2023-12-19 14:52:13

by Aleksa Savic

Subject: [PATCH] hwmon: (gigabyte_waterforce) Mark status report as received under a spinlock

Through hidraw, userspace can cause a status report to be sent
from the device. The parsing in waterforce_raw_event() may happen in
parallel to a waterforce_get_status() call (which resets the completion
for tracking the report) if it's running on a different CPU where
bottom half interrupts are not disabled.

Add a spinlock around the complete_all() call in waterforce_raw_event()
to prevent race issues.

Fixes: d5939a793693 ("hwmon: Add driver for Gigabyte AORUS Waterforce AIO coolers")
Signed-off-by: Aleksa Savic <[email protected]>
---
Does the above make sense? The raw event parsing is a softirq AFAIK and
presumably could run on a different CPU in parallel to a waterforce_get_status()
call.

Appreciate thoughts on this.
---
drivers/hwmon/gigabyte_waterforce.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/hwmon/gigabyte_waterforce.c b/drivers/hwmon/gigabyte_waterforce.c
index 1799377fc2f1..85e523775714 100644
--- a/drivers/hwmon/gigabyte_waterforce.c
+++ b/drivers/hwmon/gigabyte_waterforce.c
@@ -276,8 +276,10 @@ static int waterforce_raw_event(struct hid_device *hdev, struct hid_report *repo
priv->duty_input[0] = data[WATERFORCE_FAN_DUTY];
priv->duty_input[1] = data[WATERFORCE_PUMP_DUTY];

+ spin_lock(&priv->status_report_request_lock);
if (!completion_done(&priv->status_report_received))
complete_all(&priv->status_report_received);
+ spin_unlock(&priv->status_report_request_lock);

priv->updated = jiffies;

--
2.43.0

2024-01-08 15:30:12

by Guenter Roeck

Subject: Re: [PATCH] hwmon: (gigabyte_waterforce) Mark status report as received under a spinlock

On Tue, Dec 19, 2023 at 03:36:19PM +0100, Aleksa Savic wrote:
> Through hidraw, userspace can cause a status report to be sent
> from the device. The parsing in waterforce_raw_event() may happen in
> parallel to a waterforce_get_status() call (which resets the completion
> for tracking the report) if it's running on a different CPU where
> bottom half interrupts are not disabled.
>
> Add a spinlock around the complete_all() call in waterforce_raw_event()
> to prevent race issues.
>
> Fixes: d5939a793693 ("hwmon: Add driver for Gigabyte AORUS Waterforce AIO coolers")
> Signed-off-by: Aleksa Savic <[email protected]>
> ---
> Does the above make sense? The raw event parsing is a softirq AFAIK and
> presumably could run on a different CPU in parallel to a waterforce_get_status()
> call.
>
> Appreciate thoughts on this.

I think it does. Applied.

Thanks,
Guenter

> ---
> drivers/hwmon/gigabyte_waterforce.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/hwmon/gigabyte_waterforce.c b/drivers/hwmon/gigabyte_waterforce.c
> index 1799377fc2f1..85e523775714 100644
> --- a/drivers/hwmon/gigabyte_waterforce.c
> +++ b/drivers/hwmon/gigabyte_waterforce.c
> @@ -276,8 +276,10 @@ static int waterforce_raw_event(struct hid_device *hdev, struct hid_report *repo
> priv->duty_input[0] = data[WATERFORCE_FAN_DUTY];
> priv->duty_input[1] = data[WATERFORCE_PUMP_DUTY];
>
> + spin_lock(&priv->status_report_request_lock);
> if (!completion_done(&priv->status_report_received))
> complete_all(&priv->status_report_received);
> + spin_unlock(&priv->status_report_request_lock);
>
> priv->updated = jiffies;
>