Re: [PATCH net-next v9 7/9] libeth: add Rx buffer management

From: Alexander Lobakin
Date: Wed Apr 10 2024 - 07:54:05 EST


From: Przemek Kitszel <przemyslaw.kitszel@xxxxxxxxx>
Date: Tue, 9 Apr 2024 12:58:33 +0200

> On 4/8/24 11:09, Alexander Lobakin wrote:
>> From: Przemek Kitszel <przemyslaw.kitszel@xxxxxxxxx>
>> Date: Fri, 5 Apr 2024 12:32:55 +0200
>>
>>> On 4/4/24 17:44, Alexander Lobakin wrote:
>>>> Add a couple intuitive helpers to hide Rx buffer implementation details
>>
>> [...]
>>
>>>> +struct libeth_fqe {
>>>> +    struct page        *page;
>>>> +    u32            offset;
>>>> +    u32            truesize;
>>>> +} __aligned_largest;
>>>> +
>>>> +/**
>>>> + * struct libeth_fq - structure representing a buffer queue
>>>> + * @fp: hotpath part of the structure
>>>> + * @pp: &page_pool for buffer management
>>>> + * @fqes: array of Rx buffers
>>>> + * @truesize: size to allocate per buffer, w/overhead
>>>> + * @count: number of descriptors/buffers the queue has
>>>> + * @buf_len: HW-writeable length per each buffer
>>>> + * @nid: ID of the closest NUMA node with memory
>>>> + */
>>>> +struct libeth_fq {
>>>> +    struct_group_tagged(libeth_fq_fp, fp,
>>>> +        struct page_pool    *pp;
>>>> +        struct libeth_fqe    *fqes;
>>>> +
>>>> +        u32            truesize;
>>>> +        u32            count;
>>>> +    );
>>>> +
>>>> +    /* Cold fields */
>>>> +    u32            buf_len;
>>>> +    int            nid;
>>>> +};
>>>
>>> [...]
>>>
>>> Could you please unpack the meaning of `fq` and `fqe` acronyms here?
>>
>> Rx:
>>
>> RQ -- receive queue, on which you get Rx DMA complete descriptors
>> FQ -- fill queue, the one you fill with free buffers
>>    FQE -- fill queue element, i.e. smth like "iavf_rx_buffer" or whatever
>>
>> Tx:
>>
>> SQ -- send queue, the one you fill with buffers to transmit
>>    SQE -- send queue element, i.e. "iavf_tx_buffer"
>> CQ -- completion queue, on which you get Tx DMA complete descriptors
>>
>> XDPSQ, XSkRQ etc. -- same as above, but for XDP / XSk
>>
>> I know that rxq, txq, bufq, complq is more common since it's been used
>> for years, but I like these "new" ones more :>
>>
>
> Thank you, that sounds right. If you happen to sent v10, a bit of code
> comment with this info would be useful ;)

The current kdoc in front of each struct and function declaration is not
enough? :D

Thanks,
Olek