fix(fuse) support FUSE submounts and fix virtiofs mount semantics#1968
Open
xboHodx wants to merge 21 commits into
Open
fix(fuse) support FUSE submounts and fix virtiofs mount semantics#1968xboHodx wants to merge 21 commits into
xboHodx wants to merge 21 commits into
Conversation
github-actions
Bot
added
the
ambiguous
xboHodx
force-pushed
the
kata-split/fuse-vfs
branch
2 times, most recently
from
c3e8983 to
b029c4e
Compare
xboHodx
changed the title
github-actions
Bot
removed
the
ambiguous
xboHodx
force-pushed
the
kata-split/fuse-vfs
branch
from
83398cc to
b7eb76c
Compare
xboHodx
changed the title
github-actions
Bot
added
the
ambiguous
Contributor
Author
当前pr的临时修复
这可以修复一部分问题，尤其是避免把 |
- Enhance the sys_mount implementation to support various mount operations including bind mounts and remounts, with improved handling of mount flags and target paths. - Introduce new tests for FUSE to validate behavior with zero file handles, fsync operations without open, and correct handling of open flags in accordance with Linux standards. - Update the simple filesystem structure to accommodate open file handles and flags, ensuring compatibility with the extended FUSE operations. - Refactor FUSE daemon handling to track and validate file handle usage during open, read, write, and release operations, improving robustness and correctness of the FUSE interface. Signed-off-by: xboHodx <2679933924@qq.com>
xboHodx
force-pushed
the
kata-split/fuse-vfs
branch
from
b7eb76c to
5177f9e
Compare
xboHodx
changed the title
github-actions
Bot
removed
the
ambiguous
Member
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5177f9e70b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
Signed-off-by: longjin <longjin@dragonos.org>
…, and direct I/O lock-owner - Cache ENOSYS for FUSE_FLUSH, FUSE_FSYNC, FUSE_FSYNCDIR to avoid repeated round-trips when the daemon does not implement these ops - Move FUSE_FLUSH from IndexNode::close to new flush_file(), called at fd-close time so errors propagate to userspace via close(2) - Send FUSE_RELEASE asynchronously via request_nocreds_background to prevent close(2) from blocking on daemon release replies - Add DroppedFd deferred-close mechanism: flush + posix-lock release happen after fd-table lock is dropped, fixing lock-ordering issues - Pass lock_owner in FUSE_READ/FUSE_WRITE with FUSE_READ_LOCKOWNER / FUSE_WRITE_LOCKOWNER flags for direct I/O - Add FuseOpenContext capturing request cred, lock owner, node generation, and writeback error cursor at open time - Validate node generation on writeback writes to detect stale handles - Add direct_io_lock per FuseNode and prepare_direct_io_range for cache invalidation before direct I/O writes - Check and advance per-open writeback errors in fsync/flush paths - Add dunitest coverage for ENOSYS caching, flush error propagation, and close semantics Signed-off-by: longjin <longjin@dragonos.org>
fslongjin
force-pushed
the
kata-split/fuse-vfs
branch
from
8bce03f to
8eb9d40
Compare
Member
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8eb9d407d6
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
- Add request_with_cred to FuseConn so write requests carry the open-time credentials (pid/uid/gid) to the FUSE daemon, matching Linux semantics. - Replace invalidate_read + update_clean_page with update_ready_page in the write path: instead of discarding the whole page cache on every write, merge written data into existing up-to-date pages so that overlapping mmap reads see the latest bytes. - Acquire the page-cache invalidate-write lock around cached writes and wait for any in-flight writeback in the overlapping range before issuing the FUSE_WRITE to prevent stale writeback from overwriting newer data. - Move invalidate_read into prepare_*_writeback_entry so that page faults are blocked during the state transition, not after the page contents have already been read. - Add dunitest SharedMmapDirtyThenPwriteKeepsLatestData verifying that pwrite over a dirty mmap page updates the page cache and that the subsequent msync writes back the latest (pwrite) data. - Extend the msync writeback test to assert that write requests carry the same pid as the preceding open. Signed-off-by: longjin <longjin@dragonos.org>
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.
主要改动
FUSE_SUBMOUNTS和FUSE_ATTR_SUBMOUNT支持，在 lookup 到子挂载标记目录时自动创建 DragonOS 子挂载。FuseNode状态：ESTALE；FUSE_WRITEBACK_CACHE，避免语义不完整。O_CREAT/O_EXCL/O_NOCTTY；FUSE_ATOMIC_O_TRUNC不支持时过滤O_TRUNC；fh=0；fuse:<nodeid>之类 synthetic path 破坏路径重写。测试
新增/扩展 FUSE dunitest 覆盖：
FuseExtended.OpenReturnsZeroFhIsValidFuseExtended.NoOpenFsyncUsesZeroFhFuseExtended.OpenFlagsMatchLinuxMaskFuseExtended.InitRequestsLinuxNoOpenSupport同时扩展 simple FUSE test daemon，支持记录 open/read/write/fsync/release 的 fh、flags 和计数，用于验证 Linux 兼容行为。