Re: [PATCH bpf-next v6 5/7] bpf: Support specifying linear xdp packet data size for BPF_PROG_TEST_RUN

[Martin KaFai Lau <martin.lau@linux.dev>]

On 9/22/25 12:48 PM, Amery Hung wrote:
>>> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
>>> index 4a862d605386..0cbd3b898c45 100644
>>> --- a/net/bpf/test_run.c
>>> +++ b/net/bpf/test_run.c
>>> @@ -665,7 +665,7 @@ static void *bpf_test_init(const union bpf_attr *kattr, u32 user_size,
>>>        void __user *data_in = u64_to_user_ptr(kattr->test.data_in);
>>>        void *data;
>>>
>>> -     if (user_size < ETH_HLEN || user_size > PAGE_SIZE - headroom - tailroom)
>>> +     if (user_size > PAGE_SIZE - headroom - tailroom)
>>>                return ERR_PTR(-EINVAL);
>>>
>>>        size = SKB_DATA_ALIGN(size);
>>> @@ -1001,6 +1001,9 @@ int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
>>>            kattr->test.cpu || kattr->test.batch_size)
>>>                return -EINVAL;
>>>
>>> +     if (size < ETH_HLEN)
>>> +             return -EINVAL;
>>> +
>>>        data = bpf_test_init(kattr, kattr->test.data_size_in,
>>>                             size, NET_SKB_PAD + NET_IP_ALIGN,
>>>                             SKB_DATA_ALIGN(sizeof(struct skb_shared_info)));
>>> @@ -1246,13 +1249,15 @@ int bpf_prog_test_run_xdp(struct bpf_prog *prog, const union bpf_attr *kattr,
>>
>> I just noticed it. It still needs a "size < ETH_HLEN" test at the beginning of
>> test_run_xdp. At least the do_live mode should still needs to have ETH_HLEN bytes.
> 
> Make sense. I will add the check for live mode.

The earlier comment wasn't clear, my bad. no need to limit the ETH_HLEN test to 
live mode only. multi-frags or not, kattr->test.data_size_in should not be < 
ETH_HLEN.