* [PATCH] platform/x86/amd/pmf: check APTS buffer before reading header
@ 2026-11-03 11:41 Kang Chen
2026-11-03 14:11 ` Mario Limonciello
0 siblings, 1 reply; 2+ messages in thread
From: Kang Chen @ 2026-11-03 11:41 UTC (permalink / raw)
To: Shyam Sundar S K
Cc: Hans de Goede, Ilpo Järvinen, Patil Rajesh Reddy,
platform-driver-x86, linux-kernel
apts_if_call_store_buffer() reads a u16 length from the APTS buffer
without checking that the buffer has at least two bytes. The sibling
apmf_if_call_store_buffer() performs this check. A Buffer(0) reply
from the APTS method makes the driver read two bytes past the ACPI
allocation.
Add the missing length check.
Fixes: 3eecb434d7f2 ("platform/x86/amd/pmf: Add support to get sps default APTS index values")
Signed-off-by: Kang Chen <k.chen@smail.nju.edu.cn>
---
drivers/platform/x86/amd/pmf/acpi.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/drivers/platform/x86/amd/pmf/acpi.c b/drivers/platform/x86/amd/pmf/acpi.c
index 4392bf0bc0bf..8e0d9be8c042 100644
--- a/drivers/platform/x86/amd/pmf/acpi.c
+++ b/drivers/platform/x86/amd/pmf/acpi.c
@@ -137,6 +137,12 @@ int apts_if_call_store_buffer(struct amd_pmf_dev *pdev, u32 index, void *data, s
goto out;
}
+ if (info->buffer.length < sizeof(u16)) {
+ dev_err(pdev->dev, "buffer too small\n");
+ err = -EINVAL;
+ goto out;
+ }
+
size = *(u16 *)info->buffer.pointer;
if (info->buffer.length < size) {
dev_err(pdev->dev, "buffer smaller than header size %u < %zu\n",
--
2.34.1
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH] platform/x86/amd/pmf: check APTS buffer before reading header
2026-11-03 11:41 [PATCH] platform/x86/amd/pmf: check APTS buffer before reading header Kang Chen
@ 2026-11-03 14:11 ` Mario Limonciello
0 siblings, 0 replies; 2+ messages in thread
From: Mario Limonciello @ 2026-11-03 14:11 UTC (permalink / raw)
To: Kang Chen, Shyam Sundar S K
Cc: Hans de Goede, Ilpo Järvinen, Patil Rajesh Reddy,
platform-driver-x86, linux-kernel
On 11/3/26 05:41, Kang Chen wrote:
> apts_if_call_store_buffer() reads a u16 length from the APTS buffer
> without checking that the buffer has at least two bytes. The sibling
> apmf_if_call_store_buffer() performs this check. A Buffer(0) reply
> from the APTS method makes the driver read two bytes past the ACPI
> allocation.
>
> Add the missing length check.
>
> Fixes: 3eecb434d7f2 ("platform/x86/amd/pmf: Add support to get sps default APTS index values")
> Signed-off-by: Kang Chen <k.chen@smail.nju.edu.cn>
> ---
> drivers/platform/x86/amd/pmf/acpi.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/drivers/platform/x86/amd/pmf/acpi.c b/drivers/platform/x86/amd/pmf/acpi.c
> index 4392bf0bc0bf..8e0d9be8c042 100644
> --- a/drivers/platform/x86/amd/pmf/acpi.c
> +++ b/drivers/platform/x86/amd/pmf/acpi.c
> @@ -137,6 +137,12 @@ int apts_if_call_store_buffer(struct amd_pmf_dev *pdev, u32 index, void *data, s
> goto out;
> }
>
> + if (info->buffer.length < sizeof(u16)) {
> + dev_err(pdev->dev, "buffer too small\n");
> + err = -EINVAL;
> + goto out;
> + }
> +
> size = *(u16 *)info->buffer.pointer;
> if (info->buffer.length < size) {
> dev_err(pdev->dev, "buffer smaller than header size %u < %zu\n",
Thanks for the patch. But I would say there is another on the mailing
list that does very similar and covers more:
https://lore.kernel.org/platform-driver-x86/20261005072811.1378555-1-Shyam-sundar.S-k@amd.com/
It just needs to be rebased.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-03 14:11 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-03 11:41 [PATCH] platform/x86/amd/pmf: check APTS buffer before reading header Kang Chen
2026-11-03 14:11 ` Mario Limonciello
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®