Re: [PATCH v2 2/3] net: stmmac: fix l3l4 filter rejecting unsupported offload requests

["Nazle Asmade, Muhammad Nazim Amirul" <muhammad.nazim.amirul.nazle.asmade@altera.com>]

On 10/6/2026 8:43 am, Jakub Kicinski wrote:
> On Fri,  5 Jun 2026 02:01:13 -0700
> muhammad.nazim.amirul.nazle.asmade@altera.com wrote:
>> +    /* Both network proto and transport proto not present in the key */
>> +    if (!match.mask || !(match.mask->n_proto || match.mask->ip_proto))
>> +            return -EOPNOTSUPP;
>> +
>> +    /* If the proto is present in the key and is not full mask */
>> +    if ((match.mask->n_proto && match.mask->n_proto != ETHER_TYPE_FULL_MASK) ||
>> +        (match.mask->ip_proto && match.mask->ip_proto != IP_PROTO_FULL_MASK))
>> +            return -EOPNOTSUPP;
>> +
>> +    /* Network proto is present in the key and is not IPv4 */
>> +    if (match.mask->n_proto && match.key->n_proto != cpu_to_be16(ETH_P_IP))
>> +            return -EOPNOTSUPP;
>> +
>> +    /* Transport proto is present in the key and is not TCP or UDP */
>> +    if (match.mask->ip_proto &&
>> +        match.key->ip_proto != IPPROTO_TCP &&
>> +        match.key->ip_proto != IPPROTO_UDP)
>> +            return -EOPNOTSUPP;
>
> Please add extack messages to let user know which part of the match is
> unsupported. Extack pointer is somewhere inside struct flow_cls_offload
>
> FWIW Sashiko points out a bunch of other potential issues, not sure if
> they matter
> https://sashiko.dev/#/patchset/20260605090114.16028-2-muhammad.nazim.amirul.nazle.asmade@altera.com
Hi Jakub,
Thank you for the review! After going through the Shashiko review, there
are some that valid and related. It was quite helpful! Here are the
changes in v3 of patch 2/3:
1. Added extack messages to all four -EOPNOTSUPP return paths in
    tc_add_basic_flow() so users know exactly which part of the match
    is unsupported.
2. Fixed a bug in v2: using break when -EOPNOTSUPP was returned
    silently discarded the error on FLOW_CLS_REPLACE - entry->in_use
    is already true on replace, so tc_add_flow() would fall through
    and return 0 (success) for an unsupported rule. Fixed by returning
    ret directly instead.
Patches 1/3 and 3/3 are unchanged.

Posted patch set with the updates above.

https://lore.kernel.org/all/20260616042655.7782-1-muhammad.nazim.amirul.nazle.asmade@altera.com/


BR,
Nazim