containerd/mount/mountinfo_linux.go
Kir Kolyshkin 8eec9259e6 mount/mountinfo_linux: parser speed up
The mountinfo parser implemented via `fmt.Sscanf()` is slower than the one
using `strings.Split()` and `strconv.Atoi()`. This rewrite helps to speed it
up to a factor of 8x, here is a result from `go bench`:

> BenchmarkParsingScanf-4            300          22294112 ns/op
> BenchmarkParsingSplit-4           3000           2780703 ns/op

I tried other approaches, such as using `fmt.Sscanf()` for the first
three (integer) fields and `strings.Split()` for the rest, but it slows
things down considerably:

> BenchmarkParsingMixed-4           1000           8827058 ns/op

Note the old code uses `fmt.Sscanf` first, then a linear search for the
'-' field, then a split for the last 3 fields. The new code relies
on a single split.

One other thing is, the new code is more future proof as it skips
extra optional fields before the separator (currently there are none).

I have also added more comments to aid in future development.

Finally, the test data is fixed to not have white space before
the first field.

Based on a similar change in Moby,
 https://github.com/moby/moby/pull/36091

[v2: remove no-op break statement to silence staticcheck]
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
2018-05-07 11:38:37 -07:00

136 lines
4.2 KiB
Go

// +build linux
/*
Copyright The containerd Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package mount
import (
"bufio"
"fmt"
"io"
"os"
"strconv"
"strings"
)
// Self retrieves a list of mounts for the current running process.
func Self() ([]Info, error) {
f, err := os.Open("/proc/self/mountinfo")
if err != nil {
return nil, err
}
defer f.Close()
return parseInfoFile(f)
}
func parseInfoFile(r io.Reader) ([]Info, error) {
s := bufio.NewScanner(r)
out := []Info{}
for s.Scan() {
if err := s.Err(); err != nil {
return nil, err
}
/*
36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
(1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
(1) mount ID: unique identifier of the mount (may be reused after umount)
(2) parent ID: ID of parent (or of self for the top of the mount tree)
(3) major:minor: value of st_dev for files on filesystem
(4) root: root of the mount within the filesystem
(5) mount point: mount point relative to the process's root
(6) mount options: per mount options
(7) optional fields: zero or more fields of the form "tag[:value]"
(8) separator: marks the end of the optional fields
(9) filesystem type: name of filesystem of the form "type[.subtype]"
(10) mount source: filesystem specific information or "none"
(11) super options: per super block options
*/
text := s.Text()
fields := strings.Split(text, " ")
numFields := len(fields)
if numFields < 10 {
// should be at least 10 fields
return nil, fmt.Errorf("Parsing '%s' failed: not enough fields (%d)", text, numFields)
}
p := Info{}
// ignore any numbers parsing errors, as there should not be any
p.ID, _ = strconv.Atoi(fields[0])
p.Parent, _ = strconv.Atoi(fields[1])
mm := strings.Split(fields[2], ":")
if len(mm) != 2 {
return nil, fmt.Errorf("Parsing '%s' failed: unexpected minor:major pair %s", text, mm)
}
p.Major, _ = strconv.Atoi(mm[0])
p.Minor, _ = strconv.Atoi(mm[1])
p.Root = fields[3]
p.Mountpoint = fields[4]
p.Options = fields[5]
// one or more optional fields, when a separator (-)
i := 6
for ; i < numFields && fields[i] != "-"; i++ {
switch i {
case 6:
p.Optional = fields[6]
default:
/* NOTE there might be more optional fields before the separator
such as fields[7]...fields[N] (where N < separatorIndex),
although as of Linux kernel 4.15 the only known ones are
mount propagation flags in fields[6]. The correct
behavior is to ignore any unknown optional fields.
*/
}
}
if i == numFields {
return nil, fmt.Errorf("Parsing '%s' failed: missing separator ('-')", text)
}
// There should be 3 fields after the separator...
if i+4 > numFields {
return nil, fmt.Errorf("Parsing '%s' failed: not enough fields after a separator", text)
}
// ... but in Linux <= 3.9 mounting a cifs with spaces in a share name
// (like "//serv/My Documents") _may_ end up having a space in the last field
// of mountinfo (like "unc=//serv/My Documents"). Since kernel 3.10-rc1, cifs
// option unc= is ignored, so a space should not appear. In here we ignore
// those "extra" fields caused by extra spaces.
p.FSType = fields[i+1]
p.Source = fields[i+2]
p.VFSOptions = fields[i+3]
out = append(out, p)
}
return out, nil
}
// PID collects the mounts for a specific process ID. If the process
// ID is unknown, it is better to use `Self` which will inspect
// "/proc/self/mountinfo" instead.
func PID(pid int) ([]Info, error) {
f, err := os.Open(fmt.Sprintf("/proc/%d/mountinfo", pid))
if err != nil {
return nil, err
}
defer f.Close()
return parseInfoFile(f)
}