Re: [PATCH v2] iio: proximity: hx9023s: validate firmware size

[Jonathan Cameron]

On Fri, 10 Jul 2026 21:13:42 +0545
Laxman Acharya Padhya <acharyalaxman8848@xxxxxxxxx> wrote:

> hx9023s_send_cfg() copies the firmware into a counted flexible array and
> then reads fixed offsets from the copied data before walking register/value
> pairs starting at FW_DATA_OFFSET. A truncated firmware image can therefore
> make the driver read past the copied buffer during probe-time configuration
> loading.
> 
> Reject firmware images that cannot contain the fixed header, reject images
> too large for the u16 fw_size field, and validate that the advertised
> register count fits in the remaining payload.
> 
> Fixes: e9ed97be4fcc ("iio: proximity: hx9023s: Added firmware file parsing functionality")
> Cc: stable@xxxxxxxxxxxxxxx
> Reviewed-by: Joshua Crofts <joshua.crofts1@xxxxxxxxx>
> Signed-off-by: Laxman Acharya Padhya <acharyalaxman8848@xxxxxxxxx>
A few minor comments.

Thanks,

Jonathan

> ---
>  drivers/iio/proximity/hx9023s.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/iio/proximity/hx9023s.c b/drivers/iio/proximity/hx9023s.c
> index a6ff7cbe9e6..9d91ce681ac 100644
> --- a/drivers/iio/proximity/hx9023s.c
> +++ b/drivers/iio/proximity/hx9023s.c
> @@ -18,6 +18,7 @@
>  #include <linux/i2c.h>
>  #include <linux/interrupt.h>
>  #include <linux/irqreturn.h>
> +#include <linux/limits.h>
>  #include <linux/math64.h>
>  #include <linux/module.h>
>  #include <linux/mutex.h>
> @@ -1031,8 +1032,11 @@ static int hx9023s_bin_load(struct hx9023s_data *data, struct hx9023s_bin *bin)
>  
>  static int hx9023s_send_cfg(const struct firmware *fw, struct hx9023s_data *data)
>  {
> +	if (fw->size < FW_DATA_OFFSET || fw->size > U16_MAX)

Add a comment on why you've picked that upper limit.

> +		return -EINVAL;
> +
>  	struct hx9023s_bin *bin __free(kfree) =
> -		kzalloc(fw->size + sizeof(*bin), GFP_KERNEL);
> +		kzalloc(sizeof(*bin) + fw->size, GFP_KERNEL);

This doesn't belong in the fix given it is just a reorder. Maybe it makes sense
but if it does, separate patch with an explanation of why.

>  	if (!bin)
>  		return -ENOMEM;
>  
> @@ -1041,7 +1045,8 @@ static int hx9023s_send_cfg(const struct firmware *fw, struct hx9023s_data *data
>  	bin->fw_ver = bin->data[FW_VER_OFFSET];
>  	bin->reg_count = get_unaligned_le16(bin->data + FW_REG_CNT_OFFSET);
>  
> -	release_firmware(fw);
> +	if (bin->reg_count > (bin->fw_size - FW_DATA_OFFSET) / 2)
> +		return -EINVAL;
>  
>  	return hx9023s_bin_load(data, bin);
>  }
> @@ -1058,6 +1063,7 @@ static void hx9023s_cfg_update(const struct firmware *fw, void *context)
>  	}
>  
>  	ret = hx9023s_send_cfg(fw, data);
> +	release_firmware(fw);
>  	if (ret) {
>  		dev_warn(dev, "Firmware update failed: %d\n", ret);
>  		goto no_fw;