perf bench mem: Always memset source before memcpy

[ Upstream commit 1beaef2 ]

For memcpy, the source pages are memset to zero only when --cycles is
used.  This leads to wildly different results with or without --cycles,
since all sources pages are likely to be mapped to the same zero page
without explicit writes.

Before this fix:

$ export cmd="./perf stat -e LLC-loads -- ./perf bench \
  mem memcpy -s 1024MB -l 100 -f default"
$ $cmd

         2,935,826      LLC-loads
       3.821677452 seconds time elapsed

$ $cmd --cycles

       217,533,436      LLC-loads
       8.616725985 seconds time elapsed

After this fix:

$ $cmd

       214,459,686      LLC-loads
       8.674301124 seconds time elapsed

$ $cmd --cycles

       214,758,651      LLC-loads
       8.644480006 seconds time elapsed

Fixes: 47b5757 ("perf bench mem: Move boilerplate memory allocation to the infrastructure")
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: kernel@axis.com
Link: http://lore.kernel.org/lkml/20200810133404.30829-1-vincent.whitchurch@axis.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

tools/perf/bench/mem-functions.c

@@ -222,12 +222,8 @@ static int bench_mem_common(int argc, const char **argv, struct bench_mem_info *
 	return 0;
 }
 
-static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, void *dst)
+static void memcpy_prefault(memcpy_t fn, size_t size, void *src, void *dst)
 {
-	u64 cycle_start = 0ULL, cycle_end = 0ULL;
-	memcpy_t fn = r->fn.memcpy;
-	int i;
-
 	/* Make sure to always prefault zero pages even if MMAP_THRESH is crossed: */
 	memset(src, 0, size);
 
@@ -236,6 +232,15 @@ static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, vo
 	 * to not measure page fault overhead:
 	 */
 	fn(dst, src, size);
+}
+
+static u64 do_memcpy_cycles(const struct function *r, size_t size, void *src, void *dst)
+{
+	u64 cycle_start = 0ULL, cycle_end = 0ULL;
+	memcpy_t fn = r->fn.memcpy;
+	int i;
+
+	memcpy_prefault(fn, size, src, dst);
 
 	cycle_start = get_cycles();
 	for (i = 0; i < nr_loops; ++i)
@@ -251,11 +256,7 @@ static double do_memcpy_gettimeofday(const struct function *r, size_t size, void
 	memcpy_t fn = r->fn.memcpy;
 	int i;
 
-	/*
-	 * We prefault the freshly allocated memory range here,
-	 * to not measure page fault overhead:
-	 */
-	fn(dst, src, size);
+	memcpy_prefault(fn, size, src, dst);
 
 	BUG_ON(gettimeofday(&tv_start, NULL));
 	for (i = 0; i < nr_loops; ++i)