mirror of
https://github.com/libfuse/libfuse.git
synced 2026-07-07 14:47:41 +08:00
Benjamin Peterson
57954f4a5f
As discussed on https://github.com/libfuse/libfuse/issues/1515, libfuse has a very light dependency on libnuma, only using very light wrappers over mmap, mbind, and munmap. Remove the libnuma dependency by calling those syscalls directly. Populate memory after mmap to avoid SIGBUS later. Signed-off-by: Benjamin Peterson <benjamin@python.org>
38 lines
1.3 KiB
Plaintext
38 lines
1.3 KiB
Plaintext
fuse-over-io-uring uses io-uring for transport of kernel/userspace
|
|
messages. See also https://docs.kernel.org/filesystems/fuse/fuse-io-uring.html
|
|
|
|
In order to enable it, the kernel module needs to have it enabled:
|
|
echo 1 > /sys/module/fuse/parameters/enable_uring
|
|
|
|
Additionally, FUSE_CAP_OVER_IO_URING needs to be set and
|
|
se->uring.enable has to be true. The latter can be
|
|
achieved with the libfuse option '-o io_uring'.
|
|
|
|
Default queue-depth is 8 and can be changed with the parameter
|
|
'-oio_uring_q_depth'.
|
|
|
|
As of now there is always one queue per core. A reduced number
|
|
of cores in development.
|
|
|
|
Benefits:
|
|
- Improved performance by using io_uring for kernel/userspace communication
|
|
- Reduced system call overhead compared to traditional FUSE
|
|
- Asynchronous I/O operations
|
|
|
|
Usage:
|
|
To enable io_uring support when mounting a FUSE filesystem:
|
|
1. Enable kernel support: echo 1 > /sys/module/fuse/parameters/enable_uring
|
|
2. Mount with io_uring option: -o io_uring
|
|
3. Optionally adjust queue depth: -o io_uring_q_depth=<depth>
|
|
|
|
Example:
|
|
./my_fuse_fs /source /mountpoint -o io_uring -o io_uring_q_depth=16
|
|
|
|
Requirements:
|
|
- Linux kernel with io_uring and FUSE io_uring support enabled
|
|
- libfuse compiled with io_uring support
|
|
|
|
Build Dependencies:
|
|
- liburing (for io_uring support)
|
|
- meson build system with option: -Denable-io-uring=true
|