2015-05-15 15:42:16

by Shailendra Verma

Subject: [PATCH 10/10] arm: perf: Fix memory leak when probing PMU PPIs

Commit 338d9dd3e2ae ("ARM: 8351/1: perf: don't warn about missing
interrupt-affinity property for PPIs") added a check for PPIs so that
we avoid parsing the interrupt-affinity property for these naturally
affine interrupts.

Unfortunately, this check can trigger an early (successful) return and
we will leak the irqs array. This patch fixes the issue by reordering
the code so that the check is performed before any independent
allocation.

Signed-off-by: Shailendra Verma <[email protected]>
---
arch/arm/kernel/perf_event_cpu.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/arch/arm/kernel/perf_event_cpu.c b/arch/arm/kernel/perf_event_cpu.c
index 213919b..a7099ee 100644
--- a/arch/arm/kernel/perf_event_cpu.c
+++ b/arch/arm/kernel/perf_event_cpu.c
@@ -303,17 +303,17 @@ static int probe_current_pmu(struct arm_pmu *pmu)

static int of_pmu_irq_cfg(struct platform_device *pdev)
{
- int i, irq;
- int *irqs = kcalloc(pdev->num_resources, sizeof(*irqs), GFP_KERNEL);
-
- if (!irqs)
- return -ENOMEM;
+ int i, irq, *irqs;

/* Don't bother with PPIs; they're already affine */
irq = platform_get_irq(pdev, 0);
if (irq >= 0 && irq_is_percpu(irq))
return 0;

+ irqs = kcalloc(pdev->num_resources, sizeof(*irqs), GFP_KERNEL);
+ if (!irqs)
+ return -ENOMEM;
+
for (i = 0; i < pdev->num_resources; ++i) {
struct device_node *dn;
int cpu;
--
1.7.9.5

2015-05-15 17:07:15

by Mark Rutland

Subject: Re: [PATCH 10/10] arm: perf: Fix memory leak when probing PMU PPIs

On Fri, May 15, 2015 at 04:41:54PM +0100, Shailendra Verma wrote:
> Commit 338d9dd3e2ae ("ARM: 8351/1: perf: don't warn about missing
> interrupt-affinity property for PPIs") added a check for PPIs so that
> we avoid parsing the interrupt-affinity property for these naturally
> affine interrupts.
>
> Unfortunately, this check can trigger an early (successful) return and
> we will leak the irqs array. This patch fixes the issue by reordering
> the code so that the check is performed before any independent
> allocation.
>
> Signed-off-by: Shailendra Verma <[email protected]>
> ---
> arch/arm/kernel/perf_event_cpu.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/arch/arm/kernel/perf_event_cpu.c b/arch/arm/kernel/perf_event_cpu.c
> index 213919b..a7099ee 100644
> --- a/arch/arm/kernel/perf_event_cpu.c
> +++ b/arch/arm/kernel/perf_event_cpu.c
> @@ -303,17 +303,17 @@ static int probe_current_pmu(struct arm_pmu *pmu)
>
> static int of_pmu_irq_cfg(struct platform_device *pdev)
> {
> - int i, irq;
> - int *irqs = kcalloc(pdev->num_resources, sizeof(*irqs), GFP_KERNEL);
> -
> - if (!irqs)
> - return -ENOMEM;
> + int i, irq, *irqs;
>
> /* Don't bother with PPIs; they're already affine */
> irq = platform_get_irq(pdev, 0);
> if (irq >= 0 && irq_is_percpu(irq))
> return 0;
>
> + irqs = kcalloc(pdev->num_resources, sizeof(*irqs), GFP_KERNEL);
> + if (!irqs)
> + return -ENOMEM;
> +

Will Deacon posted a patch for this a few days ago [1], and it's
allready in Russell's patch system as 8357/1.

Thanks,
Mark.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2015-May/342367.html
[2] http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=8357/1