfs/xfs_inode_buf.c | 2 +- fs/xfs/libxfs/xfs_inode_fork.c | 2 +- fs/xfs/libxfs/xfs_inode_fork.h | 9 ++------- fs/xfs/libxfs/xfs_log_format.h | 10 ++++------ fs/xfs/xfs_inode_item.c | 4 ++-- fs/xfs/xfs_log_recover.c | 2 +- fs/xfs/xfs_symlink.c | 2 +- 11 files changed, 27 insertions(+), 37 deletions(-)
fs/xfs/libxfs/xfs_attr_leaf.c +2 -3
@@ -456,7 +456,7 @@ xfs_attr_shortform_bytesfit(int offset;/* rounded down */ - offset = (XFS_LITINO (mp, dp->i_d.di_version) - bytes) >>3 ; + offset = (XFS_LITINO (mp) - bytes) >>3 ;if (dp->i_d.di_format ==XFS_DINODE_FMT_DEV ) { minforkoff =roundup (sizeof (xfs_dev_t),8 ) >>3 ;
@@ -523,8 +523,7 @@ xfs_attr_shortform_bytesfit( minforkoff =roundup (minforkoff,8 ) >>3 ;/* attr fork btree root can have at least this many key/ptr pairs */ - maxforkoff =XFS_LITINO (mp, dp->i_d.di_version) - -XFS_BMDR_SPACE_CALC (MINABTPTRS ); + maxforkoff =XFS_LITINO (mp) -XFS_BMDR_SPACE_CALC (MINABTPTRS ); maxforkoff = maxforkoff >>3 ;/* rounded down */ if (offset >= maxforkoff)
fs/xfs/libxfs/xfs_bmap.c +4 -6
@@ -192,14 +192,12 @@ xfs_default_attroffset(struct xfs_mount *mp = ip->i_mount;uint offset; -if (mp->m_sb.sb_inodesize ==256 ) { - offset =XFS_LITINO (mp, ip->i_d.di_version) - -XFS_BMDR_SPACE_CALC (MINABTPTRS ); - }else { +if (mp->m_sb.sb_inodesize ==256 ) + offset =XFS_LITINO (mp) -XFS_BMDR_SPACE_CALC (MINABTPTRS ); +else offset =XFS_BMDR_SPACE_CALC (6 *MINABTPTRS ); - } -ASSERT (offset <XFS_LITINO (mp, ip->i_d.di_version)); +ASSERT (offset <XFS_LITINO (mp));return offset; }
fs/xfs/libxfs/xfs_format.h +8 -8
@@ -963,8 +963,12 @@ typedef enum xfs_dinode_fmt {/* * Inode size for given fs. */ -#define XFS_LITINO (mp, version) \ - ((int)(((mp)->m_sb.sb_inodesize) - xfs_dinode_size(version))) +#define XFS_DINODE_SIZE (sbp) \ + (xfs_sb_version_has_v3inode(sbp) ? \ + sizeof(struct xfs_dinode) : \ + offsetof(struct xfs_dinode, di_crc)) +#define XFS_LITINO (mp) \ + ((mp)->m_sb.sb_inodesize - XFS_DINODE_SIZE(&(mp)->m_sb))/* * Inode data & attribute fork sizes, per inode.
@@ -973,13 +977,9 @@ typedef enum xfs_dinode_fmt {#define XFS_DFORK_BOFF (dip) ((int)((dip)->di_forkoff << 3))#define XFS_DFORK_DSIZE (dip,mp) \ - (XFS_DFORK_Q(dip) ? \ - XFS_DFORK_BOFF(dip) : \ - XFS_LITINO(mp, (dip)->di_version)) + (XFS_DFORK_Q(dip) ? XFS_DFORK_BOFF(dip) : XFS_LITINO(mp))#define XFS_DFORK_ASIZE (dip,mp) \ - (XFS_DFORK_Q(dip) ? \ - XFS_LITINO(mp, (dip)->di_version) - XFS_DFORK_BOFF(dip) : \ - 0) + (XFS_DFORK_Q(dip) ? XFS_LITINO(mp) - XFS_DFORK_BOFF(dip) : 0)#define XFS_DFORK_SIZE (dip,mp,w) \ ((w) == XFS_DATA_FORK ? \ XFS_DFORK_DSIZE(dip, mp) : \
fs/xfs/libxfs/xfs_ialloc.c +1 -1
@@ -339,7 +339,7 @@ xfs_ialloc_inode_init(xfs_buf_zero (fbuf,0 ,BBTOB (fbuf->b_length));for (i =0 ; i <M_IGEO (mp)->inodes_per_cluster; i++) {int ioffset = i << mp->m_sb.sb_inodelog; -uint isize =xfs_dinode_size (version); +uint isize =XFS_DINODE_SIZE (&mp->m_sb); free =xfs_make_iptr (mp, fbuf, i); free->di_magic =cpu_to_be16 (XFS_DINODE_MAGIC );
fs/xfs/libxfs/xfs_inode_buf.c +1 -1
@@ -417,7 +417,7 @@ xfs_dinode_verify_forkoff(case XFS_DINODE_FMT_LOCAL :/* fall through ... */ case XFS_DINODE_FMT_EXTENTS :/* fall through ... */ case XFS_DINODE_FMT_BTREE : -if (dip->di_forkoff >= (XFS_LITINO (mp, dip->di_version) >>3 )) +if (dip->di_forkoff >= (XFS_LITINO (mp) >>3 ))return __this_address;break ;default :
fs/xfs/libxfs/xfs_inode_fork.c +1 -1
@@ -183,7 +183,7 @@ xfs_iformat_local(*/ if (unlikely (size >XFS_DFORK_SIZE (dip, ip->i_mount, whichfork))) {xfs_warn (ip->i_mount, -"corrupt inode %Lu (bad size %d for local fork, size = %d)." , +"corrupt inode %Lu (bad size %d for local fork, size = %zd)." , (unsigned long long ) ip->i_ino, size,XFS_DFORK_SIZE (dip, ip->i_mount, whichfork)); xfs_inode_verifier_error(ip, -EFSCORRUPTED,
fs/xfs/libxfs/xfs_inode_fork.h +2 -7
@@ -46,14 +46,9 @@ struct xfs_ifork { (ip)->i_afp : \ (ip)->i_cowfp))#define XFS_IFORK_DSIZE (ip) \ - (XFS_IFORK_Q(ip) ? \ - XFS_IFORK_BOFF(ip) : \ - XFS_LITINO((ip)->i_mount, (ip)->i_d.di_version)) + (XFS_IFORK_Q(ip) ? XFS_IFORK_BOFF(ip) : XFS_LITINO((ip)->i_mount))#define XFS_IFORK_ASIZE (ip) \ - (XFS_IFORK_Q(ip) ? \ - XFS_LITINO((ip)->i_mount, (ip)->i_d.di_version) - \ - XFS_IFORK_BOFF(ip) : \ - 0) + (XFS_IFORK_Q(ip) ? XFS_LITINO((ip)->i_mount) - XFS_IFORK_BOFF(ip) : 0)#define XFS_IFORK_SIZE (ip,w) \ ((w) == XFS_DATA_FORK ? \ XFS_IFORK_DSIZE(ip) : \
fs/xfs/libxfs/xfs_log_format.h +4 -6
@@ -424,12 +424,10 @@ struct xfs_log_dinode {/* structure must be padded to 64 bit alignment */ }; -static inline uint xfs_log_dinode_size (int version) -{ -if (version ==3 ) -return sizeof (struct xfs_log_dinode ); -return offsetof(struct xfs_log_dinode , di_next_unlinked); -} +#define xfs_log_dinode_size (mp) \ + (xfs_sb_version_has_v3inode(&(mp)->m_sb) ? \ + sizeof(struct xfs_log_dinode) : \ + offsetof(struct xfs_log_dinode, di_next_unlinked))/* * Buffer Log Format defintions
fs/xfs/xfs_inode_item.c +2 -2
@@ -125,7 +125,7 @@ xfs_inode_item_size( *nvecs +=2 ; *nbytes +=sizeof (struct xfs_inode_log_format ) + -xfs_log_dinode_size (ip->i_d.di_version); +xfs_log_dinode_size (ip->i_mount);xfs_inode_item_data_fork_size (iip, nvecs, nbytes);if (XFS_IFORK_Q (ip))
@@ -370,7 +370,7 @@ xfs_inode_item_format_core( dic =xlog_prepare_iovec (lv, vecp,XLOG_REG_TYPE_ICORE );xfs_inode_to_log_dinode (ip, dic, ip->i_itemp->ili_item.li_lsn); -xlog_finish_iovec (lv, *vecp,xfs_log_dinode_size (ip->i_d.di_version)); +xlog_finish_iovec (lv, *vecp,xfs_log_dinode_size (ip->i_mount)); }/*
fs/xfs/xfs_log_recover.c +1 -1
@@ -3089,7 +3089,7 @@ xlog_recover_inode_pass2( error = -EFSCORRUPTED ; goto out_release; } - isize =xfs_log_dinode_size (ldip->di_version); + isize =xfs_log_dinode_size (mp);if (unlikely (item->ri_buf[1 ].i_len > isize)) { XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(7)", XFS_ERRLEVEL_LOW, mp, ldip,
fs/xfs/xfs_symlink.c +1 -1
@@ -201,7 +201,7 @@ xfs_symlink(* The symlink will fit into the inode data fork? * There can't be any attributes so we get the whole variable part. */ -if (pathlen <=XFS_LITINO (mp, dp->i_d.di_version)) +if (pathlen <=XFS_LITINO (mp)) fs_blocks =0 ;else fs_blocks =xfs_symlink_blocks (mp, pathlen);
Signature · 1 line
2.39.1
PATCH 8/17 xfs: simplify di_flags2 inheritance in xfs_ialloc
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit b3d1d37544d8c98be610df0ed66c884ff18748d5 upstream.
di_flags2 is initialized to zero for v4 and earlier file systems. This means di_flags2 can only be non-zero for a v5 file systems, in which case both the parent and child inodes can store the field. Remove the extra di_version check, and also remove the rather pointless local di_flags2 variable while at it.
--- fs/xfs/xfs_inode.c | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-)
fs/xfs/xfs_inode.c +3 -10
@@ -902,20 +902,13 @@ xfs_ialloc( ip->i_d.di_flags |= di_flags; } -if (pip && - (pip->i_d.di_flags2 &XFS_DIFLAG2_ANY ) && - pip->i_d.di_version ==3 && - ip->i_d.di_version ==3 ) { -uint64_t di_flags2 =0 ; - +if (pip && (pip->i_d.di_flags2 &XFS_DIFLAG2_ANY )) {if (pip->i_d.di_flags2 &XFS_DIFLAG2_COWEXTSIZE ) { - di_flags2 |=XFS_DIFLAG2_COWEXTSIZE ; + ip->i_d.di_flags2 |=XFS_DIFLAG2_COWEXTSIZE ; ip->i_d.di_cowextsize = pip->i_d.di_cowextsize; }if (pip->i_d.di_flags2 &XFS_DIFLAG2_DAX ) - di_flags2 |=XFS_DIFLAG2_DAX ; - - ip->i_d.di_flags2 |= di_flags2; + ip->i_d.di_flags2 |=XFS_DIFLAG2_DAX ; }/* FALLTHROUGH */ case S_IFLNK :
Signature · 1 line
2.39.1
PATCH 9/17 xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 5e28aafe708ba3e388f92a7148093319d3521c2f uptream.
Only v5 file systems can have the reflink feature, and those will always use the large dinode format. Remove the extra check for the inode version.
--- fs/xfs/xfs_ioctl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
fs/xfs/xfs_ioctl.c +1 -2
@@ -1510,8 +1510,7 @@ xfs_ioctl_setattr_check_cowextsize(if (!(fa->fsx_xflags &FS_XFLAG_COWEXTSIZE ))return 0 ; -if (!xfs_sb_version_hasreflink (&ip->i_mount->m_sb) || - ip->i_d.di_version !=3 ) +if (!xfs_sb_version_hasreflink (&ip->i_mount->m_sb))return -EINVAL ;if (fa->fsx_cowextsize ==0 )
Signature · 1 line
2.39.1
kernel test robot
build botRe: [PATCH 5.4 09/17] xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
FYI: kernel test robot notices the stable kernel rule is not satisfied.
Hi,
Thanks for your patch.
FYI: kernel test robot notices the stable kernel rule is not satisfied.
Rule: 'Cc: [email protected]' or 'commit <sha1> upstream.'
Subject: [PATCH 5.4 09/17] xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
The check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
Signature · 2 lines
0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests
Re: [PATCH 5.4 09/17] xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
Quoted 7 lines · kernel test robot
Hi,
Thanks for your patch.
FYI: kernel test robot notices the stable kernel rule is not satisfied.
Rule: 'Cc: [email protected]' or 'commit <sha1> upstream.'
I have CCed [email protected] and also mentioned the upstream commit ID in the commit message.
Quoted 4 lines
Subject: [PATCH 5.4 09/17] xfs: simplify a check in
xfs_ioctl_setattr_check_cowextsizeThe check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
Signature · 1 line
chandan
Re: [PATCH 5.4 09/17] xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
Quoted 11 lines · Chandan Babu R
Quoted 7 lines · kernel test robot
Hi,
Thanks for your patch.
FYI: kernel test robot notices the stable kernel rule is not satisfied.
Rule: 'Cc: [email protected]' or 'commit <sha1> upstream.'
I have CCed [email protected] and also mentioned the upstream commit ID in the commit message.
Hi Chandan, this looks to be a typo of "uptream" in your changelog
commit 5e28aafe708ba3e388f92a7148093319d3521c2f uptream.
Thanks
Quoted 9 lines
Quoted 5 lines
Subject: [PATCH 5.4 09/17] xfs: simplify a check in
xfs_ioctl_setattr_check_cowextsizeThe check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
Signature · 1 line
chandan
Re: [PATCH 5.4 09/17] xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
Quoted 8 lines · Philip Li
Quoted 7 lines · Chandan Babu R
Quoted 6 lines · kernel test robot
Hi,
Thanks for your patch.
FYI: kernel test robot notices the stable kernel rule is not satisfied.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
Rule: 'Cc: [email protected]' or 'commit <sha1> upstream.'
I have CCed [email protected] and also mentioned the upstream commit ID in the commit message.
Hi Chandan, this looks to be a typo of "uptream" in your changelog
commit
5e28aafe708ba3e388f92a7148093319d3521c2fuptream.
Ah! I am sorry. Thanks for pointing that out.
Quoted 9 lines
Thanks
Quoted 6 lines
Quoted 5 lines
Subject: [PATCH 5.4 09/17] xfs: simplify a check in
xfs_ioctl_setattr_check_cowextsizeThe check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
Signature · 1 line
chandan
PATCH v2 1/1 xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 5e28aafe708ba3e388f92a7148093319d3521c2f upstream.
Only v5 file systems can have the reflink feature, and those will always use the large dinode format. Remove the extra check for the inode version.
---
Changelog:
V1 -> V2:
Fix spelling mistake.
fs/xfs/xfs_ioctl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
fs/xfs/xfs_ioctl.c +1 -2
@@ -1510,8 +1510,7 @@ xfs_ioctl_setattr_check_cowextsize(if (!(fa->fsx_xflags &FS_XFLAG_COWEXTSIZE ))return 0 ; -if (!xfs_sb_version_hasreflink (&ip->i_mount->m_sb) || - ip->i_d.di_version !=3 ) +if (!xfs_sb_version_hasreflink (&ip->i_mount->m_sb))return -EINVAL ;if (fa->fsx_cowextsize ==0 )
Signature · 1 line
2.39.1
PATCH 10/17 xfs: remove the di_version field from struct icdinode
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 6471e9c5e7a109a952be8e3e80b8d9e262af239d upstream.
We know the version is 3 if on a v5 file system. For earlier file
systems formats we always upgrade the remaining v1 inodes to v2 and thus only use v2 inodes. Use the xfs_sb_version_has_large_dinode helper to check if we deal with small or large dinodes, and thus remove the need for the di_version field in struct icdinode.
--- fs/xfs/libxfs/xfs_inode_buf.c | 16 ++++++---------- fs/xfs/libxfs/xfs_inode_buf.h | 1 - fs/xfs/xfs_bmap_util.c | 16 ++++++++-------- fs/xfs/xfs_inode.c | 16 ++-------------- fs/xfs/xfs_inode_item.c | 8 +++----- fs/xfs/xfs_ioctl.c | 5 ++--- fs/xfs/xfs_iops.c | 2 +- fs/xfs/xfs_itable.c | 2 +- fs/xfs/xfs_log_recover.c | 2 +- 9 files changed, 24 insertions(+), 44 deletions(-)
fs/xfs/libxfs/xfs_inode_buf.c +6 -10
@@ -194,16 +194,14 @@ xfs_inode_from_disk(struct xfs_icdinode *to = &ip->i_d;struct inode *inode =VFS_I (ip); -/* * Convert v1 inodes immediately to v2 inode format as this is the * minimum inode version format we support in the rest of the code. +* They will also be unconditionally written back to disk as v2 inodes. */ - to->di_version = from->di_version; -if (to->di_version ==1 ) { +if (unlikely (from->di_version ==1 )) {set_nlink (inode,be16_to_cpu (from->di_onlink)); to->di_projid =0 ; - to->di_version =2 ; }else {set_nlink (inode,be32_to_cpu (from->di_nlink)); to->di_projid = (prid_t)be16_to_cpu(from->di_projid_hi) << 16 |
@@ -241,7 +239,7 @@ xfs_inode_from_disk( to->di_dmstate =be16_to_cpu (from->di_dmstate); to->di_flags =be16_to_cpu (from->di_flags); -if (to->di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&ip->i_mount->m_sb)) {inode_set_iversion_queried (inode,be64_to_cpu (from->di_changecount)); to->di_crtime.t_sec =be32_to_cpu (from->di_crtime.t_sec);
@@ -263,7 +261,6 @@ xfs_inode_to_disk( to->di_magic =cpu_to_be16 (XFS_DINODE_MAGIC ); to->di_onlink =0 ; - to->di_version = from->di_version; to->di_format = from->di_format; to->di_uid =cpu_to_be32 (i_uid_read (inode)); to->di_gid =cpu_to_be32 (i_gid_read (inode));
@@ -292,7 +289,8 @@ xfs_inode_to_disk( to->di_dmstate =cpu_to_be16 (from->di_dmstate); to->di_flags =cpu_to_be16 (from->di_flags); -if (from->di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&ip->i_mount->m_sb)) { + to->di_version =3 ; to->di_changecount =cpu_to_be64 (inode_peek_iversion (inode)); to->di_crtime.t_sec =cpu_to_be32 (from->di_crtime.t_sec); to->di_crtime.t_nsec =cpu_to_be32 (from->di_crtime.t_nsec);
@@ -304,6 +302,7 @@ xfs_inode_to_disk(uuid_copy (&to->di_uuid, &ip->i_mount->m_sb.sb_meta_uuid); to->di_flushiter =0 ; }else { + to->di_version =2 ; to->di_flushiter =cpu_to_be16 (from->di_flushiter); } }
@@ -623,7 +622,6 @@ xfs_iread(/* initialise the on-disk inode core */ memset (&ip->i_d,0 ,sizeof (ip->i_d));VFS_I (ip)->i_generation =prandom_u32 (); - ip->i_d.di_version =3 ;return 0 ; }
@@ -665,7 +663,6 @@ xfs_iread(* Partial initialisation of the in-core inode. Just the bits * that xfs_ialloc won't overwrite or relies on being correct. */ - ip->i_d.di_version = dip->di_version;VFS_I (ip)->i_generation =be32_to_cpu (dip->di_gen); ip->i_d.di_flushiter =be16_to_cpu (dip->di_flushiter);
@@ -679,7 +676,6 @@ xfs_iread(VFS_I (ip)->i_mode =0 ; } -ASSERT (ip->i_d.di_version >=2 ); ip->i_delayed_blks =0 ;/*
fs/xfs/libxfs/xfs_inode_buf.h +0 -1
@@ -16,7 +16,6 @@ struct xfs_dinode;* format specific structures at the appropriate time. */ struct xfs_icdinode { -int8_t di_version;/* inode version */ int8_t di_format;/* format of di_c data */ uint16_t di_flushiter;/* incremented on flush */ uint32_t di_projid;/* owner's project id */
fs/xfs/xfs_bmap_util.c +8 -8
@@ -1624,12 +1624,12 @@ xfs_swap_extent_forks(* event of a crash. Set the owner change log flags now and leave the * bmbt scan as the last step. */ -if (ip->i_d.di_version ==3 && - ip->i_d.di_format ==XFS_DINODE_FMT_BTREE ) - (*target_log_flags) |=XFS_ILOG_DOWNER ; -if (tip->i_d.di_version ==3 && - tip->i_d.di_format ==XFS_DINODE_FMT_BTREE ) - (*src_log_flags) |=XFS_ILOG_DOWNER ; +if (xfs_sb_version_has_v3inode (&ip->i_mount->m_sb)) { +if (ip->i_d.di_format ==XFS_DINODE_FMT_BTREE ) + (*target_log_flags) |=XFS_ILOG_DOWNER ; +if (tip->i_d.di_format ==XFS_DINODE_FMT_BTREE ) + (*src_log_flags) |=XFS_ILOG_DOWNER ; + }/* * Swap the data forks of the inodes
@@ -1664,7 +1664,7 @@ xfs_swap_extent_forks( (*src_log_flags) |=XFS_ILOG_DEXT ;break ;case XFS_DINODE_FMT_BTREE : -ASSERT (ip->i_d.di_version <3 || +ASSERT (!xfs_sb_version_has_v3inode (&ip->i_mount->m_sb) || (*src_log_flags &XFS_ILOG_DOWNER )); (*src_log_flags) |=XFS_ILOG_DBROOT ;break ;
@@ -1676,7 +1676,7 @@ xfs_swap_extent_forks(break ;case XFS_DINODE_FMT_BTREE : (*target_log_flags) |=XFS_ILOG_DBROOT ; -ASSERT (tip->i_d.di_version <3 || +ASSERT (!xfs_sb_version_has_v3inode (&ip->i_mount->m_sb) || (*target_log_flags &XFS_ILOG_DOWNER ));break ; }
fs/xfs/xfs_inode.c +2 -14
@@ -795,15 +795,6 @@ xfs_ialloc(return error;ASSERT (ip !=NULL ); inode =VFS_I (ip); - -/* -* We always convert v1 inodes to v2 now - we only support filesystems -* with >= v2 inode capability, so there is no reason for ever leaving -* an inode in v1 format. -*/ -if (ip->i_d.di_version ==1 ) - ip->i_d.di_version =2 ; - inode->i_mode = mode;set_nlink (inode, nlink); inode->i_uid =current_fsuid ();
@@ -841,7 +832,7 @@ xfs_ialloc( ip->i_d.di_dmstate =0 ; ip->i_d.di_flags =0 ; -if (ip->i_d.di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&mp->m_sb)) {inode_set_iversion (inode,1 ); ip->i_d.di_flags2 =0 ; ip->i_d.di_cowextsize =0 ;
@@ -849,7 +840,6 @@ xfs_ialloc( ip->i_d.di_crtime.t_nsec = (int32_t )tv.tv_nsec; } - flags =XFS_ILOG_CORE ;switch (mode &S_IFMT ) {case S_IFIFO :
@@ -1110,7 +1100,6 @@ xfs_bumplink( {xfs_trans_ichgtime (tp, ip,XFS_ICHGTIME_CHG ); -ASSERT (ip->i_d.di_version >1 );inc_nlink (VFS_I (ip));xfs_trans_log_inode (tp, ip,XFS_ILOG_CORE ); }
@@ -3822,7 +3811,6 @@ xfs_iflush_int(ASSERT (ip->i_d.di_format !=XFS_DINODE_FMT_BTREE || ip->i_d.di_nextents >XFS_IFORK_MAXEXT (ip,XFS_DATA_FORK ));ASSERT (iip !=NULL && iip->ili_fields !=0 ); -ASSERT (ip->i_d.di_version >1 );/* set *dip = inode's place in the buffer */ dip =xfs_buf_offset (bp, ip->i_imap.im_boffset);
@@ -3883,7 +3871,7 @@ xfs_iflush_int(* backwards compatibility with old kernels that predate logging all * inode changes. */ -if (ip->i_d.di_version <3 ) +if (!xfs_sb_version_has_v3inode (&mp->m_sb)) ip->i_d.di_flushiter++;/* Check the inline fork data before we write out. */
fs/xfs/xfs_inode_item.c +3 -5
@@ -305,8 +305,6 @@ xfs_inode_to_log_dinode(struct inode *inode =VFS_I (ip); to->di_magic =XFS_DINODE_MAGIC ; - - to->di_version = from->di_version; to->di_format = from->di_format; to->di_uid =i_uid_read (inode); to->di_gid =i_gid_read (inode);
@@ -339,7 +337,8 @@ xfs_inode_to_log_dinode(/* log a dummy value to ensure log structure is fully initialised */ to->di_next_unlinked =NULLAGINO ; -if (from->di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&ip->i_mount->m_sb)) { + to->di_version =3 ; to->di_changecount =inode_peek_iversion (inode); to->di_crtime.t_sec = from->di_crtime.t_sec; to->di_crtime.t_nsec = from->di_crtime.t_nsec;
@@ -351,6 +350,7 @@ xfs_inode_to_log_dinode(uuid_copy (&to->di_uuid, &ip->i_mount->m_sb.sb_meta_uuid); to->di_flushiter =0 ; }else { + to->di_version =2 ; to->di_flushiter = from->di_flushiter; } }
@@ -395,8 +395,6 @@ xfs_inode_item_format(struct xfs_log_iovec *vecp =NULL ;struct xfs_inode_log_format *ilf; -ASSERT (ip->i_d.di_version >1 ); - ilf =xlog_prepare_iovec (lv, &vecp,XLOG_REG_TYPE_IFORMAT ); ilf->ilf_type =XFS_LI_INODE ; ilf->ilf_ino = ip->i_ino;
fs/xfs/xfs_ioctl.c +2 -3
@@ -1299,7 +1299,7 @@ xfs_ioctl_setattr_xflags(/* diflags2 only valid for v3 inodes. */ di_flags2 =xfs_flags2diflags2 (ip, fa->fsx_xflags); -if (di_flags2 && ip->i_d.di_version <3 ) +if (di_flags2 && !xfs_sb_version_has_v3inode (&mp->m_sb))return -EINVAL ; ip->i_d.di_flags =xfs_flags2diflags (ip, fa->fsx_xflags);
@@ -1638,7 +1638,6 @@ xfs_ioctl_setattr( olddquot =xfs_qm_vop_chown (tp, ip, &ip->i_pdquot, pdqp); } -ASSERT (ip->i_d.di_version >1 ); ip->i_d.di_projid = fa->fsx_projid; }
@@ -1651,7 +1650,7 @@ xfs_ioctl_setattr( ip->i_d.di_extsize = fa->fsx_extsize >> mp->m_sb.sb_blocklog; else ip->i_d.di_extsize = 0; -if (ip->i_d.di_version ==3 && +if (xfs_sb_version_has_v3inode (&mp->m_sb) && (ip->i_d.di_flags2 &XFS_DIFLAG2_COWEXTSIZE )) ip->i_d.di_cowextsize = fa->fsx_cowextsize >> mp->m_sb.sb_blocklog;
fs/xfs/xfs_iops.c +1 -1
@@ -517,7 +517,7 @@ xfs_vn_getattr( stat->blocks =XFS_FSB_TO_BB (mp, ip->i_d.di_nblocks + ip->i_delayed_blks); -if (ip->i_d.di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&mp->m_sb)) {if (request_mask &STATX_BTIME ) { stat->result_mask |=STATX_BTIME ; stat->btime.tv_sec = ip->i_d.di_crtime.t_sec;
fs/xfs/xfs_itable.c +1 -1
@@ -110,7 +110,7 @@ xfs_bulkstat_one_int( buf->bs_forkoff =XFS_IFORK_BOFF (ip); buf->bs_version =XFS_BULKSTAT_VERSION_V5 ; -if (dic->di_version ==3 ) { +if (xfs_sb_version_has_v3inode (&mp->m_sb)) {if (dic->di_flags2 &XFS_DIFLAG2_COWEXTSIZE ) buf->bs_cowextsize_blks = dic->di_cowextsize; }
fs/xfs/xfs_log_recover.c +1 -1
@@ -2879,8 +2879,8 @@ xfs_recover_inode_owner_change(return -ENOMEM ;/* instantiate the inode */ +ASSERT (dip->di_version >=3 );xfs_inode_from_disk (ip, dip); -ASSERT (ip->i_d.di_version >=3 ); error =xfs_iformat_fork (ip, dip);if (error)
Signature · 1 line
2.39.1
PATCH 11/17 xfs: fix up non-directory creation in SGID directories
In reply to Chandan Babu RFrom: Christoph Hellwig <[email protected]>
commit 01ea173e103edd5ec41acec65b9261b87e123fc2 upstream.
XFS always inherits the SGID bit if it is set on the parent inode, while the generic inode_init_owner does not do this in a few cases where it can create a possible security problem, see commit 0fa3ecd87848 ("Fix up non-directory creation in SGID directories") for details.
Switch XFS to use the generic helper for the normal path to fix this, just keeping the simple field inheritance open coded for the case of the non-sgid case with the bsdgrpid mount option.
--- fs/xfs/xfs_inode.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
fs/xfs/xfs_inode.c +7 -7
@@ -750,6 +750,7 @@ xfs_ialloc( xfs_buf_t **ialloc_context, xfs_inode_t **ipp) { +struct inode *dir = pip ?VFS_I (pip) :NULL ;struct xfs_mount *mp = tp->t_mountp;xfs_ino_t ino;xfs_inode_t *ip;
@@ -795,18 +796,17 @@ xfs_ialloc(return error;ASSERT (ip !=NULL ); inode =VFS_I (ip); - inode->i_mode = mode;set_nlink (inode, nlink); - inode->i_uid =current_fsuid (); inode->i_rdev = rdev; ip->i_d.di_projid = prid; -if (pip &&XFS_INHERIT_GID (pip)) { - inode->i_gid =VFS_I (pip)->i_gid; -if ((VFS_I (pip)->i_mode &S_ISGID ) &&S_ISDIR (mode)) - inode->i_mode |=S_ISGID ; +if (dir && !(dir->i_mode &S_ISGID ) && + (mp->m_flags &XFS_MOUNT_GRPID )) { + inode->i_uid =current_fsuid (); + inode->i_gid = dir->i_gid; + inode->i_mode = mode; }else { - inode->i_gid =current_fsgid (); +inode_init_owner (inode, dir, mode); }/*
Signature · 1 line
2.39.1
PATCH 12/17 xfs: set inode size after creating symlink
In reply to Chandan Babu RFrom: Jeffrey Mitchell <[email protected]>
commit 8aa921a95335d0a8c8e2be35a44467e7c91ec3e4 upstream.
When XFS creates a new symlink, it writes its size to disk but not to the VFS inode. This causes i_size_read() to return 0 for that symlink until it is re-read from disk, for example when the system is rebooted.
I found this inconsistency while protecting directories with eCryptFS. The command "stat path/to/symlink/in/ecryptfs" will report "Size: 0" if the symlink was created after the last reboot on an XFS root.
Call i_size_write() in xfs_symlink()
--- fs/xfs/xfs_symlink.c | 1 + 1 file changed, 1 insertion(+)
fs/xfs/xfs_symlink.c +1 -0
@@ -314,6 +314,7 @@ xfs_symlink( }ASSERT (pathlen ==0 ); } +i_size_write (VFS_I (ip), ip->i_d.di_size);/* * Create the directory entry for the symlink.
Signature · 1 line
2.39.1
PATCH 13/17 xfs: report corruption only as a regular error
In reply to Chandan Babu RFrom: "Darrick J. Wong" <[email protected]>
commit 6519f708cc355c4834edbe1885c8542c0e7ef907 uptream.
[ Slightly modify fs/xfs/xfs_linux.h to resolve merge conflicts ]
Redefine XFS_IS_CORRUPT so that it reports corruptions only via xfs_corruption_report. Since these are on-disk contents (and not checks of internal state), we don't ever want to panic the kernel. This also amends the corruption report to recommend unmounting and running xfs_repair.
--- fs/xfs/xfs_error.c | 2 +- fs/xfs/xfs_linux.h | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletion(-)
fs/xfs/xfs_error.c +1 -1
@@ -335,7 +335,7 @@ xfs_corruption_error( int linenum, xfs_failaddr_t failaddr) { -if (level <= xfs_error_level) +if (buf && level <= xfs_error_level)xfs_hex_dump (buf, bufsize);xfs_error_report (tag, level, mp, filename, linenum, failaddr);xfs_alert (mp,"Corruption detected. Unmount and run xfs_repair" );
fs/xfs/xfs_linux.h +6 -0
@@ -217,6 +217,12 @@ int xfs_rw_bdev(struct block_device *bdev, sector_t sector, unsigned int count,#endif /* XFS_WARN */ #endif /* DEBUG */ +#define XFS_IS_CORRUPT (mp, expr) \ + (unlikely(expr) ? xfs_corruption_error(#expr, XFS_ERRLEVEL_LOW, (mp), \ + NULL, 0, __FILE__, __LINE__, \ + __this_address), \ + true : false) +#define STATIC static noinline#ifdef CONFIG_XFS_RT
Signature · 1 line
2.39.1
PATCH 14/17 xfs: shut down the filesystem if we screw up quota reservation
In reply to Chandan Babu RFrom: "Darrick J. Wong" <[email protected]>
commit 2a4bdfa8558ca2904dc17b83497dc82aa7fc05e9 upstream.
If we ever screw up the quota reservations enough to trip the assertions, something's wrong with the quota code. Shut down the filesystem when this happens, because this is corruption.
--- fs/xfs/xfs_trans_dquot.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-)
fs/xfs/xfs_trans_dquot.c +13 -3
@@ -15,6 +15,7 @@#include "xfs_trans_priv.h" #include "xfs_quota.h" #include "xfs_qm.h" +#include "xfs_error.h" STATIC void xfs_trans_alloc_dqinfo(xfs_trans_t *);
@@ -700,9 +701,14 @@ xfs_trans_dqresv(XFS_TRANS_DQ_RES_INOS , ninos); } -ASSERT (dqp->q_res_bcount >=be64_to_cpu (dqp->q_core.d_bcount)); -ASSERT (dqp->q_res_rtbcount >=be64_to_cpu (dqp->q_core.d_rtbcount)); -ASSERT (dqp->q_res_icount >=be64_to_cpu (dqp->q_core.d_icount)); + +if (XFS_IS_CORRUPT (mp, + dqp->q_res_bcount <be64_to_cpu (dqp->q_core.d_bcount)) || +XFS_IS_CORRUPT (mp, + dqp->q_res_rtbcount <be64_to_cpu (dqp->q_core.d_rtbcount)) || +XFS_IS_CORRUPT (mp, + dqp->q_res_icount <be64_to_cpu (dqp->q_core.d_icount))) + goto error_corrupt;xfs_dqunlock (dqp);return 0 ;
@@ -712,6 +718,10 @@ xfs_trans_dqresv(if (flags &XFS_QMOPT_ENOSPC )return -ENOSPC ;return -EDQUOT ; +error_corrupt: +xfs_dqunlock (dqp); +xfs_force_shutdown (mp,SHUTDOWN_CORRUPT_INCORE ); +return -EFSCORRUPTED ; }
Signature · 1 line
2.39.1
PATCH 15/17 xfs: consider shutdown in bmapbt cursor delete assert
In reply to Chandan Babu RFrom: Brian Foster <[email protected]>
commit 1cd738b13ae9b29e03d6149f0246c61f76e81fcf upstream.
[ Slightly modify fs/xfs/libxfs/xfs_btree.c to resolve merge conflicts ]
The assert in xfs_btree_del_cursor() checks that the bmapbt block allocation field has been handled correctly before the cursor is freed. This field is used for accurate calculation of indirect block reservation requirements (for delayed allocations), for example. generic/019 reproduces a scenario where this assert fails because the filesystem has shutdown while in the middle of a bmbt record insertion. This occurs after a bmbt block has been allocated via the cursor but before the higher level bmap function (i.e. xfs_bmap_add_extent_hole_real()) completes and resets the field.
Update the assert to accommodate the transient state if the filesystem has shutdown. While here, clean up the indentation and comments in the function.
--- fs/xfs/libxfs/xfs_btree.c | 30 +++++++++++------------------- 1 file changed, 11 insertions(+), 19 deletions(-)
fs/xfs/libxfs/xfs_btree.c +11 -19
@@ -354,20 +354,17 @@ xfs_btree_free_block(*/ void xfs_btree_del_cursor ( -xfs_btree_cur_t *cur,/* btree cursor */ -int error)/* del because of error */ +struct xfs_btree_cur *cur,/* btree cursor */ +int error)/* del because of error */ { -int i;/* btree level */ +int i;/* btree level */ /* -* Clear the buffer pointers, and release the buffers. -* If we're doing this in the face of an error, we -* need to make sure to inspect all of the entries -* in the bc_bufs array for buffers to be unlocked. -* This is because some of the btree code works from -* level n down to 0, and if we get an error along -* the way we won't have initialized all the entries -* down to 0. +* Clear the buffer pointers and release the buffers. If we're doing +* this because of an error, inspect all of the entries in the bc_bufs +* array for buffers to be unlocked. This is because some of the btree +* code works from level n down to 0, and if we get an error along the +* way we won't have initialized all the entries down to 0. */ for (i =0 ; i < cur->bc_nlevels; i++) {if (cur->bc_bufs[i])
@@ -375,15 +372,10 @@ xfs_btree_del_cursor(else if (!error)break ; } -/* -* Can't free a bmap cursor without having dealt with the -* allocated indirect blocks' accounting. -*/ +ASSERT (cur->bc_btnum !=XFS_BTNUM_BMAP || - cur->bc_private.b.allocated ==0 ); -/* -* Free the cursor. -*/ + cur->bc_private.b.allocated ==0 || +XFS_FORCED_SHUTDOWN (cur->bc_mp));kmem_zone_free (xfs_btree_cur_zone, cur); }
Signature · 1 line
2.39.1
PATCH 16/17 xfs: don't reuse busy extents on extent trim
In reply to Chandan Babu RFrom: Brian Foster <[email protected]>
commit 06058bc40534530e617e5623775c53bb24f032cb upstream.
Freed extents are marked busy from the point the freeing transaction commits until the associated CIL context is checkpointed to the log. This prevents reuse and overwrite of recently freed blocks before the changes are committed to disk, which can lead to corruption after a crash. The exception to this rule is that metadata allocation is allowed to reuse busy extents because metadata changes are also logged.
As of commit 97d3ac75e5e0 ("xfs: exact busy extent tracking"), XFS has allowed modification or complete invalidation of outstanding busy extents for metadata allocations. This implementation assumes that use of the associated extent is imminent, which is not always the case. For example, the trimmed extent might not satisfy the minimum length of the allocation request, or the allocation algorithm might be involved in a search for the optimal result based on locality.
generic/019 reproduces a corruption caused by this scenario. First, a metadata block (usually a bmbt or symlink block) is freed from an inode. A subsequent bmbt split on an unrelated inode attempts a near mode allocation request that invalidates the busy block during the search, but does not ultimately allocate it. Due to the busy state invalidation, the block is no longer considered busy to subsequent allocation. A direct I/O write request immediately allocates the block and writes to it. Finally, the filesystem crashes while in a state where the initial metadata block free had not committed to the on-disk log. After recovery, the original metadata block is in its original location as expected, but has been corrupted by the aforementioned dio.
This demonstrates that it is fundamentally unsafe to modify busy extent state for extents that are not guaranteed to be allocated. This applies to pretty much all of the code paths that currently trim busy extents for one reason or another. Therefore to address this problem, drop the reuse mechanism from the busy extent trim path. This code already knows how to return partial non-busy ranges of the targeted free extent and higher level code tracks the busy state of the allocation attempt. If a block allocation fails where one or more candidate extents is busy, we force the log and retry the allocation.
--- fs/xfs/xfs_extent_busy.c | 14 -------------- 1 file changed, 14 deletions(-)
fs/xfs/xfs_extent_busy.c +0 -14
@@ -344,7 +344,6 @@ xfs_extent_busy_trim(ASSERT (*len >0 );spin_lock (&args->pag->pagb_lock); -restart: fbno = *bno; flen = *len; rbp = args->pag->pagb_tree.rb_node;
@@ -363,19 +362,6 @@ xfs_extent_busy_trim(continue ; } -/* -* If this is a metadata allocation, try to reuse the busy -* extent instead of trimming the allocation. -*/ -if (!xfs_alloc_is_userdata (args->datatype) && - !(busyp->flags &XFS_EXTENT_BUSY_DISCARDED )) { -if (!xfs_extent_busy_update_extent (args->mp, args->pag, - busyp, fbno, flen, - false)) - goto restart; -continue ; - } -if (bbno <= fbno) {/* start overlap */
Signature · 1 line
2.39.1
PATCH 17/17 xfs: force log and push AIL to clear pinned inodes when aborting mount
In reply to Chandan Babu RFrom: "Darrick J. Wong" <[email protected]>
commit d336f7ebc65007f5831e2297e6f3383ae8dbf8ed upstream.
[ Slightly modify fs/xfs/xfs_mount.c to resolve merge conflicts ]
If we allocate quota inodes in the process of mounting a filesystem but then decide to abort the mount, it's possible that the quota inodes are sitting around pinned by the log. Now that inode reclaim relies on the AIL to flush inodes, we have to force the log and push the AIL in between releasing the quota inodes and kicking off reclaim to tear down all the incore inodes. Do this by extracting the bits we need from the unmount path and reusing them. As an added bonus, failed writes during a failed mount will not retry forever now.
This was originally found during a fuzz test of metadata directories (xfs/1546), but the actual symptom was that reclaim hung up on the quota inodes.
--- fs/xfs/xfs_mount.c | 90 +++++++++++++++++++++++----------------------- 1 file changed, 44 insertions(+), 46 deletions(-)
fs/xfs/xfs_mount.c +44 -46
@@ -674,6 +674,47 @@ xfs_check_summary_counts(return xfs_initialize_perag_data (mp, mp->m_sb.sb_agcount); } +/* +* Flush and reclaim dirty inodes in preparation for unmount. Inodes and +* internal inode structures can be sitting in the CIL and AIL at this point, +* so we need to unpin them, write them back and/or reclaim them before unmount +* can proceed. +* +* An inode cluster that has been freed can have its buffer still pinned in +* memory because the transaction is still sitting in a iclog. The stale inodes +* on that buffer will be pinned to the buffer until the transaction hits the +* disk and the callbacks run. Pushing the AIL will skip the stale inodes and +* may never see the pinned buffer, so nothing will push out the iclog and +* unpin the buffer. +* +* Hence we need to force the log to unpin everything first. However, log +* forces don't wait for the discards they issue to complete, so we have to +* explicitly wait for them to complete here as well. +* +* Then we can tell the world we are unmounting so that error handling knows +* that the filesystem is going away and we should error out anything that we +* have been retrying in the background. This will prevent never-ending +* retries in AIL pushing from hanging the unmount. +* +* Finally, we can push the AIL to clean all the remaining dirty objects, then +* reclaim the remaining inodes that are still in memory at this point in time. +*/ +static void +xfs_unmount_flush_inodes ( +struct xfs_mount *mp) +{ +xfs_log_force (mp,XFS_LOG_SYNC ); +xfs_extent_busy_wait_all (mp); +flush_workqueue (xfs_discard_wq); + + mp->m_flags |=XFS_MOUNT_UNMOUNTING ; + +xfs_ail_push_all_sync (mp->m_ail); +cancel_delayed_work_sync (&mp->m_reclaim_work); +xfs_reclaim_inodes (mp,SYNC_WAIT ); +xfs_health_unmount (mp); +} +/* * This function does the following on an initial mount of a file system: * - reads the superblock from disk and init the mount struct
@@ -1047,7 +1088,7 @@ xfs_mountfs(/* Clean out dquots that might be in memory after quotacheck. */ xfs_qm_unmount (mp);/* -* Cancel all delayed reclaim work and reclaim the inodes directly. +* Flush all inode reclamation work and flush the log. * We have to do this /after/ rtunmount and qm_unmount because those * two will have scheduled delayed reclaim for the rt/quota inodes. *
@@ -1057,11 +1098,8 @@ xfs_mountfs(* qm_unmount_quotas and therefore rely on qm_unmount to release the * quota inodes. */ -cancel_delayed_work_sync (&mp->m_reclaim_work); -xfs_reclaim_inodes (mp,SYNC_WAIT ); -xfs_health_unmount (mp); +xfs_unmount_flush_inodes (mp); out_log_dealloc: - mp->m_flags |=XFS_MOUNT_UNMOUNTING ;xfs_log_mount_cancel (mp); out_fail_wait:if (mp->m_logdev_targp && mp->m_logdev_targp != mp->m_ddev_targp)
@@ -1102,47 +1140,7 @@ xfs_unmountfs(xfs_rtunmount_inodes (mp);xfs_irele (mp->m_rootip); -/* -* We can potentially deadlock here if we have an inode cluster -* that has been freed has its buffer still pinned in memory because -* the transaction is still sitting in a iclog. The stale inodes -* on that buffer will have their flush locks held until the -* transaction hits the disk and the callbacks run. the inode -* flush takes the flush lock unconditionally and with nothing to -* push out the iclog we will never get that unlocked. hence we -* need to force the log first. -*/ -xfs_log_force (mp,XFS_LOG_SYNC ); - -/* -* Wait for all busy extents to be freed, including completion of -* any discard operation. -*/ -xfs_extent_busy_wait_all (mp); -flush_workqueue (xfs_discard_wq); - -/* -* We now need to tell the world we are unmounting. This will allow -* us to detect that the filesystem is going away and we should error -* out anything that we have been retrying in the background. This will -* prevent neverending retries in AIL pushing from hanging the unmount. -*/ - mp->m_flags |=XFS_MOUNT_UNMOUNTING ; - -/* -* Flush all pending changes from the AIL. -*/ -xfs_ail_push_all_sync (mp->m_ail); - -/* -* And reclaim all inodes. At this point there should be no dirty -* inodes and none should be pinned or locked, but use synchronous -* reclaim just to be sure. We can stop background inode reclaim -* here as well if it is still running. -*/ -cancel_delayed_work_sync (&mp->m_reclaim_work); -xfs_reclaim_inodes (mp,SYNC_WAIT ); -xfs_health_unmount (mp); +xfs_unmount_flush_inodes (mp);xfs_qm_unmount (mp);
Signature · 1 line
2.39.1
Greg KH
maintainerQuoted 7 lines · Chandan Babu R
Hi Greg,
This 5.4.y backport series contains XFS fixes from v5.11 & v5.12. The patchset has been acked by Darrick.
As a side note, where applicable, patches have been cherry picked from 5.10.y rather than from v5.11/v5.12.
all now queued up, thanks.
greg k-h
Greg KH
maintainerQuoted 10 lines · Greg KH
Quoted 7 lines · Chandan Babu R
Hi Greg,
This 5.4.y backport series contains XFS fixes from v5.11 & v5.12. The patchset has been acked by Darrick.
As a side note, where applicable, patches have been cherry picked from 5.10.y rather than from v5.11/v5.12.
all now queued up, thanks.
Wait, did you miss commit ada49d64fb35 ("xfs: fix forkoff miscalculation related to XFS_LITINO(mp)")?
Quoted 7 lines · Greg KH
Quoted 6 lines · Greg KH
Quoted 5 lines · Chandan Babu R
Hi Greg,
This 5.4.y backport series contains XFS fixes from v5.11 & v5.12. The patchset has been acked by Darrick.
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines
As a side note, where applicable, patches have been cherry picked from 5.10.y rather than from v5.11/v5.12.
all now queued up, thanks.
Wait, did you miss commit
ada49d64fb35("xfs: fix forkoff miscalculation related toXFS_LITINO(mp)")?
Yes, unfortunately I missed that commit.
I always invoked my "find fixes" script as shown below,
# find-fixes-for-xfs-fixes.sh v5.11 v6.2 ~/junk/commits-fixes.log
The commits to be searched for "fixes for backported commits" spanned from the kernel from which the fixes were backported to the latest upstream kernel. I passed v5.11 as the argument since I was backporting patches from v5.11 and v5.12.
However, e9e2eae89ddb ("xfs: only check the superblock version for dinode size calculation") was from v5.7 and hence I missed ada49d64fb35.
# find-fixes-for-xfs-fixes.sh v5.4 v6.2 ~/junk/commits-fixes.log ada49d64fb3538144192181db05de17e2ffc3551 e9e2eae89ddb
I will test the fix and send it out soon.
Signature · 1 line
chandan