[lvc-project] [PATCH] wifi: ath9k: cleanup ath9k_hw_get_nf_hist_mid()

Toke Høiland-Jørgensen toke at toke.dk
Thu Jan 2 18:58:51 MSK 2025


Dmitry Antipov <dmantipov at yandex.ru> writes:

> In 'ath9k_hw_get_nf_hist_mid()', prefer 'memcpy()' and 'sort()'
> over an ad-hoc things. Briefly tested as a separate module.
>
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
>
> Signed-off-by: Dmitry Antipov <dmantipov at yandex.ru>

Thank you for the cleanup; a few nits, see below:

> ---
>  drivers/net/wireless/ath/ath9k/calib.c | 24 ++++++++++--------------
>  1 file changed, 10 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath9k/calib.c b/drivers/net/wireless/ath/ath9k/calib.c
> index 4b331c85509c..26465c1a59e2 100644
> --- a/drivers/net/wireless/ath/ath9k/calib.c
> +++ b/drivers/net/wireless/ath/ath9k/calib.c
> @@ -16,29 +16,25 @@
>  
>  #include "hw.h"
>  #include "hw-ops.h"
> +#include <linux/sort.h>
>  #include <linux/export.h>
>  
>  /* Common calibration code */
>  
> +static inline int rcmp_i16(const void *x, const void *y)

We generally don't mark static functions as inline, but instead let the
compiler make the decision (which in this case is to *not* inline, as
the function address needs to be passed to sort()).

> +{
> +	/* Sort in reverse order. */
> +	return *(int16_t *)y - *(int16_t *)x;
> +}
>  
>  static int16_t ath9k_hw_get_nf_hist_mid(int16_t *nfCalBuffer)
>  {
> -	int16_t nfval;
> -	int16_t sort[ATH9K_NF_CAL_HIST_MAX];
> -	int i, j;
> -
> -	for (i = 0; i < ATH9K_NF_CAL_HIST_MAX; i++)
> -		sort[i] = nfCalBuffer[i];
> +	int16_t nfcal[ATH9K_NF_CAL_HIST_MAX];
>  
> -	for (i = 0; i < ATH9K_NF_CAL_HIST_MAX - 1; i++) {
> -		for (j = 1; j < ATH9K_NF_CAL_HIST_MAX - i; j++) {
> -			if (sort[j] > sort[j - 1])
> -				swap(sort[j], sort[j - 1]);
> -		}
> -	}
> -	nfval = sort[(ATH9K_NF_CAL_HIST_MAX - 1) >> 1];
> +	memcpy(nfcal, nfCalBuffer, ATH9K_NF_CAL_HIST_MAX * sizeof(int16_t));

The third argument to memcpy() can just be sizeof(nfcal).


-Toke



More information about the lvc-project mailing list