tildefriends/deps/libuv/test/test-thread-affinity.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

155 lines
4.1 KiB
C
Raw Normal View History

/* Copyright libuv project contributors. All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include "uv.h"
#include "task.h"
#include <string.h>
#ifndef NO_CPU_AFFINITY
static void check_affinity(void* arg) {
int r;
char* cpumask;
int cpumasksize;
uv_thread_t tid;
cpumask = (char*)arg;
cpumasksize = uv_cpumask_size();
ASSERT_GT(cpumasksize, 0);
tid = uv_thread_self();
r = uv_thread_setaffinity(&tid, cpumask, NULL, cpumasksize);
ASSERT_OK(r);
r = uv_thread_setaffinity(&tid, cpumask + cpumasksize, cpumask, cpumasksize);
ASSERT_OK(r);
}
TEST_IMPL(thread_affinity) {
int t1first;
int t1second;
int t2first;
int t2second;
int cpumasksize;
char* cpumask;
int ncpus;
int r;
int c;
int i;
uv_thread_t threads[3];
#ifdef _WIN32
/* uv_thread_self isn't defined for the main thread on Windows */
threads[0] = GetCurrentThread();
#else
threads[0] = uv_thread_self();
#endif
cpumasksize = uv_cpumask_size();
ASSERT_GT(cpumasksize, 0);
cpumask = calloc(4 * cpumasksize, 1);
ASSERT(cpumask);
r = uv_thread_getaffinity(&threads[0], cpumask, cpumasksize);
ASSERT_OK(r);
ASSERT(cpumask[0] && "test must be run with cpu 0 affinity");
ncpus = 0;
while (cpumask[++ncpus]) { }
memset(cpumask, 0, 4 * cpumasksize);
t1first = cpumasksize * 0;
t1second = cpumasksize * 1;
t2first = cpumasksize * 2;
t2second = cpumasksize * 3;
cpumask[t1second + 0] = 1;
cpumask[t2first + 0] = 1;
cpumask[t1first + (ncpus >= 2)] = 1;
cpumask[t2second + (ncpus >= 2)] = 1;
#ifdef __linux__
cpumask[t1second + 2] = 1;
cpumask[t2first + 2] = 1;
cpumask[t1first + 3] = 1;
cpumask[t2second + 3] = 1;
#else
if (ncpus >= 3) {
cpumask[t1second + 2] = 1;
cpumask[t2first + 2] = 1;
}
if (ncpus >= 4) {
cpumask[t1first + 3] = 1;
cpumask[t2second + 3] = 1;
}
#endif
ASSERT_OK(uv_thread_create(threads + 1,
check_affinity,
&cpumask[t1first]));
ASSERT_OK(uv_thread_create(threads + 2,
check_affinity,
&cpumask[t2first]));
ASSERT_OK(uv_thread_join(threads + 1));
ASSERT_OK(uv_thread_join(threads + 2));
ASSERT(cpumask[t1first + 0] == (ncpus == 1));
ASSERT(cpumask[t1first + 1] == (ncpus >= 2));
ASSERT_OK(cpumask[t1first + 2]);
ASSERT(cpumask[t1first + 3] == (ncpus >= 4));
ASSERT_EQ(1, cpumask[t2first + 0]);
ASSERT_OK(cpumask[t2first + 1]);
ASSERT(cpumask[t2first + 2] == (ncpus >= 3));
ASSERT_OK(cpumask[t2first + 3]);
c = uv_thread_getcpu();
ASSERT_GE(c, 0);
memset(cpumask, 0, cpumasksize);
cpumask[c] = 1;
r = uv_thread_setaffinity(&threads[0], cpumask, NULL, cpumasksize);
ASSERT_OK(r);
memset(cpumask, 0, cpumasksize);
r = uv_thread_getaffinity(&threads[0], cpumask, cpumasksize);
ASSERT_OK(r);
for (i = 0; i < cpumasksize; i++) {
if (i == c)
ASSERT_EQ(1, cpumask[i]);
else
ASSERT_OK(cpumask[i]);
}
free(cpumask);
return 0;
}
#else
TEST_IMPL(thread_affinity) {
int cpumasksize;
cpumasksize = uv_cpumask_size();
ASSERT_EQ(cpumasksize, UV_ENOTSUP);
return 0;
}
#endif