test[CI][littlefs]: add RT-Thread QEMU smoke coverage#36
Merged
Conversation
add DFS v2 mount, dentry, vnode and filesystem type registration paths for littlefs centralize DFS ABI differences across DFS v2, RT-Thread 5.x DFS v1 and legacy DFS layouts adapt read/write/lseek signatures and explicit offset handling for DFS v2 keep legacy DFS v1 behavior through compatibility macros verify basic build and QEMU littlefs DFS v2 smoke flow on bsp/qemu-vexpress-a9
avoid copying DFS_PATH_MAX bytes into struct dirent d_name use the actual d_name buffer size when filling littlefs directory entries keep d_namlen consistent with the copied name length after truncation fix stack corruption observed when listing littlefs directories on DFS v2
add RT-Thread qemu-vexpress-a9 compile checks for littlefs DFS profiles cover legacy DFS, RT-Thread 5.x DFS v1, DFS v2, and master builds reuse a prepared RT-Thread checkout across profile builds to reduce repeated downloads upload per-profile build logs for CI failure diagnosis
WalkthroughThis PR refactors dfs_lfs.c to support DFS v1/v2 ABIs and adds end-to-end CI: a composite action to set up toolchain/env, a build script that clones/prepares RT-Thread and runs SCons, a GitHub Actions workflow, pinned CI Python deps, QEMU MTD emulation, and smoke-test applications. ChangesDFS/LittleFS Multi-ABI Support & CI Integration
Sequence DiagramsequenceDiagram
actor User
participant GH as GitHub Actions
participant Cache as Cache/Artifacts
participant Action as setup-rtthread-build
participant Build as build.sh
participant RTT as RT-Thread
participant QEMU as QEMU Emulator
User->>GH: push / open PR / dispatch
GH->>Action: run composite action (setup toolchain/env)
Action->>Cache: restore ARM GCC cache
alt cache miss
Action->>Cache: download+verify toolchain -> install
end
GH->>Build: invoke build.sh
Build->>RTT: prepare sources / copy package / apply Kconfig
Build->>RTT: run SCons -> produce ELF
Build->>Build: verify ELF symbols
alt rtthread-run-smoke enabled
Build->>QEMU: launch emulator
QEMU->>QEMU: run littlefs_smoke_run (write/read/list/remount)
QEMU-->>Build: emit PASS/FAIL marker in log
Build->>QEMU: stop emulator
end
Build->>Cache: upload logs/artifacts
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (1)
.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c (1)
73-86: ⚡ Quick winAlign macro guards with
dfs_lfs.cpattern for consistency.Lines 73 and 82 use
RT_VERSION_CHECKwithout adefined(RT_VERSION_CHECK)guard, whiledfs_lfs.cin the same codebase consistently wraps similar checks. Consider adding the guard for defensive programming consistency, even thoughRT_VERSION_CHECKis defined in all tested RT-Thread versions (v4.1.1 and later).Suggested fix
-#if RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0) +#if defined(RT_VERSION_CHECK) && (RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0)) typedef rt_ssize_t qemu_lfs_mtd_rw_ret_t; `#else` typedef rt_size_t qemu_lfs_mtd_rw_ret_t; -#endif /* RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0) */ +#endif /* defined(RT_VERSION_CHECK) && (RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0)) */ /** * `@brief` Length type used by RT-Thread MTD NOR callbacks. */ -#if RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0) +#if defined(RT_VERSION_CHECK) && (RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0)) typedef rt_size_t qemu_lfs_mtd_len_t; `#else` typedef rt_uint32_t qemu_lfs_mtd_len_t; -#endif /* RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0) */ +#endif /* defined(RT_VERSION_CHECK) && (RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0)) */🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c around lines 73 - 86, The RT_VERSION_CHECK macro uses should be guarded like in dfs_lfs.c to avoid undefined-macro issues: update the two typedef conditionals around qemu_lfs_mtd_rw_ret_t and qemu_lfs_mtd_len_t to use a combined preprocessor check (e.g., `#if` defined(RT_VERSION_CHECK) && RTTHREAD_VERSION >= RT_VERSION_CHECK(5, 0, 0)) and keep the existing else branches unchanged so behavior is identical when RT_VERSION_CHECK is unavailable.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/actions/setup-rtthread-build/action.yml:
- Around line 21-24: Change the input name and verification algorithm from MD5
to SHA-256: rename the input key arm-gcc-md5 to arm-gcc-sha256, update its
description to say "Expected SHA-256 checksum for the ARM GCC toolchain
archive", and replace the default MD5 hex with the correct SHA-256 hex. Then
update all places that reference arm-gcc-md5 (workflow steps, scripts, and
checksum checks) to use arm-gcc-sha256 and change the verification command to
use sha256sum (or an equivalent SHA-256 check) instead of md5sum so the CI
actually verifies using SHA-256.
In @.github/ci/rtthread-littlefs/build.sh:
- Around line 667-683: Wrap the blocking git clone in a timeout so stalled
clones don't hang the retry loop: replace the plain git clone call inside the
while loop with a timed invocation such as timeout --preserve-status
"$CLONE_TIMEOUT" git clone --depth 1 --branch "$rtthread_ref"
https://github.com/RT-Thread/rt-thread.git "$rtthread_dir" and ensure a sensible
default CLONE_TIMEOUT is defined (e.g. export
CLONE_TIMEOUT="${CLONE_TIMEOUT:-300}") so the loop reacts to timeout failures
and proceeds to the rm -rf / retry logic and ultimately calls fail if all
attempts exhaust.
In `@dfs_lfs.c`:
- Around line 1249-1255: The _dfs_lfs_dentry_rename handler currently always
calls _dfs_lfs_rename using old_dentry->mnt and will forward
new_dentry->pathname into the wrong FS when the destination is on a different
mount; modify _dfs_lfs_dentry_rename to compare old_dentry->mnt and
new_dentry->mnt and if they differ return -EXDEV, otherwise call
_dfs_lfs_rename(old_dentry->mnt, old_dentry->pathname, new_dentry->pathname).
Ensure you reference the dfs_dentry struct fields mnt and pathname and return
the POSIX cross-device error -EXDEV for cross-mount renames.
- Around line 1135-1147: The code currently converts offset into an unsigned
directory index before calling lfs_dir_seek, which permits negative or unaligned
offsets to wrap or desynchronize; before computing off and calling lfs_dir_seek
(in the DFS_LFS_FILE_TYPE(file) == FT_DIRECTORY branch), validate offset is
non-negative and aligned to sizeof(struct dirent) and return -EINVAL if not;
then compute off = offset / sizeof(struct dirent) and call lfs_dir_seek as
before, ensuring any casts to lfs_off_t/lfs_soff_t are done after the
validation.
---
Nitpick comments:
In @.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c:
- Around line 73-86: The RT_VERSION_CHECK macro uses should be guarded like in
dfs_lfs.c to avoid undefined-macro issues: update the two typedef conditionals
around qemu_lfs_mtd_rw_ret_t and qemu_lfs_mtd_len_t to use a combined
preprocessor check (e.g., `#if` defined(RT_VERSION_CHECK) && RTTHREAD_VERSION >=
RT_VERSION_CHECK(5, 0, 0)) and keep the existing else branches unchanged so
behavior is identical when RT_VERSION_CHECK is unavailable.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 489be611-da60-4f28-b0a8-7b939a837a43
📒 Files selected for processing (8)
.github/actions/setup-rtthread-build/action.yml.github/ci/rtthread-littlefs/build.sh.github/ci/rtthread-littlefs/requirements.txt.github/ci/rtthread-littlefs/smoke/littlefs_compile_check.c.github/ci/rtthread-littlefs/smoke/littlefs_smoke.c.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c.github/workflows/rtthread.ymldfs_lfs.c
Comment on lines
+21
to
+24
| arm-gcc-md5: | ||
| description: Expected MD5 checksum for the ARM GCC toolchain archive. | ||
| required: false | ||
| default: 2383e4eb4ea23f248d33adc70dc3227e |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
cat -n .github/actions/setup-rtthread-build/action.ymlRepository: RT-Thread-packages/littlefs
Length of output: 9548
Use SHA-256 instead of MD5 for the toolchain archive check.
The downloaded compiler is a high-trust artifact in the CI supply chain. MD5 is cryptographically broken and should not be used for integrity verification of build tools, as it cannot guarantee protection against intentional modification.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.github/actions/setup-rtthread-build/action.yml around lines 21 - 24, Change
the input name and verification algorithm from MD5 to SHA-256: rename the input
key arm-gcc-md5 to arm-gcc-sha256, update its description to say "Expected
SHA-256 checksum for the ARM GCC toolchain archive", and replace the default MD5
hex with the correct SHA-256 hex. Then update all places that reference
arm-gcc-md5 (workflow steps, scripts, and checksum checks) to use arm-gcc-sha256
and change the verification command to use sha256sum (or an equivalent SHA-256
check) instead of md5sum so the CI actually verifies using SHA-256.
- add QEMU littlefs smoke sources for RT-Thread BSP validation - split RT-Thread tag and master checks into separate GitHub Actions matrix jobs - add RT-Thread source archive/cache handling to reduce repeated clone cost - run compile and smoke verification against qemu-vexpress-a9 DFS v1/v2 profiles
reject cross-mount dentry rename with -EXDEV before using the old mount context validate directory lseek offsets before converting them to littlefs directory positions return -EINVAL for negative or non-dirent-aligned directory offsets
Collaborator
Author
|
@coderabbitai review |
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
.github/actions/setup-rtthread-build/action.yml (1)
21-24:⚠️ Potential issue | 🟠 Major | ⚡ Quick winMD5 integrity check still flagged.
The
arm-gcc-md5input and correspondingmd5sum -cverification at line 201 still rely on MD5 for a high-trust build artifact. SHA-256 is the appropriate algorithm here; ARM publishes SHA-256 sums alongside the toolchain archives.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/actions/setup-rtthread-build/action.yml around lines 21 - 24, The workflow still uses MD5: update the action input and verification to SHA-256 by renaming the input arm-gcc-md5 to arm-gcc-sha256 (and update its description), replace any md5sum -c usage with sha256sum -c in the verification step (the place currently invoking md5sum -c), and replace the default checksum value with the published SHA-256 for the ARM GCC archive so the integrity check uses SHA-256 end-to-end.
🧹 Nitpick comments (2)
.github/workflows/rtthread.yml (1)
70-77: 💤 Low valueSet a short
retention-dayson the inter-job source artifact.
rtthread-master-sourceis only consumed bytest-littlefs-masterin the same run, but defaults to 90-day retention. Settingretention-days: 1(or a similar small value) avoids accumulating large tarballs in the repository's artifact storage.♻️ Proposed change
- name: Upload RT-Thread master source uses: actions/upload-artifact@v4 with: name: rtthread-master-source path: | _ci/artifacts/rt-thread-master.tar _ci/artifacts/rt-thread-master.sha if-no-files-found: error + retention-days: 1🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/workflows/rtthread.yml around lines 70 - 77, The artifact upload step "Upload RT-Thread master source" currently leaves artifacts with the default retention and should be updated to avoid long-term storage; add a retention-days field under the step's with block (for the artifact named rtthread-master-source) — e.g., add retention-days: 1 so the actions/upload-artifact@v4 invocation that publishes rtthread-master-source keeps the tarball only for the short desired period..github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c (1)
25-29: 💤 Low valueDrop the unused
lfsrootMTD instance.Only
lfsnoris consumed bylittlefs_smoke.c; thelfsrootdevice is registered, zero-filled, and never mounted. Each instance carries a 512 KiB staticstorage[]buffer in BSS, so this wastes ~512 KiB and adds maintenance noise. ReduceQEMU_LFS_MTD_DEVICE_COUNTto 1 and drop thelfsrootentry unless a future test plans to use it.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c around lines 25 - 29, Remove the unused "lfsroot" MTD instance and shrink the static device table to only the used "lfsnor" entry: update QEMU_LFS_MTD_DEVICE_COUNT from 2 to 1 and change the qemu_lfs_mtd_devices initializer to contain just { .name = "lfsnor" }; ensure any code that iterates or indexes qemu_lfs_mtd_devices still respects the new count (symbols: QEMU_LFS_MTD_DEVICE_COUNT, qemu_lfs_mtd_devices, "lfsroot", "lfsnor", and any consumers such as littlefs_smoke.c).
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/ci/rtthread-littlefs/smoke/littlefs_smoke.c:
- Around line 270-274: The mkdir call creating LFS_SMOKE_DIR uses mode 0 which
yields no permission bits; change the mode passed to mkdir in the littlefs_smoke
setup to a sensible permission like 0755 so the directory is created
readable/executable by owner/group/others and subsequent filesystem layers can
traverse it; update the call that currently returns
littlefs_smoke_fail("mkdir-dir", ret) on error (the mkdir invocation referencing
LFS_SMOKE_DIR) to use 0755 as the mode argument.
---
Duplicate comments:
In @.github/actions/setup-rtthread-build/action.yml:
- Around line 21-24: The workflow still uses MD5: update the action input and
verification to SHA-256 by renaming the input arm-gcc-md5 to arm-gcc-sha256 (and
update its description), replace any md5sum -c usage with sha256sum -c in the
verification step (the place currently invoking md5sum -c), and replace the
default checksum value with the published SHA-256 for the ARM GCC archive so the
integrity check uses SHA-256 end-to-end.
---
Nitpick comments:
In @.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c:
- Around line 25-29: Remove the unused "lfsroot" MTD instance and shrink the
static device table to only the used "lfsnor" entry: update
QEMU_LFS_MTD_DEVICE_COUNT from 2 to 1 and change the qemu_lfs_mtd_devices
initializer to contain just { .name = "lfsnor" }; ensure any code that iterates
or indexes qemu_lfs_mtd_devices still respects the new count (symbols:
QEMU_LFS_MTD_DEVICE_COUNT, qemu_lfs_mtd_devices, "lfsroot", "lfsnor", and any
consumers such as littlefs_smoke.c).
In @.github/workflows/rtthread.yml:
- Around line 70-77: The artifact upload step "Upload RT-Thread master source"
currently leaves artifacts with the default retention and should be updated to
avoid long-term storage; add a retention-days field under the step's with block
(for the artifact named rtthread-master-source) — e.g., add retention-days: 1 so
the actions/upload-artifact@v4 invocation that publishes rtthread-master-source
keeps the tarball only for the short desired period.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 3af99aa9-59b4-43d4-986d-e316627f5114
📒 Files selected for processing (8)
.github/actions/setup-rtthread-build/action.yml.github/ci/rtthread-littlefs/build.sh.github/ci/rtthread-littlefs/requirements.txt.github/ci/rtthread-littlefs/smoke/littlefs_compile_check.c.github/ci/rtthread-littlefs/smoke/littlefs_smoke.c.github/ci/rtthread-littlefs/smoke/qemu_lfs_mtd.c.github/workflows/rtthread.ymldfs_lfs.c
🚧 Files skipped from review as they are similar to previous changes (1)
- dfs_lfs.c
Collaborator
Author
Rbb666
approved these changes
May 7, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
拉取/合并请求描述：(PR description)
[
为什么提交这份PR (why to submit this PR)
当前 RT-Thread littlefs 包 CI 主要覆盖多版本编译检查，缺少在 QEMU BSP 上执行 littlefs 文件系统基本读写、目录遍历和卸载重挂载流程的自动化 smoke test。
该 PR 增加 qemu-vexpress-a9 场景下的 littlefs smoke test，并将 RT-Thread tag 与 master 的测试拆分为独立矩阵任务，提升 CI 覆盖度和失败定位效率。
你的解决方案是什么 (what is your solution)
.github/ci/rtthread-littlefs/build.sh中增加可选的RTTHREAD_RUN_SMOKE流程。.github/workflows/rtthread.yml拆分为 tag profile 和 master profile 两类 job，并加入 ARM GCC、Python 依赖、RT-Thread source 和 QEMU 相关准备步骤。.github/ci/rtthread-littlefs/requirements.txt固定kconfiglib与scons版本。请提供验证的bsp和config (provide the config and bsp)
BSP:
bsp/qemu-vexpress-a9.config:
RTTHREAD_DFS_VERSION=v1RTTHREAD_DFS_VERSION=v2RTTHREAD_RUN_SMOKE=1RT_USING_DFSDFS_USING_POSIXDFS_USING_WORKDIRRT_USING_DFS_DEVFSRT_USING_DFS_ROMFSRT_USING_DFS_TMPFSRT_USING_MTD_NORPKG_USING_LITTLEFSRT_USING_DFS_V1或RT_USING_DFS_V2action: 本地输入中未提供可引用的 GitHub Actions 成功链接；提交 PR 前请替换为当前 PR 分支对应的成功 workflow 链接。
]
当前拉取/合并请求的状态 Intent for your PR
必须选择一项 Choose one (Mandatory):
代码质量 Code Quality：
我在这个拉取/合并请求中已经考虑了 As part of this pull request, I've considered the following:
#if 0代码，不包含已经被注释了的代码 All redundant code is removed and cleaned upSummary by CodeRabbit