From: Jake Hillion <jake@hillion.co.uk>

Currently hsmp_send_message uses down_timeout with a 100ms timeout to
take the semaphore. However __hsmp_send_message, the content of the
critical section, has a sleep in it. On systems with significantly
delayed scheduling behaviour this may take over 100ms.

Convert this method to down_interruptible. Leave the error handling the
same as the documentation currently is not specific about what error is
returned.

Previous behaviour: a caller who competes with another caller stuck in
the critical section due to scheduler delays would receive -ETIMEDOUT.

New behaviour: a caller who competes with another caller stuck in the
critical section due to scheduler delays will complete successfully.

Signed-off-by: Jake Hillion <jake@hillion.co.uk>
---
 drivers/platform/x86/amd/hsmp/hsmp.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/drivers/platform/x86/amd/hsmp/hsmp.c b/drivers/platform/x86/amd/hsmp/hsmp.c
index f35c639457ac425e79dead2515c0eddea0759323..6c30bb3edc1d77939b10047b771a5c574e5f2a1e 100644
--- a/drivers/platform/x86/amd/hsmp/hsmp.c
+++ b/drivers/platform/x86/amd/hsmp/hsmp.c
@@ -216,13 +216,7 @@ int hsmp_send_message(struct hsmp_message *msg)
 		return -ENODEV;
 	sock = &hsmp_pdev.sock[msg->sock_ind];
 
-	/*
-	 * The time taken by smu operation to complete is between
-	 * 10us to 1ms. Sometime it may take more time.
-	 * In SMP system timeout of 100 millisecs should
-	 * be enough for the previous thread to finish the operation
-	 */
-	ret = down_timeout(&sock->hsmp_sem, msecs_to_jiffies(HSMP_MSG_TIMEOUT));
+	ret = down_interruptible(&sock->hsmp_sem);
 	if (ret < 0)
 		return ret;
 

-- 
2.47.2

Hi,


On 5/30/2025 9:45 PM, Jake Hillion via B4 Relay wrote:
> Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
>
>
> From: Jake Hillion <jake@hillion.co.uk>
>
> Currently hsmp_send_message uses down_timeout with a 100ms timeout to
> take the semaphore. However __hsmp_send_message, the content of the
> critical section, has a sleep in it. On systems with significantly
> delayed scheduling behaviour this may take over 100ms.
>
> Convert this method to down_interruptible. Leave the error handling the
> same as the documentation currently is not specific about what error is
> returned.
>
> Previous behaviour: a caller who competes with another caller stuck in
> the critical section due to scheduler delays would receive -ETIMEDOUT.


down_timeout() will return -ETIME,  Please replace -ETIMEDOUT with -ETIME.

>
> New behaviour: a caller who competes with another caller stuck in the
> critical section due to scheduler delays will complete successfully.
>
> Signed-off-by: Jake Hillion <jake@hillion.co.uk>


Reviewed-by: Suma Hegde <suma.hegde@amd.com>
Tested-by: Suma Hegde <suma.hegde@amd.com>


> ---
>   drivers/platform/x86/amd/hsmp/hsmp.c | 8 +-------
>   1 file changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/drivers/platform/x86/amd/hsmp/hsmp.c b/drivers/platform/x86/amd/hsmp/hsmp.c
> index f35c639457ac425e79dead2515c0eddea0759323..6c30bb3edc1d77939b10047b771a5c574e5f2a1e 100644
> --- a/drivers/platform/x86/amd/hsmp/hsmp.c
> +++ b/drivers/platform/x86/amd/hsmp/hsmp.c
> @@ -216,13 +216,7 @@ int hsmp_send_message(struct hsmp_message *msg)
>                  return -ENODEV;
>          sock = &hsmp_pdev.sock[msg->sock_ind];
>
> -       /*
> -        * The time taken by smu operation to complete is between
> -        * 10us to 1ms. Sometime it may take more time.
> -        * In SMP system timeout of 100 millisecs should
> -        * be enough for the previous thread to finish the operation
> -        */
> -       ret = down_timeout(&sock->hsmp_sem, msecs_to_jiffies(HSMP_MSG_TIMEOUT));
> +       ret = down_interruptible(&sock->hsmp_sem);
>          if (ret < 0)
>                  return ret;
>
>
> --
> 2.47.2

Thanks and Regards,

Suma

>
>