[PATCH 0/8] U-boot: fs: add generic unaligned read offset handling
wqu at suse.com
Wed Jun 29 21:38:21 AEST 2022
- More (manual) testing
Unfortunately, in the latest master (75967970850a), the fs-tests.sh
always seems to hang at preparing the fs image.
Thus still has to do manual testing, tested btrfs, ext4 and fat, with
aligned and unaligned read, also added soft link read, all looks fine here.
Extra testing is still appreciated.
- Two more btrfs specific bug fixes
All exposed during manual tests
- Remove the tailing unaligned block handling
In fact, all fses can easily handle such case, just a min() call is
- Remove the support for sandboxfs
Since it's using read() calls, really no need to do block alignment
- Enhanced blocksize check
Ensure the returned blocksize is not only non-error, but also
This patchset can be fetched from github:
Unlike FUSE/Kernel which always pass aligned read range, U-boot fs code
just pass the request range to underlying fses.
Under most case, this works fine, as U-boot only really needs to read
the whole file (aka, 0 for both offset and len, len will be later
determined using file size).
But if some advanced user/script wants to extract kernel/initramfs from
combined image, we may need to do unaligned read in that case.
This patchset will handle unaligned read range in _fs_read():
- Get blocksize of the underlying fs
- Read the leading block contianing the unaligned range
The full block will be stored in a local buffer, then only copy
the bytes in the unaligned range into the destination buffer.
If the first block covers the whole range, we just call it aday.
- Read the remaining range which starts at block aligned offset
For most common case, which is 0 offset and 0 len, the code will not
be changed at all.
Just one extra get_blocksize() call, and for FAT/Btrfs/EROFS they all have
cached blocksize, thus it takes almost no extra cost.
Although for EXT4, it doesn't seem to cache the blocksize globally,
thus has to do a path resolve and grab the blocksize.
The involved problem is:
- Extra path resolving
All those supported fs may have to do one extra path resolve if the
read offset is not aligned.
For EXT4, it will do one extra path resolve just to grab the
For data read which starts at offset 0 (the most common case), it
should cause *NO* difference in performance.
As the extra handling is only for unaligned offset.
The common path is not really modified.
- Btrfs (manually tested*)
- Ext4 (manually tested)
- FAT (manually tested)
- ubifs (unable to test, due to compile failure)
*: Failed to get the test cases run, thus have to go sandbox mode, and
attach an image with target fs, load the target file (with unaligned
range) and compare the result using md5sum.
For EXT4/FAT, they may need extra cleanup, as their existing unaligned
range handling is no longer needed anymore, cleaning them up should free
more code lines than the added one.
Just not confident enough to modify them all by myself.
They don't support non-zero offset, thus it can not handle the block
Need extra help to add block aligned offset support.
They all use read() directly, no need to do alignment check at all.
Extra testing/feedback is always appreciated.
Qu Wenruo (8):
fs: fat: unexport file_fat_read_at()
fs: btrfs: fix a bug which no data get read if the length is not 0
fs: btrfs: fix a crash if specified range is beyond file size
fs: btrfs: move the unaligned read code to _fs_read() for btrfs
fs: ext4: rely on _fs_read() to handle leading unaligned block read
fs: fat: rely on higher layer to get block aligned read range
fs: ubifs: rely on higher layer to do unaligned read
fs: erofs: add unaligned read range handling
fs/btrfs/btrfs.c | 33 ++++++++---
fs/btrfs/inode.c | 89 +++--------------------------
fs/erofs/internal.h | 1 +
fs/erofs/super.c | 6 ++
fs/ext4/ext4fs.c | 22 +++++++
fs/fat/fat.c | 17 +++++-
fs/fs.c | 130 +++++++++++++++++++++++++++++++++++++++---
fs/ubifs/ubifs.c | 13 +++--
include/btrfs.h | 1 +
include/erofs.h | 1 +
include/ext4fs.h | 1 +
include/fat.h | 3 +-
include/ubifs_uboot.h | 1 +
13 files changed, 212 insertions(+), 106 deletions(-)
More information about the Linux-erofs