From dc6d2bc2d893a878e7b58578ff01b4738708deb4 Mon Sep 17 00:00:00 2001 From: Ian Rogers Date: Mon, 13 Jan 2025 11:43:45 -0800 Subject: perf sample: Make user_regs and intr_regs optional The struct dump_regs contains 512 bytes of cache_regs, meaning the two values in perf_sample contribute 1088 bytes of its total 1384 bytes size. Initializing this much memory has a cost reported by Tavian Barnes as about 2.5% when running `perf script --itrace=i0`: https://lore.kernel.org/lkml/d841b97b3ad2ca8bcab07e4293375fb7c32dfce7.1736618095.git.tavianator@tavianator.com/ Adrian Hunter replied that the zero initialization was necessary and couldn't simply be removed. This patch aims to strike a middle ground of still zeroing the perf_sample, but removing 79% of its size by make user_regs and intr_regs optional pointers to zalloc-ed memory. To support the allocation accessors are created for user_regs and intr_regs. To support correct cleanup perf_sample__init and perf_sample__exit functions are created and added throughout the code base. Signed-off-by: Ian Rogers Link: https://lore.kernel.org/r/20250113194345.1537821-1-irogers@google.com Signed-off-by: Namhyung Kim --- tools/perf/builtin-trace.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'tools/perf/builtin-trace.c') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index dbf513f9e478..f55a8a6481f2 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -4066,13 +4066,16 @@ static int __trace__deliver_event(struct trace *trace, union perf_event *event) { struct evlist *evlist = trace->evlist; struct perf_sample sample; - int err = evlist__parse_sample(evlist, event, &sample); + int err; + perf_sample__init(&sample, /*all=*/false); + err = evlist__parse_sample(evlist, event, &sample); if (err) fprintf(trace->output, "Can't parse sample, err = %d, skipping...\n", err); else trace__handle_event(trace, event, &sample); + perf_sample__exit(&sample); return 0; } -- cgit