1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
|
#!/usr/bin/perl
# **** License ****
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# General Public License for more details.
#
# This code was originally developed by Vyatta, Inc.
# Portions created by Vyatta are Copyright (C) 2009,2010 Vyatta, Inc.
# All Rights Reserved.
#
# **** End License ****
#
use warnings;
use strict;
use Sys::Syslog qw(:standard :macros);
die "Usage: $0 ifname {auto | mask} { debug }\n" if ($#ARGV < 1);
my ($ifname, $mask, $debug) = @ARGV;
die "Error: Interface $ifname does not exist\n"
unless -d "/sys/class/net/$ifname";
openlog("irq-affinity","",LOG_LOCAL0);
my ( $cpus, $cores, $processors ) = cpuinfo();
printf("cpus = $cpus, cores = $cores, processors = $processors.\n")
if (defined($debug));
if ($mask eq 'auto') {
affinity_auto($ifname);
} else {
affinity_mask($ifname, $mask);
}
exit 0;
sub do_log {
syslog(LOG_INFO, @_);
if (defined($debug)) {
printf(@_);
printf("\n");
}
}
# Get current irq assignments by reading /proc/interrupts
# returns reference to hash of interrupt infromation for given interface
# i.e. {'eth1'} => 22, {'eth1-tx-1'} => 31, ...
sub irqinfo {
my $ifname = shift;
my $irqmap;
open( my $f, '<', "/proc/interrupts" )
or die "Can't read /proc/interrupts";
while (<$f>) {
chomp;
my @cols = split;
# First column is IRQ number (and colon)
next unless /^\s*(\d+):\s/;
my $irq = $1;
# Skip columns for IRQ's per CPU
foreach my $name ( @cols[ $cpus+1 .. $#cols ] ) {
$name =~ s/,$//;
next unless ($name eq $ifname || $name =~ /^$ifname-/ );
$irqmap->{$name} = $irq;
}
}
close $f;
return $irqmap;
}
# Determine number of cpus and cores
sub cpuinfo {
my ( $cpu, $core, $packages );
open( my $f, '<', "/proc/cpuinfo" )
or die "Can't read /proc/cpuinfo";
while (<$f>) {
chomp;
if (/^cpu cores\s+:\s(\d+)$/) {
$core = $1;
}
elsif (/^processor\s+:\s+(\d+)$/) {
$cpu = $1;
}
elsif (/^physical id\s+:\s+(\d+)$/) {
$packages = $1;
}
}
close $f;
$packages++;
$core *= $packages;
return ( $cpu + 1, $core, $packages );
}
# Determine hyperthreading factor
# most CPU's have either 1 or 2 threads per core
sub threads_per_core {
return 1 unless defined($cores);
return $cpus / $cores;
}
# Set affinity value for a irq
sub set_affinity {
my ( $ifname, $irq, $mask ) = @_;
my $smp_affinity = "/proc/irq/$irq/smp_affinity";
do_log("%s: irq %d affinity set to 0x%x", $ifname, $irq, $mask);
open( my $f, '>', $smp_affinity )
or die "Can't open: $smp_affinity : $!\n";
printf {$f} "%x\n", $mask;
close $f;
}
# set Receive Packet Steering mask
sub set_rps {
my ( $ifname, $q, $mask ) = @_;
# ignore if older kernel without RPS
my $rxq = "/sys/class/net/$ifname/queues";
return unless ( -d $rxq );
do_log("%s: receive queue %d cpus set to 0x%x",
$ifname, $q, $mask);
my $rps_cpus = "$rxq/rx-$q/rps_cpus";
open( my $f, '>', $rps_cpus )
or die "Can't open: $rps_cpus : $!\n";
printf {$f} "%x\n", $mask;
close $f;
}
# Check if the current if this cpu is in the banned mask
# Uses environment variable VYATTA_IRQAFFINITY_BANNED_CPUS
# to mask cpus which irq affinity script should ignore
sub skip_cpu {
my $cpu = shift;
my $banned = $ENV{'VYATTA_IRQAFFINITY_BANNED_CPUS'};
return unless defined($banned); # false
return ((1 << $cpu) & hex($banned)) != 0;
}
# For multi-queue NIC choose next cpu to be on next core
sub next_cpu {
my $origcpu = shift;
my $threads = threads_per_core();
my $cpu = $origcpu;
do {
$cpu += $threads;
if ( $cpu >= $cpus ) {
# wraparound to next thread on core 0
$cpu = ($cpu + 1) % $threads;
}
} while ($cpu != $origcpu && skip_cpu($cpu));
return $cpu;
}
# Get cpu to assign for the queues for single queue nic
sub choose_cpu {
my $ifname = shift;
# For single-queue nic choose IRQ based on name
# Ideally should make decision on least loaded CPU
my ($ifunit) = ($ifname =~ m/^[a-z]*(\d+)$/);
die "can't find number for $ifname\n"
unless defined($ifunit);
my $threads = threads_per_core();
# Give the load first to one CPU of each hyperthreaded core, then
# if there are enough NICs, give the load to the other CPU of
# each core.
my $ht_wrap = (($ifunit * $threads) / $cpus) % $threads;
my $cpu = ((($ifunit * $threads) + $ht_wrap) % $cpus);
$cpu = next_cpu($cpu) if skip_cpu($cpu);
return $cpu;
}
# Assignment for multi-queue NICs
sub assign_multiqueue {
my $ifname = shift;
my $irqmap = shift;
my $numq = scalar(@_);
my $cpu;
if ($numq == 1) {
# This is a single-queue NIC using the multi-queue naming
# format. In this case, we use the same algorithm to select
# the CPU as we use for standard single-queue NICs. This
# algorithm spreads the work of different NICs accross
# different CPUs.
$cpu = choose_cpu($ifname);
} else {
# For multi-queue nic's always starts with CPU 0
# This is less than ideal when there are more core's available
# than number of queues (probably should barber pole);
# but the Intel IXGBE needs CPU 0 <-> queue 0
# because of flow director bug.
$cpu = 0;
}
foreach my $name (sort @_) {
my $irq = $irqmap->{$name};
die "Can't find irq in map for $name\n" unless $irq;
do_log("%s: assign %s to cpu %d",
$ifname, $name, $cpu );
# Assign CPU affinity for both IRQs
set_affinity( $ifname, $irq, 1 << $cpu );
# TODO use RPS to steer data if cores > queues?
$cpu = next_cpu($cpu);
}
}
# Affinity assignment function for single-queue NICs. The strategy
# here is to just spread the interrupts of different NICs evenly
# across all CPUs. That is the best we can do without monitoring the
# load and traffic patterns. So we just directly map the NIC unit
# number into a CPU number.
sub assign_single {
my ( $ifname, $irq ) = @_;
my $cpu = choose_cpu($ifname);
do_log("%s: assign irq %d to cpu %d", $ifname, $irq, $cpu );
set_affinity( $ifname, $irq, 1 << $cpu );
my $threads = threads_per_core();
if ($threads > 1) {
# Use both threads on this cpu if hyperthreading
my $mask = ((1 << $threads) - 1) << $cpu;
set_rps($ifname, 0, $mask);
}
# MAYBE - Use all cpu's if no HT
}
# Mask must contain at least one CPU and
# no bits outside of range of available CPU's
sub check_mask {
my ($ifname, $name, $mask) = @_;
my $m = hex($mask);
die "$ifname: $name mask $mask has no bits set\n"
if ($m == 0);
die "$ifname: $name mask $mask too large for number of CPU's: $cpus\n"
if ($m >= 1 << $cpus);
}
# Set affinity (and RPS) based on mask
sub affinity_mask {
my ($ifname, $mask) = @_;
# match on <hex> or <hex>,<hex>
unless ($mask =~ /^([0-9a-f]+)(|,([0-9a-f]+))$/) {
die "$ifname: irq mask $mask is not a valid affinity mask\n"
}
my $irqmsk = $1;
my $rpsmsk = $3;
check_mask($ifname, "irq", $irqmsk);
check_mask($ifname, "rps", $rpsmsk) if $rpsmsk;
my $irqmap = irqinfo($ifname);
while (my ($name, $irq) = each (%{$irqmap})) {
do_log("%s: assign irq %d mask %s", $name, $irq, $irqmsk);
set_affinity($name, $irq, hex($irqmsk));
}
set_rps($ifname, 0, hex($rpsmsk)) if $rpsmsk;
}
# The auto strategy involves trying to achieve the following goals:
#
# - Spread the receive load among as many CPUs as possible.
#
# - For all multi-queue NICs in the system that provide both tx and
# rx queues, keep all of the queues that share the same queue
# number on same CPUs. I.e. tx and rx queue 0 of all such NICs
# should interrupt one CPU; tx and rx queue 1 should interrupt a
# different CPU, etc.
#
# - If hyperthreading is supported and enabled, avoid assigning
# queues to both CPUs of a hyperthreaded pair if there are enough
# CPUs available to do that.
sub affinity_auto {
my $ifname = shift;
my $irqmap = irqinfo($ifname);
my @irqnames = keys %{$irqmap};
my $numirq = scalar(@irqnames);
# Figure out what style of irq naming is being used
if ( $numirq == 1 ) {
my $irq = $irqmap->{$ifname};
assign_single( $ifname, $irq) if $irq;
} elsif ($numirq > 1) {
# Special case for paired Rx and Tx
my @mirq = grep { /^$ifname-rx-/ } @irqnames;
my $num_mirq = scalar(@mirq);
if ( $num_mirq > 0 ) {
assign_multiqueue( $ifname, $irqmap, @mirq );
@mirq = grep { /^$ifname-tx-/ } @irqnames;
assign_multiqueue( $ifname, $irqmap, @mirq );
return;
}
# Normal case for single irq per queue
@mirq = grep { /^$ifname-/ } @irqnames;
$num_mirq = scalar(@mirq);
if ( $num_mirq > 0 ) {
assign_multiqueue( $ifname, $irqmap, @mirq );
return;
}
# Netxen thought up yet another convention
@mirq = grep { /^$ifname\[/ } @irqnames;
if ( $#mirq > 0 ) {
assign_multiqueue( $ifname, $irqmap, @mirq );
return;
}
do_log("%s: Unknown multiqueue irq naming: %s\n", $ifname,
join(' ', @irqnames));
}
}
|