Minor release: v2.11 #1103
Merged
Minor release: v2.11 #1103
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
"make test" on macOS: ``` using runner: ./runners/test_runner found 19 suites, 188 cases, 11242/11770 permutations running test_alloc: 12/12 cases, 207/207 perms running test_attrs: 4/4 cases, 20/20 perms running test_badblocks: 4/4 cases, 300/300 perms running test_bd: 5/5 cases, 85/85 perms running test_compat: 17/17 cases, 205/205 perms running test_dirs: 15/15 cases, 450/450 perms, 1756pls! running test_entries: 8/8 cases, 32/32 perms running test_evil: 8/8 cases, 105/105 perms running test_exhaustion: 5/5 cases, 85/85 perms running test_files: 10/10 cases, 7155/7155 perms, 9410pls! running test_interspersed: 4/4 cases, 190/190 perms, 2835pls! running test_move: 17/17 cases, 161/161 perms, 157pls! running test_orphans: 6/6 cases, 50/50 perms, 846pls! running test_paths: 33/33 cases, 325/325 perms running test_powerloss: 2/2 cases, 21/21 perms running test_relocations: 4/4 cases, 68/68 perms, 1612pls! running test_seek: 10/10 cases, 195/195 perms, 1050pls! running test_superblocks: 17/17 cases, 318/318 perms, 1437pls! running test_truncate: 7/7 cases, 1270/1270 perms, 9691pls! done: 11242/11242 passed, 0/11242 failed, 28794pls!, in 585.76s ```
This PR adds a new `lfs_fs_shrink`, which functions similarly to `lfs_fs_grow`, but supports reducing the block count. This functions first checks that none of the removed block are in use. If it is the case, it will fail.
These are the same as the related reentrant variants, but by opting out of powerloss testing, we can test a much larger number of states without having to worry about the impact on powerloss testing runtime. Bumped CYCLES from 20 -> 2000. This reveals an orphan remove bug found by Hugh-Baoa.
Long story short: There is a specific case where removing a directory
can trigger a deorphan pass, but lfs_remove did not check for this,
would try to clean up the (already cleaned) directory orphan, and
trigger an assert:
lfs.c:4890:assert: assert failed with false, expected eq true
LFS_ASSERT(lfs_tag_size(lfs->gstate.tag) > 0x000 || orphans >= 0);
The specific case being a remove commit that triggers a relocation that
creates an orphan.
This is also possible in lfs_rename, but only if you're renaming a
directory that implies a remove, which is a pretty rare operation.
---
This was probably an oversight introduced in the non-recursive commit
logic rework.
Fortunately the fix is to just check if we even have an orphan before
trying to remove it. We can rely on this instead of the file type, so
this fix shouldn't even increase the code size.
Found and root-caused by Hugh-Baoa
In 16 bit and 8 bit architectures, overflow and underflow issues were occuring while using functions lfs_frombe32 and lfs_fromle32
This is the same as the implicit Clang => NO_GCC behavior introduced by yamt, but with an explicit variable that can be assigned by users using other, non-gcc, compilers: $ NO_GCC=1 make Note, stack measurements are currently GCC specific: $ NO_GCC=1 make stack ... snip ... FileNotFoundError: [Errno 2] No such file or directory: 'lfs.ci' make: *** [Makefile:494: lfs.stack.csv] Error 1
Thanks to yamt, GCC-specific flags should now be disabled if compiling with clang. Dropping the explicit flags also doubles as a test that the NO_GCC inference works.
lfs_crc should be static if LFS_CRC is defined
fix: added uint32_t cast to the bitshift places
…phan Fix double deorphan caused by relocation mid dir remove
drop a few unsupported CFLAGS for clang
adapt the linker sections usage to mach-o
Add support for shrinking a filesystem
Tests passed ✓, Code: 17104 B (-0.1%), Stack: 1448 B (+0.0%), Structs: 812 B (+0.0%)
|
This prevents a "OSError: [Errno 18] Invalid cross-device link" if the temporary file was created on different filesystem (such as a tmpfs mount).
use shutil.move instead of os.rename to move file
Tests passed ✓, Code: 17104 B (-0.1%), Stack: 1448 B (+0.0%), Structs: 812 B (+0.0%)
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
A small minor release, driven by features proposed by @sosthene-nitrokey and @yamt, with bug fixes from @DvdGiessen, @selimkeles, and yours truly.
Bringing in:
Draft of release notes follows:
A small minor release, driven by features proposed by @sosthene-nitrokey and @yamt, with bug fixes from @DvdGiessen, @selimkeles, and yours truly.
What's new?
Thanks to @sosthene-nitrokey,
lfs_fs_grownow supports a limited form of filesystem shrinking (#1094)If compiled with
LFS_SHRINKNONRELOCATING,lfs_fs_growcan be used to reduce theblock_count, as long is it does not require block relocation. This provides a useful path for migrating from old filesystem images created when littlefs ignored the on-diskblock_count.Note that this is unreliable for general use. littlefs allocates blocks pseudorandomly as a part of wear-leveling, so
LFS_SHRINKNONRELOCATINGis unlikely to succeed even with small filesystems.Thanks to @yamt,
make testshould now work on Mac out of the box. This required linker and Makefile changes due to littlefs's test-runner quirks (#1090, #1091)As a side-effect, littlefs's Makefile can now compile with Clang. The only limitation being that stack measurements are currently broken. If anyone knows a good Clang alternative to GCC's
-fcallgraph-info=su, feel free to create an issue (or PR if you're willing to write a bit of python).littlefs's Makefile can also compile with other non-Clang non-GCC compilers if you set the
NO_GCCenvironment variable.@DvdGiessen fixed several bugs related to LFS_CRC (1095)
lfs_util.his included in more than one file.Thanks to @selimkeles,
lfs_fromle32/lfs_frombe32no longer risk overflow on 16-bit devices (#1100)Found by @Hugh-Baoa, fixed a double deorphan that can occur if an mdir relocates during a remove operation (#1099)
Thanks to @DvdGiessen,
changeprefix.pyno longer breaks if /tmp lives on a different filesystem (#1104)