diff options
| author | Ian Rogers <irogers@google.com> | 2025-01-13 11:43:45 -0800 |
|---|---|---|
| committer | Namhyung Kim <namhyung@kernel.org> | 2025-02-12 20:06:11 -0800 |
| commit | dc6d2bc2d893a878e7b58578ff01b4738708deb4 (patch) | |
| tree | b5524ff4429f1e88533f518753091462bc4bd8b4 /tools/perf/util/unwind-libdw.c | |
| parent | perf test stat_all_metrics: Ensure missing events fail test (diff) | |
| download | linux-dc6d2bc2d893a878e7b58578ff01b4738708deb4.tar.gz linux-dc6d2bc2d893a878e7b58578ff01b4738708deb4.zip | |
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 <tavianator@tavianator.com> as about 2.5% when running `perf
script --itrace=i0`:
https://lore.kernel.org/lkml/d841b97b3ad2ca8bcab07e4293375fb7c32dfce7.1736618095.git.tavianator@tavianator.com/
Adrian Hunter <adrian.hunter@intel.com> 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 <irogers@google.com>
Link: https://lore.kernel.org/r/20250113194345.1537821-1-irogers@google.com
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Diffstat (limited to 'tools/perf/util/unwind-libdw.c')
| -rw-r--r-- | tools/perf/util/unwind-libdw.c | 9 |
1 files changed, 6 insertions, 3 deletions
diff --git a/tools/perf/util/unwind-libdw.c b/tools/perf/util/unwind-libdw.c index bde216e630d2..793d11832694 100644 --- a/tools/perf/util/unwind-libdw.c +++ b/tools/perf/util/unwind-libdw.c @@ -190,7 +190,10 @@ static bool memory_read(Dwfl *dwfl __maybe_unused, Dwarf_Addr addr, Dwarf_Word * int offset; int ret; - ret = perf_reg_value(&start, &ui->sample->user_regs, + if (!ui->sample->user_regs) + return false; + + ret = perf_reg_value(&start, ui->sample->user_regs, perf_arch_reg_sp(arch)); if (ret) return false; @@ -273,7 +276,7 @@ int unwind__get_entries(unwind_entry_cb_t cb, void *arg, Dwarf_Word ip; int err = -EINVAL, i; - if (!data->user_regs.regs) + if (!data->user_regs || !data->user_regs->regs) return -EINVAL; ui = zalloc(sizeof(ui_buf) + sizeof(ui_buf.entries[0]) * max_stack); @@ -286,7 +289,7 @@ int unwind__get_entries(unwind_entry_cb_t cb, void *arg, if (!ui->dwfl) goto out; - err = perf_reg_value(&ip, &data->user_regs, perf_arch_reg_ip(arch)); + err = perf_reg_value(&ip, data->user_regs, perf_arch_reg_ip(arch)); if (err) goto out; |