NACK: [SRU][F:linux-bluefield][PATCH] UBUNTU: SAUCE: net/sched: Increase support of reclassification to 10

Tim Gardner tim.gardner at canonical.com
Fri May 7 12:57:45 UTC 2021


I hate to be nit picky, but the subject line reclassify loop limit 
doesn't match the value in the patch.

What kind of performance impact will this incur ? Are unclassifiable 
packets common ?

On 5/5/21 10:43 AM, Daniel Jurgens wrote:
> From: Roi Dayan <roid at nvidia.com>
> 
> BugLink: https://bugs.launchpad.net/bugs/1927257
> 
> To support packet reclassification of more than 4 chains
> increase max_reclassify_loop from 4 to 15.
> 
> Change-Id: Ie18c37fb0255a737dc8f1b2fcf10a1163266dde6
> Signed-off-by: Roi Dayan <roid at nvidia.com>
> Reviewed-by: Oz Shlomo <ozsh at nvidia.com>
> Signed-off-by: Daniel Jurgens <danielj at nvidia.com>
> ---
>   net/sched/cls_api.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
> index b6e0e15..9c9afc0 100644
> --- a/net/sched/cls_api.c
> +++ b/net/sched/cls_api.c
> @@ -1582,7 +1582,7 @@ static inline int __tcf_classify(struct sk_buff *skb,
>   				 u32 *last_executed_chain)
>   {
>   #ifdef CONFIG_NET_CLS_ACT
> -	const int max_reclassify_loop = 4;
> +	const int max_reclassify_loop = 15;
>   	const struct tcf_proto *first_tp;
>   	int limit = 0;
>   
> 

-- 
-----------
Tim Gardner
Canonical, Inc



More information about the kernel-team mailing list