- 05 Dec, 2022 40 commits
-
-
Josef Bacik authored
These helpers use functions that are in multiple places, which makes it tricky to sync them into btrfs-progs. Move them to file-item.h and then include file-item.h in places that use these helpers. Signed-off-by:
Josef Bacik <josef@toxicpanda.com> Signed-off-by:
David Sterba <dsterba@suse.com>
-
Josef Bacik authored
This is only used in ctree.c, with the exception of zero'ing out extent buffers we're getting ready to write out. In theory we shouldn't have an extent buffer with 0 items that we're writing out, however I'd rather be safe than sorry so open code it in extent_io.c, and then copy the helper into ctree.c. This will make it easier to sync accessors.[ch] into btrfs-progs, as this requires a helper that isn't defined in accessors.h. Signed-off-by:
-
Josef Bacik authored
These accidentally got brought into accessors.h, but belong with the btrfs_root definitions which are currently in ctree.h. Move these to make it easier to sync accessors.[ch] into btrfs-progs. Signed-off-by:
-
Christoph Hellwig authored
repair_io_failure ties directly into all the glory low-level details of mapping a bio with a logic address to the actual physical location. Move it right below btrfs_submit_bio to keep all the related logic together. Also move btrfs_repair_eb_io_failure to its caller in disk-io.c now that repair_io_failure is available in a header. Reviewed-by:
Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
-
Christoph Hellwig authored
The code used by btrfs_submit_bio only interacts with the rest of volumes.c through __btrfs_map_block (which itself is a more generic version of two exported helpers) and does not really have anything to do with volumes.c. Create a new bio.c file and a bio.h header going along with it for the btrfs_bio-based storage layer, which will grow even more going forward. Also update the file with my copyright notice given that a large part of the moved code was written or rewritten by me. Reviewed-by:
-
Christoph Hellwig authored
Move struct btrfs_tree_parent_check out of disk-io.h so that volumes.h an various .c files don't have to include disk-io.h just for it. Reviewed-by:
Qu Wenruo <wqu@suse.com> Signed-off-by:
-
Qu Wenruo authored
[BUG] For the following small script, btrfs will be unable to recover the content of file1: mkfs.btrfs -f -m raid1 -d raid5 -b 1G $dev1 $dev2 $dev3 mount $dev1 $mnt xfs_io -f -c "pwrite -S 0xff 0 64k" -c sync $mnt/file1 md5sum $mnt/file1 umount $mnt # Corrupt the above 64K data stripe. xfs_io -f -c "pwrite -S 0x00 323026944 64K" -c sync $dev3 mount $dev1 $mnt # Write a new 64K, which should be in the other data stripe # And this is a sub-stripe write, which will cause RMW xfs_io -f -c "pwrite 0 64k" -c sync $mnt/file2 md5sum $mnt/file1 umount $mnt Above md5sum would fail. [CAUSE] There is a long existing problem for raid56 (not limited to btrfs raid56) that, if we already have some corrupted on-disk data, and then trigger a sub-stripe write (which needs RMW cycle), it can cause further damage into P/Q stripe. Disk 1: data 1 |0x000000000000| <- Corrupted Disk 2: data 2 |0x000000000000| Disk 2: parity |0xffffffffffff| In above case, data 1 is already corrupted, the original data should be 64KiB of 0xff. At this stage, if we read data 1, and it has data checksum, we can still recovery going via the regular RAID56 recovery path. But if now we decide to write some data into data 2, then we need to go RMW. Let's say we want to write 64KiB of '0x00' into data 2, then we read the on-disk data of data 1, calculate the new parity, resulting the following layout: Disk 1: data 1 |0x000000000000| <- Corrupted Disk 2: data 2 |0x000000000000| <- New '0x00' writes Disk 2: parity |0x000000000000| <- New Parity. But the new parity is calculated using the *corrupted* data 1, we can no longer recover the correct data of data1. Thus the corruption is forever there. [FIX] To solve above problem, this patch will do a full stripe data checksum verification at RMW time. This involves the following changes: - Always read the full stripe (including data/P/Q) when doing RMW Before we only read the missing data sectors, but since we may do a data csum verification and recovery, we need to read everything out. Please note that, if we have a cached rbio, we don't need to read anything, and can treat it the same as full stripe write. As only stripe with all its csum matches can be cached. - Verify the data csum during read. The goal is only the rbio stripe sectors, and only if the rbio already has csum_buf/csum_bitmap filled. And sectors which cannot pass csum verification will have their bit set in error_bitmap. - Always call recovery_sectors() after we read out all the sectors Since error_bitmap will be updated during read, recover_sectors() can easily find out all the bad sectors and try to recover (if still under tolerance). And since recovery_sectors() is already migrated to use error_bitmap, it can skip vertical stripes which don't have any error. - Verify the repaired sectors against its csum in recover_vertical() - Rename rmw_read_and_wait() to rmw_read_wait_recover() Since we will always recover the sectors, the old name is no longer accurate. Furthermore since recovery is already done in rmw_read_wait_recover(), we no longer need to call recovery_sectors() inside rmw_rbio(). Obviously this will have a performance impact, as we are doing more work during RMW cycle: - Fetch the data checksums - Do checksum verification for all data stripes - Do checksum verification again after repair But for full stripe write or cached rbio we won't have the overhead all, thus for fully optimized RAID56 workload (always full stripe write), there should be no extra overhead. To me, the extra overhead looks reasonable, as data consistency is way more important than performance. Signed-off-by:
-
Qu Wenruo authored
This is for later data checksum verification at RMW time. This patch will try to allocate the needed memory for a locked rbio if the rbio is for data exclusively (we don't want to handle mixed bg yet). The memory will be released when the rbio is finished. Signed-off-by:
-
Qu Wenruo authored
Although we have an existing function, btrfs_lookup_csums_range(), to find all data checksums for a range, it's based on a btrfs_ordered_sum list. For the incoming RAID56 data checksum verification at RMW time, we don't want to waste time by allocating temporary memory. So this patch will introduce a new helper, btrfs_lookup_csums_bitmap(). It will use bitmap based result, which will be a perfect fit for later RAID56 usage. Signed-off-by:
-
Qu Wenruo authored
The refactoring involves the following parts: - Introduce bytes_to_csum_size() and csum_size_to_bytes() helpers As we have quite some open-coded calculations, some of them are even split into two assignments just to fit 80 chars limit. - Remove the @csum_size parameter from max_ordered_sum_bytes() Csum size can be fetched from @fs_info. And we will use the csum_size_to_bytes() helper anyway. - Add a comment explaining how we handle the first search result - Use newly introduced helpers to cleanup btrfs_lookup_csums_range() - Move variables declaration to the minimal scope - Never mix number of sectors with bytes There are several locations doing things like: size = min_t(size_t, csum_end - start, max_ordered_sum_bytes(fs_info)); ... size >>= fs_info->sectorsize_bits Or offset = (start - key.offset) >> fs_info->sectorsize_bits; offset *= csum_size; Make sure these variables can only represent BYTES inside the function, by using the above bytes_to_csum_size() helpers. Signed-off-by:
-
Li zeming authored
The __GFP_NOFAIL flag could loop indefinitely when allocation memory in alloc_btrfs_io_context. The callers starting from __btrfs_map_block already handle errors so it's safe to drop the flag. Signed-off-by:
Li zeming <zeming@nfschina.com> Reviewed-by:
-
Qu Wenruo authored
[BUG] If dev-replace failed to re-construct its data/metadata, the kernel message would be incorrect for the missing device: BTRFS info (device dm-1): dev_replace from <missing disk> (devid 2) to /dev/mapper/test-scratch2 started BTRFS error (device dm-1): failed to rebuild valid logical 38862848 for dev (efault) Note the above "dev (efault)" of the second line. While the first line is properly reporting "<missing disk>". [CAUSE] Although dev-replace is using btrfs_dev_name(), the heavy lifting work is still done by scrub (scrub is reused by both dev-replace and regular scrub). Unfortunately scrub code never uses btrfs_dev_name() helper, as it's only declared locally inside dev-replace.c. [FIX] Fix the output by: - Move the btrfs_dev_name() helper to volumes.h - Use btrfs_dev_name() to replace open-coded rcu_str_deref() calls Only zoned code is not touched, as I'm not familiar with degraded zoned code. - Constify return value and parameter Now the output looks pretty sane: BTRFS info (device dm-1): dev_replace from <missing disk> (devid 2) to /dev/mapper/test-scratch2 started BTRFS error (device dm-1): failed to rebuild valid logical 38862848 for dev <missing disk> Reviewed-by:
Anand Jain <anand.jain@oracle.com> Signed-off-by:
-
Filipe Manana authored
During lseek (SEEK_HOLE/DATA), whenever we find a hole or prealloc extent, we will look for delalloc in that range, and one of the things we do for that is to find out ranges in the inode's io_tree marked with EXTENT_DELALLOC, using calls to count_range_bits(). Typically there's a single, or few, searches in the io_tree for delalloc per lseek call. However it's common for applications to keep calling lseek with SEEK_HOLE and SEEK_DATA to find where extents and holes are in a file, read the extents and skip holes in order to avoid unnecessary IO and save disk space by preserving holes. One popular user is the cp utility from coreutils. Starting with coreutils 9.0, cp uses SEEK_HOLE and SEEK_DATA to iterate over the extents of a file. Before 9.0, it used fiemap to figure out where holes and extents are in the source file. Another popular user is the tar utility when used with the --sparse / -S option to detect and preserve holes. Given that the pattern is to keep calling lseek with a start offset that matches the returned offset from the previous lseek call, we can benefit from caching the last extent state visited in count_range_bits() and use it for the next count_range_bits() from the next lseek call. Example, the following strace excerpt from running tar: $ strace tar cJSvf foo.tar.xz qemu_disk_file.raw (...) lseek(5, 125019574272, SEEK_HOLE) = 125024989184 lseek(5, 125024989184, SEEK_DATA) = 125024993280 lseek(5, 125024993280, SEEK_HOLE) = 125025239040 lseek(5, 125025239040, SEEK_DATA) = 125025255424 lseek(5, 125025255424, SEEK_HOLE) = 125025353728 lseek(5, 125025353728, SEEK_DATA) = 125025357824 lseek(5, 125025357824, SEEK_HOLE) = 125026766848 lseek(5, 125026766848, SEEK_DATA) = 125026770944 lseek(5, 125026770944, SEEK_HOLE) = 125027053568 (...) Shows that pattern, which is the same as with cp from coreutils 9.0+. So start using a cached state for the delalloc searches in lseek, and store it in struct file's private data so that it can be reused across lseek calls. This change is part of a patchset that is comprised of the following patches: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek The following test was run before and after applying the whole patchset: $ cat test-cp.sh #!/bin/bash DEV=/dev/sdh MNT=/mnt/sdh # coreutils 8.32, cp uses fiemap to detect holes and extents #CP_PROG=/usr/bin/cp # coreutils 9.1, cp uses SEEK_HOLE/DATA to detect holes and extents CP_PROG=/home/fdmanana/git/hub/coreutils/src/cp umount $DEV &> /dev/null mkfs.btrfs -f $DEV mount $DEV $MNT FILE_SIZE=$((1024 * 1024 * 1024)) echo "Creating file with a size of $((FILE_SIZE / 1024 / 1024))M" # Create a very sparse file, where each extent has a length of 4K and # is preceded by a 4K hole and followed by another 4K hole. start=$(date +%s%N) echo -n > $MNT/foobar for ((off = 0; off < $FILE_SIZE; off += 8192)); do xfs_io -c "pwrite -S 0xab $off 4K" $MNT/foobar > /dev/null echo -ne "\r$off / $FILE_SIZE ..." done end=$(date +%s%N) echo -e "\nFile created ($(( (end - start) / 1000000 )) milliseconds)" start=$(date +%s%N) $CP_PROG $MNT/foobar /dev/null end=$(date +%s%N) dur=$(( (end - start) / 1000000 )) echo "cp took $dur milliseconds with data/metadata cached and delalloc" # Flush all delalloc. sync start=$(date +%s%N) $CP_PROG $MNT/foobar /dev/null end=$(date +%s%N) dur=$(( (end - start) / 1000000 )) echo "cp took $dur milliseconds with data/metadata cached and no delalloc" # Unmount and mount again to test the case without any metadata # loaded in memory. umount $MNT mount $DEV $MNT start=$(date +%s%N) $CP_PROG $MNT/foobar /dev/null end=$(date +%s%N) dur=$(( (end - start) / 1000000 )) echo "cp took $dur milliseconds without data/metadata cached and no delalloc" umount $MNT The results, running on a box with a non-debug kernel (Debian's default kernel config), were the following: 128M file, before patchset: cp took 16574 milliseconds with data/metadata cached and delalloc cp took 122 milliseconds with data/metadata cached and no delalloc cp took 20144 milliseconds without data/metadata cached and no delalloc 128M file, after patchset: cp took 6277 milliseconds with data/metadata cached and delalloc cp took 109 milliseconds with data/metadata cached and no delalloc cp took 210 milliseconds without data/metadata cached and no delalloc 512M file, before patchset: cp took 14369 milliseconds with data/metadata cached and delalloc cp took 429 milliseconds with data/metadata cached and no delalloc cp took 88034 milliseconds without data/metadata cached and no delalloc 512M file, after patchset: cp took 12106 milliseconds with data/metadata cached and delalloc cp took 427 milliseconds with data/metadata cached and no delalloc cp took 824 milliseconds without data/metadata cached and no delalloc 1G file, before patchset: cp took 10074 milliseconds with data/metadata cached and delalloc cp took 886 milliseconds with data/metadata cached and no delalloc cp took 181261 milliseconds without data/metadata cached and no delalloc 1G file, after patchset: cp took 3320 milliseconds with data/metadata cached and delalloc cp took 880 milliseconds with data/metadata cached and no delalloc cp took 1801 milliseconds without data/metadata cached and no delalloc Reported-by:Wang Yugui <wangyugui@e16-tech.com> Link: https://lore.kernel.org/linux-btrfs/20221106073028.71F9.409509F4@e16-tech.com/ Link: https://lore.kernel.org/linux-btrfs/CAL3q7H5NSVicm7nYBJ7x8fFkDpno8z3PYt5aPU43Bajc1H0h1Q@mail.gmail.com/Signed-off-by:
Filipe Manana <fdmanana@suse.com> Signed-off-by:
-
Filipe Manana authored
During fiemap, whenever we find a hole or prealloc extent, we will look for delalloc in that range, and one of the things we do for that is to find out ranges in the inode's io_tree marked with EXTENT_DELALLOC, using calls to count_range_bits(). Since we process file extents from left to right, if we have a file with several holes or prealloc extents, we benefit from keeping a cached extent state record for calls to count_range_bits(). Most of the time the last extent state record we visited in one call to count_range_bits() matches the first extent state record we will use in the next call to count_range_bits(), so there's a benefit here. So use an extent state record to cache results from count_range_bits() calls during fiemap. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Filipe Manana authored
The comment for count_range_bits() mentions that the search is fast if we are asking for a range with the EXTENT_DIRTY bit set. However that is no longer true since we don't use that bit and the optimization for that was removed in: commit 71528e9e ("btrfs: get rid of extent_io_tree::dirty_bytes") So remove that part of the comment mentioning the no longer existing optimized case, and, while at it, add proper documentation describing the purpose, arguments and return value of the function. Signed-off-by:
-
Filipe Manana authored
An inode's io_tree can be quite large and there are cases where due to delalloc it can have thousands of extent state records, which makes the red black tree have a depth of 10 or more, making the operation of count_range_bits() slow if we repeatedly call it for a range that starts where, or after, the previous one we called it for. Such use cases are when searching for delalloc in a file range that corresponds to a hole or a prealloc extent, which is done during lseek SEEK_HOLE/DATA and fiemap. So introduce a cached state parameter to count_range_bits() which we use to store the last extent state record we visited, and then allow the caller to pass it again on its next call to count_range_bits(). The next patches in the series will make fiemap and lseek use the new parameter. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Filipe Manana authored
There are no more users of btrfs_next_extent_map(), the previous patch in the series ("btrfs: search for delalloc more efficiently during lseek/fiemap") removed the last usage of the function, so delete it. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by: -
Filipe Manana authored
During lseek (SEEK_HOLE/DATA) and fiemap, when processing a file range that corresponds to a hole or a prealloc extent, we have to check if there's any delalloc in the range. We do it by searching for delalloc ranges in the inode's io_tree (for unflushed delalloc) and in the inode's extent map tree (for delalloc that is flushing). We avoid searching the extent map tree if the number of outstanding extents is 0, as in that case we can't have extent maps for our search range in the tree that correspond to delalloc that is flushing. However if we have any unflushed delalloc, due to buffered writes or mmap writes, then the outstanding extents counter is not 0 and we'll search the extent map tree. The tree may be large because it can have lots of extent maps that were loaded by reads or created by previous writes, therefore taking a significant time to search the tree, specially if have a file with a lot of holes and/or prealloc extents. We can improve on this by instead of searching the extent map tree, searching the ordered extents tree of the inode, since when delalloc is flushing we create an ordered extent along with the new extent map, while holding the respective file range locked in the inode's io_tree. The ordered extents tree is typically much smaller, since ordered extents have a short life and get removed from the tree once they are completed, while extent maps can stay for a very long time in the extent map tree, either created by previous writes or loaded by read operations. So use the ordered extents tree instead of the extent maps tree. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Filipe Manana authored
During lseek (SEEK_HOLE/DATA) and fiemap, when processing a file range that corresponds to a hole or a prealloc extent, if we find that there is no delalloc marked in the inode's io_tree but there is delalloc due to an extent map in the io tree, then on the next iteration that calls find_delalloc_subrange() we can skip searching the io tree again, since on the first call we had no delalloc in the io tree for the whole range. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Filipe Manana authored
During fiemap and lseek (SEEK_HOLE/DATA), when looking for delalloc in a range corresponding to a hole or a prealloc extent, if we found the whole range marked as delalloc in the inode's io_tree, then we can terminate immediately and avoid searching the extent map tree. If not, and if the found delalloc starts at the same offset of our search start but ends before our search range's end, then we can adjust the search range for the search in the extent map tree. So implement those changes. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Filipe Manana authored
We don't need to set the EXTENT_UPDATE bit in an inode's io_tree to mark a range as uptodate, we rely on the pages themselves being uptodate - page reading is not triggered for already uptodate pages. Recently we removed most use of the EXTENT_UPTODATE for buffered IO with commit 52b029f4 ("btrfs: remove unnecessary EXTENT_UPTODATE state in buffered I/O path"), but there were a few leftovers, namely when reading from holes and successfully finishing read repair. These leftovers are unnecessarily making an inode's tree larger and deeper, slowing down searches on it. So remove all the leftovers. This change is part of a patchset that has the goal to make performance better for applications that use lseek's SEEK_HOLE and SEEK_DATA modes to iterate over the extents of a file. Two examples are the cp program from coreutils 9.0+ and the tar program (when using its --sparse / -S option). A sample test and results are listed in the changelog of the last patch in the series: 1/9 btrfs: remove leftover setting of EXTENT_UPTODATE state in an inode's io_tree 2/9 btrfs: add an early exit when searching for delalloc range for lseek/fiemap 3/9 btrfs: skip unnecessary delalloc searches during lseek/fiemap 4/9 btrfs: search for delalloc more efficiently during lseek/fiemap 5/9 btrfs: remove no longer used btrfs_next_extent_map() 6/9 btrfs: allow passing a cached state record to count_range_bits() 7/9 btrfs: update stale comment for count_range_bits() 8/9 btrfs: use cached state when looking for delalloc ranges with fiemap 9/9 btrfs: use cached state when looking for delalloc ranges with lseek Reported-by:
-
Qu Wenruo authored
[BACKGROUND] Although both btrfs metadata and data has their read time verification done at endio time (btrfs_validate_metadata_buffer() and btrfs_verify_data_csum()), metadata has extra verification, mostly parentness check including first key/transid/owner_root/level, done at read_tree_block() and btrfs_read_extent_buffer(). On the other hand, all the data verification is done at endio context. [ENHANCEMENT] This patch will make a new union in btrfs_bio, taking the space of the old data checksums, thus it will not increase the memory usage. With that extra btrfs_tree_parent_check inside btrfs_bio, we can just pass the check parameter into read_extent_buffer_pages(), and before submitting the bio, we can copy the check structure into btrfs_bio. And finally at endio time, we can grab btrfs_bio::parent_check and pass it to validate_extent_buffer(), to move the remaining checks into it. This brings the following benefits: - Much simpler btrfs_read_extent_buffer() Now it only needs to iterate through all mirrors. - Simpler read-time transid check Previously we go verify_parent_transid() after reading out the extent buffer. Now the transid check is done inside the endio function, no other code can modify the content. Thus no need to use the extent lock anymore. Signed-off-by:
-
Qu Wenruo authored
There are several different tree block parentness check parameters used across several helpers: - level Mandatory - transid Under most cases it's mandatory, but there are several backref cases which skips this check. - owner_root - first_key Utilized by most top-down tree search routine. Otherwise can be skipped. Those four members are not always mandatory checks, and some of them are the same u64, which means if some arguments got swapped compiler will not catch it. Furthermore if we're going to further expand the parentness check, we need to modify quite some helpers just to add one more parameter. This patch will concentrate all these members into a structure called btrfs_tree_parent_check, and pass that structure for the following helpers: - btrfs_read_extent_buffer() - read_tree_block() Signed-off-by:
-
Anand Jain authored
There is a repeating code section in the parent function after calling btrfs_alloc_device(), as below: name = rcu_string_strdup(path, GFP_...); if (!name) { btrfs_free_device(device); return ERR_PTR(-ENOMEM); } rcu_assign_pointer(device->name, name); Except in add_missing_dev() for obvious reasons. This patch consolidates that repeating code into the btrfs_alloc_device() itself so that the parent function doesn't have to duplicate code. This consolidation also helps to review issues regarding RCU lock violation with device->name. Parent function device_list_add() and add_missing_dev() use GFP_NOFS for the allocation, whereas the rest of the parent functions use GFP_KERNEL, so bring the NOFS allocation context using memalloc_nofs_save() in the function device_list_add() and add_missing_dev() is already doing it. Signed-off-by: -
David Sterba authored
The input buffers passed down to compression must never be changed, switch type to u8 as it's a raw byte buffer and use const. Reviewed-by:
-
Qu Wenruo authored
Since all the recovery paths have been migrated to the new error bitmap based system, we can remove the old stripe number based system. This cleanup involves one behavior change: - Rebuild rbio can no longer be merged Previously a rebuild rbio (caused by retry after data csum mismatch) can be merged, if the error happens in the same stripe. But with the new error bitmap based solution, it's much harder to compare error bitmaps. So here we just don't merge rebuild rbio at all. This may introduce some performance impact at extreme corner cases, but we're willing to take it. Other than that, this patch will cleanup the following members: - rbio::faila - rbio::failb They will be replaced by per-vertical stripe check, which is more accurate. - rbio::error It will be replace by per-vertical stripe error bitmap check. - Allow get_rbio_vertical_errors() to accept NULL pointers for @faila and @failb Some call sites only want to check if we have errors beyond the tolerance. Signed-off-by:
-
Qu Wenruo authored
Since we have rbio::error_bitmap to indicate exactly where the errors are (including read error and csum mismatch error), we can make recovery path more accurate. For example: 0 32K 64K Data 1 |XXXXXXXX| | Data 2 | |XXXXXXXXX| Parity | | | 1) Get csum mismatch when reading data 1 [0, 32K) 2) Mark corresponding range error The old code will mark the whole data 1 stripe as error. While the new code will only mark data 1 [0, 32K) as error. 3) Recovery path The old code will recover data 1 [0, 64K), all using Data 2 and parity. This means, Data 1 [32K, 64K) will be corrupted data, as data 2 [32K, 64K) is already corrupted. While the new code will only recover data 1 [0, 32K), as only that range has error so far. This new behavior can avoid populating rbio cache with incorrect data. Signed-off-by: -
Qu Wenruo authored
Currently btrfs raid56 uses btrfs_raid_bio::faila and failb to indicate which stripe(s) had IO errors. But that has some problems: - If one sector failed csum check, the whole stripe where the corruption is will be marked error. This can reduce the chance we do recover, like this: 0 4K 8K Data 1 |XX| | Data 2 | |XX| Parity | | | In above case, 0~4K in data 1 should be recovered using data 2 and parity, while 4K~8K in data 2 should be recovered using data 1 and parity. Currently if we trigger read on 0~4K of data 1, we will also recover 4K~8K of data 1 using corrupted data 2 and parity, causing wrong result in rbio cache. - Harder to expand for future M-N scheme As we're limited to just faila/b, two corruptions. - Harder to expand to handle extra csum errors This can be problematic if we start to do csum verification. This patch will introduce an extra @error_bitmap, where one bit represents error that happened for that sector. The choice to introduce a new error bitmap other than reusing sector_ptr, is to avoid extra search between rbio::stripe_sectors[] and rbio::bio_sectors[]. Since we can submit bio using sectors from both sectors, doing proper search on both array will more complex. Although the new bitmap will take extra memory, later we can remove things like @error and faila/b to save some memory. Currently the new error bitmap and failab mechanism coexists, the error bitmap is only updated at endio time and recover entrance. Signed-off-by: -
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is mostly using internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is mostly using internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The async_chunk::inode structure is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
David Sterba authored
The function is for internal interfaces so we should use the btrfs_inode. Reviewed-by:
-
