Applied, but...
> --- /dev/null > +++ b/drivers/infiniband/hw/nes/nes_mgt.c > +void nes_replenish_mgt_rq(struct nes_vnic_mgt *mgtvnic) Most of these functions are only used in nes_mgt.c and can be made static. I fixed that up. > + barrier(); barrier() is only a compiler optimization barrier. Does this need to be a full memory barrier to make sure that the CPU and NES see the same thing? (ie wmb() or something?) > + u32tmp = (fpdu_info->frags[1].frag_len << 16) | > fpdu_info->frags[0].frag_len; > + set_wqe_32bit_value(cqp_wqe->wqe_words, > NES_NIC_SQ_WQE_LENGTH_2_1_IDX, > + cpu_to_le32(u32tmp)); set_wqe_32bit_value does the cpu_to_le32() internally -- I assume this will swap twice on big-endian (which I'm sure you never tested). I got rid of this superfluous call here. Please retest my for-next tree and make sure I didn't break anything in my cleanups. - R. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html