On 31/07/15 12:57, Zoltan Kiss wrote: >> >> Another thing, that I just thought about: >> Right now we invoke ixgbe_rxq_rearm() only at the start of >> _recv_raw_pkts_vec(). >> Before it was ok, as _recv_raw_pkts_vec() would never try to read more >> then 32 RXDs. >> But what would happen if nb_pkts > rxq->nb_desc and rxq->rxrearm_nb == 0? > Yes, that call would deplete the RX ring, the card wouldn't be able to > receive more, so the receive function wouldn't be called again to rearm > the ring. > Actually not, the problem is that the recv function would probably overran the the descriptor ring. But anyway, we should limit nb_pkts indeed.
- [dpdk-dev] [PATCH v1] ixgbe: remove vector pmd burst s... Cunming Liang
- [dpdk-dev] [PATCH v1] ixgbe: remove vector pmd bu... Ananyev, Konstantin
- [dpdk-dev] [PATCH v1] ixgbe: remove vector pmd bu... Zoltan Kiss
- [dpdk-dev] [PATCH v1] ixgbe: remove vector pm... Ananyev, Konstantin
- [dpdk-dev] [PATCH v1] ixgbe: remove vecto... Zoltan Kiss
- [dpdk-dev] [PATCH v1] ixgbe: remove v... Zoltan Kiss
- [dpdk-dev] [PATCH v1] ixgbe: remove vecto... Liang, Cunming
- [dpdk-dev] [PATCH v1] ixgbe: remove v... Liang, Cunming
- [dpdk-dev] [PATCH v1] ixgbe: remove vector pm... Liang, Cunming
- [dpdk-dev] [PATCH v2] ixgbe: remove vector pmd bu... Cunming Liang
- [dpdk-dev] [PATCH v2] ixgbe: remove vector pm... Zoltan Kiss
- [dpdk-dev] [PATCH v2] ixgbe: remove vecto... Liang, Cunming
- [dpdk-dev] [PATCH v3] ixgbe: remove vector pm... Cunming Liang
- [dpdk-dev] [PATCH v3] ixgbe: remove vecto... Zoltan Kiss
- [dpdk-dev] [PATCH v3] ixgbe: remove v... Liang, Cunming
- [dpdk-dev] [PATCH v3] ixgbe: rem... Zoltan Kiss