This is similar to the golang.org/x/tools/internal/fastwalk I'd
previously written but not recursive and using mem.RO.
The metrics package already had some Linux-specific directory reading
code in it. Move that out to a new general package that can be reused
by portlist too, which helps its scanning of all /proc files:
name old time/op new time/op delta
FindProcessNames-8 2.79ms ± 6% 2.45ms ± 7% -12.11% (p=0.000 n=10+10)
name old alloc/op new alloc/op delta
FindProcessNames-8 62.9kB ± 0% 33.5kB ± 0% -46.76% (p=0.000 n=9+10)
name old allocs/op new allocs/op delta
FindProcessNames-8 2.25k ± 0% 0.38k ± 0% -82.98% (p=0.000 n=9+10)
Change-Id: I75db393032c328f12d95c39f71c9742c375f207a
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Add an osImpl interface that can be stateful and thus more efficient
between calls. It will later be implemented by all OSes but for now
this change only adds a Linux implementation.
Remove Port.inode. It was only used by Linux and moves into its osImpl.
Don't reopen /proc/net/* files on each run. Turns out you can just
keep then open and seek to the beginning and reread and the contents
are fresh.
name old time/op new time/op delta
GetListIncremental-8 7.29ms ± 2% 6.53ms ± 1% -10.50% (p=0.000 n=9+9)
name old alloc/op new alloc/op delta
GetListIncremental-8 1.30kB ±13% 0.70kB ± 5% -46.38% (p=0.000 n=9+10)
name old allocs/op new allocs/op delta
GetListIncremental-8 33.2 ±11% 18.0 ± 0% -45.82% (p=0.000 n=9+10)
Updates #5958
Change-Id: I4be83463cbd23c2e2fa5d0bdf38560004f53401b
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Make Linux parsePorts also an append-style API and attach it to
caller's provided append base memory.
And add a little string intern pool in front of the []byte to string
for inode names.
name old time/op new time/op delta
GetList-8 11.1ms ± 4% 9.8ms ± 6% -11.68% (p=0.000 n=9+10)
name old alloc/op new alloc/op delta
GetList-8 92.8kB ± 2% 79.7kB ± 0% -14.11% (p=0.000 n=10+9)
name old allocs/op new allocs/op delta
GetList-8 2.94k ± 1% 2.76k ± 0% -6.16% (p=0.000 n=10+10)
More coming. (the bulk of the allocations are in addProcesses and
filesystem operations, most of which we should usually be able to
skip)
Updates #5958
Change-Id: I3f0c03646d314a16fef7f8346aefa7d5c96701e7
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
In prep for reducing garbage, being able to reuse memory. So far this
doesn't actually reuse much. This is just changing signatures around.
But some improvement in any case:
bradfitz@tsdev:~/src/tailscale.com$ ~/go/bin/benchstat before after
name old time/op new time/op delta
GetList-8 11.8ms ± 9% 9.9ms ± 3% -15.98% (p=0.000 n=10+10)
name old alloc/op new alloc/op delta
GetList-8 99.5kB ± 2% 91.9kB ± 0% -7.62% (p=0.000 n=9+9)
name old allocs/op new allocs/op delta
GetList-8 3.05k ± 1% 2.93k ± 0% -3.83% (p=0.000 n=8+9)
More later, once parsers can reuse strings from previous parses.
Updates #5958
Change-Id: I76cd5048246dd24d11c4e263d8bb8041747fb2b0
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Avoid splitting fields in the common case. Field splitting was 84% of
the overall CPU.
name old time/op new time/op delta
ParsePorts-6 33.3ms ± 2% 6.3ms ± 4% -80.97% (p=0.000 n=9+10)
name old alloc/op new alloc/op delta
ParsePorts-6 520B ±79% 408B ± 0% -21.49% (p=0.046 n=10+8)
name old allocs/op new allocs/op delta
ParsePorts-6 7.00 ± 0% 7.00 ± 0% ~ (all equal)
Updates tailscale/corp#2566
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Notably, it no longer allocates proportional to the number of open
sockets on the machine. Any alloc reduction numbers are a little
contrived with such a reduction but e.g. on a machine with 50,000
connections open:
name old time/op new time/op delta
ParsePorts-6 57.7ms ± 6% 32.8ms ± 3% -43.04% (p=0.000 n=9+10)
name old alloc/op new alloc/op delta
ParsePorts-6 24.0MB ± 0% 0.0MB ± 0% -100.00% (p=0.000 n=10+9)
name old allocs/op new allocs/op delta
ParsePorts-6 100k ± 0% 0k ± 0% -99.99% (p=0.000 n=10+10)
Updates tailscale/corp#2566
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
This is important because some of those v6 sockets are actually
dual-stacked sockets, so this is our only chance of discovering
some services.
Fixes#1443.
Signed-off-by: David Anderson <danderson@tailscale.com>
Brad Fitzpatrick
David Anderson