Re: [PATCH] hw: Fix permissions for OCRDMA_RESET_STATS

[Selvin Xavier]

On Tue, Dec 19, 2017 at 6:00 PM, Leon Romanovsky <leon@xxxxxxxxxx> wrote:
>> status = ocrdma_nonemb_mbx_cmd(dev, mqe, dev->stats_mem.va);
>
> It still doesn't make a lot of sense to me:
>
> ocrdma_mbx_rdma_stats():
> 1315        if (reset)
> 1316                 req->reset_stats = reset;
> 1317
> 1318         status = ocrdma_nonemb_mbx_cmd(dev, mqe, dev->stats_mem.va);
>
>
> 1117 static int ocrdma_nonemb_mbx_cmd(struct ocrdma_dev *dev, struct ocrdma_mqe *mqe,
> 1118                                  void *payload_va)
> 1119 {
> 1120         int status;
> 1121         struct ocrdma_mbx_rsp *rsp = payload_va;
> 1122
> 1123         if ((mqe->hdr.spcl_sge_cnt_emb & OCRDMA_MQE_HDR_EMB_MASK)
> 1124                                 OCRDMA_MQE_HDR_EMB_SHIFT)
> 1125                 BUG();
> 1126
> 1127         status = ocrdma_mbx_cmd(dev, mqe);
>
> You are not using rsp and/or payload_va to provide data to ocrdma_mbx_cmd.
>
> How are you passing reset_stats information to FW?
>
This is a command where the response size is more than  236 bytes. So this is
sent as a non-embedded command where the req/resp will be available in
payload_va
and phys address of payload_va (payload_pa) will be send down to
firmware as MQE. FW will dma the mailbox
from the payload_pa and response is transferred back.

req->reset_stats = reset; is actually setting the corresponding bit in
the request structure @payload_va.
FW receives reset_stats once the request is DMA-ed from payload_pa.

> Thanks
>

Thanks,
Selvin