[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZIb/WKKNlzjTIu2h@shell.armlinux.org.uk>
Date: Mon, 12 Jun 2023 12:19:52 +0100
From: "Russell King (Oracle)" <linux@...linux.org.uk>
To: Daniel Golle <daniel@...rotopia.org>
Cc: netdev@...r.kernel.org, linux-mediatek@...ts.infradead.org,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
devicetree@...r.kernel.org,
AngeloGioacchino Del Regno
<angelogioacchino.delregno@...labora.com>,
Matthias Brugger <matthias.bgg@...il.com>,
Lorenzo Bianconi <lorenzo@...nel.org>,
Mark Lee <Mark-MC.Lee@...iatek.com>,
Sean Wang <sean.wang@...iatek.com>,
John Crispin <john@...ozen.org>, Felix Fietkau <nbd@....name>,
Conor Dooley <conor+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Rob Herring <robh+dt@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Jakub Kicinski <kuba@...nel.org>,
Eric Dumazet <edumazet@...gle.com>,
"David S. Miller" <davem@...emloft.net>,
Sam Shih <Sam.Shih@...iatek.com>
Subject: Re: [PATCH net-next 5/8] net: ethernet: mtk_eth_soc: add
MTK_NETSYS_V3 capability bit
On Sun, Jun 11, 2023 at 01:38:23AM +0100, Daniel Golle wrote:
> @@ -1333,8 +1354,13 @@ static int mtk_tx_map(struct sk_buff *skb, struct net_device *dev,
> mtk_tx_set_dma_desc(dev, itxd, &txd_info);
>
> itx_buf->flags |= MTK_TX_FLAGS_SINGLE0;
> - itx_buf->flags |= (!mac->id) ? MTK_TX_FLAGS_FPORT0 :
> - MTK_TX_FLAGS_FPORT1;
> + if (mac->id == MTK_GMAC1_ID)
> + itx_buf->flags |= MTK_TX_FLAGS_FPORT0;
> + else if (mac->id == MTK_GMAC2_ID)
> + itx_buf->flags |= MTK_TX_FLAGS_FPORT1;
> + else
> + itx_buf->flags |= MTK_TX_FLAGS_FPORT2;
There appears to be two places that this code structure appears, and
this is in the path for packet transmission. I wonder if it would be
more efficient to instead do:
itx_buf->flags |= MTK_TX_FLAGS_SINGLE0 | mac->tx_flags;
with mac->tx_flags appropriately initialised?
> @@ -2170,7 +2214,9 @@ static int mtk_poll_tx_qdma(struct mtk_eth *eth, int budget,
> tx_buf = mtk_desc_to_tx_buf(ring, desc,
> eth->soc->txrx.txd_size);
> if (tx_buf->flags & MTK_TX_FLAGS_FPORT1)
> - mac = 1;
> + mac = MTK_GMAC2_ID;
> + else if (tx_buf->flags & MTK_TX_FLAGS_FPORT2)
> + mac = MTK_GMAC3_ID;
This has me wondering whether the flags are used for hardware or just
for the driver's purposes. If it's the latter, can we instead store the
MAC index in tx_buf, rather than having to decode a bitfield?
I suspect these are just for the driver given that the addition of
MTK_TX_FLAGS_FPORT2 changes all subsequent bit numbers in this struct
member.
>
> if (!tx_buf->data)
> break;
> @@ -3783,7 +3829,26 @@ static int mtk_hw_init(struct mtk_eth *eth, bool reset)
> mtk_w32(eth, eth->soc->txrx.rx_irq_done_mask, reg_map->qdma.int_grp + 4);
> mtk_w32(eth, 0x21021000, MTK_FE_INT_GRP);
>
> - if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V2)) {
> + if (MTK_HAS_CAPS(eth->soc->caps, MTK_NETSYS_V3)) {
> + /* PSE should not drop port1, port8 and port9 packets */
> + mtk_w32(eth, 0x00000302, PSE_DROP_CFG);
> +
> + /* GDM and CDM Threshold */
> + mtk_w32(eth, 0x00000707, MTK_CDMW0_THRES);
> + mtk_w32(eth, 0x00000077, MTK_CDMW1_THRES);
> +
> + /* Disable GDM1 RX CRC stripping */
> + val = mtk_r32(eth, MTK_GDMA_FWD_CFG(0));
> + val &= ~MTK_GDMA_STRP_CRC;
> + mtk_w32(eth, val, MTK_GDMA_FWD_CFG(0));
mtk_m32() ?
Thanks!
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!
Powered by blists - more mailing lists