summaryrefslogtreecommitdiff
path: root/fs
AgeCommit message (Collapse)Author
2017-11-13ceph: disable cached readdir after dropping positive dentryYan, Zheng
Ideally CEPH_CAP_FILE_SHARED should have been revoked before postive dentry get dropped. But if something goes wrong, later cached readdir may dereference the dropped dentry. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13ceph: fix bool initialization/comparisonThomas Meyer
Bool initializations should use true and false. Bool tests don't need comparisons. Signed-off-by: Thomas Meyer <thomas@m3y3r.de> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13ceph: handle 'session get evicted while there are file locks'Yan, Zheng
When session get evicted, all file locks associated with the session get released remotely by mds. File locks tracked by kernel become stale. In this situation, set an error flag on inode. The flag makes further file locks return -EIO. Another option to handle this situation is cleanup file locks tracked kernel. I do not choose it because it is inconvenient to notify user program about the error. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13ceph: optimize flock encoding during reconnectYan, Zheng
Don't malloc if there is no flock. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13ceph: make lock_to_ceph_filelock() staticYan, Zheng
Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13ceph: keep auth cap when inode has flocks or posix locksYan, Zheng
file locks are tracked by inode's auth mds. dropping auth caps is equivalent to releasing all file locks. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
2017-11-13quota: be aware of error from dquot_initializeChao Yu
Commit 6184fc0b8dd7 ("quota: Propagate error from ->acquire_dquot()") missed to handle error from dquot_initialize in dquot_file_open, fix it. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-11-12pstore: Use ktime_get_real_fast_ns() instead of __getnstimeofday()Arnd Bergmann
__getnstimeofday() is a rather odd interface, with a number of quirks: - The caller may come from NMI context, but the implementation is not NMI safe, one way to get there from NMI is NMI handler: something bad panic() kmsg_dump() pstore_dump() pstore_record_init() __getnstimeofday() - The calling conventions are different from any other timekeeping functions, to deal with returning an error code during suspended timekeeping. Address the above issues by using a completely different method to get the time: ktime_get_real_fast_ns() is NMI safe and has a reasonable behavior when timekeeping is suspended: it returns the time at which it got suspended. As Thomas Gleixner explained, this is safe, as ktime_get_real_fast_ns() does not call into the clocksource driver that might be suspended. The result can easily be transformed into a timespec structure. Since ktime_get_real_fast_ns() was not exported to modules, add the export. The pstore behavior for the suspended case changes slightly, as it now stores the timestamp at which timekeeping was suspended instead of storing a zero timestamp. This change is not addressing y2038-safety, that's subject to a more complex follow up patch. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Kees Cook <keescook@chromium.org> Cc: Tony Luck <tony.luck@intel.com> Cc: Anton Vorontsov <anton@enomsg.org> Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: John Stultz <john.stultz@linaro.org> Cc: Colin Cross <ccross@android.com> Link: https://lkml.kernel.org/r/20171110152530.1926955-1-arnd@arndb.de
2017-11-10fs: guard_bio_eod() needs to consider partitionsGreg Edwards
guard_bio_eod() needs to look at the partition capacity, not just the capacity of the whole device, when determining if truncation is necessary. [ 60.268688] attempt to access beyond end of device [ 60.268690] unknown-block(9,1): rw=0, want=67103509, limit=67103506 [ 60.268693] buffer_io_error: 2 callbacks suppressed [ 60.268696] Buffer I/O error on dev md1p7, logical block 4524305, async page read Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index") Cc: stable@vger.kernel.org # v4.13 Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Greg Edwards <gedwards@ddn.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2017-11-10block, scsi: Make SCSI quiesce and resume work reliablyBart Van Assche
The contexts from which a SCSI device can be quiesced or resumed are: * Writing into /sys/class/scsi_device/*/device/state. * SCSI parallel (SPI) domain validation. * The SCSI device power management methods. See also scsi_bus_pm_ops. It is essential during suspend and resume that neither the filesystem state nor the filesystem metadata in RAM changes. This is why while the hibernation image is being written or restored that SCSI devices are quiesced. The SCSI core quiesces devices through scsi_device_quiesce() and scsi_device_resume(). In the SDEV_QUIESCE state execution of non-preempt requests is deferred. This is realized by returning BLKPREP_DEFER from inside scsi_prep_state_check() for quiesced SCSI devices. Avoid that a full queue prevents power management requests to be submitted by deferring allocation of non-preempt requests for devices in the quiesced state. This patch has been tested by running the following commands and by verifying that after each resume the fio job was still running: for ((i=0; i<10; i++)); do ( cd /sys/block/md0/md && while true; do [ "$(<sync_action)" = "idle" ] && echo check > sync_action sleep 1 done ) & pids=($!) for d in /sys/class/block/sd*[a-z]; do bdev=${d#/sys/class/block/} hcil=$(readlink "$d/device") hcil=${hcil#../../../} echo 4 > "$d/queue/nr_requests" echo 1 > "/sys/class/scsi_device/$hcil/device/queue_depth" fio --name="$bdev" --filename="/dev/$bdev" --buffered=0 --bs=512 \ --rw=randread --ioengine=libaio --numjobs=4 --iodepth=16 \ --iodepth_batch=1 --thread --loops=$((2**31)) & pids+=($!) done sleep 1 echo "$(date) Hibernating ..." >>hibernate-test-log.txt systemctl hibernate sleep 10 kill "${pids[@]}" echo idle > /sys/block/md0/md/sync_action wait echo "$(date) Done." >>hibernate-test-log.txt done Reported-by: Oleksandr Natalenko <oleksandr@natalenko.name> References: "I/O hangs after resuming from suspend-to-ram" (https://marc.info/?l=linux-block&m=150340235201348). Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Tested-by: Martin Steigerwald <martin@lichtvoll.de> Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name> Cc: Martin K. Petersen <martin.petersen@oracle.com> Cc: Ming Lei <ming.lei@redhat.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Johannes Thumshirn <jthumshirn@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2017-11-10f2fs: validate before set/clear free nat bitmapLiFan
In flush_nat_entries, all dirty nats will be flushed and if their new address isn't NULL_ADDR, their bitmaps will be updated, the free_nid_count of the bitmaps will be increaced regardless of whether the nats have already been occupied before. This could lead to wrong free_nid_count. So this patch checks the status of the bits beforeactually set/clear them. Fixes: 586d1492f301 ("f2fs: skip scanning free nid bitmap of full NAT blocks") Signed-off-by: Fan li <fanofcode.li@samsung.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-10f2fs: avoid opened loop codes in __add_ino_entryChao Yu
We will keep __add_ino_entry success all the time, for ENOMEM failure case, we have already handled it by using __GFP_NOFAIL flag, so we don't have to use additional opened loop codes here, remove them. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-10coredump: call do_unlinkat directly instead of sys_unlinkChristoph Hellwig
And stop messing with the address limit. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2017-11-10fs: expose do_unlinkat for built-in callersChristoph Hellwig
And make it take a struct filename instead of a user pointer. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2017-11-10ext4: take handling of EXT4_IOC_GROUP_ADD into a helper, get rid of set_fs()Al Viro
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2017-11-10ovl: remove unneeded arg from ovl_verify_origin()Amir Goldstein
Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: Put upperdentry if ovl_check_origin() failsVivek Goyal
If ovl_check_origin() fails, we should put upperdentry. We have a reference on it by now. So goto out_put_upper instead of out. Fixes: a9d019573e88 ("ovl: lookup non-dir copy-up-origin by file handle") Cc: <stable@vger.kernel.org> #4.12 Signed-off-by: Vivek Goyal <vgoyal@redhat.com> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: rename ufs to ofsMiklos Szeredi
Rename all "struct ovl_fs" pointers to "ofs". The "ufs" name is historical and can only be found in overlayfs/super.c. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: clean up getting lower layersMiklos Szeredi
Move calling ovl_get_lower_layers() into ovl_get_lowerstack(). ovl_get_lowerstack() now returns the root dentry's filled in ovl_entry. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: clean up workdir creationMiklos Szeredi
Move calling ovl_get_workdir() into ovl_get_workpath(). Rename ovl_get_workdir() to ovl_make_workdir() and ovl_get_workpath() to ovl_get_workdir(). Workpath is now not needed outside ovl_get_workdir(). Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: clean up getting upper layerMiklos Szeredi
Merge ovl_get_upper() and ovl_get_upperpath(). The resulting function is named ovl_get_upper(), though it still returns upperpath as well. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: move ovl_get_workdir() and ovl_get_lower_layers()Miklos Szeredi
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: reduce the number of arguments for ovl_workdir_create()Miklos Szeredi
Remove "sb" and "dentry" arguments of ovl_workdir_create() and related functions. Move setting MS_RDONLY flag to callers. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: change order of setup in ovl_fill_super()Miklos Szeredi
Move ovl_get_upper() immediately after ovl_get_upperpath(), ovl_get_workdir() immediately after ovl_get_workdir() and ovl_get_lower_layers() immediately after ovl_get_lowerstack(). Also move prepare_creds() up to where other allocations are happening. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-10ovl: factor out ovl_free_fs() helperMiklos Szeredi
This can be called both from ovl_put_super() and in the error cleanup path from ovl_fill_super(). Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09xfs: on failed mount, force-reclaim inodes after unmounting quota controlsDarrick J. Wong
When mounting fails, we must force-reclaim inodes (and disable delayed reclaim) /after/ the realtime and quota control have let go of the realtime and quota inodes. Without this, we corrupt the timer list and cause other weird problems. Found by xfs/376 fuzzing u3.bmbt[0].lastoff on an rmap filesystem to force a bogus post-eof extent reclaim that causes the fs to go down. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: check the uniqueness of the AGFL entriesDarrick J. Wong
Make sure we don't list a block twice in the agfl by copying the contents of the AGFL to an array, sorting it, and looking for duplicates. We can easily check that the number of agfl entries we see actually matches the flcount, so do that too. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: remove u_int* type usageDarrick J. Wong
Use the uint* types instead of the u_int* types. This will (hopefully) pair with an xfsprogs cleanup. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: handle zero entries case in xfs_iext_rebalance_leafChristoph Hellwig
And also rename fill to nr_entries to match the rest of the code. Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: add comments documenting the rebalance algorithmChristoph Hellwig
Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: trivial indentation fixup for xfs_iext_remove_nodeChristoph Hellwig
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: remove a superflous assignment in xfs_iext_remove_nodeChristoph Hellwig
Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: add some comments to xfs_iext_insert/xfs_iext_insert_nodeChristoph Hellwig
Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: fix number of records handling in xfs_iext_split_leafChristoph Hellwig
Fix to check the correct value, and remove a duplicate handling of the uneven record number split algorith, Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09f2fs: apply write hints to select the type of segments for buffered writeHyunchul Lee
Write hints helps F2FS to determine which type of segments would be selected for buffered write. This patch implements the mapping from write hints to segment types as shown below. hints segment type ----- ------------ WRITE_LIFE_SHORT CURSEG_HOT_DATA WRITE_LIFE_EXTREME CURSEG_COLD_DATA others CURSEG_WARM_DATA the F2FS poliy for hot/cold seperation has precedence over this hints. And hints are not applied in in-place update. Signed-off-by: Hyunchul Lee <cheol.lee@lge.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-09f2fs: introduce scan_curseg_cache for cleanupChao Yu
Commit 4ac912427c42 ("f2fs: introduce free nid bitmap") copied codes from __build_free_nids() into scan_free_nid_bits(), they are redundant, introduce one common function scan_curseg_cache for cleanup. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-09f2fs: optimize the way of traversing free_nid_bitmapFan Li
We call scan_free_nid_bits only when there isn't many free nids left, it means that marked bits in free_nid_bitmap are supposed to be few, use find_next_bit_le is more efficient in such case. According to my tests, use find_next_bit_le instead of test_bit_le will cut down the traversal time to one third of its original. Signed-off-by: Fan li <fanofcode.li@samsung.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-09f2fs: keep scanning until enough free nids are acquiredFan Li
In current version, after scan_free_nid_bits, the scan is over if nid_cnt[FREE_NID] != 0. In most cases, there are still free nids in the free list during the scan, and scan_free_nid_bits usually can't increase nid_cnt[FREE_NID]. It causes that __build_free_nids is called many times without solving the shortage of the free nids. This patch fixes that. Signed-off-by: Fan li <fanofcode.li@samsung.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-11-09fs/xfs: Remove NULL check before kmem_cache_destroyTim Hansen
kmem_cache_destroy already checks for null values. Signed-off-by: Tim Hansen <devtimhansen@gmail.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-11-09xfs: only check da node header padding on v5 filesystemsDarrick J. Wong
It turns out that we only started zeroing a new da btree node's block header on v5 filesystems. Prior to that, we just wouldn't set anything at all, which means that the pad field never got set and would retain whatever happened to be in memory. Therefore, we can only check the pad for zeroness on v5 filesystems. shared/006 on a v4 filesystem exposes this scrub bug. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: fix btree scrub deref checkDarrick J. Wong
The btree scrubber has some custom code to retrieve and check a btree block via xfs_btree_lookup_get_block. This function will either return an error code (verifiers failed) or a *pblock will be untouched (bad pointer). Since we previously set *pblock to NULL, we need to check *pblock, not pblock, to trigger the early bailout. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: fix uninitialized return values in scrub codeDarrick J. Wong
Fix smatch complaints about uninitialized return codes. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: pass inode number to xfs_scrub_ino_set_{preen,warning}Darrick J. Wong
There are two ways to scrub an inode -- calling xfs_iget and checking the raw inode core, or by loading the inode cluster buffer and checking the on-disk contents directly. The second method is only useful if _iget fails the verifiers; when this is the case, sc->ip is NULL and calling the tracepoint will cause a system crash. Therefore, pass the raw inode number directly into the _preen and _warning functions. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09xfs: refactor the directory data block bestfree checksDarrick J. Wong
In a directory data block, the zeroth bestfree item must point to the longest free space. Therefore, when we check the bestfree block's records against the data blocks, we only need to compare with bf[0] and don't need the loop. The weird loop was most probably the result of an earlier refactoring gone bad. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2017-11-09ovl: grab reference to workbasedir earlyMiklos Szeredi
and related cleanups. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09ovl: split out ovl_get_indexdir() from ovl_fill_super()Miklos Szeredi
It's okay to get rid of the intermediate error label due to ufs being zeroed on allocation. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09ovl: split out ovl_get_lower_layers() from ovl_fill_super()Miklos Szeredi
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09ovl: split out ovl_get_workdir() from ovl_fill_super()Miklos Szeredi
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09ovl: split out ovl_get_upper() from ovl_fill_super()Miklos Szeredi
And don't clobber ufs->upper_mnt on error. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
2017-11-09ovl: split out ovl_get_lowerstack() from ovl_fill_super()Miklos Szeredi
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>