diff options
| author | Constantine Shulyupin | 2018-09-06 14:37:06 +0200 |
|---|---|---|
| committer | Miklos Szeredi | 2018-10-01 10:07:06 +0200 |
| commit | 5da784cce4308ae10a79e3c8c41b13fb9568e4e0 (patch) | |
| tree | f55416dec7532814d30c1dbd415ef11baec9fc7a /include/uapi/linux/fuse.h | |
| parent | fuse: allocate page array more efficiently (diff) | |
| download | kernel-qcow2-linux-5da784cce4308ae10a79e3c8c41b13fb9568e4e0.tar.gz kernel-qcow2-linux-5da784cce4308ae10a79e3c8c41b13fb9568e4e0.tar.xz kernel-qcow2-linux-5da784cce4308ae10a79e3c8c41b13fb9568e4e0.zip | |
fuse: add max_pages to init_out
Replace FUSE_MAX_PAGES_PER_REQ with the configurable parameter max_pages to
improve performance.
Old RFC with detailed description of the problem and many fixes by Mitsuo
Hayasaka (mitsuo.hayasaka.hu@hitachi.com):
- https://lkml.org/lkml/2012/7/5/136
We've encountered performance degradation and fixed it on a big and complex
virtual environment.
Environment to reproduce degradation and improvement:
1. Add lag to user mode FUSE
Add nanosleep(&(struct timespec){ 0, 1000 }, NULL); to xmp_write_buf in
passthrough_fh.c
2. patch UM fuse with configurable max_pages parameter. The patch will be
provided latter.
3. run test script and perform test on tmpfs
fuse_test()
{
cd /tmp
mkdir -p fusemnt
passthrough_fh -o max_pages=$1 /tmp/fusemnt
grep fuse /proc/self/mounts
dd conv=fdatasync oflag=dsync if=/dev/zero of=fusemnt/tmp/tmp \
count=1K bs=1M 2>&1 | grep -v records
rm fusemnt/tmp/tmp
killall passthrough_fh
}
Test results:
passthrough_fh /tmp/fusemnt fuse.passthrough_fh \
rw,nosuid,nodev,relatime,user_id=0,group_id=0 0 0
1073741824 bytes (1.1 GB) copied, 1.73867 s, 618 MB/s
passthrough_fh /tmp/fusemnt fuse.passthrough_fh \
rw,nosuid,nodev,relatime,user_id=0,group_id=0,max_pages=256 0 0
1073741824 bytes (1.1 GB) copied, 1.15643 s, 928 MB/s
Obviously with bigger lag the difference between 'before' and 'after'
will be more significant.
Mitsuo Hayasaka, in 2012 (https://lkml.org/lkml/2012/7/5/136),
observed improvement from 400-550 to 520-740.
Signed-off-by: Constantine Shulyupin <const@MakeLinux.com>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Diffstat (limited to 'include/uapi/linux/fuse.h')
| -rw-r--r-- | include/uapi/linux/fuse.h | 7 |
1 files changed, 6 insertions, 1 deletions
diff --git a/include/uapi/linux/fuse.h b/include/uapi/linux/fuse.h index 31a504f1ee60..76f46f159992 100644 --- a/include/uapi/linux/fuse.h +++ b/include/uapi/linux/fuse.h @@ -120,6 +120,7 @@ * 7.28 * - add FUSE_COPY_FILE_RANGE * - add FOPEN_CACHE_DIR + * - add FUSE_MAX_PAGES, add max_pages to init_out */ #ifndef _LINUX_FUSE_H @@ -255,6 +256,7 @@ struct fuse_file_lock { * FUSE_HANDLE_KILLPRIV: fs handles killing suid/sgid/cap on write/chown/trunc * FUSE_POSIX_ACL: filesystem supports posix acls * FUSE_ABORT_ERROR: reading the device after abort returns ECONNABORTED + * FUSE_MAX_PAGES: init_out.max_pages contains the max number of req pages */ #define FUSE_ASYNC_READ (1 << 0) #define FUSE_POSIX_LOCKS (1 << 1) @@ -278,6 +280,7 @@ struct fuse_file_lock { #define FUSE_HANDLE_KILLPRIV (1 << 19) #define FUSE_POSIX_ACL (1 << 20) #define FUSE_ABORT_ERROR (1 << 21) +#define FUSE_MAX_PAGES (1 << 22) /** * CUSE INIT request/reply flags @@ -617,7 +620,9 @@ struct fuse_init_out { uint16_t congestion_threshold; uint32_t max_write; uint32_t time_gran; - uint32_t unused[9]; + uint16_t max_pages; + uint16_t padding; + uint32_t unused[8]; }; #define CUSE_INIT_INFO_MAX 4096 |