In most (likely all) cases, INIT_HAL_MSG() is called before
PREPARE_HAL_BUF().
In such cases calling memset() is useless because:
msg_body.header.len = sizeof(msg_body)
So, instead of writing twice the memory, we just have a sanity check to
make sure that some potential trailing memory is zeroed.
It even gives the opportunity to see that by itself and optimize it away.
Signed-off-by: Christophe JAILLET <[email protected]>
---
drivers/net/wireless/ath/wcn36xx/smd.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/ath/wcn36xx/smd.c b/drivers/net/wireless/ath/wcn36xx/smd.c
index 566f0b9c1584..17e1919d1cd8 100644
--- a/drivers/net/wireless/ath/wcn36xx/smd.c
+++ b/drivers/net/wireless/ath/wcn36xx/smd.c
@@ -475,8 +475,8 @@ static int wcn36xx_smd_send_and_wait(struct wcn36xx *wcn, size_t len)
#define PREPARE_HAL_BUF(send_buf, msg_body) \
do { \
- memset(send_buf, 0, msg_body.header.len); \
- memcpy(send_buf, &msg_body, sizeof(msg_body)); \
+ memcpy_and_pad(send_buf, msg_body.header.len, \
+ &msg_body, sizeof(msg_body), 0); \
} while (0) \
#define PREPARE_HAL_PTT_MSG_BUF(send_buf, p_msg_body) \
--
2.34.1
On Thu, 2 Feb 2023 at 21:58, Christophe JAILLET
<[email protected]> wrote:
>
> In most (likely all) cases, INIT_HAL_MSG() is called before
> PREPARE_HAL_BUF().
> In such cases calling memset() is useless because:
> msg_body.header.len = sizeof(msg_body)
>
> So, instead of writing twice the memory, we just have a sanity check to
> make sure that some potential trailing memory is zeroed.
> It even gives the opportunity to see that by itself and optimize it away.
>
> Signed-off-by: Christophe JAILLET <[email protected]>
Acked-by: Loic Poulain <[email protected]>
> ---
> drivers/net/wireless/ath/wcn36xx/smd.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/wcn36xx/smd.c b/drivers/net/wireless/ath/wcn36xx/smd.c
> index 566f0b9c1584..17e1919d1cd8 100644
> --- a/drivers/net/wireless/ath/wcn36xx/smd.c
> +++ b/drivers/net/wireless/ath/wcn36xx/smd.c
> @@ -475,8 +475,8 @@ static int wcn36xx_smd_send_and_wait(struct wcn36xx *wcn, size_t len)
>
> #define PREPARE_HAL_BUF(send_buf, msg_body) \
> do { \
> - memset(send_buf, 0, msg_body.header.len); \
> - memcpy(send_buf, &msg_body, sizeof(msg_body)); \
> + memcpy_and_pad(send_buf, msg_body.header.len, \
> + &msg_body, sizeof(msg_body), 0); \
> } while (0) \
>
> #define PREPARE_HAL_PTT_MSG_BUF(send_buf, p_msg_body) \
> --
> 2.34.1
>
Christophe JAILLET <[email protected]> wrote:
> In most (likely all) cases, INIT_HAL_MSG() is called before
> PREPARE_HAL_BUF().
> In such cases calling memset() is useless because:
> msg_body.header.len = sizeof(msg_body)
>
> So, instead of writing twice the memory, we just have a sanity check to
> make sure that some potential trailing memory is zeroed.
> It even gives the opportunity to see that by itself and optimize it away.
>
> Signed-off-by: Christophe JAILLET <[email protected]>
> Acked-by: Loic Poulain <[email protected]>
> Signed-off-by: Kalle Valo <[email protected]>
Patch applied to ath-next branch of ath.git, thanks.
4a51e66fe96d wifi: wcn36xx: Slightly optimize PREPARE_HAL_BUF()
--
https://patchwork.kernel.org/project/linux-wireless/patch/7d8ab7fee45222cdbaf80c507525f2d3941587c1.1675371372.git.christophe.jaillet@wanadoo.fr/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches