Repository navigation
DAOS-19915 build: Update to libfuse 3.18.3 - #12
Merged
Merged
Conversation
Bumps [github/codeql-action](https://github.com/github/codeql-action) from 3.28.9 to 3.28.11. - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@9e8d078...6bb031a) --- updated-dependencies: - dependency-name: github/codeql-action dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
The main goal of this new file is to warn about the deadlocking risks when using the out-of-band notify operations while replying to other requests. Signed-off-by: Luis Henriques <luis@igalia.com>
Added PanFSto whitelist. This should allow us to mount gocryptfs onto the parallel filesystem. Signed-off-by: Jacob Riley <jnr0006@uah.edu>
Some applications use that for detection of features between distributions/libfuse version. Closes: libfuse#1163 Signed-off-by: Bernd Schubert <bschubert@ddn.com>
This fixes dynamic buffer allocation in commit 0e0f43b ("Reallocate fuse_session buffer...") I noticed that when I increased the default fuse buf size as possible in recent kernels. Signed-off-by: Bernd Schubert <bschubert@ddn.com>
se->buf_reallocable is true when reading /dev/fuse is handled from internal functions - we can set the variable in fuse_session_receive_buf_internal(). With that we also don't need to have it an _Atomic variable anymore. In _fuse_session_receive_buf() we can use "bool internal" to check if the buffer can be re-allocated. Signed-off-by: Bernd Schubert <bschubert@ddn.com>
Redirect stdin/stdout/stderr to /dev/null to prevent newly opened file descriptors from reusing these low numbers (0,1,2) and potential issues with that. Signed-off-by: Bernd Schubert <bschubert@ddn.com>
We have a non ideal configuration for different editors. The project uses linux style - emacs config should also correspond to that. Taken from https://www.kernel.org/doc/html/v6.13/process/coding-style.html#you-ve-made-a-mess-of-it with the attempt to update for c++. Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Closes: libfuse#1168 Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
This is to avoid checkpatch reports and might potentially avoid future issues if these macros would be part of conditions. Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
The file is going to be updated - should be more conform Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
32-bit conn->want flags been left to be ABI compatible to 3.10, even though the so version was changed. The more recent way is to use fuse_set_feature_flag(), which will use conn->want_ext. Given that we now have two flags (want and want_ext), we need to convert and that brought several issues - If the application sets conn->want, that needs to be set into the lower 32 bit of conn->want_ext. As the application might actually unset values, it really has to be a copy and not just 'or' - fixed now. - convert_to_conn_want_ext() actually needs to check for _modified_ conn->want and conn->want_ext - convert_to_conn_want_ext() must consider being called from high and lowlevel interfact, with different want_ext_default and want_default values. It is only a failure, if the application changed both, conn->want and conn->want_ext. This function was failing in issue libfuse#1171, because high level fuse_fs_init() was changing values and then lowlevel do_init() was incorrectly failing on that. This also adds a new test (test_want_conversion) and sets values into example/{hello.c,hello_ll.c} Also some more internal users of conn->want are converted to fuse_{set,unset}_feature_flag(). Closes: libfuse#1171 Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Thes commits have the habit to persistently fail, mostly with long lines. We don't have control over these commits, so let's ignore them. Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Bumps [github/codeql-action](https://github.com/github/codeql-action) from 3.28.11 to 3.28.13. - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@6bb031a...1b549b9) --- updated-dependencies: - dependency-name: github/codeql-action dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
When building and installing FUSE3 using Meson on CentOS8, the fuse3.pc file is installed in the /usr/local/lib64/pkgconfig directory. However, pkg-config does not search for fuse3.pc in this directory, leading to GCC compilation failures. This patch improves the compilation comments, helping users successfully execute the GCC command. Signed-off-by: izxl007 <zeng.zheng@zte.com.cn>
When building FUSE3 with Meson on CentOS 8, the fuse3.pc file gets installed in /usr/local/lib64/pkgconfig. Since pkg-config doesn't search this path by default, GCC compilation fails due to missing FUSE3 flags. This patch adds an instruction for setting PKG_CONFIG_PATH variable in README.compile to fix GCC compilation issues. Signed-off-by: izxl007 <zeng.zheng@zte.com.cn>
Copy over usdt.h from [1] as part of the prepatory work for adding tracepoints for requests. [1] https://github.com/libbpf/usdt Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
… replying
Add user statically-defined tracepoints for request lifecycle.
Verified by:
[machine]$ readelf -n /home/libfuse/build/lib/libfuse3.so | grep -A 1000 '.note.stapsdt'
Displaying notes found in: .note.stapsdt
Owner Data size Description
stapsdt 0x00000035 NT_STAPSDT (SystemTap probe descriptors)
Provider: libfuse
Name: request_receive
Location: 0x0000000000016360, Base: 0x00000000000002fc, Semaphore: 0x0000000000000000
Arguments: -4@%edi
stapsdt 0x00000055 NT_STAPSDT (SystemTap probe descriptors)
Provider: libfuse
Name: request_reply
Location: 0x0000000000017c0f, Base: 0x00000000000002fc, Semaphore: 0x0000000000000000
Arguments: 8@8(%r12) 4@(%r12) -4@4(%r12) -4@%r13d
stapsdt 0x0000003e NT_STAPSDT (SystemTap probe descriptors)
Provider: libfuse
Name: request_process
Location: 0x000000000001acf7, Base: 0x00000000000002fc, Semaphore: 0x0000000000000000
Arguments: 4@%r8d 4@%edx
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Add example bpftrace file for monitoring tracepoints. Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
linux/close_range.h is only available since kernel 5.9 and torvalds/linux@60997c3 resulting in the following build failure: ../util/fusermount.c:40:10: fatal error: linux/close_range.h: No such file or directory So let's check for header presence and emit HAVE_LINUX_CLOSE_RANGE_H accordingly and check for it when including <linux/close_range.h> and calling close_range() instead of checking for close_range() function in meson and check against HAVE_CLOSE_RANGE. Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
At the moment build fails due to lack of static_assert: https://gitlab.com/jolivain/buildroot/-/jobs/9606292537 this means that the check per date is not enough, so let's use meson to check if static_assert() is present or not and simplify fuse_static_assert() definition by only checking HAVE_STATIC_ASSERT. Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Default is having usdt disabled. Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
These are several buf size fixes 1) Set bufsize when HAVE_SPLICE is not defined. Addresses libfuse#1184 2) Check in the read retry condition for bufsize, i.e. the value passed to read and not for the buf->mem_size. Using buf->mem_size can be startup racy. Additionally we now also set bufsize on allocation to avoid these races. 3) Allocation and value assigned needs to follow the pattern bufsize = se->bufsize; buf_alloc(bufsize, internal) buf->mem_size = bufsize; I.e. bufsize has to be retrieved first, as se->bufsize might change at anytime - the value used for allocation and must not differ from the value actually used. This also tries to set large sizes in passthrough_hp, to catch issues in xfstests - though requires to set /proc/sys/fs/fuse/max_pages_limit Signed-off-by: Bernd Schubert <bschubert@ddn.com>
Bumps [github/codeql-action](https://github.com/github/codeql-action) from 3.28.13 to 3.28.15. - [Release notes](https://github.com/github/codeql-action/releases) - [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md) - [Commits](github/codeql-action@1b549b9...45775bd) --- updated-dependencies: - dependency-name: github/codeql-action dependency-version: 3.28.15 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>
Newer gccs now use -Werror=redundant-decls which means that anyone
including fuse.h is getting an error of:
/usr/include/fuse3/fuse.h:959:5: error: redundant redeclaration of ‘fuse_main_real_versioned’ [-Werror=redundant-decls]
959 | int fuse_main_real_versioned(int argc, char *argv[],
| ^~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/fuse3/fuse.h:885:5: note: previous declaration of ‘fuse_main_real_versioned’ with type ‘int(int, char **, const struct fuse_operations *, size_t, struct libfuse_version *, void *)’ {aka ‘int(int, char **, const struct fuse_operations *, long unsigned int, struct libfuse_version *, void *)’}
885 | int fuse_main_real_versioned(int argc, char *argv[],
| ^~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
For kernel development it might be necessary to load a module with renamed symbols and renamed /dev/<devicenode>. Reason is that for example ubuntu kernels have fuse compiled in and it is not possible to replace it at run time. And fuse might also be used for other file systems - a different device node is then needed. Also consolidate device path handling and remove unnecessary string duplication in mount_fuse() in fusermount.c. Signed-off-by: Meng Lu Wang <mwang@ddn.com>
The high-level interface triggers flag conversion twice: once in the high-level init and once in the low-level init. This caused false "both 'want' and 'want_ext' are set" errors when using fuse_set_feature_flag() or fuse_unset_feature_flag(). The existing check for duplicate conversion only worked when 32-bit flags were set directly. When using the preferred flag manipulation functions, conn->want and the lower 32 bits of conn->want_ext would differ, triggering the error. Fix this by synchronizing conn->want with the lower 32 bits of conn->want_ext after conversion, ensuring consistent state for subsequent calls. Closes: libfuse#1171 Signed-off-by: Bernd Schubert <bschubert@ddn.com>
Since pthread_setname_np() is the only pthread function that requires NPTL and it basically only set thread name, let's check if pthread_setname_np() does exist, otherwise let's not call pthread_setname_np() to shrink dependencies. Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
Signed-off-by: swj <1186093704@qq.com>
Commit dae1184 ("Add syslog and fatal signal handler feature") added fuse_set_fail_signal_handlers() which can store "se". But as fuse_set_signal_handlers() also stores the object storing it was made conditionally if not set already. As per libfuse#1182 this breaks some applications like osspd, that have multiple sessions and rely on the right order and that the last call of fuse_set_signal_handlers() wins. Special thanks to Sébastien Noel to debug this issue. Closes: libfuse#1182 Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Do not make sentences more complicated than they need to be. Signed-off-by: Jan Engelhardt <jengelh@inai.de> (cherry picked from commit fa03307)
Signed-off-by: wilkis <marvin@alruno.eu> (cherry picked from commit 1cb5878)
The previous one 404ed. Signed-off-by: Benjamin Peterson <benjamin@python.org> (cherry picked from commit ff7aa45)
Reported-by: Yaroslav Ibrahimov <51054594+yaroslavibr@users.noreply.github.com> Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 173cf15)
Signed-off-by: Ashish Ahuja <shivashish.ahuja@gmail.com> (cherry picked from commit ba68751)
Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit b2852e0)
The sourceforge address in the README is the only contact the project gives, and the release announcement template mails it too. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 22f3036)
A case that exercises the fusermount3 mountpoint checks needs the kernel module but never completes a mount, so it must not be gated on a setuid fusermount3. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 0622998)
Nothing exercised check_perm()'s refusals, so the symlink rule and the write-access rule could regress silently, and nothing pinned that only the final path component is protected -- a mountpoint under a symlinked parent has to keep working. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 7e69009)
Don't know how it can happen, but apparently it can. Check NULL cmsg to prevent crash. Reporter is content with that, since it's a stray fuse startup anyway. If it happens "for real", then we'll have better diagnostic. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> (cherry picked from commit 9d78424)
Replies are not submitted as they are produced: while a queue thread is in fuse_uring_queue_handle_cqes() (cqe_processing set), each reply only prepares a COMMIT_AND_FETCH SQE and the per-reply io_uring_submit() is skipped. The whole batch, plus the eventfd poll re-arm, is flushed by the next io_uring_submit_and_wait(). Without IORING_SETUP_SUBMIT_ALL the kernel aborts io_submit_sqes() at the first SQE whose prep fails, leaving the rest of the batch unconsumed until the following loop iteration -- head-of-line blocking the other in-flight replies behind one bad commit. Set the flag so the kernel drains the whole batch and only the failing SQE gets an error CQE. This might have resulted in fuse-over-io-uring stalls, if applications were waiting for results and libfuse was waiting for CQEs, but not all SQEs had been submitted yet. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit a19bb4f)
Over io_uring a ring entry only goes back to the kernel via COMMIT_AND_FETCH. fuse_reply_none() merely freed the request, so a FORGET answered this way leaked the entry and left the kernel-side request outstanding. Signed-off-by: Robert Doebbelin <robert@quobyte.com> (cherry picked from commit 758c7aa)
A daemon that dies with io-uring registered leaves the mountpoint returning ECONNABORTED, not ENOTCONN: the probe open() is queued on a ring the kernel has not finished tearing down yet, and the teardown ends the request rather than refusing it. Accepting only ENOTCONN skips the auto-unmount and leaves the mount behind. The cherry pick to 3.18 differs in respect to should_auto_unmount(), because commit a486808 removed the privileged open() from should_auto_unmount(); only check_ENOTCONN_as_owner() needs ECONNABORTED. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 60d2180)
This is safer than using umount2() directly, since it does the all the "fusermount -u" checks. Signed-off-by: Miklos Szeredi <mszeredi@redhat.com> (cherry picked from commit ff27278)
struct fuse_file_info reserves padding bits so new flags can be added without an ABI break, but abidiff reports the resulting member insertion and padding offset shift as a change and fails the check. Bound the suppression to size-preserving changes so a real size change still fails. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit a2c34a7)
With FUSE_CAP_HANDLE_KILLPRIV_V2 the kernel stops clearing suid/sgid itself and leaves that to the filesystem. Clearing them depends on CAP_FSETID, held by the application that issued the syscall - the filesystem's own privileges do not matter, and it cannot see the application's, so the kernel signals the decision per request. Of the three flags carrying it only FATTR_KILL_SUIDGID made it through, because FUSE_SET_ATTR_KILL_SUID happens to be the same bit. FUSE_OPEN_KILL_SUIDGID and FUSE_WRITE_KILL_SUIDGID were dropped, so a server that negotiated KILLPRIV_V2 never saw the request on O_TRUNC open and on write, and suid/sgid survived. Decode both into a new fuse_file_info::kill_suidgid, taken from the padding bits, in open, create, write and write_buf. Gate it on the negotiated capability and not on proto_minor: a filesystem that did not ask for KILLPRIV_V2 must never see fi->kill_suidgid set, and negotiating KILLPRIV_V2 implies 7.31, so fuse_create_in::open_flags is there to read. Added by Bernd: - fix the comment style for checkpatch - drop the XXX markers for unhandled open_flags, this patch handles them - replace the proto_minor guard with the capability check - drop security.capability from the field's documentation The cherry pick to 3.18 differs in respect to ChangeLog.rst, because prepare-3.18.3-rc1 records fuse_file_info::kill_suidgid in the 3.18.3 release notes. Signed-off-by: Robert Doebbelin <robert@quobyte.com> Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit 9895469)
Bump the version and record the backported fixes. Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
fuse_session_custom_io() moves the FUSE protocol onto a descriptor whose peer is not the kernel, and the only writeup of that is an example. Say what the callbacks owe the session, what a session loses without a mount, and which request fields reach the handlers with no bounds check, since a peer that skips those checks crashes the filesystem process. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit f1df035)
fuse_session_custom_io() puts a peer of the caller's choosing where the kernel would be, and libfuse parses what arrives trusting that a kernel has already validated it -- unvalidated lengths and counts reach the handlers otherwise. No in-tree user asks for that transport but the hello_ll_uds example, so it is built only where the option is. The symbols stay exported and refuse at runtime instead of being compiled out: dropping fuse_session_custom_io@FUSE_3.17 from the default build would break every binary already linked against it. The cherry pick to 3.18 differs in respect to the summary() hunk in meson.build, which is dropped because 3.18 has no configuration summary block. The ChangeLog entry goes into the Security Fixes section of 3.18.3-rc1 rather than under Unreleased Changes. Signed-off-by: Bernd Schubert <bernd@bsbernd.com> (cherry picked from commit cbd6679)
Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Replace the FSEL_FILES macro and explicit char array with a string literal for hex_map, deriving the file count from sizeof. Use size_t for loop variables and fix the printf format specifier accordingly. Signed-off-by: wangrzneu <wangrzneu@gmail.com> (cherry picked from commit 22d0fcd)
Compilation fails on Ubutu 26.06, we need another backport.
Signed-off-by: Bernd Schubert <bernd@bsbernd.com>
Take upstream libfuse 3.18.3 with the fuse3 -> fused rename. The fused-only packaging files are kept from master. Signed-off-by: Mohamad Chaarawi <mohamad.chaarawi@hpe.com>
Signed-off-by: Mohamad Chaarawi <mohamad.chaarawi@hpe.com>
posix_spawn() creates its child with clone(CLONE_VM|CLONE_VFORK). Under valgrind (as used by DAOS NLT) the child never reaches execve(), so fusermount3 is not run and the mount fails. Go back to fork+exec, as fused v1.0.0/v1.1.0 did, keeping the rest of upstream 3.18.3 mount.c unchanged. Signed-off-by: Mohamad Chaarawi <mohamad.chaarawi@hpe.com>
mchaarawi
changed the title
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
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.
No description provided.