Re: [dpdk-dev] [PATCH v4 4/5] compress/zlib: support burst enqueue/dequeue

["Verma, Shally" <Shally.Verma@cavium.com>]

>-----Original Message-----
>From: De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>
>Sent: 24 July 2018 03:55
>To: Verma, Shally <Shally.Verma@cavium.com>
>Cc: dev@dpdk.org; Athreya, Narayana Prasad <NarayanaPrasad.Athreya@cavium.com>; Challa, Mahipal
><Mahipal.Challa@cavium.com>; Sahu, Sunila <Sunila.Sahu@cavium.com>; Sahu, Sunila <Sunila.Sahu@cavium.com>; Gupta, Ashish
><Ashish.Gupta@cavium.com>
>Subject: RE: [PATCH v4 4/5] compress/zlib: support burst enqueue/dequeue
>
>External Email
>
>> -----Original Message-----
>> From: Shally Verma [mailto:shally.verma@caviumnetworks.com]
>> Sent: Monday, July 23, 2018 3:51 PM
>> To: De Lara Guarch, Pablo <pablo.de.lara.guarch@intel.com>
>> Cc: dev@dpdk.org; pathreya@caviumnetworks.com;
>> mchalla@caviumnetworks.com; Sunila Sahu <ssahu@caviumnetworks.com>;
>> Sunila Sahu <sunila.sahu@caviumnetworks.com>; Ashish Gupta
>> <ashish.gupta@caviumnetworks.com>
>> Subject: [PATCH v4 4/5] compress/zlib: support burst enqueue/dequeue
>>
>> From: Sunila Sahu <ssahu@caviumnetworks.com>
>>
>> Signed-off-by: Sunila Sahu <sunila.sahu@caviumnetworks.com>
>> Signed-off-by: Shally Verma <shally.verma@caviumnetworks.com>
>> Signed-off-by: Ashish Gupta <ashish.gupta@caviumnetworks.com>
>> ---
>>  drivers/compress/zlib/zlib_pmd.c | 255
>> ++++++++++++++++++++++++++++++++++++++-
>>  1 file changed, 254 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/compress/zlib/zlib_pmd.c b/drivers/compress/zlib/zlib_pmd.c
>> index 47bc73d..dc1e230 100644
>> --- a/drivers/compress/zlib/zlib_pmd.c
>> +++ b/drivers/compress/zlib/zlib_pmd.c
>> @@ -7,7 +7,214 @@
>>
>>  #include "zlib_pmd_private.h"
>>
>> -/** Parse comp xform and set private xform/stream parameters */
>> +/** Compute next mbuf in the list, assign data buffer and length,
>> + *  returns 0 if mbuf is NULL
>> + */
>> +#define COMPUTE_BUF(mbuf, data, len)         \
>> +             ((mbuf = mbuf->next) ?          \
>> +             (data = rte_pktmbuf_mtod(mbuf, uint8_t *)),     \
>> +             (len = rte_pktmbuf_data_len(mbuf)) : 0)
>> +
>> +static void
>> +process_zlib_deflate(struct rte_comp_op *op, z_stream *strm) {
>
>...
>
>> +     /* Update z_stream with the inputs provided by application */
>> +     strm->next_in = rte_pktmbuf_mtod_offset(mbuf_src, uint8_t *,
>> +                     op->src.offset);
>
>This is assuming that src buffer is a linear buffer or that offset won't be large enough to cross boundaries between segments.
>If an SGL is passed and offset is bigger than the first segment, next_in should point at a different segment, with the remaining part of
>the offset in that segment
>(look at ISA-L SGL patch: http://patches.dpdk.org/patch/43283/). Same applies to avail_in, next_out and avail_out.

[Shally] as per my last knowledge, offset was expected to be belonging only to the first segment in chained mbuf. Isn't that the case anymore? Did I miss any update on its definition?
We had added the logic earlier that you're suggesting but removed that later, as I understood clarification about offset falling into any segment is still pending.

Thanks
Shally

>
>> +
>> +     strm->avail_in = rte_pktmbuf_data_len(mbuf_src) - op->src.offset;
>> +
>> +     strm->next_out = rte_pktmbuf_mtod_offset(mbuf_dst, uint8_t *,
>> +                     op->dst.offset);
>> +
>> +     strm->avail_out = rte_pktmbuf_data_len(mbuf_dst) - op->dst.offset;
>> +
>
>...
>
>> +     strm->next_in = rte_pktmbuf_mtod_offset(mbuf_src, uint8_t *,
>> +                     op->src.offset);
>> +
>> +     strm->avail_in = rte_pktmbuf_data_len(mbuf_src) - op->src.offset;
>> +
>> +     strm->next_out = rte_pktmbuf_mtod_offset(mbuf_dst, uint8_t *,
>> +                     op->dst.offset);
>> +
>> +     strm->avail_out = rte_pktmbuf_data_len(mbuf_dst) - op->dst.offset;
>
>Same comments as above (compression).
>
>...
>
>> +static inline int
>> +process_zlib_op(struct zlib_qp *qp, struct rte_comp_op *op) {
>> +     struct zlib_stream *stream;
>> +     struct zlib_priv_xform *private_xform;
>> +
>> +     if ((op->op_type == RTE_COMP_OP_STATEFUL) ||
>> +         (op->src.offset > rte_pktmbuf_data_len(op->m_src)) ||
>> +         (op->dst.offset > rte_pktmbuf_data_len(op->m_dst))) {
>
>Since m_src and m_dst could be SGLs, pkt_len should be checked, instead of data_len (which would be only for single segment).
>Also, you should check the length too, in case of source buffers (src.offset + src.length > m_src->pkt_len).
>
>Lastly, the two lines after the first if line should have double indentation to distinguish clearly against the body of the if.
>If line is too long, consider storing the length of the buffers in variables.
>
>
>> +             op->status = RTE_COMP_OP_STATUS_INVALID_ARGS;
>> +             ZLIB_PMD_ERR("Invalid source or destination buffers or "
>> +                          "invalid Operation requested\n");