Orange Pi5 kernel

^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   3)  * mem-memcpy.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   4)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   5)  * Simple memcpy() and memset() benchmarks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   6)  *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   7)  * Written by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   8)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   9) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  10) #include "debug.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  11) #include "../perf-sys.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  12) #include <subcmd/parse-options.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  13) #include "../util/header.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  14) #include "../util/cloexec.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  15) #include "../util/string2.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  16) #include "bench.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  17) #include "mem-memcpy-arch.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  18) #include "mem-memset-arch.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  19) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  20) #include <stdio.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  21) #include <stdlib.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  22) #include <string.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  23) #include <unistd.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  24) #include <sys/time.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  25) #include <errno.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  26) #include <linux/time64.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  27) #include <linux/zalloc.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  28) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  29) #define K 1024
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  30) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  31) static const char	*size_str	= "1MB";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  32) static const char	*function_str	= "all";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  33) static int		nr_loops	= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  34) static bool		use_cycles;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  35) static int		cycles_fd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  36) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  37) static const struct option options[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  38) 	OPT_STRING('s', "size", &size_str, "1MB",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  39) 		    "Specify the size of the memory buffers. "
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  40) 		    "Available units: B, KB, MB, GB and TB (case insensitive)"),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  41) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  42) 	OPT_STRING('f', "function", &function_str, "all",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  43) 		    "Specify the function to run, \"all\" runs all available functions, \"help\" lists them"),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  44) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  45) 	OPT_INTEGER('l', "nr_loops", &nr_loops,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  46) 		    "Specify the number of loops to run. (default: 1)"),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  47) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  48) 	OPT_BOOLEAN('c', "cycles", &use_cycles,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  49) 		    "Use a cycles event instead of gettimeofday() to measure performance"),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  50) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  51) 	OPT_END()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  52) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  53) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  54) typedef void *(*memcpy_t)(void *, const void *, size_t);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  55) typedef void *(*memset_t)(void *, int, size_t);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  56) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  57) struct function {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  58) 	const char *name;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  59) 	const char *desc;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  60) 	union {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  61) 		memcpy_t memcpy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  62) 		memset_t memset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  63) 	} fn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  64) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  65) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  66) static struct perf_event_attr cycle_attr = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  67) 	.type		= PERF_TYPE_HARDWARE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  68) 	.config		= PERF_COUNT_HW_CPU_CYCLES
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  69) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  70) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  71) static int init_cycles(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  72) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  73) 	cycles_fd = sys_perf_event_open(&cycle_attr, getpid(), -1, -1, perf_event_open_cloexec_flag());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  74) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  75) 	if (cycles_fd < 0 && errno == ENOSYS) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  76) 		pr_debug("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  77) 		return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  78) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  79) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  80) 	return cycles_fd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  81) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  82) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  83) static u64 get_cycles(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  84) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  85) 	int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  86) 	u64 clk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  87) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  88) 	ret = read(cycles_fd, &clk, sizeof(u64));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  89) 	BUG_ON(ret != sizeof(u64));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  90) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  91) 	return clk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  92) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  93) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  94) static double timeval2double(struct timeval *ts)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  95) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  96) 	return (double)ts->tv_sec + (double)ts->tv_usec / (double)USEC_PER_SEC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  97) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  98) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  99) #define print_bps(x) do {						\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) 		if (x < K)						\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) 			printf(" %14lf bytes/sec\n", x);		\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) 		else if (x < K * K)					\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) 			printf(" %14lfd KB/sec\n", x / K);		\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) 		else if (x < K * K * K)					\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) 			printf(" %14lf MB/sec\n", x / K / K);		\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) 		else							\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) 			printf(" %14lf GB/sec\n", x / K / K / K);	\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) 	} while (0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) struct bench_mem_info {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) 	const struct function *functions;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) 	u64 (*do_cycles)(const struct function *r, size_t size, void *src, void *dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) 	double (*do_gettimeofday)(const struct function *r, size_t size, void *src, void *dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) 	const char *const *usage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) 	bool alloc_src;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) static void __bench_mem_function(struct bench_mem_info *info, int r_idx, size_t size, double size_total)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) 	const struct function *r = &info->functions[r_idx];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) 	double result_bps = 0.0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) 	u64 result_cycles = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) 	void *src = NULL, *dst = zalloc(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) 	printf("# function '%s' (%s)\n", r->name, r->desc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) 	if (dst == NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) 		goto out_alloc_failed;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) 	if (info->alloc_src) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) 		src = zalloc(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) 		if (src == NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) 			goto out_alloc_failed;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) 	if (bench_format == BENCH_FORMAT_DEFAULT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) 		printf("# Copying %s bytes ...\n\n", size_str);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) 	if (use_cycles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) 		result_cycles = info->do_cycles(r, size, src, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) 	} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) 		result_bps = info->do_gettimeofday(r, size, src, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) 	switch (bench_format) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) 	case BENCH_FORMAT_DEFAULT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) 		if (use_cycles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) 			printf(" %14lf cycles/byte\n", (double)result_cycles/size_total);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) 		} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) 			print_bps(result_bps);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) 		}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) 		break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) 	case BENCH_FORMAT_SIMPLE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) 		if (use_cycles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) 			printf("%lf\n", (double)result_cycles/size_total);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) 		} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) 			printf("%lf\n", result_bps);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) 		}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) 		break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) 	default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) 		BUG_ON(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) 		break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) out_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) 	free(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) 	free(dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) 	return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) out_alloc_failed:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) 	printf("# Memory allocation failed - maybe size (%s) is too large?\n", size_str);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) 	goto out_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) static int bench_mem_common(int argc, const char **argv, struct bench_mem_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) 	size_t size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) 	double size_total;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) 	argc = parse_options(argc, argv, options, info->usage, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) 	if (use_cycles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) 		i = init_cycles();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) 		if (i < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) 			fprintf(stderr, "Failed to open cycles counter\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) 			return i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) 		}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 	size = (size_t)perf_atoll((char *)size_str);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) 	size_total = (double)size * nr_loops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) 	if ((s64)size <= 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) 		fprintf(stderr, "Invalid size:%s\n", size_str);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) 		return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) 	if (!strncmp(function_str, "all", 3)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) 		for (i = 0; info->functions[i].name; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) 			__bench_mem_function(info, i, size, size_total);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) 		return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) 	for (i = 0; info->functions[i].name; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) 		if (!strcmp(info->functions[i].name, function_str))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) 			break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) 	if (!info->functions[i].name) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) 		if (strcmp(function_str, "help") && strcmp(function_str, "h"))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) 			printf("Unknown function: %s\n", function_str);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) 		printf("Available functions:\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) 		for (i = 0; info->functions[i].name; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) 			printf("\t%s ... %s\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) 			       info->functions[i].name, info->functions[i].desc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) 		}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) 		return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) 	__bench_mem_function(info, i, size, size_total);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) 	return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) static void memcpy_prefault(memcpy_t fn, size_t size, void *src, void *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) 	/* Make sure to always prefault zero pages even if MMAP_THRESH is crossed: */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) 	memset(src, 0, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) 	 * We prefault the freshly allocated memory range here,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) 	 * to not measure page fault overhead:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) 	fn(dst, src, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, void *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) 	u64 cycle_start = 0ULL, cycle_end = 0ULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) 	memcpy_t fn = r->fn.memcpy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) 	memcpy_prefault(fn, size, src, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) 	cycle_start = get_cycles();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) 	for (i = 0; i < nr_loops; ++i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) 		fn(dst, src, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) 	cycle_end = get_cycles();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) 	return cycle_end - cycle_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) static double do_memcpy_gettimeofday(const struct function *r, size_t size, void *src, void *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) 	struct timeval tv_start, tv_end, tv_diff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) 	memcpy_t fn = r->fn.memcpy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) 	memcpy_prefault(fn, size, src, dst);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) 	BUG_ON(gettimeofday(&tv_start, NULL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) 	for (i = 0; i < nr_loops; ++i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) 		fn(dst, src, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) 	BUG_ON(gettimeofday(&tv_end, NULL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) 	timersub(&tv_end, &tv_start, &tv_diff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) 	return (double)(((double)size * nr_loops) / timeval2double(&tv_diff));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) struct function memcpy_functions[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) 	{ .name		= "default",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) 	  .desc		= "Default memcpy() provided by glibc",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) 	  .fn.memcpy	= memcpy },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) #ifdef HAVE_ARCH_X86_64_SUPPORT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) # define MEMCPY_FN(_fn, _name, _desc) {.name = _name, .desc = _desc, .fn.memcpy = _fn},
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) # include "mem-memcpy-x86-64-asm-def.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) # undef MEMCPY_FN
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) 	{ .name = NULL, }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) static const char * const bench_mem_memcpy_usage[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) 	"perf bench mem memcpy <options>",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) 	NULL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) int bench_mem_memcpy(int argc, const char **argv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) 	struct bench_mem_info info = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) 		.functions		= memcpy_functions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) 		.do_cycles		= do_memcpy_cycles,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) 		.do_gettimeofday	= do_memcpy_gettimeofday,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) 		.usage			= bench_mem_memcpy_usage,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) 		.alloc_src              = true,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) 	};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) 	return bench_mem_common(argc, argv, &info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) static u64 do_memset_cycles(const struct function *r, size_t size, void *src __maybe_unused, void *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) 	u64 cycle_start = 0ULL, cycle_end = 0ULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) 	memset_t fn = r->fn.memset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) 	 * We prefault the freshly allocated memory range here,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) 	 * to not measure page fault overhead:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) 	fn(dst, -1, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) 	cycle_start = get_cycles();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) 	for (i = 0; i < nr_loops; ++i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) 		fn(dst, i, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) 	cycle_end = get_cycles();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) 	return cycle_end - cycle_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) static double do_memset_gettimeofday(const struct function *r, size_t size, void *src __maybe_unused, void *dst)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) 	struct timeval tv_start, tv_end, tv_diff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) 	memset_t fn = r->fn.memset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) 	 * We prefault the freshly allocated memory range here,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) 	 * to not measure page fault overhead:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) 	fn(dst, -1, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) 	BUG_ON(gettimeofday(&tv_start, NULL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) 	for (i = 0; i < nr_loops; ++i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) 		fn(dst, i, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) 	BUG_ON(gettimeofday(&tv_end, NULL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) 	timersub(&tv_end, &tv_start, &tv_diff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) 	return (double)(((double)size * nr_loops) / timeval2double(&tv_diff));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) static const char * const bench_mem_memset_usage[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) 	"perf bench mem memset <options>",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) 	NULL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) static const struct function memset_functions[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) 	{ .name		= "default",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) 	  .desc		= "Default memset() provided by glibc",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) 	  .fn.memset	= memset },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) #ifdef HAVE_ARCH_X86_64_SUPPORT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) # define MEMSET_FN(_fn, _name, _desc) { .name = _name, .desc = _desc, .fn.memset = _fn },
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) # include "mem-memset-x86-64-asm-def.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) # undef MEMSET_FN
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) 	{ .name = NULL, }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) int bench_mem_memset(int argc, const char **argv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) 	struct bench_mem_info info = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) 		.functions		= memset_functions,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) 		.do_cycles		= do_memset_cycles,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) 		.do_gettimeofday	= do_memset_gettimeofday,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) 		.usage			= bench_mem_memset_usage,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) 	};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) 	return bench_mem_common(argc, argv, &info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) }