[PATCH] watchdog: Prefer use "ref-cycles" for NMI watchdog

From: Song Liu
Date: Tue May 09 2023 - 18:18:00 EST


NMI watchdog permanently consumes one hardware counters per CPU on the
system. For systems that use many hardware counters, this causes more
aggressive time multiplexing of perf events.

OTOH, some CPUs (mostly Intel) support "ref-cycles" event, which is rarely
used. Try use "ref-cycles" for the watchdog. If the CPU supports it, so
that one more hardware counter is available to the user. If the CPU doesn't
support "ref-cycles", fall back to "cycles".

The downside of this change is that users of "ref-cycles" need to disable
nmi_watchdog.

Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Signed-off-by: Song Liu <song@xxxxxxxxxx>
---
kernel/watchdog_hld.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c
index 247bf0b1582c..f77109d98641 100644
--- a/kernel/watchdog_hld.c
+++ b/kernel/watchdog_hld.c
@@ -100,7 +100,7 @@ static inline bool watchdog_check_timestamp(void)

static struct perf_event_attr wd_hw_attr = {
.type = PERF_TYPE_HARDWARE,
- .config = PERF_COUNT_HW_CPU_CYCLES,
+ .config = PERF_COUNT_HW_REF_CPU_CYCLES,
.size = sizeof(struct perf_event_attr),
.pinned = 1,
.disabled = 1,
@@ -286,6 +286,12 @@ int __init hardlockup_detector_perf_init(void)
{
int ret = hardlockup_detector_event_create();

+ if (ret) {
+ /* Failed to create "ref-cycles", try "cycles" instead */
+ wd_hw_attr.config = PERF_COUNT_HW_CPU_CYCLES;
+ ret = hardlockup_detector_event_create();
+ }
+
if (ret) {
pr_info("Perf NMI watchdog permanently disabled\n");
} else {
--
2.34.1