[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <ed30ed57584e09b8492c0ceab20c1a7f16d27726.1710257512.git.robin.murphy@arm.com>
Date: Tue, 12 Mar 2024 17:34:03 +0000
From: Robin Murphy <robin.murphy@....com>
To: Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Will Deacon <will@...nel.org>
Cc: Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...nel.org>,
Ian Rogers <irogers@...gle.com>,
Adrian Hunter <adrian.hunter@...el.com>,
linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org,
x86@...nel.org,
linux-perf-users@...r.kernel.org,
jialong.yang@...ngroup.cn,
Shuai Xue <xueshuai@...ux.alibaba.com>
Subject: [PATCH 01/10] perf/alibaba_uncore_drw: Use correct CPU affinity
Although this driver is trying to use the IRQ-multiplexing mechanism to
handle context migration correctly, it's failing to make events follow
the IRQ instance, instead forcing them onto a per-PMU CPU which is set
once at probe time then never maintained. Fix this by using the correct
CPU from the IRQ instance, and remove the erroneous per-PMU data plus
the redundant check which does not mean what it thinks it means.
CC: Shuai Xue <xueshuai@...ux.alibaba.com>
Fixes: cf7b61073e45 ("drivers/perf: add DDR Sub-System Driveway PMU driver for Yitian 710 SoC")
Signed-off-by: Robin Murphy <robin.murphy@....com>
---
drivers/perf/alibaba_uncore_drw_pmu.c | 12 ++----------
1 file changed, 2 insertions(+), 10 deletions(-)
diff --git a/drivers/perf/alibaba_uncore_drw_pmu.c b/drivers/perf/alibaba_uncore_drw_pmu.c
index 19d459a36be5..b37e9794823a 100644
--- a/drivers/perf/alibaba_uncore_drw_pmu.c
+++ b/drivers/perf/alibaba_uncore_drw_pmu.c
@@ -96,8 +96,6 @@ struct ali_drw_pmu {
struct list_head pmus_node;
struct ali_drw_pmu_irq *irq;
- int irq_num;
- int cpu;
DECLARE_BITMAP(used_mask, ALI_DRW_PMU_COMMON_MAX_COUNTERS);
struct perf_event *events[ALI_DRW_PMU_COMMON_MAX_COUNTERS];
int evtids[ALI_DRW_PMU_COMMON_MAX_COUNTERS];
@@ -221,7 +219,7 @@ static ssize_t ali_drw_pmu_cpumask_show(struct device *dev,
{
struct ali_drw_pmu *drw_pmu = to_ali_drw_pmu(dev_get_drvdata(dev));
- return cpumap_print_to_pagebuf(true, buf, cpumask_of(drw_pmu->cpu));
+ return cpumap_print_to_pagebuf(true, buf, cpumask_of(drw_pmu->irq->cpu));
}
static struct device_attribute ali_drw_pmu_cpumask_attr =
@@ -550,11 +548,7 @@ static int ali_drw_pmu_event_init(struct perf_event *event)
return -EOPNOTSUPP;
}
- event->cpu = drw_pmu->cpu;
- if (event->cpu < 0) {
- dev_err(dev, "Per-task mode not supported!\n");
- return -EOPNOTSUPP;
- }
+ event->cpu = drw_pmu->irq->cpu;
if (event->group_leader != event &&
!is_software_event(event->group_leader)) {
@@ -701,8 +695,6 @@ static int ali_drw_pmu_probe(struct platform_device *pdev)
/* clearing interrupt status */
writel(0xffffff, drw_pmu->cfg_base + ALI_DRW_PMU_OV_INTR_CLR);
- drw_pmu->cpu = smp_processor_id();
-
ret = ali_drw_pmu_init_irq(drw_pmu, pdev);
if (ret)
return ret;
--
2.39.2.101.g768bb238c484.dirty
Powered by blists - more mailing lists