Re: [PATCH net-next] octeontx2-pf: flower: check for unsupported control flags

[Jacob Keller]

On 4/22/2024 8:27 AM, Asbjørn Sloth Tønnesen wrote:
> Use flow_rule_is_supp_control_flags() to reject filters with
> unsupported control flags.
> 
> In case any unsupported control flags are masked,
> flow_rule_is_supp_control_flags() sets a NL extended
> error message, and we return -EOPNOTSUPP.
> 
> Remove FLOW_DIS_FIRST_FRAG specific error message,
> and treat it as any other unsupported control flag.
> 
> Only compile-tested.
> 
> Signed-off-by: Asbjørn Sloth Tønnesen <ast@xxxxxxxxxxx>
> ---
>  drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c
> index 6d4ce2ece8d0..e63cc1eb6d89 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c
> @@ -700,10 +700,6 @@ static int otx2_tc_prepare_flow(struct otx2_nic *nic, struct otx2_tc_flow *node,
>  		u32 val;
>  
>  		flow_rule_match_control(rule, &match);
> -		if (match.mask->flags & FLOW_DIS_FIRST_FRAG) {
> -			NL_SET_ERR_MSG_MOD(extack, "HW doesn't support frag first/later");
> -			return -EOPNOTSUPP;
> -		}
>  
>  		if (match.mask->flags & FLOW_DIS_IS_FRAGMENT) {
>  			val = match.key->flags & FLOW_DIS_IS_FRAGMENT;
> @@ -721,6 +717,10 @@ static int otx2_tc_prepare_flow(struct otx2_nic *nic, struct otx2_tc_flow *node,
>  				return -EOPNOTSUPP;
>  			}
>  		}
> +
> +		if (!flow_rule_is_supp_control_flags(FLOW_DIS_IS_FRAGMENT,
> +						     match.mask->flags, extack))
> +			return -EOPNOTSUPP;

This confuses me since you pass FLOW_DIS_IS_FRAGMENT here, but you
removed the check for FLOW_DIS_FIRST_FRAG??

Am I misunderstanding how flow_rule_is_supp_control_flags works?

The code just above this appears to support FLOW_DIS_IS_FRAGMENT.

Here is the implementation of flow_rule_is_supp_control_flags for reference:

> /**
>  * flow_rule_is_supp_control_flags() - check for supported control flags
>  * @supp_flags: control flags supported by driver
>  * @ctrl_flags: control flags present in rule
>  * @extack: The netlink extended ACK for reporting errors.
>  *
>  * Return: true if only supported control flags are set, false otherwise.
>  */
> static inline bool flow_rule_is_supp_control_flags(const u32 supp_flags,
>                                                    const u32 ctrl_flags,
>                                                    struct netlink_ext_ack *extack)
> {
>         if (likely((ctrl_flags & ~supp_flags) == 0))
>                 return true;
> 
>         NL_SET_ERR_MSG_FMT_MOD(extack,
>                                "Unsupported match on control.flags %#x",
>                                ctrl_flags);
> 
>         return false;
> }
> 

This seems to me that it you accidentally passed FLOW_DIS_IS_FRAGMENT
when you meant FLOW_DIS_FIRST_FRAG??

I also think its a bit strange that you moved the placement of the check
instead of replacing in the same location as where the previous check was.


>  	}
>  
>  	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ETH_ADDRS)) {