On 3/6/2026 3:03 PM, Burakov, Anatoly wrote:
CodeRabbit picked up on something here, and I think it's worth
highlighting.
When we're splitting segments, we assign txe->mbuf to the first
segment...
<snip>
+ txe = &sw_ring[sw_id];
+ /* sub-descriptor slots do not own the mbuf */
+ txe->mbuf = NULL;
...then set subsequent segments to NULL...
+ }
- /* Setup TX descriptor */
- txd->buf_addr =
- rte_cpu_to_le_64(rte_mbuf_data_iova(tx_pkt));
- cmd_dtype |= IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE;
+ /* Write the final (or only) descriptor for this segment */
+ txd = &txr[tx_id];
+ txd->buf_addr = rte_cpu_to_le_64(buf_dma_addr);
txd->qw1.cmd_dtype = cmd_dtype;
- txd->qw1.rxr_bufsize = tx_pkt->data_len;
+ txd->qw1.rxr_bufsize = slen;
txd->qw1.compl_tag = sw_id;
...and we're supposed to write the final descriptor here, but we've
stored the mbuf pointer in the *first* descriptor, not in the *last*
one, which means when this descriptor gets to processing completions,
the mbuf pointer of that descriptor will be NULL? Is that intended?
Actually, digging in, I don't see where we free mbufs at all in splitq
path? Am I missing something here?
Yes I am, RS bit is set by the hardware. I guess the question is then,
does it set RS bits for *all* segments, or just the ones we marked with
EOP? If it's the latter then it's definitely a bug. If all segments get
RS bit set, then it's not.
--
Thanks,
Anatoly
