Received-SPF: pass (google.com: domain of linux-kernel+bounces-129755-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
From: Dawei Li <dawei.li@shingroup.cn>
To: will@kernel.org,
	mark.rutland@arm.com,
	yury.norov@gmail.com,
	linux@rasmusvillemoes.dk
Cc: xueshuai@linux.alibaba.com,
	renyu.zj@linux.alibaba.com,
	yangyicong@hisilicon.com,
	jonathan.cameron@huawei.com,
	andersson@kernel.org,
	konrad.dybcio@linaro.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org,
	linux-arm-msm@vger.kernel.org,
	Dawei Li <dawei.li@shingroup.cn>
Subject: [PATCH v2 05/10] perf/arm_dsu: Avoid placing cpumask var on stack
Date: Wed,  3 Apr 2024 20:51:04 +0800
Message-Id: <20240403125109.2054881-6-dawei.li@shingroup.cn>
In-Reply-To: <20240403125109.2054881-1-dawei.li@shingroup.cn>
References: <20240403125109.2054881-1-dawei.li@shingroup.cn>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Feedback-ID: bizesmtpsz:shingroup.cn:qybglogicsvrgz:qybglogicsvrgz5a-1

For CONFIG_CPUMASK_OFFSTACK=y kernel, explicit allocation of cpumask
variable on stack is not recommended since it can cause potential stack
overflow.

Instead, kernel code should always use *cpumask_var API(s) to allocate
cpumask var in config-neutral way, leaving allocation strategy to
CONFIG_CPUMASK_OFFSTACK.

But dynamic allocation in cpuhp's teardown callback is somewhat problematic
for if allocation fails(which is unlikely but still possible):
- If -ENOMEM is returned to caller, kernel crashes for non-bringup
  teardown;
- If callback pretends nothing happened and returns 0 to caller, it may
  trap system into an in-consisitent/compromised state;

Use newly-introduced cpumask_any_and_but() to address all issues above.
It eliminates usage of temporary cpumask var in generic way, no matter how
the cpumask var is allocated.

Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Dawei Li <dawei.li@shingroup.cn>
---
 drivers/perf/arm_dsu_pmu.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/drivers/perf/arm_dsu_pmu.c b/drivers/perf/arm_dsu_pmu.c
index bae3ca37f846..adc0bbb5fafe 100644
--- a/drivers/perf/arm_dsu_pmu.c
+++ b/drivers/perf/arm_dsu_pmu.c
@@ -230,15 +230,6 @@ static const struct attribute_group *dsu_pmu_attr_groups[] = {
 	NULL,
 };
 
-static int dsu_pmu_get_online_cpu_any_but(struct dsu_pmu *dsu_pmu, int cpu)
-{
-	struct cpumask online_supported;
-
-	cpumask_and(&online_supported,
-			 &dsu_pmu->associated_cpus, cpu_online_mask);
-	return cpumask_any_but(&online_supported, cpu);
-}
-
 static inline bool dsu_pmu_counter_valid(struct dsu_pmu *dsu_pmu, u32 idx)
 {
 	return (idx < dsu_pmu->num_counters) ||
@@ -827,14 +818,16 @@ static int dsu_pmu_cpu_online(unsigned int cpu, struct hlist_node *node)
 
 static int dsu_pmu_cpu_teardown(unsigned int cpu, struct hlist_node *node)
 {
-	int dst;
-	struct dsu_pmu *dsu_pmu = hlist_entry_safe(node, struct dsu_pmu,
-						   cpuhp_node);
+	struct dsu_pmu *dsu_pmu;
+	unsigned int dst;
+
+	dsu_pmu = hlist_entry_safe(node, struct dsu_pmu, cpuhp_node);
 
 	if (!cpumask_test_and_clear_cpu(cpu, &dsu_pmu->active_cpu))
 		return 0;
 
-	dst = dsu_pmu_get_online_cpu_any_but(dsu_pmu, cpu);
+	dst = cpumask_any_and_but(&dsu_pmu->associated_cpus,
+				  cpu_online_mask, cpu);
 	/* If there are no active CPUs in the DSU, leave IRQ disabled */
 	if (dst >= nr_cpu_ids)
 		return 0;
-- 
2.27.0