2024-04-02 11:02:50

by Dawei Li

Subject: [PATCH 2/9] perf/arm-cmn: Avoid explicit cpumask var allocation from stack

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.

Use *cpumask_var API(s) to address it.

Signed-off-by: Dawei Li <[email protected]>
---
drivers/perf/arm-cmn.c | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/perf/arm-cmn.c b/drivers/perf/arm-cmn.c
index 7ef9c7e4836b..7278fd72d3da 100644
--- a/drivers/perf/arm-cmn.c
+++ b/drivers/perf/arm-cmn.c
@@ -1949,21 +1949,26 @@ static int arm_cmn_pmu_offline_cpu(unsigned int cpu, struct hlist_node *cpuhp_no
{
struct arm_cmn *cmn;
unsigned int target;
+ cpumask_var_t mask;
int node;
- cpumask_t mask;

cmn = hlist_entry_safe(cpuhp_node, struct arm_cmn, cpuhp_node);
if (cpu != cmn->cpu)
return 0;

+ if (!alloc_cpumask_var(&mask, GFP_KERNEL))
+ return 0;
+
node = dev_to_node(cmn->dev);
- if (cpumask_and(&mask, cpumask_of_node(node), cpu_online_mask) &&
- cpumask_andnot(&mask, &mask, cpumask_of(cpu)))
- target = cpumask_any(&mask);
+ if (cpumask_and(mask, cpumask_of_node(node), cpu_online_mask) &&
+ cpumask_andnot(mask, mask, cpumask_of(cpu)))
+ target = cpumask_any(mask);
else
target = cpumask_any_but(cpu_online_mask, cpu);
if (target < nr_cpu_ids)
arm_cmn_migrate(cmn, target);
+
+ free_cpumask_var(mask);
return 0;
}

--
2.27.0

2024-04-05 14:32:07

by Robin Murphy

Subject: Re: [PATCH 2/9] perf/arm-cmn: Avoid explicit cpumask var allocation from stack

On 2024-04-02 11:56 am, Dawei Li wrote:
> 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.
>
> Use *cpumask_var API(s) to address it.

I think the temporary mask may simply be redundant anyway. It seems like
I may have misunderstood, and cpumask_of_node() actually only covers
online CPUs already.

Thanks,
Robin.

> Signed-off-by: Dawei Li <[email protected]>
> ---
> drivers/perf/arm-cmn.c | 13 +++++++++----
> 1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/perf/arm-cmn.c b/drivers/perf/arm-cmn.c
> index 7ef9c7e4836b..7278fd72d3da 100644
> --- a/drivers/perf/arm-cmn.c
> +++ b/drivers/perf/arm-cmn.c
> @@ -1949,21 +1949,26 @@ static int arm_cmn_pmu_offline_cpu(unsigned int cpu, struct hlist_node *cpuhp_no
> {
> struct arm_cmn *cmn;
> unsigned int target;
> + cpumask_var_t mask;
> int node;
> - cpumask_t mask;
>
> cmn = hlist_entry_safe(cpuhp_node, struct arm_cmn, cpuhp_node);
> if (cpu != cmn->cpu)
> return 0;
>
> + if (!alloc_cpumask_var(&mask, GFP_KERNEL))
> + return 0;
> +
> node = dev_to_node(cmn->dev);
> - if (cpumask_and(&mask, cpumask_of_node(node), cpu_online_mask) &&
> - cpumask_andnot(&mask, &mask, cpumask_of(cpu)))
> - target = cpumask_any(&mask);
> + if (cpumask_and(mask, cpumask_of_node(node), cpu_online_mask) &&
> + cpumask_andnot(mask, mask, cpumask_of(cpu)))
> + target = cpumask_any(mask);
> else
> target = cpumask_any_but(cpu_online_mask, cpu);
> if (target < nr_cpu_ids)
> arm_cmn_migrate(cmn, target);
> +
> + free_cpumask_var(mask);
> return 0;
> }
>