Re: [PATCH v2] HID: wacom: avoid overflows in sequence number tracking

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] hid: wacom: 64-bit modulo operation breaks 32-bit builds
--

commit 07dd42ab3d43fb513fdcf25e5070f8dd85e2b11c
Author: Aldo Ariel Panzardo <qwe.aldo@gmail.com>

HID: wacom: avoid overflows in sequence number tracking

This commit updates the dropped-packet calculation to use signed 64-bit 
arithmetic to prevent divide exceptions caused by 32-bit wrap-arounds. It also 
normalizes the sequence values according to the HID parser's signedness rules.

Fixes: 359673ea3a20 ("HID: wacom: Support sequence numbers smaller than 16-bit")
Cc: stable@vger.kernel.org

> diff --git a/drivers/hid/wacom_wac.c b/drivers/hid/wacom_wac.c
> index 8feb8027be95e..ff99ba259b030 100644
> --- a/drivers/hid/wacom_wac.c
> +++ b/drivers/hid/wacom_wac.c
> @@ -2542,18 +2542,34 @@ static void wacom_wac_pen_event(struct hid_device *hdev, struct hid_field *field
>  	case WACOM_HID_WD_BARRELSWITCH3:
>  		wacom_wac->hid_data.barrelswitch3 = value;
>  		return;
> -	case WACOM_HID_WD_SEQUENCENUMBER:
> -		if (wacom_wac->hid_data.sequence_number != value &&
> +	case WACOM_HID_WD_SEQUENCENUMBER: {
> +		s64 sequence_minimum = field->logical_minimum;
> +		s64 sequence_maximum = sequence_minimum < 0 ?
> +			(s64)field->logical_maximum :
> +			(s64)(u32)field->logical_maximum;
> +		s64 sequence_value = sequence_minimum < 0 ?
> +			(s64)value : (s64)(u32)value;
> +
> +		if (wacom_wac->hid_data.sequence_number != sequence_value &&
>  		    wacom_wac->hid_data.sequence_number >= 0) {
> -			int sequence_size = field->logical_maximum - field->logical_minimum + 1;
> -			int drop_count = (value - wacom_wac->hid_data.sequence_number) % sequence_size;
> -			hid_warn(hdev, "Dropped %d packets", drop_count);
> +			s64 sequence_size = sequence_maximum -
> +					    sequence_minimum + 1;
> +			s64 drop_count = sequence_value -
> +					 wacom_wac->hid_data.sequence_number;
> +
> +			drop_count %= sequence_size;

[Severity: Low]
Will this raw 64-bit modulo operation cause link-time build failures on 32-bit
architectures?

The variables drop_count and sequence_size are both s64. On 32-bit systems
(like i386 or arm32), the kernel does not link __moddi3, so using the raw
modulo operator on 64-bit integers typically results in an unresolved symbol
build failure.

Should this use a helper from <linux/math64.h>, such as div64_s64_rem(), to
safely perform the calculation while ensuring 32-bit compatibility?

> +			if (drop_count < 0)
> +				drop_count += sequence_size;
> +			hid_warn(hdev, "Dropped %lld packets",
> +				 (long long)drop_count);
>  		}

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260915141621.3011011-1-qwe.aldo@gmail.com?part=1