[PATCH] ACPI, APEI, EINJ: Refuse to inject into the zero page

[PATCH] ACPI, APEI, EINJ: Refuse to inject into the zero page

[Tony Luck]

Some validation tests dynamically inject errors into memory used by
applications to check that the system can recover from a variety of
poison consumption sceenarios.

But sometimes the virtual address picked by these tests is mapped to
the zero page.

This causes additional unexpected machine checks as other processes that
map the zero page also consume the poison.

Disallow injection to the zero page.

Signed-off-by: Tony Luck <tony.luck@intel.com>

---
I picked -EADDRINUSE as a somewhat descriptive error code distinct
fromm -EINVAL used elsewhere in EINJ, but I'm not strongly attached
to it. Any other non -EINVAL value would be as good an indicator to
the validation tests that they shouldn't inject to this address.
---
 drivers/acpi/apei/einj.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj.c
index 95cc2a9f3e05..d4326ec12d29 100644
--- a/drivers/acpi/apei/einj.c
+++ b/drivers/acpi/apei/einj.c
@@ -549,6 +549,9 @@ static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2,
 	     !arch_is_platform_page(base_addr)))
 		return -EINVAL;
 
+	if (is_zero_pfn(base_addr >> PAGE_SHIFT))
+		return -EADDRINUSE;
+
 inject:
 	mutex_lock(&einj_mutex);
 	rc = __einj_error_inject(type, flags, param1, param2, param3, param4);
-- 
2.35.1

Re: [PATCH] ACPI, APEI, EINJ: Refuse to inject into the zero page

[Rafael J. Wysocki]

On Tue, Apr 19, 2022 at 11:19 PM Tony Luck <tony.luck@intel.com> wrote:
>
> Some validation tests dynamically inject errors into memory used by
> applications to check that the system can recover from a variety of
> poison consumption sceenarios.
>
> But sometimes the virtual address picked by these tests is mapped to
> the zero page.
>
> This causes additional unexpected machine checks as other processes that
> map the zero page also consume the poison.
>
> Disallow injection to the zero page.
>
> Signed-off-by: Tony Luck <tony.luck@intel.com>
>
> ---
> I picked -EADDRINUSE as a somewhat descriptive error code distinct
> fromm -EINVAL used elsewhere in EINJ, but I'm not strongly attached
> to it. Any other non -EINVAL value would be as good an indicator to
> the validation tests that they shouldn't inject to this address.
> ---
>  drivers/acpi/apei/einj.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj.c
> index 95cc2a9f3e05..d4326ec12d29 100644
> --- a/drivers/acpi/apei/einj.c
> +++ b/drivers/acpi/apei/einj.c
> @@ -549,6 +549,9 @@ static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2,
>              !arch_is_platform_page(base_addr)))
>                 return -EINVAL;
>
> +       if (is_zero_pfn(base_addr >> PAGE_SHIFT))
> +               return -EADDRINUSE;
> +
>  inject:
>         mutex_lock(&einj_mutex);
>         rc = __einj_error_inject(type, flags, param1, param2, param3, param4);
> --

Applied as 5.19 material, thanks!