- 18 Mar, 2020 6 commits
-
-
Darrick J. Wong authored
Add a new btree function that enables us to bulk load a btree cursor. This will be used by the upcoming online repair patches to generate new btrees. This avoids the programmatic inefficiency of calling xfs_btree_insert in a loop (which generates a lot of log traffic) in favor of stamping out new btree blocks with ordered buffers, and then committing both the new root and scheduling the removal of the old btree blocks in a single transaction commit. The design of this new generic code is based off the btree rebuilding code in xfs_repair's phase 5 code, with the explicit goal of enabling us to share that code between scrub and repair. It has the additional feature of being able to control btree block loading factors. Signed-off-by:
Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by:
Brian Foster <bfoster@redhat.com>
-
Darrick J. Wong authored
Create an in-core fake root for inode-rooted btree types so that callers can generate a whole new btree using the upcoming btree bulk load function without making the new tree accessible from the rest of the filesystem. It is up to the individual btree type to provide a function to create a staged cursor (presumably with the appropriate callouts to update the fakeroot) and then commit the staged root back into the filesystem. Signed-off-by:
-
Darrick J. Wong authored
Create an in-core fake root for AG-rooted btree types so that callers can generate a whole new btree using the upcoming btree bulk load function without making the new tree accessible from the rest of the filesystem. It is up to the individual btree type to provide a function to create a staged cursor (presumably with the appropriate callouts to update the fakeroot) and then commit the staged root back into the filesystem. Signed-off-by:
-
Darrick J. Wong authored
Add a xbitmap_hweight helper function so that we can get rid of the open-coded loop. Signed-off-by:
-
Darrick J. Wong authored
Shorten the name of xfs_bitmap to xbitmap since the scrub bitmap has nothing to do with the libxfs bitmap. Signed-off-by:
-
Darrick J. Wong authored
Remove the xfs_bitmap_destroy call from the end of xrep_reap_extents because this sort of violates our rule that the function initializing a structure should destroy it. Signed-off-by:
-
- 17 Mar, 2020 1 commit
-
-
Darrick J. Wong authored
When I lifted the code in xfs_alloc_ag_vextent_lastblock out of a loop, I forgot to convert all the accesses to len to be pointer dereferences. Coverity-id: 1457918 Fixes: 5113f8ec ("xfs: clean up weird while loop in xfs_alloc_ag_vextent_near") Signed-off-by:
Christoph Hellwig <hch@lst.de>
-
- 15 Mar, 2020 1 commit
-
-
Darrick J. Wong authored
If the xfs_buf_map array allocation in xfs_dabuf_map fails for whatever reason, we bail out with error code zero. This will confuse callers, so make sure that we return ENOMEM. Allocation failure should never happen with the small size of the array, but code defensively anyway. Fixes: 45feef8f ("xfs: refactor xfs_dabuf_map") Signed-off-by:
Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
-
- 13 Mar, 2020 13 commits
-
-
Christoph Hellwig authored
Move the code for verifying the iclog state on a clean unmount into a helper, and instead of checking the iclog state just rely on the shutdown check as they are equivalent. Also remove the ifdef DEBUG as the compiler is smart enough to eliminate the dead code for non-debug builds. Signed-off-by:
-
Christoph Hellwig authored
When the log is shut down all iclogs are in the XLOG_STATE_IOERROR state, which means that xlog_state_want_sync and xlog_state_release_iclog are no-ops. Remove the whole section of code. Signed-off-by:
-
Christoph Hellwig authored
Remove the ignored return value from xfs_log_unmount_write, and also remove a rather pointless assert on the return value from xfs_log_force. Signed-off-by:
-
Christoph Hellwig authored
Signed-off-by:
-
Christoph Hellwig authored
A shutdown log is a slow failure path. Add an unlikely annotation to it. Signed-off-by:
-
Dave Chinner authored
This is much less widely used than the bc_private union was, so this is done as a single patch. The named union xfs_btree_cur_private goes away and is embedded into the struct xfs_btree_cur_ag as an anonymous union, and the code is modified via this script: $ sed -i 's/priv\.\([abt|refc]\)/\1/g' fs/xfs/*[ch] fs/xfs/*/*[ch] Signed-off-by:
Dave Chinner <dchinner@redhat.com> Reviewed-by:
-
Dave Chinner authored
we need to name the btree cursor private structures to be able to pull them out of the deeply nested structure definition they are in now. Based on code extracted from a patchset by Darrick Wong. Signed-off-by:
-
Dave Chinner authored
Rename the union and it's internal structures to the new name and remove the temporary defines that facilitated the change. Signed-off-by:
-
Dave Chinner authored
BPRV is not longer appropriate because bc_private is going away. Script: $ sed -i 's/BTCUR_BPRV/BTCUR_BMBT/g' fs/xfs/*[ch] fs/xfs/*/*[ch] With manual cleanup to the definitions in fs/xfs/libxfs/xfs_btree.h Signed-off-by:
-
Dave Chinner authored
bc_private.b -> bc_ino conversion via script: $ sed -i 's/bc_private\.b/bc_ino/g' fs/xfs/*[ch] fs/xfs/*/*[ch] And then revert the change to the bc_ino #define in fs/xfs/libxfs/xfs_btree.h manually. Signed-off-by:
-
Dave Chinner authored
bc_private.a -> bc_ag conversion via script: `sed -i 's/bc_private\.a/bc_ag/g' fs/xfs/*[ch] fs/xfs/*/*[ch]` And then revert the change to the bc_ag #define in fs/xfs/libxfs/xfs_btree.h manually. Signed-off-by:
-
Dave Chinner authored
Just the defines of the new names - the conversion will be in scripted commits after this. Signed-off-by:
-
Tommi Rantala authored
Commit 263dde86 ("xfs: cleanup xfs_dir2_block_getdents") introduced a getdents regression, when it converted the pointer arithmetics to offset calculations: offset is updated in the loop already for the next iteration, but the updated offset value is used incorrectly in two places, where we should have used the not-yet-updated value. This caused for example "git clean -ffdx" failures to cleanup certain directory structures when running in a container. Fix the regression by making sure we use proper offset in the loop body. Thanks to Christoph Hellwig for suggestion how to best fix the code. Cc: Christoph Hellwig <hch@lst.de> Fixes: 263dde86 ("xfs: cleanup xfs_dir2_block_getdents") Signed-off-by:
Tommi Rantala <tommi.t.rantala@nokia.com> Reviewed-by:
-
- 12 Mar, 2020 12 commits
-
-
Takashi Iwai authored
Since snprintf() returns the would-be-output size instead of the actual output size, the succeeding calls may go beyond the given buffer limit. Fix it by replacing with scnprintf(). Signed-off-by:
Takashi Iwai <tiwai@suse.de> Reviewed-by:
-
Darrick J. Wong authored
In xchk_xattr_listent, we attempt to validate the extended attribute hash structures by performing a attr lookup by (hashed) name. If the lookup returns ENODATA, that means that the hash information is corrupt. The _process_error functions don't catch this, so we have to add that explicitly. Signed-off-by:
-
Darrick J. Wong authored
In xchk_dir_actor, we attempt to validate the directory hash structures by performing a directory entry lookup by (hashed) name. If the lookup returns ENOENT, that means that the hash information is corrupt. The _process_error functions don't catch this, so we have to add that explicitly. Signed-off-by:
-
Darrick J. Wong authored
Check the owner field of dir3 block headers. If it's corrupt, release the buffer and return EFSCORRUPTED. All callers handle this properly. Signed-off-by:
-
Darrick J. Wong authored
Check the owner field of dir3 data block headers. If it's corrupt, release the buffer and return EFSCORRUPTED. All callers handle this properly. Signed-off-by:
-
Darrick J. Wong authored
Check the owner field of dir3 free block headers and reject the metadata if there's something wrong with it. Signed-off-by:
Allison Collins <allison.henderson@oracle.com> Reviewed-by:
-
Darrick J. Wong authored
If we decide that a directory free block is corrupt, we must take care not to leak a buffer pointer to the caller. After xfs_trans_brelse returns, the buffer can be freed or reused, which means that we have to set *bpp back to NULL. Callers are supposed to notice the nonzero return value and not use the buffer pointer, but we should code more defensively, even if all current callers handle this situation correctly. Fixes: de14c5f5 ("xfs: verify free block header fields") Signed-off-by:
-
Darrick J. Wong authored
xfs_verifier_error is supposed to be called on a corrupt metadata buffer from within a buffer verifier function, whereas xfs_buf_mark_corrupt is the function to be called when a piece of code has read a buffer and catches something that a read verifier cannot. The first function sets b_error anticipating that the low level buffer handling code will see the nonzero b_error and clear XBF_DONE on the buffer, whereas the second function does not. Since xfs_dir3_free_header_check examines fields in the dir free block header that require more context than can be provided to read verifiers, we must call xfs_buf_mark_corrupt when it finds a problem. Switching the calls has a secondary effect that we no longer corrupt the buffer state by setting b_error and leaving XBF_DONE set. When /that/ happens, we'll trip over various state assertions (most commonly the b_error check in xfs_buf_reverify) on a subsequent attempt to read the buffer. Fixes: bc1a09b8 ("xfs: refactor verifier callers to print address of failing check") Signed-off-by:
-
Darrick J. Wong authored
Add a xfs_failaddr_t parameter to this function so that callers can potentially pass in (and therefore report) the exact point in the code where we decided that a metadata buffer was corrupt. This enables us to wire it up to checking functions that have to run outside of verifiers. Signed-off-by:
-
Darrick J. Wong authored
Add a helper function to get rid of buffers that we have decided are corrupt after the verifiers have run. This function is intended to handle metadata checks that can't happen in the verifiers, such as inter-block relationship checking. Note that we now mark the buffer stale so that it will not end up on any LRU and will be purged on release. Signed-off-by:
-
Darrick J. Wong authored
In e7ee96df, we converted all ITER_ABORT users to use ECANCELED instead, but we forgot to teach xfs_rmap_has_other_keys not to return that magic value to callers. Fix it now by using ECANCELED both to abort the iteration and to signal that we found another reverse mapping. This enables us to drop the separate boolean flag. Fixes: e7ee96df ("xfs: remove all *_ITER_ABORT values") Signed-off-by:
-
Darrick J. Wong authored
Log the corrupt buffer before we release the buffer. Fixes: a5155b87 ("xfs: always log corruption errors") Signed-off-by:
-
- 11 Mar, 2020 6 commits
-
-
Christoph Hellwig authored
Just dereference bp->b_addr directly and make the code a little simpler and more clear. Signed-off-by:
Eric Sandeen <sandeen@redhat.com> Reviewed-by:
-
Christoph Hellwig authored
Just dereference bp->b_addr directly and make the code a little simpler and more clear. Signed-off-by:
-
Christoph Hellwig authored
Just dereference bp->b_addr directly and make the code a little simpler and more clear. Signed-off-by:
-
Christoph Hellwig authored
There is just a single user left, so remove it. Signed-off-by:
-
Christoph Hellwig authored
struct xfs_agfl is a header in front of the AGFL entries that exists for CRC enabled file systems. For not CRC enabled file systems the AGFL is simply a list of agbno. Make the CRC case similar to that by just using the list behind the new header. This indirectly solves a problem with modern gcc versions that warn about taking addresses of packed structures (and we have to pack the AGFL given that gcc rounds up structure sizes). Also replace the helper macro to get from a buffer with an inline function in xfs_alloc.h to make the code easier to read. Signed-off-by:
-
Eric Biggers authored
Leaving PF_MEMALLOC set when exiting a kthread causes it to remain set during do_exit(). That can confuse things. In particular, if BSD process accounting is enabled, then do_exit() writes data to an accounting file. If that file has FS_SYNC_FL set, then this write occurs synchronously and can misbehave if PF_MEMALLOC is set. For example, if the accounting file is located on an XFS filesystem, then a WARN_ON_ONCE() in iomap_do_writepage() is triggered and the data doesn't get written when it should. Or if the accounting file is located on an ext4 filesystem without a journal, then a WARN_ON_ONCE() in ext4_write_inode() is triggered and the inode doesn't get written. Fix this in xfsaild() by using the helper functions to save and restore PF_MEMALLOC. This can be reproduced as follows in the kvm-xfstests test appliance modified to add the 'acct' Debian package, and with kvm-xfstests's recommended kconfig modified to add CONFIG_BSD_PROCESS_ACCT=y: mkfs.xfs -f /dev/vdb mount /vdb touch /vdb/file chattr +S /vdb/file accton /vdb/file mkfs.xfs -f /dev/vdc mount /vdc umount /vdc It causes: WARNING: CPU: 1 PID: 336 at fs/iomap/buffered-io.c:1534 CPU: 1 PID: 336 Comm: xfsaild/vdc Not tainted 5.6.0-rc5 #3 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20191223_100556-anatol 04/01/2014 RIP: 0010:iomap_do_writepage+0x16b/0x1f0 fs/iomap/buffered-io.c:1534 [...] Call Trace: write_cache_pages+0x189/0x4d0 mm/page-writeback.c:2238 iomap_writepages+0x1c/0x33 fs/iomap/buffered-io.c:1642 xfs_vm_writepages+0x65/0x90 fs/xfs/xfs_aops.c:578 do_writepages+0x41/0xe0 mm/page-writeback.c:2344 __filemap_fdatawrite_range+0xd2/0x120 mm/filemap.c:421 file_write_and_wait_range+0x71/0xc0 mm/filemap.c:760 xfs_file_fsync+0x7a/0x2b0 fs/xfs/xfs_file.c:114 generic_write_sync include/linux/fs.h:2867 [inline] xfs_file_buffered_aio_write+0x379/0x3b0 fs/xfs/xfs_file.c:691 call_write_iter include/linux/fs.h:1901 [inline] new_sync_write+0x130/0x1d0 fs/read_write.c:483 __kernel_write+0x54/0xe0 fs/read_write.c:515 do_acct_process+0x122/0x170 kernel/acct.c:522 slow_acct_process kernel/acct.c:581 [inline] acct_process+0x1d4/0x27c kernel/acct.c:607 do_exit+0x83d/0xbc0 kernel/exit.c:791 kthread+0xf1/0x140 kernel/kthread.c:257 ret_from_fork+0x27/0x50 arch/x86/entry/entry_64.S:352 This bug was originally reported by syzbot at https://lore.kernel.org/r/0000000000000e7156059f751d7b@google.com. Reported-by: syzbot+1f9dc49e8de2582d90c2@syzkaller.appspotmail.com Signed-off-by:Eric Biggers <ebiggers@google.com> Reviewed-by:
-
- 03 Mar, 2020 1 commit
-
-
Christoph Hellwig authored
Let the low-level attr code only allocate the needed buffer size for xfs_attrmulti_attr_get instead of allocating the upper bound at the top of the call chain. Suggested-by:
-
