Hi Linus,
Please pull this set of new code for 5.5. For this release, we changed
quite a few things. Highlights:
- Fixed some long tail latency problems in the block allocator;
- Removed some long deprecated (and for the past several years no-op)
mount options and ioctls;
- Strengthened the extended attribute and directory verifiers;
- Audited and fixed all the places where we could return EFSCORRUPTED
without logging anything;
- Refactored the old SGI space allocation ioctls to make the equivalent
fallocate calls;
- Fixed a race between fallocate and directio;
- Fixed an integer overflow when files have more than a few billion(!)
extents;
- Fixed a longstanding bug where quota accounting could be incorrect
when performing unwritten extent conversion on a freshly mounted fs;
- Fixed various complaints in scrub about soft lockups and
unresponsiveness to signals;
- De-vtable'd the directory handling code, which should make it faster;
- Converted to the new mount api, for better or for worse;
- Cleaned up some memory leaks;
and quite a lot of other smaller fixes and cleanups.
The branch has survived several days of xfstests runs and merges cleanly
with this morning's master. Please let me know if anything strange
happens.
FYI, Stephen Rothwell reported a merge conflict with the y2038 tree at
the end of October[1]. His resolution looked pretty straightforward,
though the current y2038 for-next branch no longer changes fs/ioctl.c
(and the changes that were in it are not in upstream master), so that
may not be necessary.
[1] https://lore.kernel.org/linux-xfs/[email protected]/
--D
The following changes since commit c039b99792726346ad46ff17c5a5bcb77a5edac4:
iomap: use a srcmap for a read-modify-write I/O (2019-10-21 08:51:59 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git tags/xfs-5.5-merge-16
for you to fetch changes up to 8feb4732ff9f2732354b44c4418569974e2f949c:
xfs: allow parent directory scans to be interrupted with fatal signals (2019-11-27 08:23:14 -0800)
----------------------------------------------------------------
New code for 5.5:
- Fill out the build string
- Prevent inode fork extent count overflows
- Refactor the allocator to reduce long tail latency
- Rework incore log locking a little to reduce spinning
- Break up the xfs_iomap_begin functions into smaller more cohesive
parts
- Fix allocation alignment being dropped too early when the allocation
request is for more blocks than an AG is large
- Other small cleanups
- Clean up file buftarg retrieval helpers
- Hoist the resvsp and unresvsp ioctls to the vfs
- Remove the undocumented biosize mount option, since it has never been
mentioned as existing or supported on linux
- Clean up some of the mount option printing and parsing
- Enhance attr leaf verifier to check block structure
- Check dirent and attr names for invalid characters before passing them
to the vfs
- Refactor open-coded bmbt walking
- Fix a few places where we return EIO instead of EFSCORRUPTED after
failing metadata sanity checks
- Fix a synchronization problem between fallocate and aio dio corrupting
the file length
- Clean up various loose ends in the iomap and bmap code
- Convert to the new mount api
- Make sure we always log something when returning EFSCORRUPTED
- Fix some problems where long running scrub loops could trigger soft
lockup warnings and/or fail to exit due to fatal signals pending
- Fix various Coverity complaints
- Remove most of the function pointers from the directory code to reduce
indirection penalties
- Ensure that dquots are attached to the inode when performing unwritten
extent conversion after io
- Deuglify incore projid and crtime types
- Fix another AGI/AGF locking order deadlock when renaming
- Clean up some quota typedefs
- Remove the FSSETDM ioctls which haven't done anything in 20 years
- Fix some memory leaks when mounting the log fails
- Fix an underflow when updating an xattr leaf freemap
- Remove some trivial wrappers
- Report metadata corruption as an error, not a (potentially) fatal
assertion
- Clean up the dir/attr buffer mapping code
- Allow fatal signals to kill scrub during parent pointer checks
----------------------------------------------------------------
Arnd Bergmann (1):
xfs: avoid time_t in user api
Ben Dooks (Codethink) (1):
xfs: add mising include of xfs_pnfs.h for missing declarations
Brian Foster (13):
xfs: track active state of allocation btree cursors
xfs: introduce allocation cursor data structure
xfs: track allocation busy state in allocation cursor
xfs: track best extent from cntbt lastblock scan in alloc cursor
xfs: refactor cntbt lastblock scan best extent logic into helper
xfs: reuse best extent tracking logic for bnobt scan
xfs: refactor allocation tree fixup code
xfs: refactor and reuse best extent scanning logic
xfs: refactor near mode alloc bnobt scan into separate function
xfs: factor out tree fixup logic into helper
xfs: optimize near mode bnobt scans with concurrent cntbt lookups
xfs: don't set bmapi total block req where minleft is
xfs: fix attr leaf header freemap.size underflow
Carlos Maiolino (3):
xfs: Remove slab init wrappers
xfs: Remove kmem_zone_destroy() wrapper
xfs: Remove kmem_zone_free() wrapper
Christoph Hellwig (112):
xfs: ignore extent size hints for always COW inodes
xfs: pass the correct flag to xlog_write_iclog
xfs: remove the unused ic_io_size field from xlog_in_core
xfs: move the locking from xlog_state_finish_copy to the callers
xfs: call xlog_state_release_iclog with l_icloglock held
xfs: remove dead ifdef XFSERRORDEBUG code
xfs: remove the unused XLOG_STATE_ALL and XLOG_STATE_UNUSED flags
xfs: turn ic_state into an enum
xfs: remove the XLOG_STATE_DO_CALLBACK state
xfs: also call xfs_file_iomap_end_delalloc for zeroing operations
xfs: remove xfs_reflink_dirty_extents
xfs: pass two imaps to xfs_reflink_allocate_cow
xfs: refactor xfs_file_iomap_begin_delay
xfs: fill out the srcmap in iomap_begin
xfs: factor out a helper to calculate the end_fsb
xfs: split out a new set of read-only iomap ops
xfs: move xfs_file_iomap_begin_delay around
xfs: split the iomap ops for buffered vs direct writes
xfs: rename the whichfork variable in xfs_buffered_write_iomap_begin
xfs: cleanup xfs_direct_write_iomap_begin
xfs: improve the IOMAP_NOWAIT check for COW inodes
xfs: mark xfs_buf_free static
xfs: add a xfs_inode_buftarg helper
xfs: use xfs_inode_buftarg in xfs_file_dio_aio_write
xfs: use xfs_inode_buftarg in xfs_file_ioctl
xfs: don't implement XFS_IOC_RESVSP / XFS_IOC_RESVSP64
fs: add generic UNRESVSP and ZERO_RANGE ioctl handlers
xfs: disable xfs_ioc_space for always COW inodes
xfs: consolidate preallocation in xfs_file_fallocate
xfs: simplify setting bio flags
xfs: remove the dsunit and dswidth variables in
xfs: cleanup calculating the stat optimal I/O size
xfs: don't use a different allocsice for -o wsync
xfs: remove the m_readio_* fields in struct xfs_mount
xfs: rename the m_writeio_* fields in struct xfs_mount
xfs: simplify parsing of allocsize mount option
xfs: rename the XFS_MOUNT_DFLT_IOSIZE option to
xfs: reverse the polarity of XFS_MOUNT_COMPAT_IOSIZE
xfs: clean up printing the allocsize option in
xfs: clean up printing inode32/64 in xfs_showargs
xfs: merge xfs_showargs into xfs_fs_show_options
xfs: simplify xfs_iomap_eof_align_last_fsb
xfs: mark xfs_eof_alignment static
xfs: remove the extsize argument to xfs_eof_alignment
xfs: slightly tweak an assert in xfs_fs_map_blocks
xfs: don't log the inode in xfs_fs_map_blocks if it
xfs: simplify the xfs_iomap_write_direct calling
xfs: refactor xfs_bmapi_allocate
xfs: move extent zeroing to xfs_bmapi_allocate
xfs: cleanup use of the XFS_ALLOC_ flags
xfs: move incore structures out of xfs_da_format.h
xfs: use unsigned int for all size values in struct xfs_da_geometry
xfs: refactor btree node scrubbing
xfs: devirtualize ->node_hdr_from_disk
xfs: devirtualize ->node_hdr_to_disk
xfs: add a btree entries pointer to struct xfs_da3_icnode_hdr
xfs: move the node header size to struct xfs_da_geometry
xfs: devirtualize ->leaf_hdr_from_disk
xfs: devirtualize ->leaf_hdr_to_disk
xfs: add an entries pointer to struct xfs_dir3_icleaf_hdr
xfs: move the dir2 leaf header size to struct xfs_da_geometry
xfs: move the max dir2 leaf entries count to struct xfs_da_geometry
xfs: devirtualize ->free_hdr_from_disk
xfs: devirtualize ->free_hdr_to_disk
xfs: make the xfs_dir3_icfree_hdr available to xfs_dir2_node_addname_int
xfs: add a bests pointer to struct xfs_dir3_icfree_hdr
xfs: move the dir2 free header size to struct xfs_da_geometry
xfs: move the max dir2 free bests count to struct xfs_da_geometry
xfs: devirtualize ->db_to_fdb and ->db_to_fdindex
xfs: devirtualize ->sf_get_parent_ino and ->sf_put_parent_ino
xfs: devirtualize ->sf_entsize and ->sf_nextentry
xfs: devirtualize ->sf_get_ino and ->sf_put_ino
xfs: devirtualize ->sf_get_ftype and ->sf_put_ftype
xfs: remove the unused ->data_first_entry_p method
xfs: remove the data_dot_offset field in struct xfs_dir_ops
xfs: remove the data_dotdot_offset field in struct xfs_dir_ops
xfs: remove the ->data_dot_entry_p and ->data_dotdot_entry_p methods
xfs: remove the ->data_unused_p method
xfs: cleanup xfs_dir2_block_getdents
xfs: cleanup xfs_dir2_leaf_getdents
xfs: cleanup xchk_dir_rec
xfs: cleanup xchk_directory_data_bestfree
xfs: cleanup xfs_dir2_block_to_sf
xfs: cleanup xfs_dir2_data_freescan_int
xfs: cleanup __xfs_dir3_data_check
xfs: remove the now unused ->data_entry_p method
xfs: replace xfs_dir3_data_endp with xfs_dir3_data_end_offset
xfs: devirtualize ->data_entsize
xfs: devirtualize ->data_entry_tag_p
xfs: move the dir2 data block fixed offsets to struct xfs_da_geometry
xfs: cleanup xfs_dir2_data_entsize
xfs: devirtualize ->data_bestfree_p
xfs: devirtualize ->data_get_ftype and ->data_put_ftype
xfs: remove the now unused dir ops infrastructure
xfs: merge xfs_dir2_data_freescan and xfs_dir2_data_freescan_int
xfs: always pass a valid hdr to xfs_dir3_leaf_check_int
xfs: remove the unused m_chsize field
xfs: devirtualize ->m_dirnameops
xfs: use a struct timespec64 for the in-core crtime
xfs: merge the projid fields in struct xfs_icdinode
xfs: don't reset the "inode core" in xfs_iread
xfs: remove XFS_IOC_FSSETDM and XFS_IOC_FSSETDM_BY_HANDLE
xfs: simplify mappedbno handling in xfs_da_{get,read}_buf
xfs: refactor xfs_dabuf_map
xfs: improve the xfs_dabuf_map calling conventions
xfs: remove the mappedbno argument to xfs_da_reada_buf
xfs: remove the mappedbno argument to xfs_attr3_leaf_read
xfs: remove the mappedbno argument to xfs_dir3_leaf_read
xfs: remove the mappedbno argument to xfs_dir3_leafn_read
xfs: split xfs_da3_node_read
xfs: remove the mappedbno argument to xfs_da_read_buf
xfs: remove the mappedbno argument to xfs_da_get_buf
Colin Ian King (1):
xfs: remove redundant assignment to variable error
Dan Carpenter (1):
xfs: remove a stray tab in xfs_remount_rw()
Darrick J. Wong (32):
xfs: check attribute leaf block structure
xfs: namecheck attribute names before listing them
xfs: namecheck directory entry names before listing them
xfs: replace -EIO with -EFSCORRUPTED for corrupt metadata
xfs: refactor xfs_bmap_count_blocks using newer btree helpers
xfs: refactor xfs_iread_extents to use xfs_btree_visit_blocks
xfs: relax shortform directory size checks
xfs: constify the buffer pointer arguments to error functions
xfs: always log corruption errors
xfs: decrease indenting problems in xfs_dabuf_map
xfs: add missing assert in xfs_fsmap_owner_from_rmap
xfs: make the assertion message functions take a mount parameter
xfs: add missing early termination checks to record scrubbing functions
xfs: periodically yield scrub threads to the scheduler
xfs: fix missing header includes
xfs: null out bma->prev if no previous extent
xfs: "optimize" buffer item log segment bitmap setting
xfs: range check ri_cnt when recovering log items
xfs: annotate functions that trip static checker locking checks
xfs: clean up weird while loop in xfs_alloc_ag_vextent_near
xfs: refactor "does this fork map blocks" predicate
xfs: convert EIO to EFSCORRUPTED when log contents are invalid
xfs: actually check xfs_btree_check_block return in xfs_btree_islastblock
xfs: attach dquots and reserve quota blocks during unwritten conversion
xfs: attach dquots before performing xfs_swap_extents
xfs: add a XFS_IS_CORRUPT macro
xfs: kill the XFS_WANT_CORRUPT_* macros
xfs: convert open coded corruption check to use XFS_IS_CORRUPT
xfs: fix another missing include
xfs: fix some memory leaks in log recovery
xfs: report corruption only as a regular error
xfs: allow parent directory scans to be interrupted with fatal signals
Dave Chinner (3):
xfs: fix inode fork extent count overflow
xfs: cap longest free extent to maximum allocatable
xfs: properly serialise fallocate against AIO+DIO
Eric Sandeen (2):
xfs: remove unused typedef definitions
xfs: remove unused structure members & simple typedefs
Ian Kent (18):
xfs: remove the biosize mount option
xfs: remove unused struct xfs_mount field m_fsname_len
xfs: use super s_id instead of struct xfs_mount m_fsname
xfs: dont use XFS_IS_QUOTA_RUNNING() for option check
xfs: use kmem functions for struct xfs_mount
xfs: merge freeing of mp names and mp
xfs: add xfs_remount_rw() helper
xfs: add xfs_remount_ro() helper
xfs: refactor suffix_kstrtoint()
xfs: avoid redundant checks when options is empty
xfs: refactor xfs_parseags()
xfs: move xfs_parseargs() validation to a helper
xfs: dont set sb in xfs_mount_alloc()
xfs: switch to use the new mount-api
xfs: move xfs_fc_reconfigure() above xfs_fc_free()
xfs: move xfs_fc_get_tree() above xfs_fc_reconfigure()
xfs: move xfs_fc_parse_param() above xfs_fc_get_tree()
xfs: fold xfs_mount-alloc() into xfs_init_fs_context()
Jan Kara (1):
xfs: Sanity check flags of Q_XQUOTARM call
Joe Perches (1):
xfs: Correct comment tyops -> typos
Pavel Reichl (5):
xfs: remove the xfs_disk_dquot_t and xfs_dquot_t
xfs: remove the xfs_quotainfo_t typedef
xfs: remove the xfs_dq_logitem_t typedef
xfs: remove the xfs_qoff_logitem_t typedef
xfs: Replace function declaration by actual definition
YueHaibing (1):
xfs: remove duplicated include from xfs_dir2_data.c
kaixuxia (2):
xfs: remove the duplicated inode log fieldmask set
xfs: Fix deadlock between AGI and AGF when target_ip exists in xfs_rename()
yu kuai (1):
xfs: include QUOTA, FATAL ASSERT build options in XFS_BUILD_OPTIONS
fs/compat_ioctl.c | 31 +-
fs/ioctl.c | 12 +-
fs/xfs/Makefile | 1 -
fs/xfs/kmem.c | 2 +-
fs/xfs/kmem.h | 30 -
fs/xfs/libxfs/xfs_ag_resv.c | 2 +
fs/xfs/libxfs/xfs_alloc.c | 1236 ++++++++++++++++++++-------------
fs/xfs/libxfs/xfs_alloc.h | 16 +-
fs/xfs/libxfs/xfs_alloc_btree.c | 1 +
fs/xfs/libxfs/xfs_attr.c | 24 +-
fs/xfs/libxfs/xfs_attr_leaf.c | 134 +++-
fs/xfs/libxfs/xfs_attr_leaf.h | 30 +-
fs/xfs/libxfs/xfs_attr_remote.c | 1 +
fs/xfs/libxfs/xfs_bit.c | 1 +
fs/xfs/libxfs/xfs_bmap.c | 686 ++++++++++--------
fs/xfs/libxfs/xfs_btree.c | 97 +--
fs/xfs/libxfs/xfs_btree.h | 37 +-
fs/xfs/libxfs/xfs_da_btree.c | 668 +++++++++---------
fs/xfs/libxfs/xfs_da_btree.h | 73 +-
fs/xfs/libxfs/xfs_da_format.c | 888 ------------------------
fs/xfs/libxfs/xfs_da_format.h | 59 +-
fs/xfs/libxfs/xfs_dir2.c | 72 +-
fs/xfs/libxfs/xfs_dir2.h | 90 +--
fs/xfs/libxfs/xfs_dir2_block.c | 131 ++--
fs/xfs/libxfs/xfs_dir2_data.c | 282 ++++----
fs/xfs/libxfs/xfs_dir2_leaf.c | 307 ++++----
fs/xfs/libxfs/xfs_dir2_node.c | 431 ++++++------
fs/xfs/libxfs/xfs_dir2_priv.h | 114 ++-
fs/xfs/libxfs/xfs_dir2_sf.c | 424 ++++++-----
fs/xfs/libxfs/xfs_dquot_buf.c | 8 +-
fs/xfs/libxfs/xfs_format.h | 14 +-
fs/xfs/libxfs/xfs_fs.h | 4 +-
fs/xfs/libxfs/xfs_ialloc.c | 117 +++-
fs/xfs/libxfs/xfs_iext_tree.c | 2 +-
fs/xfs/libxfs/xfs_inode_buf.c | 21 +-
fs/xfs/libxfs/xfs_inode_buf.h | 5 +-
fs/xfs/libxfs/xfs_inode_fork.c | 22 +-
fs/xfs/libxfs/xfs_inode_fork.h | 18 +-
fs/xfs/libxfs/xfs_log_format.h | 4 +-
fs/xfs/libxfs/xfs_log_recover.h | 4 +-
fs/xfs/libxfs/xfs_refcount.c | 174 +++--
fs/xfs/libxfs/xfs_rmap.c | 377 +++++++---
fs/xfs/libxfs/xfs_rtbitmap.c | 4 +-
fs/xfs/libxfs/xfs_sb.c | 1 +
fs/xfs/libxfs/xfs_trans_inode.c | 8 +-
fs/xfs/libxfs/xfs_trans_resv.c | 6 +-
fs/xfs/libxfs/xfs_types.h | 2 -
fs/xfs/scrub/attr.c | 11 +-
fs/xfs/scrub/bitmap.c | 3 +-
fs/xfs/scrub/common.h | 9 +-
fs/xfs/scrub/dabtree.c | 62 +-
fs/xfs/scrub/dabtree.h | 3 +-
fs/xfs/scrub/dir.c | 132 ++--
fs/xfs/scrub/fscounters.c | 8 +-
fs/xfs/scrub/health.c | 1 +
fs/xfs/scrub/parent.c | 27 +-
fs/xfs/scrub/quota.c | 7 +
fs/xfs/scrub/scrub.c | 1 +
fs/xfs/xfs_acl.c | 18 +-
fs/xfs/xfs_aops.c | 43 +-
fs/xfs/xfs_aops.h | 3 -
fs/xfs/xfs_attr_inactive.c | 76 +-
fs/xfs/xfs_attr_list.c | 75 +-
fs/xfs/xfs_bmap_item.c | 11 +-
fs/xfs/xfs_bmap_util.c | 255 ++-----
fs/xfs/xfs_bmap_util.h | 4 -
fs/xfs/xfs_buf.c | 32 +-
fs/xfs/xfs_buf.h | 1 -
fs/xfs/xfs_buf_item.c | 6 +-
fs/xfs/xfs_dir2_readdir.c | 137 ++--
fs/xfs/xfs_discard.c | 6 +-
fs/xfs/xfs_dquot.c | 46 +-
fs/xfs/xfs_dquot.h | 98 +--
fs/xfs/xfs_dquot_item.h | 34 +-
fs/xfs/xfs_error.c | 31 +-
fs/xfs/xfs_error.h | 33 +-
fs/xfs/xfs_extent_busy.c | 2 +-
fs/xfs/xfs_extfree_item.c | 9 +-
fs/xfs/xfs_file.c | 104 ++-
fs/xfs/xfs_filestream.c | 3 +-
fs/xfs/xfs_fsmap.c | 1 +
fs/xfs/xfs_icache.c | 8 +-
fs/xfs/xfs_icreate_item.c | 2 +-
fs/xfs/xfs_inode.c | 48 +-
fs/xfs/xfs_inode.h | 31 +-
fs/xfs/xfs_inode_item.c | 15 +-
fs/xfs/xfs_ioctl.c | 203 +-----
fs/xfs/xfs_ioctl.h | 7 -
fs/xfs/xfs_ioctl32.c | 49 +-
fs/xfs/xfs_ioctl32.h | 13 +-
fs/xfs/xfs_iomap.c | 862 ++++++++++++-----------
fs/xfs/xfs_iomap.h | 11 +-
fs/xfs/xfs_iops.c | 70 +-
fs/xfs/xfs_itable.c | 6 +-
fs/xfs/xfs_iwalk.c | 3 +-
fs/xfs/xfs_linux.h | 14 +-
fs/xfs/xfs_log.c | 434 +++++-------
fs/xfs/xfs_log_cil.c | 6 +-
fs/xfs/xfs_log_priv.h | 33 +-
fs/xfs/xfs_log_recover.c | 148 ++--
fs/xfs/xfs_message.c | 22 +-
fs/xfs/xfs_message.h | 6 +-
fs/xfs/xfs_mount.c | 58 +-
fs/xfs/xfs_mount.h | 57 +-
fs/xfs/xfs_pnfs.c | 56 +-
fs/xfs/xfs_qm.c | 67 +-
fs/xfs/xfs_qm.h | 6 +-
fs/xfs/xfs_qm_bhv.c | 8 +-
fs/xfs/xfs_qm_syscalls.c | 139 ++--
fs/xfs/xfs_quotaops.c | 3 +
fs/xfs/xfs_refcount_item.c | 9 +-
fs/xfs/xfs_reflink.c | 138 +---
fs/xfs/xfs_reflink.h | 4 +-
fs/xfs/xfs_rmap_item.c | 13 +-
fs/xfs/xfs_rtalloc.c | 3 +-
fs/xfs/xfs_super.c | 1466 +++++++++++++++++++--------------------
fs/xfs/xfs_super.h | 10 +
fs/xfs/xfs_symlink.c | 1 +
fs/xfs/xfs_symlink.h | 2 +-
fs/xfs/xfs_trace.h | 35 +-
fs/xfs/xfs_trans.c | 2 +-
fs/xfs/xfs_trans_ail.c | 10 +-
fs/xfs/xfs_trans_dquot.c | 56 +-
fs/xfs/xfs_xattr.c | 1 +
include/linux/falloc.h | 3 +
include/linux/fs.h | 2 +-
126 files changed, 5833 insertions(+), 6270 deletions(-)
delete mode 100644 fs/xfs/libxfs/xfs_da_format.c
The pull request you sent on Sun, 1 Dec 2019 10:48:14 -0800:
> git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git tags/xfs-5.5-merge-16
has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/97eeb4d9d755605385fa329da9afa38729f3413c
Thank you!
--
Deet-doot-dot, I am a bot.
https://korg.wiki.kernel.org/userdoc/prtracker
On Sun, Dec 1, 2019 at 10:48 AM Darrick J. Wong <[email protected]> wrote:
>
> FYI, Stephen Rothwell reported a merge conflict with the y2038 tree at
> the end of October[1]. His resolution looked pretty straightforward,
> though the current y2038 for-next branch no longer changes fs/ioctl.c
> (and the changes that were in it are not in upstream master), so that
> may not be necessary.
The changes and conflicts are definitely still there (now upstream),
I'm not sure what made you not see them. But thanks for the note, I
compared my end result with linux-next to verify.
My resolution is different from Stephen's. All my non-x86-64 FS_IOC_*
cases just do "goto found_handler", because the compat case is
identical for the native case outside of the special x86-64 alignment
behavior, and I think that's what Arnd meant to happen.
There was some other minor difference too, but it's also possible I
could have messed up, so cc'ing Stephen and Arnd on this just in case
they have comments.
Linus
On Mon, Dec 02, 2019 at 03:22:31PM -0800, Linus Torvalds wrote:
> On Sun, Dec 1, 2019 at 10:48 AM Darrick J. Wong <[email protected]> wrote:
> >
> > FYI, Stephen Rothwell reported a merge conflict with the y2038 tree at
> > the end of October[1]. His resolution looked pretty straightforward,
> > though the current y2038 for-next branch no longer changes fs/ioctl.c
> > (and the changes that were in it are not in upstream master), so that
> > may not be necessary.
>
> The changes and conflicts are definitely still there (now upstream),
> I'm not sure what made you not see them. But thanks for the note, I
> compared my end result with linux-next to verify.
Aha! I pulled master yesterday morning, tried a test merge with xfs,
saw the lack of merge conflicts, and sent you the xfs pull request. A
few hours later you pulled in the compat ioctl changes from Arnd's git
tree, but the branch in his repo that feeds the -next tree doesn't
contain the compat ioctl changes, so I assumed that meant he wasn't
going to send them for 5.5... and then thought better of myself and
attached an FYI anyway.
> My resolution is different from Stephen's. All my non-x86-64 FS_IOC_*
> cases just do "goto found_handler", because the compat case is
> identical for the native case outside of the special x86-64 alignment
> behavior, and I think that's what Arnd meant to happen.
Yeah, that looks correct to me. Stephen's solution backed out the
changes that Arnd made for the !x86_64 compat ioctl case, so I or
someone would have had to re-apply them.
> There was some other minor difference too, but it's also possible I
> could have messed up, so cc'ing Stephen and Arnd on this just in case
> they have comments.
<nod> Thanks for sorting this out.
--D
>
>
> Linus
On Tue, Dec 3, 2019 at 12:59 AM Darrick J. Wong <[email protected]> wrote:
> On Mon, Dec 02, 2019 at 03:22:31PM -0800, Linus Torvalds wrote:
> Yeah, that looks correct to me. Stephen's solution backed out the
> changes that Arnd made for the !x86_64 compat ioctl case, so I or
> someone would have had to re-apply them.
>
> > There was some other minor difference too, but it's also possible I
> > could have messed up, so cc'ing Stephen and Arnd on this just in case
> > they have comments.
>
> <nod> Thanks for sorting this out.
Yes, this is the right solution. Note that this part of the series originally
came from Al Viro, but I added more patches following the same idea.
While working on a follow-up series, I now also noticed that the
FS_IOC_RESVSP compat handler has always been broken for
x32 user space. I don't think anyone cares, but my series in [1]
addresses this as well by handling FS_IOC_RESVSP/
FS_IOC_UNRESVSP/FS_IOC_ZERO_RANGE on x86-64 as
well, in addition to FS_IOC_RESVSP_32 etc.
Arnd
[1] https://git.kernel.org/pub/scm/linux/kernel/git/arnd/playground.git
compat-ioctl-endgame