On 2025/1/21 2:16, Stephen Hemminger wrote:
> On Mon, 20 Jan 2025 19:14:48 +0800
> "WanRenyong" <wa...@yunsilicon.com> wrote:
>
>> +struct __rte_packed_begin xsc_send_wqe_ctrl_seg {
>> + rte_le32_t msg_opcode:8;
>> + rte_le32_t with_immdt:1;
>> + rte_le32_t csum_en:2;
>> + rte_le32_t ds_data_num:5;
>> + rte_le32_t wqe_id:16;
>> + rte_le32_t msg_len;
>> + union __rte_packed_begin {
>> + rte_le32_t opcode_data;
>> + struct __rte_packed_begin {
>> + rte_le16_t has_pph:1;
>> + rte_le16_t so_type:1;
>> + rte_le16_t so_data_size:14;
>> + rte_le16_t rsv1:8;
>> + rte_le16_t so_hdr_len:8;
>> + } __rte_packed_end;
>> + struct __rte_packed_begin {
>> + rte_le16_t desc_id;
>> + rte_le16_t is_last_wqe:1;
>> + rte_le16_t dst_qp_id:15;
>> + } __rte_packed_end;
>> + } __rte_packed_end;
>> + rte_le32_t se:1;
> If the whole structure is marked with packed attribute,
> don't think you also need to mark the sub structures and unions
> as packed.
>
> This is what is confusing the checkpatch check for __rte_packed_begin /
> __rte_packed_end
Hello, Stephen Hemminger,
Thanks for your review.
Initially, we did not mark sub structures and unions with the packed
attribute. However, when we encountered checkpatch errors due to
unmatched|__rte_packed_begin|/|__rte_packed_end|pairs,
we referred to other PMD codes in DPDK and marked the sub structures and
unions with the packed attribute as well. Unfortunately, this did not
resolve the checkpatch errors.
I think that might be a bug in the|checkpatch|tool, so I have submitted
a patch to address this
issue:https://inbox.dpdk.org/dev/20250120112654.1049456-1-wa...@yunsilicon.com/.
Regardless, I will remove the packed attribute from sub structures and
unions in the next verion.
--
Best regards,
WanRenyong
