Re: [PATCH] [v2] hinic: avoid gcc -Wrestrict warning

[Rasmus Villemoes]

On 24/03/2021 14.07, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@xxxxxxxx>
> 
> With extra warnings enabled, gcc complains that snprintf should not
> take the same buffer as source and destination:
> 
> drivers/net/ethernet/huawei/hinic/hinic_ethtool.c: In function 'hinic_set_settings_to_hw':
> drivers/net/ethernet/huawei/hinic/hinic_ethtool.c:480:9: error: 'snprintf' argument 4 overlaps destination object 'set_link_str' [-Werror=restrict]
>   480 |   err = snprintf(set_link_str, SET_LINK_STR_MAX_LEN,
>       |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   481 |           "%sspeed %d ", set_link_str, speed);
>       |           ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/net/ethernet/huawei/hinic/hinic_ethtool.c:464:7: note: destination object referenced by 'restrict'-qualified argument 1 was declared here
>   464 |  char set_link_str[SET_LINK_STR_MAX_LEN] = {0};
> 
> Rewrite this to avoid the nested sprintf and instead use separate
> buffers, which is simpler.
> 

This looks much better. Thanks.

> Cc: Rasmus Villemoes <linux@xxxxxxxxxxxxxxxxxx>
> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> ---
> v2: rework according to feedback from Rasmus. This one could
>     easily avoid most of the pitfalls
> ---
>  .../net/ethernet/huawei/hinic/hinic_ethtool.c | 25 ++++++++-----------
>  1 file changed, 10 insertions(+), 15 deletions(-)
> 
> diff --git a/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c b/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c
> index c340d9acba80..d7e20dab6e48 100644
> --- a/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c
> +++ b/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c
> @@ -34,7 +34,7 @@
>  #include "hinic_rx.h"
>  #include "hinic_dev.h"
>  
> -#define SET_LINK_STR_MAX_LEN	128
> +#define SET_LINK_STR_MAX_LEN	16
>  
>  #define GET_SUPPORTED_MODE	0
>  #define GET_ADVERTISED_MODE	1
> @@ -462,24 +462,19 @@ static int hinic_set_settings_to_hw(struct hinic_dev *nic_dev,
>  {
>  	struct hinic_link_ksettings_info settings = {0};
>  	char set_link_str[SET_LINK_STR_MAX_LEN] = {0};
> +	const char *autoneg_str;
>  	struct net_device *netdev = nic_dev->netdev;
>  	enum nic_speed_level speed_level = 0;
>  	int err;
>  
> -	err = snprintf(set_link_str, SET_LINK_STR_MAX_LEN, "%s",
> -		       (set_settings & HILINK_LINK_SET_AUTONEG) ?
> -		       (autoneg ? "autong enable " : "autong disable ") : "");
> -	if (err < 0 || err >= SET_LINK_STR_MAX_LEN) {
> -		netif_err(nic_dev, drv, netdev, "Failed to snprintf link state, function return(%d) and dest_len(%d)\n",
> -			  err, SET_LINK_STR_MAX_LEN);
> -		return -EFAULT;
> -	}
> +	autoneg_str = (set_settings & HILINK_LINK_SET_AUTONEG) ?
> +		      (autoneg ? "autong enable " : "autong disable ") : "";
>  
>  	if (set_settings & HILINK_LINK_SET_SPEED) {
>  		speed_level = hinic_ethtool_to_hw_speed_level(speed);
>  		err = snprintf(set_link_str, SET_LINK_STR_MAX_LEN,
> -			       "%sspeed %d ", set_link_str, speed);
> -		if (err <= 0 || err >= SET_LINK_STR_MAX_LEN) {
> +			       "speed %d ", speed);
> +		if (err >= SET_LINK_STR_MAX_LEN) {
>  			netif_err(nic_dev, drv, netdev, "Failed to snprintf link speed, function return(%d) and dest_len(%d)\n",
>  				  err, SET_LINK_STR_MAX_LEN);
>  			return -EFAULT;

It's not your invention of course, but this both seems needlessly harsh
and EFAULT is a weird error to return. It's just a printk() message that
might be truncated, and now that the format string only has a %d
specifier, it can actually be verified statically that overflow will
never happen (though I don't know or think gcc can do that, perhaps
there's some locale nonsense in the standard that allows using
utf16-encoded sanskrit runes). So probably that test should just be
dropped, but that's a separate thing.

Reviewed-by: Rasmus Villemoes <linux@xxxxxxxxxxxxxxxxxx>