* [PATCH] resctrl: mpam: Fix potential per-CPU data leak on probe failure
@ 2026-11-20 17:35 Guangshuo Li
2026-11-20 20:23 ` Fenghua Yu
0 siblings, 1 reply; 2+ messages in thread
From: Guangshuo Li @ 2026-11-20 17:35 UTC (permalink / raw)
To: James Morse, Ben Horgan, Reinette Chatre, Fenghua Yu
Cc: linux-arm-kernel, linux-kernel, stable
The PPI device IDs are allocated separately from the MSC. If subsequent
resource discovery fails, probe destroys the MSC without freeing its
per-CPU allocation.
Use devm_alloc_percpu() so the device releases the allocation when probe
fails. Its cleanup remains valid even if the MSC is freed beforehand.
This issue was found by our static analysis tool under development.
Fixes: 49aa621c4dca ("arm_mpam: Register and enable IRQs")
Cc: stable@vger.kernel.org
Signed-off-by: Guangshuo Li <liguangshuo@mail.sdu.edu.cn>
---
drivers/resctrl/mpam_devices.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/resctrl/mpam_devices.c b/drivers/resctrl/mpam_devices.c
index c067b6d57d71..3e89f9183041 100644
--- a/drivers/resctrl/mpam_devices.c
+++ b/drivers/resctrl/mpam_devices.c
@@ -2351,7 +2351,7 @@ static int __setup_ppi(struct mpam_msc *msc)
{
int cpu;
- msc->error_dev_id = alloc_percpu(struct mpam_msc *);
+ msc->error_dev_id = devm_alloc_percpu(&msc->pdev->dev, struct mpam_msc *);
if (!msc->error_dev_id)
return -ENOMEM;
base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH] resctrl: mpam: Fix potential per-CPU data leak on probe failure
2026-11-20 17:35 [PATCH] resctrl: mpam: Fix potential per-CPU data leak on probe failure Guangshuo Li
@ 2026-11-20 20:23 ` Fenghua Yu
0 siblings, 0 replies; 2+ messages in thread
From: Fenghua Yu @ 2026-11-20 20:23 UTC (permalink / raw)
To: Guangshuo Li, James Morse, Ben Horgan, Reinette Chatre
Cc: linux-arm-kernel, linux-kernel, stable
Hi, Guangshuo,
nits:
Subject: [PATCH] resctrl: mpam: Fix potential per-CPU data leak on probe
failure
1. s/resctrl: mpam/arm_mpam/
2. s/data leak/memory leak/. Without this fix, there is no "data" leak,
which means information is disclosed, right? Not freed memory is memory
leak, I think.
On 11/20/26 09:35, Guangshuo Li wrote:
> The PPI device IDs are allocated separately from the MSC. If subsequent
> resource discovery fails, probe destroys the MSC without freeing its
> per-CPU allocation.
>
> Use devm_alloc_percpu() so the device releases the allocation when probe
> fails. Its cleanup remains valid even if the MSC is freed beforehand.
>
> This issue was found by our static analysis tool under development.
>
> Fixes: 49aa621c4dca ("arm_mpam: Register and enable IRQs")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <liguangshuo@mail.sdu.edu.cn>
Reviewed-by: Fenghua Yu <fenghuay@nvidia.com>
Thanks.
-Fenghua
> ---
> drivers/resctrl/mpam_devices.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/resctrl/mpam_devices.c b/drivers/resctrl/mpam_devices.c
> index c067b6d57d71..3e89f9183041 100644
> --- a/drivers/resctrl/mpam_devices.c
> +++ b/drivers/resctrl/mpam_devices.c
> @@ -2351,7 +2351,7 @@ static int __setup_ppi(struct mpam_msc *msc)
> {
> int cpu;
>
> - msc->error_dev_id = alloc_percpu(struct mpam_msc *);
> + msc->error_dev_id = devm_alloc_percpu(&msc->pdev->dev, struct mpam_msc *);
> if (!msc->error_dev_id)
> return -ENOMEM;
>
>
> base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-20 20:23 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-20 17:35 [PATCH] resctrl: mpam: Fix potential per-CPU data leak on probe failure Guangshuo Li
2026-11-20 20:23 ` Fenghua Yu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®