The commit xfs: fix inode fork extent count overflow (3f8a4f1d876d3e3e49e50b0396eaffcc4ba71b08) mentions that 10 billion data fork extents should be possible to create. However the corresponding on-disk field has a signed 32-bit type. Hence this patchset extends the per-inode data fork extent counter to 64 bits (out of which 48 bits are used to store the extent count). Also, XFS has an attribute fork extent counter which is 16 bits wide. A workload that, 1. Creates 1 million 255-byte sized xattrs, 2. Deletes 50% of these xattrs in an alternating manner, 3. Tries to insert 400,000 new 255-byte sized xattrs causes the xattr extent counter to overflow. Dave tells me that there are instances where a single file has more than 100 million hardlinks. With parent pointers being stored in xattrs, we will overflow the signed 16-bits wide attribute extent counter when large number of hardlinks are created. Hence this patchset extends the on-disk field to 32-bits. The following changes are made to accomplish this, 1. A 64-bit inode field is carved out of existing di_pad and di_flushiter fields to hold the 64-bit data fork extent counter. 2. The existing 32-bit inode data fork extent counter will be used to hold the attribute fork extent counter. 3. A new incompat superblock flag to prevent older kernels from mounting the filesystem. The patchset has been tested by executing xfstests with the following mkfs.xfs options, 1. -m crc=0 -b size=1k 2. -m crc=0 -b size=4k 3. -m crc=0 -b size=512 4. -m rmapbt=1,reflink=1 -b size=1k 5. -m rmapbt=1,reflink=1 -b size=4k Each of the above test scenarios were executed on the following combinations (For V4 FS test scenario, the last combination was omitted). |---------------------------+-----------| | Xfsprogs | Kernel | |---------------------------+-----------| | Unpatched | Patched | | Patched (disable nrext64) | Unpatched | | Patched (disable nrext64) | Patched | | Patched (enable nrext64) | Patched | |---------------------------+-----------| I have also written tests to check if the correct extent counter fields are updated with/without the new incompat flag and to verify upgrading older fs instances to support large extent counters. I have also fixed xfs/270 test to work with the new code base. These patches can also be obtained from https://github.com/chandanr/linux.git at branch xfs-incompat-extend-extcnt-v8. Changelog: V7 -> V8: 1. Do not roll a transaction after upgrading an inode to "Large extent counter" feature. Any transaction which can cause an inode's extent counter to change, will have included the space required to log the inode in its transaction reservation calculation. This means that the patch "xfs: xfs_growfs_rt_alloc: Unlock inode explicitly rather than through iop_committing()" is no longer required. 2. Use XFS_MAX_EXTCNT_DATA_FORK_LARGE & XFS_MAX_EXTCNT_ATTR_FORK_LARGE to represent large extent counter limits. Similarly, use XFS_MAX_EXTCNT_DATA_FORK_SMALL & XFS_MAX_EXTCNT_ATTR_FORK_SMALL to represent previously defined extent counter limits. 3. Decouple XFS_IBULK flags from XFS_IWALK flags in a separate patch. 4. Bulkstat operation now returns XFS_MAX_EXTCNT_DATA_FORK_SMALL as the extent count if data fork extent count exceeds XFS_MAX_EXTCNT_DATA_FORK_SMALL and userspace program isn't aware of large extent counters. V6 -> V7: 1. Address the following review comments from V6, - Revert xfs_ibulk->flags to "unsigned int" type. - Fix definition of XFS_IBULK_NREXT64 to be independent of IWALK flags. - Fix possible double free of transaction handle in xfs_growfs_rt_alloc(). V5 -> V6: 1. Rebase on Linux-v5.17-rc4. 2. Upgrade inodes to use large extent counters from within a transaction context. V4 -> V5: 1. Rebase on xfs-linux/for-next. 2. Use howmany_64() to compute height of maximum bmbt tree. 3. Rename disk and log inode's di_big_dextcnt to di_big_nextents. 4. Rename disk and log inode's di_big_aextcnt to di_big_anextents. 5. Since XFS_IBULK_NREXT64 is not associated with inode walking functionality, define it as the 32nd bit and mask it when passing xfs_ibulk->flags to xfs_iwalk() function. V3 -> V4: 1. Rebase patchset on xfs-linux/for-next branch. 2. Carve out a 64-bit inode field out of the existing di_pad and di_flushiter fields to hold the 64-bit data fork extent counter. 3. Use the existing 32-bit inode data fork extent counter to hold the attr fork extent counter. 4. Verify the contents of newly introduced inode fields immediately after the inode has been read from the disk. 5. Upgrade inodes to be able to hold large extent counters when reading them from disk. 6. Use XFS_BULK_IREQ_NREXT64 as the flag that userspace can use to indicate that it can read 64-bit data fork extent counter. 7. Bulkstat ioctl returns -EOVERFLOW when userspace is not capable of working with large extent counters and inode's data fork extent count is larger than INT32_MAX. V2 -> V3: 1. Define maximum extent length as a function of BMBT_BLOCKCOUNT_BITLEN. 2. Introduce xfs_iext_max_nextents() function in the patch series before renaming MAXEXTNUM/MAXAEXTNUM. This is done to reduce proliferation of macros indicating maximum extent count for data and attribute forks. 3. Define xfs_dfork_nextents() as an inline function. 4. Use xfs_rfsblock_t as the data type for variables that hold block count. 5. xfs_dfork_nextents() now returns -EFSCORRUPTED when an invalid fork is passed as an argument. 6. The following changes are done to enable bulkstat ioctl to report 64-bit extent counters, - Carve out a new 64-bit field xfs_bulkstat->bs_extents64 from xfs_bulkstat->bs_pad[]. - Carve out a new 64-bit field xfs_bulk_ireq->bulkstat_flags from xfs_bulk_ireq->reserved[] to hold bulkstat specific operational flags. Introduce XFS_IBULK_NREXT64 flag to indicate that userspace has the necessary infrastructure to receive 64-bit extent counters. - Define the new flag XFS_BULK_IREQ_BULKSTAT for userspace to indicate that xfs_bulk_ireq->bulkstat_flags has valid flags set. 7. Rename the incompat flag from XFS_SB_FEAT_INCOMPAT_EXTCOUNT_64BIT to XFS_SB_FEAT_INCOMPAT_NREXT64. 8. Add a new helper function xfs_inode_to_disk_iext_counters() to convert from incore inode extent counters to ondisk inode extent counters. 9. Reuse XFS_ERRTAG_REDUCE_MAX_IEXTENTS error tag to skip reporting inodes with more than 10 extents when bulkstat ioctl is invoked by userspace. 10. Introduce the new per-inode XFS_DIFLAG2_NREXT64 flag to indicate that the inode uses 64-bit extent counter. This is used to allow administrators to upgrade existing filesystems. 11. Export presence of XFS_SB_FEAT_INCOMPAT_NREXT64 feature to userspace via XFS_IOC_FSGEOMETRY ioctl. V1 -> V2: 1. Rebase patches on top of Darrick's btree-dynamic-depth branch. 2. Add new bulkstat ioctl version to support 64-bit data fork extent counter field. 3. Introduce new error tag to verify if the old bulkstat ioctls skip reporting inodes with large data fork extent counters. Chandan Babu R (19): xfs: Move extent count limits to xfs_format.h xfs: Define max extent length based on on-disk format definition xfs: Introduce xfs_iext_max_nextents() helper xfs: Use xfs_extnum_t instead of basic data types xfs: Introduce xfs_dfork_nextents() helper xfs: Use basic types to define xfs_log_dinode's di_nextents and di_anextents xfs: Promote xfs_extnum_t and xfs_aextnum_t to 64 and 32-bits respectively xfs: Introduce XFS_SB_FEAT_INCOMPAT_NREXT64 and associated per-fs feature bit xfs: Introduce XFS_FSOP_GEOM_FLAGS_NREXT64 xfs: Introduce XFS_DIFLAG2_NREXT64 and associated helpers xfs: Use uint64_t to count maximum blocks that can be used by BMBT xfs: Introduce macros to represent new maximum extent counts for data/attr forks xfs: Replace numbered inode recovery error messages with descriptive ones xfs: Introduce per-inode 64-bit extent counters xfs: Directory's data fork extent counter can never overflow xfs: Conditionally upgrade existing inodes to use large extent counters xfs: Decouple XFS_IBULK flags from XFS_IWALK flags xfs: Enable bulkstat ioctl to support 64-bit per-inode extent counters xfs: Add XFS_SB_FEAT_INCOMPAT_NREXT64 to the list of supported flags fs/xfs/libxfs/xfs_alloc.c | 2 +- fs/xfs/libxfs/xfs_attr.c | 9 ++- fs/xfs/libxfs/xfs_bmap.c | 112 +++++++++++++------------- fs/xfs/libxfs/xfs_bmap_btree.c | 3 +- fs/xfs/libxfs/xfs_format.h | 80 +++++++++++++++--- fs/xfs/libxfs/xfs_fs.h | 21 ++++- fs/xfs/libxfs/xfs_ialloc.c | 2 + fs/xfs/libxfs/xfs_inode_buf.c | 80 ++++++++++++++---- fs/xfs/libxfs/xfs_inode_fork.c | 42 ++++++++-- fs/xfs/libxfs/xfs_inode_fork.h | 63 ++++++++++++++- fs/xfs/libxfs/xfs_log_format.h | 33 +++++++- fs/xfs/libxfs/xfs_sb.c | 5 ++ fs/xfs/libxfs/xfs_trans_resv.c | 11 +-- fs/xfs/libxfs/xfs_types.h | 11 +-- fs/xfs/scrub/bmap.c | 2 +- fs/xfs/scrub/inode.c | 20 ++--- fs/xfs/xfs_bmap_item.c | 8 +- fs/xfs/xfs_bmap_util.c | 57 ++++++++++--- fs/xfs/xfs_dquot.c | 9 ++- fs/xfs/xfs_inode.c | 36 ++++++++- fs/xfs/xfs_inode.h | 5 ++ fs/xfs/xfs_inode_item.c | 23 +++++- fs/xfs/xfs_inode_item_recover.c | 138 +++++++++++++++++++++++--------- fs/xfs/xfs_ioctl.c | 3 + fs/xfs/xfs_iomap.c | 45 +++++++---- fs/xfs/xfs_itable.c | 19 ++++- fs/xfs/xfs_itable.h | 4 +- fs/xfs/xfs_iwalk.h | 2 +- fs/xfs/xfs_mount.h | 2 + fs/xfs/xfs_reflink.c | 17 +++- fs/xfs/xfs_rtalloc.c | 9 ++- fs/xfs/xfs_symlink.c | 8 ++ fs/xfs/xfs_trace.h | 4 +- 33 files changed, 679 insertions(+), 206 deletions(-) -- 2.30.2
