Matthew [Sat, 24 Feb 2024 06:56:49 +0000 (22:56 -0800)]
Passthrough options starting with "x-" to mtab (#894)
This implements #651, tested with bindfs.
"x-*" options are comments meant to be interpreted by userspace.
#651 is about some 3rd party mount options like 'x-gvfs-notrash'.
This also removes the test if /etc/mtab is a symlink.
This test was added in commit 5f28cd15ab43c741f6d116be4d3a9aa5d82ab385
and the corresponding ChangeLog entry in this commit points to mount
issues for read-only mtab.
However, in all recent Linux distributions /etc/mtab is a symlink to
/proc/self/mounts and never writable. In fact, util-linux 2.39
(libmount) entirely removed support for a writable mtab.
At least since util-linux 2.19 (10-Feb-2011) /run/mount/utab is used
as replacement for userspace mount entries..
bigbrotherwei [Wed, 17 Jan 2024 15:58:27 +0000 (23:58 +0800)]
Add processing for FUSE_CAP_HANDLE_KILLPRIV and disable it by default
'FUSE_CAP_HANDLE_KILLPRIV' is not enabled by default anymore, as that
would be a sudden security issue introduced by a new ABI and API
compatible libfuse version.
Bernd Schubert [Mon, 4 Dec 2023 14:58:45 +0000 (15:58 +0100)]
Add FUSE_CAP_DIRECT_IO_ALLOW_MMAP and use in passthrough_hp
This is not called FUSE_CAP_DIRECT_IO_RELAX, as the kernel flag
FUSE_DIRECT_IO_RELAX is supposed to be renamed to
FUSE_DIRECT_IO_ALLOW_MMAP. The corresponding kernel patches just
did not land yet.
Bernd Schubert [Tue, 26 Dec 2023 10:53:55 +0000 (11:53 +0100)]
examples/notify_store_retrieve: Add a clean shutdown
On shutdown update_fs_loop() was failing frequently in github
test with ENOENT, but only ENODEV was tested for and only for the.
With clean shutdown there is no need to test for such errors at all.
George Hilliard [Sun, 19 Nov 2023 10:27:45 +0000 (04:27 -0600)]
Clarify fuse_lowlevel poll() docstring (#862)
The original docstring was confusing; it was not clear that the ph must
be retained indefinitely, nor was it clear that the client *also* needs
to reply to the poll call immediately.
Clarify this by explaining that it is only necessary to retain a single
handle, that the client must retain ph, and that it must immediately
call reply.
fdinoff [Thu, 16 Nov 2023 11:23:20 +0000 (06:23 -0500)]
Pass FUSE_PARALLEL_DIROPS to kernel (#861)
This tells the kernel that parallel lookup/readdir operations are
supported. This is enabled by default but was not passed to the kernel
so you always get the synchronized version.
Bernd Schubert [Tue, 10 Oct 2023 11:18:36 +0000 (13:18 +0200)]
Enabled parallel direct IO writes for passthrough examples
All these passthrough examples don't need writes to be serialized.
Actually, most file systems probably handle non serialized parallel
direct writes - the FOPEN_PARALLEL_DIRECT_WRITES flag is just
to avoid a regression for those file system that rely on serialized
DIO writes in fuse kernel. Passthrough file system forward the IO
to another file system, which actually handles that internally -
serialized in fuser kernel is not needed.
xfstests example: Use export in local.config and remove comment (#811)
ix passthrough mount helper for running xfstests
* The mount helper does not see the env vars exported by xfstests
* Use the NFS style format TEST_DEV=source:/$TEST_SOURCE to communicate
the sourse path to the mount helper without confusing xfstest
* Also recognise when source= is provided in mount options
* Support -o remount (e.g. for test generic/306)
Signed-off-by: Amir Goldstein <amir73il@gmail.com> Co-authored-by: Bernd Schubert <bschubert@ddn.com> Co-authored-by: Amir Goldstein <amir73il@gmail.com> Co-authored-by: Nikolaus Rath <Nikolaus@rath.org>
Allow linking with mold / fix the version script (#814)
This fixes issue https://github.com/libfuse/libfuse/issues/810
and should avoid mold linking errors.
Commit d4e294b removed made fuse_register_module() a static
function, but forgot to remove it from the version script.
Commit fe4f942 introduced copy_file_range to libfuse and
added the non-exiting (neither declared nor defined) function
fuse_reply_copy_file_range() to the version script. Kernel
side just exects an integer reply how much was copied, using
fuse_reply_write() as in fuse_lib_copy_file_range() is sufficient
and no extra function is needed.
wdlkmpx [Thu, 8 Jun 2023 10:24:28 +0000 (05:24 -0500)]
meson.build: pass -D_FILE_OFFSET_BITS=64 to C/C++ compiler (#799)
libfuse requires a 64bit off_t, it's not optional
../include/fuse_common.h:938:1: error: Kstatic assertion failed: "fuse: off_t must be 64bit"
so this only takes effect if compiling for a 32bit glibc system, it's ignored everywhere else
meson by default adds -D_FILE_OFFSET_BITS=64 to C/C++ compilers, and removes duplicate declarations
There are several opened issues in meson's git repo requesting to remove that behavior, so this makes sense even more now
this fixes compilation with muon (a C99 (mostly) meson compatible build app) in a 32bit glibc system ...
This is just to [keep address sanitizer happy](https://github.com/libfuse/libfuse/actions/runs/4730520764/jobs/8394347666?pr=784). The OS would normally clean this up anyway.
Nikolaus Rath [Mon, 8 May 2023 23:12:08 +0000 (16:12 -0700)]
Do not pass unsupported mount options to the kernel.
The filesystem daemon is responsible for implementing eg. st_atime updates, so passing
options like relatime to the kernel results in them being silently ignored. Instead, such
options need to be interpreted (and filtered out) by the filesystem daemon.
Peri [Thu, 11 May 2023 01:38:46 +0000 (02:38 +0100)]
Fix issue #746. (#782)
Added a secondary check in fuse_lib_unlink() after hide_node()
to check again under a lock if the (now hidden) file is still open.
If not then delete it.
This should synchronise fuse_lib_unlink() with fuse_lib_release(),
when nullpath_ok is set.
Direct leak of 352 byte(s) in 1 object(s) allocated from:
#0 0x7fbb19abf411 in __interceptor_calloc /usr/src/debug/gcc/gcc/libsanitizer/asan/asan_malloc_linux.cpp:77
#1 0x7fbb1a0efd3b in fuse_fs_new ../lib/fuse.c:4814
#2 0x7fbb1a0f02b5 in fuse_new_31 ../lib/fuse.c:4913
#3 0x7fbb1a10ec5e in fuse_main_real ../lib/helper.c:345
#4 0x5625db8ab418 in main ../example/hello.c:176
#5 0x7fbb1983c78f (/usr/lib/libc.so.6+0x2378f)
SUMMARY: AddressSanitizer: 352 byte(s) leaked in 1 allocation(s).
```
That's because `fuse_lowlevel.c`s `fuse_session_destroy` would only call
the user supplied `op.destroy`, if INIT had been processed, but the high
level API relied on `op.destroy` to free `f->fs`.
This patch moves the freeing into `fuse_destroy` that will always be
called by our high-level API.
```
$ doxygen doc/Doxyfile
warning: Tag 'COLS_IN_ALPHA_INDEX' at line 707 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'PERL_PATH' at line 1048 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'CLASS_DIAGRAMS' at line 1061 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'MSCGEN_PATH' at line 1070 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Tag 'DOT_TRANSPARENT' at line 1207 of file 'doc/Doxyfile' has become obsolete.
To avoid this warning please remove this line from your configuration file or upgrade it using "doxygen -u"
warning: Changing CLASS_GRAPH option to TEXT because obsolete option CLASS_DIAGRAM was found and set to NO.
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: argument 'private_data' of command @param is not found in the argument list of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data)
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: The following parameter of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data) is not documented:
parameter 'user_data'
```
So let's try `doxygen -u doc/Doxyfile`. After:
```
$ doxygen doc/Doxyfile
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: argument 'private_data' of command @param is not found in the argument list of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data)
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: The following parameter of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data) is not documented:
parameter 'user_data'
```
The generated documentation `doc/html` does not differ.
For ease of review, here's a diff with all blank lines and comments removed via `grep -vE '^#|^$' doc/Doxyfile`
Fuse mount: make auto_unmount compatible with suid/dev mount options (#762)
* Fuse mount: make auto_unmount compatible with suid/dev mount options
> When you run as root, fuse normally does not call fusermount but uses
> the mount system call directly. When you specify auto_unmount, it goes
> through fusermount instead. However, fusermount is a setuid binary that
> is normally called by regular users, so it cannot in general accept suid
> or dev options.
In this patch, we split up how fuse mounts as root when `auto_unmount`
is specified.
First, we mount using system calls directly, then we reach out to
fusermount to set up auto_unmount only (with no actual mounting done in
fusermount).
This syncs fuse_kernel.h to <linux-6.3>/include/uapi/linux/fuse.h
Special handling is done for setxattr as in linux commit 52a4c95f4d24b struct fuse_setxattr_in was extended. Extended
struct is only used when FUSE_SETXATTR_EXT is passed in FUSE_INIT
reply.
```
=============================== warnings summary ===============================
test/test_ctests.py:12
/tmp/libfuse-build-mB50ZC/build-gcc-9/test/test_ctests.py:12: DeprecationWarning: The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives
from distutils.version import LooseVersion
```
$ doxygen doc/Doxyfile
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: argument 'private_data' of command @param is not found in the argument list of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data)
/Users/matthias/prog/fuser/libfuse/include/fuse.h:934: warning: The following parameter of fuse_new_31(struct fuse_args *args, const struct fuse_operations *op, size_t op_size, void *user_data) is not documented:
parameter 'user_data'
```
```
test_examples.py::test_printcap
/usr/local/lib/python3.9/dist-packages/_pytest/python.py:199: PytestReturnNotNoneWarning: Expected None, but test_examples.py::test_printcap returned ((7, 38), {'FUSE_CAP_READDIRPLUS_AUTO', 'FUSE_CAP_ASYNC_DIO', 'FUSE_CAP_SPLICE_READ', 'FUSE_CAP_CACHE_SYMLINKS', 'FUSE_CAP_IOCTL_DIR', 'FUSE_CAP_NO_OPENDIR_SUPPORT', 'FUSE_CAP_NO_OPEN_SUPPORT', 'FUSE_CAP_POSIX_LOCKS', 'FUSE_CAP_READDIRPLUS', 'FUSE_CAP_POSIX_ACL', 'FUSE_CAP_ATOMIC_O_TRUNC', 'FUSE_CAP_SPLICE_MOVE', 'FUSE_CAP_EXPORT_SUPPORT', 'FUSE_CAP_FLOCK_LOCKS', 'FUSE_CAP_EXPLICIT_INVAL_DATA', 'FUSE_CAP_EXPIRE_ONLY', 'FUSE_CAP_DONT_MASK', 'FUSE_CAP_WRITEBACK_CACHE', 'FUSE_CAP_AUTO_INVAL_DATA', 'FUSE_CAP_PARALLEL_DIROPS', 'FUSE_CAP_SPLICE_WRITE', 'FUSE_CAP_ASYNC_READ'}), which will be an error in a future version of pytest. Did you mean to use `assert` instead of `return`?
warnings.warn(
```
```
PytestDeprecationWarning: The hookimpl pytest_pyfunc_call uses old-style configuration options (marks or attributes).
Please use the pytest.hookimpl(hookwrapper=True) decorator instead
to configure the hooks.
See https://docs.pytest.org/en/latest/deprecations.html#configuring-hook-specs-impls-using-markers
@pytest.mark.hookwrapper
```
```
WARNING: Running the setup command as `meson [options]` instead of `meson setup [options]` is ambiguous and deprecated.WARNING: Running the setup command as `meson [options]` instead of `meson setup [options]` is ambiguous and deprecated.
```
In
https://github.com/libfuse/libfuse/blob/77d662459a0fcdf358d515477d33795837e859d5/util/fusermount.c#L1219
`open` is executed as root which does not have access to the mount
point if `allow_other` was not used and the real user id is not 0. Since
`allow_other` usually cannot be specified by unprivileged users,
`auto_unmount` has no effect for unprivileged users.
In this commit, we work around this limitation:
We first try to open the mountpoint as root, and if we get `EACCES`, we
retry as the user who started fusermount, and see if we get `ENOTCONN`.
In my testing, I found that `setfsuid` and `setfsgid` don't work to get
around the lack of `allow_other`. (Sorry, I don't know enough about the
Linux kernel to tell whether that's significant.) As a workaround, I
decided to use `setresuid` and `setresgid` in a forked child process,
and communicate via its exit status.
Please give feedback on correctness, style and suggest tests.
Nikolaus Rath [Sun, 26 Mar 2023 10:35:22 +0000 (11:35 +0100)]
Revert "upgrade of fuse_kernel.h based on Miklos expire_only kernel patch https://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse.git/commit/?h=for-next&id=53e949edb7692dce02220eba926c9d75ecbb47f7"
Sarath Lakshman [Sat, 11 Mar 2023 11:28:31 +0000 (16:58 +0530)]
Fix max_threads command line parameter propagation
The fuse_main_real() method doesn't apply the max_threads parameter
parsed through the commandline arguments. This commit fixes the wiring
of max_threads argument.
Right now fuse kernel serializes direct writes on the
same file. This serialization is good for such FUSE
implementations which rely on the inode lock to
avoid any data inconsistency issues but it hurts badly
such FUSE implementations which have their own mechanism
of dealing with cache/data integrity and can handle
parallel direct writes on the same file.
This patch allows parallel direct writes on the same file to be
enabled with the help of a flag FOPEN_PARALLEL_DIRECT_WRITES.
FUSE implementations which want to use this feature can
set this flag during fuse init. Default behaviour remains
same i.e no parallel direct writes on the same file.
Bernd Schubert [Sun, 19 Feb 2023 11:52:52 +0000 (12:52 +0100)]
Avoid max-idle threads warning
If a program with API before 312 did not set
max_idle_threads the new default from
fuse_parse_cmdline_312() is applied, which sets
UINT_MAX (-1).
Later in compat fuse_session_loop_mt_32 the old
config v1 struct is converted and that conversion
prints a warning if the default unset value was used.
This could have also happened to programs using the current
API, which just apply values struct fuse_cmdline_opts,
without checking if the defaults are set.
Pierre Labastie [Sun, 19 Feb 2023 08:52:13 +0000 (09:52 +0100)]
Change define for C++ too
Commit d7560cc has split defines into private and public, and passed -DHAVE_LIBFUSE-PRIVATE_CONFIG_H to all C programs. But the arguments of C++ programs have not been changed. This leads to a test failure as reported in issue #734. Pass -DHAVE_LIBFUSE-PRIVATE_CONFIG_H to C++ programs too.
Bernd Schubert [Tue, 7 Feb 2023 22:06:42 +0000 (23:06 +0100)]
Split config.h into private and public config
This addresses https://github.com/libfuse/libfuse/issues/729
commit db35a37def14b72181f3630efeea0e0433103c41 introduced a public
config.h (rename to fuse_config.h to avoid conflicts) that
was installed with the package and included by libfuse users
through fuse_common.h. Probablem is that this file does not have
unique defines so that they are unique to libfuse - on including
the file conflicts with libfuse users came up.
In principle all defines could be prefixed, but then most of them
are internal for libfuse compilation only. So this splits out
publically required defines to a new file 'libfuse_config.h'
and changes back to include of "fuse_config.h" only when
HAVE_LIBFUSE_PRIVATE_CONFIG_H is defined.
This also renames HAVE_LIBC_VERSIONED_SYMBOLS to
LIBFUSE_BUILT_WITH_VERSIONED_SYMBOLS, as it actually
better explains for libfuse users what that variable
is for.
projects / qemu-gpiodev / libfuse.git / log