Hi Sergei,
> Subject: Re: [RFC/PATCH 14/18] ravb: Add rx_ring_format function for
> GbEthernet
>
> On 9/23/21 5:08 PM, Biju Das wrote:
>
> > This patch adds rx_ring_format function for GbEthernet found on RZ/G2L
> > SoC.
> >
> > Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > ---
> > drivers/net/ethernet/renesas/ravb.h | 1 +
> > drivers/net/ethernet/renesas/ravb_main.c | 27
> > +++++++++++++++++++++++-
> > 2 files changed, 27 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/ethernet/renesas/ravb.h
> > b/drivers/net/ethernet/renesas/ravb.h
> > index 2505de5d4a28..b0e067a6a8ee 100644
> > --- a/drivers/net/ethernet/renesas/ravb.h
> > +++ b/drivers/net/ethernet/renesas/ravb.h
> > @@ -982,6 +982,7 @@ enum CSR0_BIT {
> > #define RX_BUF_SZ (2048 - ETH_FCS_LEN + sizeof(__sum16))
> >
> > #define RGETH_RX_BUFF_MAX 8192
> > +#define RGETH_RX_DESC_DATA_SIZE 4080
> >
> > struct ravb_tstamp_skb {
> > struct list_head list;
> > diff --git a/drivers/net/ethernet/renesas/ravb_main.c
> > b/drivers/net/ethernet/renesas/ravb_main.c
> > index 038af36141bb..ee1066fedc4a 100644
> > --- a/drivers/net/ethernet/renesas/ravb_main.c
> > +++ b/drivers/net/ethernet/renesas/ravb_main.c
> > @@ -327,7 +327,32 @@ static void ravb_ring_free(struct net_device
> > *ndev, int q)
> >
> > static void ravb_rx_ring_format_rgeth(struct net_device *ndev, int q)
> > {
> > - /* Place holder */
> > + struct ravb_private *priv = netdev_priv(ndev);
> > + struct ravb_rx_desc *rx_desc;
> > + unsigned int rx_ring_size = sizeof(*rx_desc) * priv->num_rx_ring[q];
> > + dma_addr_t dma_addr;
> > + unsigned int i;
> > +
> > + memset(priv->rgeth_rx_ring[q], 0, rx_ring_size);
> > + /* Build RX ring buffer */
> > + for (i = 0; i < priv->num_rx_ring[q]; i++) {
> > + /* RX descriptor */
> > + rx_desc = &priv->rgeth_rx_ring[q][i];
>
> Looks like this patch shold come bafore the patch #12 as well...
Patch#12 is timestamp related and we agreed to merge timestamp related code to gPTP support.
>
> > + rx_desc->ds_cc = cpu_to_le16(RGETH_RX_DESC_DATA_SIZE);
> > + dma_addr = dma_map_single(ndev->dev.parent, priv-
> >rx_skb[q][i]->data,
> > + RGETH_RX_BUFF_MAX,
>
> Allocation buffer size more then the real data size? Does that make
> sense?
Allocated buffer is 8K and mapping buffer size is 8K. So there won't be any issues
for dma/cpu operations.
You can refer to cip kernel[1] for details
[1] https://github.com/renesas-rz/rz_linux-cip/blob/rzg2l-cip41/drivers/net/ethernet/renesas/ravb_main.c#L326
>
> > + DMA_FROM_DEVICE);
> > + /* We just set the data size to 0 for a failed mapping which
> > + * should prevent DMA from happening...
> > + */
> > + if (dma_mapping_error(ndev->dev.parent, dma_addr))
> > + rx_desc->ds_cc = cpu_to_le16(0);
> > + rx_desc->dptr = cpu_to_le32(dma_addr);
> > + rx_desc->die_dt = DT_FEMPTY;
> > + }
> > + rx_desc = &priv->rgeth_rx_ring[q][i];
> > + rx_desc->dptr = cpu_to_le32((u32)priv->rx_desc_dma[q]);
> > + rx_desc->die_dt = DT_LINKFIX; /* type */
> > }
> >
> > static void ravb_rx_ring_format(struct net_device *ndev, int q)
>
> MBR, Sergey
