- 11 Jun, 2020 3 commits
-
-
Jeffle Xu authored
Fix the bug when calculating the physical block number of the first block in the split extent. This bug will cause xfstests shared/298 failure on ext4 with bigalloc enabled occasionally. Ext4 error messages indicate that previously freed blocks are being freed again, and the following fsck will fail due to the inconsistency of block bitmap and bg descriptor. The following is an example case: 1. First, Initialize a ext4 filesystem with cluster size '16K', block size '4K', in which case, one cluster contains four blocks. 2. Create one file (e.g., xxx.img) on this ext4 filesystem. Now the extent tree of this file is like: ... 36864:[0]4:220160 36868:[0]14332:145408 51200:[0]2:231424 ... 3. Then execute PUNCH_HOLE fallocate on this file. The hole range is like: .. ext4_ext_remove_space: dev 254,16 ino 12 since 49506 end 49506 depth 1 ext4_ext_remove_space: dev 254,16 ino 12 since 49544 end 49546 depth 1 ext4_ext_remove_space: dev 254,16 ino 12 since 49605 end 49607 depth 1 ... 4. Then the extent tree of this file after punching is like ... 49507:[0]37:158047 49547:[0]58:158087 ... 5. Detailed procedure of punching hole [49544, 49546] 5.1. The block address space: ``` lblk ~49505 49506 49507~49543 49544~49546 49547~ ---------+------+-------------+----------------+-------- extent | hole | extent | hole | extent ---------+------+-------------+----------------+-------- pblk ~158045 158046 158047~158083 158084~158086 158087~ ``` 5.2. The detailed layout of cluster 39521: ``` cluster 39521 <-------------------------------> hole extent <----------------------><-------- lblk 49544 49545 49546 49547 +-------+-------+-------+-------+ | | | | | +-------+-------+-------+-------+ pblk 158084 1580845 158086 158087 ``` 5.3. The ftrace output when punching hole [49544, 49546]: - ext4_ext_remove_space (start 49544, end 49546) - ext4_ext_rm_leaf (start 49544, end 49546, last_extent [49507(158047), 40], partial [pclu 39522 lblk 0 state 2]) - ext4_remove_blocks (extent [49507(158047), 40], from 49544 to 49546, partial [pclu 39522 lblk 0 state 2] - ext4_free_blocks: (block 158084 count 4) - ext4_mballoc_free (extent 1/6753/1) 5.4. Ext4 error message in dmesg: EXT4-fs error (device vdb): mb_free_blocks:1457: group 1, block 158084:freeing already freed block (bit 6753); block bitmap corrupt. EXT4-fs error (device vdb): ext4_mb_generate_buddy:747: group 1, block bitmap and bg descriptor inconsistent: 19550 vs 19551 free clusters In this case, the whole cluster 39521 is freed mistakenly when freeing pblock 158084~158086 (i.e., the first three blocks of this cluster), although pblock 158087 (the last remaining block of this cluster) has not been freed yet. The root cause of this isuue is that, the pclu of the partial cluster is calculated mistakenly in ext4_ext_remove_space(). The correct partial_cluster.pclu (i.e., the cluster number of the first block in the next extent, that is, lblock 49597 (pblock 158086)) should be 39521 rather than 39522. Fixes: f4226d9e ("ext4: fix partial cluster initialization") Signed-off-by:Jeffle Xu <jefflexu@linux.alibaba.com> Reviewed-by:
Eric Whitney <enwlinux@gmail.com> Cc: stable@kernel.org # v3.19+ Link: https://lore.kernel.org/r/1590121124-37096-1-git-send-email-jefflexu@linux.alibaba.comSigned-off-by:
Theodore Ts'o <tytso@mit.edu>
-
Theodore Ts'o authored
Trying to change dax mount options when remounting could allow mount options to be enabled for a small amount of time, and then the mount option change would be reverted. In the case of "mount -o remount,dax", this can cause a race where files would temporarily treated as DAX --- and then not. Cc: stable@kernel.org Reported-by: syzbot+bca9799bf129256190da@syzkaller.appspotmail.com Signed-off-by: -
Theodore Ts'o authored
This adds the same per-file/per-directory DAX support for ext4 as was done for xfs, now that we finally have consensus over what the interface should be.
-
- 04 Jun, 2020 37 commits
-
-
Jan Kara authored
ext4_writepages() currently works in a loop like: start a transaction scan inode for pages to write map and submit these pages stop the transaction This loop results in starting transaction once more than is needed because in the last iteration we start a transaction only to scan the inode and find there are no pages to write. This can be significant increase in number of transaction starts for single-extent files or files that have all blocks already mapped. Furthermore we already know from previous iteration whether there are more pages to write or not. So propagate the information from mpage_prepare_extent_to_map() and avoid unnecessary looping in case there are no more pages to write. Signed-off-by:
Jan Kara <jack@suse.cz> Link: https://lore.kernel.org/r/20200525081215.29451-1-jack@suse.czSigned-off-by:
-
Jens Axboe authored
Running with some debug patches to detect illegal blocking triggered the extend/unaligned condition in ext4. If ext4 needs to extend the file (and hence go to buffered IO), or if the app is doing unaligned IO, then ext4 asks the iomap code to wait for IO completion. If the caller asked for no-wait semantics by setting IOCB_NOWAIT, then ext4 should return -EAGAIN instead. Signed-off-by:
Jens Axboe <axboe@kernel.dk> Link: https://lore.kernel.org/r/76152096-2bbb-7682-8fce-4cb498bcd909@kernel.dkSigned-off-by:
-
Christoph Hellwig authored
access_ok just checks we are fed a proper user pointer. We also do that in copy_to_user itself, so no need to do this early. Signed-off-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
Ritesh Harjani <riteshh@linux.ibm.com> Reviewed-by:
-
Christoph Hellwig authored
access_ok just checks we are fed a proper user pointer. We also do that in copy_to_user itself, so no need to do this early. Signed-off-by:
Darrick J. Wong <darrick.wong@oracle.com> Link: https://lore.kernel.org/r/20200523073016.2944131-9-hch@lst.deSigned-off-by:
-
Christoph Hellwig authored
By moving FIEMAP_FLAG_SYNC handling to fiemap_prep we ensure it is handled once instead of duplicated, but can still be done under fs locks, like xfs/iomap intended with its duplicate handling. Also make sure the error value of filemap_write_and_wait is propagated to user space. Signed-off-by:
Amir Goldstein <amir73il@gmail.com> Reviewed-by:
-
Christoph Hellwig authored
Replace fiemap_check_flags with a fiemap_prep helper that also takes the inode and mapped range, and performs the sanity check and truncation previously done in fiemap_check_range. This way the validation is inside the file system itself and thus properly works for the stacked overlayfs case as well. Signed-off-by:
-
Christoph Hellwig authored
iomap_fiemap should take u64 start and len arguments, just like the ->fiemap prototype. Signed-off-by:
-
Christoph Hellwig authored
No need to pull the fiemap definitions into almost every file in the kernel build. Signed-off-by:
-
Christoph Hellwig authored
There is no caller left outside of ioctl.c. Signed-off-by:
-
Christoph Hellwig authored
iomap_fiemap already calls fiemap_check_flags first thing, so this additional check is redundant. Signed-off-by:
-
Christoph Hellwig authored
The fiemap and EXT4_IOC_GET_ES_CACHE cases share almost no code, so split them into entirely separate functions. Signed-off-by:
-
Christoph Hellwig authored
Add an extra validation of the len parameter, as for ext4 some files might have smaller file size limits than others. This also means the redundant size check in ext4_ioctl_get_es_cache can go away, as all size checking is done in the shared fiemap handler. Signed-off-by:
-
Ritesh Harjani authored
ext4 supports max number of logical blocks in a file to be 0xffffffff. (This is since ext4_extent's ee_block is __le32). This means that EXT4_MAX_LOGICAL_BLOCK should be 0xfffffffe (starting from 0 logical offset). This patch fixes this. The issue was seen when ext4 moved to iomap_fiemap API and when overlayfs was mounted on top of ext4. Since overlayfs was missing filemap_check_ranges(), so it could pass a arbitrary huge length which lead to overflow of map.m_len logic. This patch fixes that. Fixes: d3b6f23f ("ext4: move ext4_fiemap to use iomap framework") Reported-by: syzbot+77fa5bdb65cc39711820@syzkaller.appspotmail.com Signed-off-by:
-
Jonathan Grant authored
Signed-off-by:
Jonathan Grant <jg@jguk.org> Reviewed-by:
Andreas Dilger <adilger@dilger.ca> Link: https://lore.kernel.org/r/ad3290d5-86af-99c1-f9d5-cd1bab710429@jguk.orgSigned-off-by:
-
Jan Kara authored
When reserved transaction handle is unused, we subtract its reserved credits in __jbd2_journal_unreserve_handle() called from jbd2_journal_stop(). However this function forgets to remove reserved credits from transaction->t_outstanding_credits and thus the transaction space that was reserved remains effectively leaked. The leaked transaction space can be quite significant in some cases and leads to unnecessarily small transactions and thus reducing throughput of the journalling machinery. E.g. fsmark workload creating lots of 4k files was observed to have about 20% lower throughput due to this when ext4 is mounted with dioread_nolock mount option. Subtract reserved credits from t_outstanding_credits as well. CC: stable@vger.kernel.org Fixes: 8f7d89f3 ("jbd2: transaction reservation support") Reviewed-by:
-
Jan Kara authored
Remove ext4_journal_free_reserved() function. It is never used. Signed-off-by:
-
Ritesh Harjani authored
Currently while doing block allocation grp->bb_free may be getting modified if discard is happening in parallel. For e.g. consider a case where there are lot of threads who have preallocated lot of blocks and there is a thread which is trying to discard all of this group's PA. Now it could happen that we see all of those group's bb_free is zero and fail the allocation while there is sufficient space if we free up all the PA. So this patch adds another flag "EXT4_MB_STRICT_CHECK" which will be set if we are unable to allocate any blocks in the first try (since we may not have considered blocks about to be discarded from PA lists). So during retry attempt to allocate blocks we will use ext4_lock_group() for checking if the group is good or not. Signed-off-by:
-
Ritesh Harjani authored
ext4_mb_good_group() definition was changed some time back and now it even initializes the buddy cache (via ext4_mb_init_group()), if in case the EXT4_MB_GRP_NEED_INIT() is true for a group. Note that ext4_mb_init_group() could sleep and so should not be called under a spinlock held. This is fine as of now because ext4_mb_good_group() is called before loading the buddy bitmap without ext4_lock_group() held and again called after loading the bitmap, only this time with ext4_lock_group() held. But still this whole thing is confusing. So this patch refactors out ext4_mb_good_group_nolock() which should be called when without holding ext4_lock_group(). Also in further patches we hold the spinlock (ext4_lock_group()) while doing any calculations which involves grp->bb_free or grp->bb_fragments. Signed-off-by:
-
Ritesh Harjani authored
There could be a race in function ext4_mb_discard_group_preallocations() where the 1st thread may iterate through group's bb_prealloc_list and remove all the PAs and add to function's local list head. Now if the 2nd thread comes in to discard the group preallocations, it will see that the group->bb_prealloc_list is empty and will return 0. Consider for a case where we have less number of groups (for e.g. just group 0), this may even return an -ENOSPC error from ext4_mb_new_blocks() (where we call for ext4_mb_discard_group_preallocations()). But that is wrong, since 2nd thread should have waited for 1st thread to release all the PAs and should have retried for allocation. Since 1st thread was anyway going to discard the PAs. The algorithm using this percpu seq counter goes below: 1. We sample the percpu discard_pa_seq counter before trying for block allocation in ext4_mb_new_blocks(). 2. We increment this percpu discard_pa_seq counter when we either allocate or free these blocks i.e. while marking those blocks as used/free in mb_mark_used()/mb_free_blocks(). 3. We also increment this percpu seq counter when we successfully identify that the bb_prealloc_list is not empty and hence proceed for discarding of those PAs inside ext4_mb_discard_group_preallocations(). Now to make sure that the regular fast path of block allocation is not affected, as a small optimization we only sample the percpu seq counter on that cpu. Only when the block allocation fails and when freed blocks found were 0, that is when we sample percpu seq counter for all cpus using below function ext4_get_discard_pa_seq_sum(). This happens after making sure that all the PAs on grp->bb_prealloc_list got freed or if it's empty. It can be well argued that why don't just check for grp->bb_free to see if there are any free blocks to be allocated. So here are the two concerns which were discussed:- 1. If for some reason the blocks available in the group are not appropriate for allocation logic (say for e.g. EXT4_MB_HINT_GOAL_ONLY, although this is not yet implemented), then the retry logic may result into infinte looping since grp->bb_free is non-zero. 2. Also before preallocation was clubbed with block allocation with the same ext4_lock_group() held, there were lot of races where grp->bb_free could not be reliably relied upon. Due to above, this patch considers discard_pa_seq logic to determine if we should retry for block allocation. Say if there are are n threads trying for block allocation and none of those could allocate or discard any of the blocks, then all of those n threads will fail the block allocation and return -ENOSPC error. (Since the seq counter for all of those will match as no block allocation/discard was done during that duration). Signed-off-by:
-
Ritesh Harjani authored
Implement ext4_mb_discard_preallocations_should_retry() which we will need in later patches to add more logic like check for sequence number match to see if we should retry for block allocation or not. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
ext4_mb_discard_preallocations() only checks for grp->bb_prealloc_list of every group to discard the group's PA to free up the space if allocation request fails. Consider below race:- Process A Process B 1. allocate blocks 1. Fails block allocation from ext4_mb_regular_allocator() ext4_lock_group() allocated blocks more than ac_o_ex.fe_len ext4_unlock_group() 2. Scans the grp->bb_prealloc_list (under ext4_lock_group()) and find nothing and thus return -ENOSPC. 2. Add the additional blocks to PA list ext4_lock_group() add blocks to grp->bb_prealloc_list ext4_unlock_group() Above race could be avoided if we add those additional blocks to grp->bb_prealloc_list at the same time with block allocation when ext4_lock_group() was still held. With this discard-PA will know if there are actually any blocks which could be freed from the PA Signed-off-by:
-
Eric Biggers authored
No one currently needs EXT4_INODE_CASEFOLD, but add it to keep the EXT4_INODE_* definitions in sync with the EXT4_*_FL definitions. Also make it clearer that the casefold flag is only for directories. Signed-off-by:
Eric Biggers <ebiggers@google.com> Link: https://lore.kernel.org/r/20200510215252.87833-1-ebiggers@kernel.orgSigned-off-by:
-
Eric Whitney authored
The path performing block allocations in ext4_ext_map_blocks() contains code trimming the length of a new extent that is repeated later in the function. This code is both redundant and unnecessary as the exact length of the new extent has already been calculated. Rewrite the instantiation of the map struct in this case to use the available values, avoiding the overhead of unnecessary conversions and improving clarity. Add another map struct instantiation tailored specifically to the separate case for an existing written extent. Remove an old comment that no longer appears applicable to the current code. Signed-off-by:
-
Ritesh Harjani authored
ext_debug() msgs could be helpful, provided those could be enabled without recompiling kernel and also if we could selectively enable only required prints for case by case debugging. So make ext_debug() implementation use pr_debug(). Also change ext_debug() to be defined with CONFIG_EXT4_DEBUG. So EXT_DEBUG macro now mostly remain for below 3 functions. ext4_ext_show_path/leaf/move() (whose print msgs use ext_debug() which again could be dynamically enabled using pr_debug()) This also changes the ext_debug() to take inode as a parameter to add inode no. in all of it's msgs. Prints additional info like process name / pid, superblock id etc. This also removes any explicit function names passed in ext_debug(). Since ext_debug() on it's own prints file, func and line no. Signed-off-by:
-
Ritesh Harjani authored
mb_debug() msg had only 1 control level for all type of msgs. And if we enable mballoc_debug then all of those msgs would be enabled. Instead of adding multiple debug levels for mb_debug() msgs, use pr_debug() with which we could have finer control to print msgs at all of different levels (i.e. at file, func, line no.). Also add process name/pid, superblk id, and other info in mb_debug() msg. This also kills the mballoc_debug module parameter, since it is not needed any more. Signed-off-by:
-
Ritesh Harjani authored
Replace EXT_DEBUG with __maybe_unused from inside ext4_ext_handle_unwritten_extents() function. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
ext4_map_blocks() has ext_debug msg early at the start of function. We also get ext_debug msg if we could allocate a block from ext4_ext_map_blocks(). But there is no ext_debug() msg in case of block allocation failure. So add one along with error code. Also add more info in ext_debug() msg like how many blocks were allocated v/s how many were requested in ext4_ext_map_blocks(). Signed-off-by:
-
Ritesh Harjani authored
Simply use BIT() macro for all BH_** state bits instead of open coding it. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
Use task_pid_nr() function instead of current->pid. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
Make sure to check for e4b->bd_info->bb_bitmap == NULL, in mb_cmp_bitmaps() and return if NULL, to avoid possible NULL ptr dereference. Similar to how we do this in other ifdef DOUBLE_CHECK functions. Also remove the BUG_ON() logic if kmalloc() or ext4_read_block_bitmap() fails. We should simply mark grp->bb_bitmap as NULL if above happens. In fact ext4_read_block_bitmap() may even return an error in case of resize ioctl. Hence remove this BUG_ON logic (fstests ext4/032 may trigger this). Link: https://lore.kernel.org/r/9a54f8a696ff17c057cd571be3d15ac3ec1407f1.1589086800.git.riteshh@linux.ibm.comSigned-off-by:
-
Ritesh Harjani authored
This patch implemets mb_group_bb_bitmap_alloc() and mb_group_bb_bitmap_free() function to remove #ifdef DOUBLE_CHECK macro and it's related code from inside ext4_mb_add_groupinfo()/ext4_mb_release(). There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
Change return type of function ext4_mb_use_preallocated() to bool to better reflect what this function can return. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
This patch simplifies error handling logic in ext4_init_mballoc(), by adding all the cleanups at one place at the end of that function. There should be no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
Fix few other format specifiers in mb_debug() msgs. As such no other functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
pa->pa_len is an integer. Fix all of the format specifier used in mb_debug() for pa_len to %d instead of %u. As such no functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
This patch adds some more debugging mb_debug() msgs to help improve mballoc code debugging. Other than adding more mb_debug() msgs at few more places, there should be no other functionality change in this patch. Signed-off-by:
-
Ritesh Harjani authored
This factors out ext4_mb_show_pa() function to show all the group's preallocation info. This could be useful info to be added in later patches. There should be no functionality change in this patch. Signed-off-by:
-
