Re: [PATCH v5] sun4i-emac.c: add dma support

Re: [PATCH v5] sun4i-emac.c: add dma support

[Jakub Kicinski]

On Tue, 28 Dec 2021 19:42:04 +0800 conleylee@foxmail.com wrote:
> +static void free_emac_dma_req(struct emac_dma_req *req)

emac_free_dma_req

> +prepare_err:
> +	dma_unmap_single(db->dev, rxbuf, count, DMA_FROM_DEVICE);
> +    return ret;

incorrect whitespace here

> @@ -599,12 +721,25 @@ static void emac_rx(struct net_device *dev)
>  			if (!skb)
>  				continue;
>  			skb_reserve(skb, 2);
> -			rdptr = skb_put(skb, rxlen - 4);
>  
>  			/* Read received packet from RX SRAM */
>  			if (netif_msg_rx_status(db))
>  				dev_dbg(db->dev, "RxLen %x\n", rxlen);
>  
> +			rdptr = skb_put(skb, rxlen - 4);

no reason to move this line

[PATCH v6] sun4i-emac.c: add dma support

[conleylee]

From: Conley Lee <conleylee@foxmail.com>

Thanks for your review. Here is the new version for this patch.

This patch adds support for the emac rx dma present on sun4i. The emac
is able to move packets from rx fifo to RAM by using dma.

Change since v4.
  - rename sbk field to skb
  - rename alloc_emac_dma_req to emac_alloc_dma_req
  - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
    sleeping
  - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
  - fix some code style issues 

Change since v5.
  - fix some code style issue

Signed-off-by: Conley Lee <conleylee@foxmail.com>
---
 drivers/net/ethernet/allwinner/sun4i-emac.c | 200 ++++++++++++++++++++
 1 file changed, 200 insertions(+)

diff --git a/drivers/net/ethernet/allwinner/sun4i-emac.c b/drivers/net/ethernet/allwinner/sun4i-emac.c
index cccf8a3ead5e..964227e342ee 100644
--- a/drivers/net/ethernet/allwinner/sun4i-emac.c
+++ b/drivers/net/ethernet/allwinner/sun4i-emac.c
@@ -29,6 +29,7 @@
 #include <linux/platform_device.h>
 #include <linux/phy.h>
 #include <linux/soc/sunxi/sunxi_sram.h>
+#include <linux/dmaengine.h>
 
 #include "sun4i-emac.h"
 
@@ -86,6 +87,16 @@ struct emac_board_info {
 	unsigned int		duplex;
 
 	phy_interface_t		phy_interface;
+	struct dma_chan	*rx_chan;
+	phys_addr_t emac_rx_fifo;
+};
+
+struct emac_dma_req {
+	struct emac_board_info *db;
+	struct dma_async_tx_descriptor *desc;
+	struct sk_buff *skb;
+	dma_addr_t rxbuf;
+	int count;
 };
 
 static void emac_update_speed(struct net_device *dev)
@@ -205,6 +216,117 @@ static void emac_inblk_32bit(void __iomem *reg, void *data, int count)
 	readsl(reg, data, round_up(count, 4) / 4);
 }
 
+static struct emac_dma_req *
+emac_alloc_dma_req(struct emac_board_info *db,
+		   struct dma_async_tx_descriptor *desc, struct sk_buff *skb,
+		   dma_addr_t rxbuf, int count)
+{
+	struct emac_dma_req *req;
+
+	req = kzalloc(sizeof(struct emac_dma_req), GFP_ATOMIC);
+	if (!req)
+		return NULL;
+
+	req->db = db;
+	req->desc = desc;
+	req->skb = skb;
+	req->rxbuf = rxbuf;
+	req->count = count;
+	return req;
+}
+
+static void emac_free_dma_req(struct emac_dma_req *req)
+{
+	kfree(req);
+}
+
+static void emac_dma_done_callback(void *arg)
+{
+	struct emac_dma_req *req = arg;
+	struct emac_board_info *db = req->db;
+	struct sk_buff *skb = req->skb;
+	struct net_device *dev = db->ndev;
+	int rxlen = req->count;
+	u32 reg_val;
+
+	dma_unmap_single(db->dev, req->rxbuf, rxlen, DMA_FROM_DEVICE);
+
+	skb->protocol = eth_type_trans(skb, dev);
+	netif_rx(skb);
+	dev->stats.rx_bytes += rxlen;
+	/* Pass to upper layer */
+	dev->stats.rx_packets++;
+
+	/* re enable cpu receive */
+	reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+	reg_val &= ~EMAC_RX_CTL_DMA_EN;
+	writel(reg_val, db->membase + EMAC_RX_CTL_REG);
+
+	/* re enable interrupt */
+	reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+	reg_val |= (0x01 << 8);
+	writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+
+	db->emacrx_completed_flag = 1;
+	emac_free_dma_req(req);
+}
+
+static int emac_dma_inblk_32bit(struct emac_board_info *db,
+		struct sk_buff *skb, void *rdptr, int count)
+{
+	struct dma_async_tx_descriptor *desc;
+	dma_cookie_t cookie;
+	dma_addr_t rxbuf;
+	struct emac_dma_req *req;
+	int ret = 0;
+
+	rxbuf = dma_map_single(db->dev, rdptr, count, DMA_FROM_DEVICE);
+	ret = dma_mapping_error(db->dev, rxbuf);
+	if (ret) {
+		dev_err(db->dev, "dma mapping error.\n");
+		return ret;
+	}
+
+	desc = dmaengine_prep_slave_single(db->rx_chan, rxbuf, count,
+					   DMA_DEV_TO_MEM,
+					   DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc) {
+		dev_err(db->dev, "prepare slave single failed\n");
+		ret = -ENOMEM;
+		goto prepare_err;
+	}
+
+	req = emac_alloc_dma_req(db, desc, skb, rxbuf, count);
+	if (!req) {
+		dev_err(db->dev, "alloc emac dma req error.\n");
+		ret = -ENOMEM;
+		goto alloc_req_err;
+	}
+
+	desc->callback_param = req;
+	desc->callback = emac_dma_done_callback;
+
+	cookie = dmaengine_submit(desc);
+	ret = dma_submit_error(cookie);
+	if (ret) {
+		dev_err(db->dev, "dma submit error.\n");
+		goto submit_err;
+	}
+
+	dma_async_issue_pending(db->rx_chan);
+	return ret;
+
+submit_err:
+	emac_free_dma_req(req);
+
+alloc_req_err:
+	dmaengine_desc_free(desc);
+
+prepare_err:
+	dma_unmap_single(db->dev, rxbuf, count, DMA_FROM_DEVICE);
+	return ret;
+}
+
 /* ethtool ops */
 static void emac_get_drvinfo(struct net_device *dev,
 			      struct ethtool_drvinfo *info)
@@ -605,6 +727,19 @@ static void emac_rx(struct net_device *dev)
 			if (netif_msg_rx_status(db))
 				dev_dbg(db->dev, "RxLen %x\n", rxlen);
 
+			if (rxlen >= dev->mtu && db->rx_chan) {
+				reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+				reg_val |= EMAC_RX_CTL_DMA_EN;
+				writel(reg_val, db->membase + EMAC_RX_CTL_REG);
+				if (!emac_dma_inblk_32bit(db, skb, rdptr, rxlen))
+					break;
+
+				/* re enable cpu receive. then try to receive by emac_inblk_32bit */
+				reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+				reg_val &= ~EMAC_RX_CTL_DMA_EN;
+				writel(reg_val, db->membase + EMAC_RX_CTL_REG);
+			}
+
 			emac_inblk_32bit(db->membase + EMAC_RX_IO_DATA_REG,
 					rdptr, rxlen);
 			dev->stats.rx_bytes += rxlen;
@@ -659,7 +794,12 @@ static irqreturn_t emac_interrupt(int irq, void *dev_id)
 		reg_val = readl(db->membase + EMAC_INT_CTL_REG);
 		reg_val |= (0xf << 0) | (0x01 << 8);
 		writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+	} else {
+		reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+		reg_val |= (0xf << 0);
+		writel(reg_val, db->membase + EMAC_INT_CTL_REG);
 	}
+
 	spin_unlock(&db->lock);
 
 	return IRQ_HANDLED;
@@ -764,6 +904,58 @@ static const struct net_device_ops emac_netdev_ops = {
 #endif
 };
 
+static int emac_configure_dma(struct emac_board_info *db)
+{
+	struct platform_device *pdev = db->pdev;
+	struct net_device *ndev = db->ndev;
+	struct dma_slave_config conf = {};
+	struct resource *regs;
+	int err = 0;
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!regs) {
+		netdev_err(ndev, "get io resource from device failed.\n");
+		err = -ENOMEM;
+		goto out_clear_chan;
+	}
+
+	netdev_info(ndev, "get io resource from device: 0x%x, size = %u\n",
+		    regs->start, resource_size(regs));
+	db->emac_rx_fifo = regs->start + EMAC_RX_IO_DATA_REG;
+
+	db->rx_chan = dma_request_chan(&pdev->dev, "rx");
+	if (IS_ERR(db->rx_chan)) {
+		netdev_err(ndev,
+			   "failed to request dma channel. dma is disabled\n");
+		err = PTR_ERR(db->rx_chan);
+		goto out_clear_chan;
+	}
+
+	conf.direction = DMA_DEV_TO_MEM;
+	conf.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	conf.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	conf.src_addr = db->emac_rx_fifo;
+	conf.dst_maxburst = 4;
+	conf.src_maxburst = 4;
+	conf.device_fc = false;
+
+	err = dmaengine_slave_config(db->rx_chan, &conf);
+	if (err) {
+		netdev_err(ndev, "config dma slave failed\n");
+		err = -EINVAL;
+		goto out_slave_configure_err;
+	}
+
+	return err;
+
+out_slave_configure_err:
+	dma_release_channel(db->rx_chan);
+
+out_clear_chan:
+	db->rx_chan = NULL;
+	return err;
+}
+
 /* Search EMAC board, allocate space and register it
  */
 static int emac_probe(struct platform_device *pdev)
@@ -806,6 +998,9 @@ static int emac_probe(struct platform_device *pdev)
 		goto out_iounmap;
 	}
 
+	if (emac_configure_dma(db))
+		netdev_info(ndev, "configure dma failed. disable dma.\n");
+
 	db->clk = devm_clk_get(&pdev->dev, NULL);
 	if (IS_ERR(db->clk)) {
 		ret = PTR_ERR(db->clk);
@@ -888,6 +1083,11 @@ static int emac_remove(struct platform_device *pdev)
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct emac_board_info *db = netdev_priv(ndev);
 
+	if (db->rx_chan) {
+		dmaengine_terminate_all(db->rx_chan);
+		dma_release_channel(db->rx_chan);
+	}
+
 	unregister_netdev(ndev);
 	sunxi_sram_release(&pdev->dev);
 	clk_disable_unprepare(db->clk);
-- 
2.31.1

Re: [PATCH v6] sun4i-emac.c: add dma support

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (master)
by Jakub Kicinski <kuba@kernel.org>:

On Wed, 29 Dec 2021 09:43:51 +0800 you wrote:
> From: Conley Lee <conleylee@foxmail.com>
> 
> Thanks for your review. Here is the new version for this patch.
> 
> This patch adds support for the emac rx dma present on sun4i. The emac
> is able to move packets from rx fifo to RAM by using dma.
> 
> [...]

Here is the summary with links:
  - [v6] sun4i-emac.c: add dma support
    https://git.kernel.org/netdev/net-next/c/47869e82c8b8

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [PATCH v6] sun4i-emac.c: add dma support

[Corentin Labbe]

Le Thu, Dec 30, 2021 at 02:00:11AM +0000, patchwork-bot+netdevbpf@kernel.org a écrit :
> Hello:
> 
> This patch was applied to netdev/net-next.git (master)
> by Jakub Kicinski <kuba@kernel.org>:
> 
> On Wed, 29 Dec 2021 09:43:51 +0800 you wrote:
> > From: Conley Lee <conleylee@foxmail.com>
> > 
> > Thanks for your review. Here is the new version for this patch.
> > 
> > This patch adds support for the emac rx dma present on sun4i. The emac
> > is able to move packets from rx fifo to RAM by using dma.
> > 
> > [...]
> 
> Here is the summary with links:
>   - [v6] sun4i-emac.c: add dma support
>     https://git.kernel.org/netdev/net-next/c/47869e82c8b8
> 
> You are awesome, thank you!
> -- 
> Deet-doot-dot, I am a bot.
> https://korg.docs.kernel.org/patchwork/pwbot.html
> 

Hello

Any news on patch which enable sun4i-emac DMA in DT ?

Regards

Re: [PATCH v6] sun4i-emac.c: add dma support

[Jakub Kicinski]

On Mon, 30 May 2022 06:51:14 +0200 Corentin Labbe wrote:
> Any news on patch which enable sun4i-emac DMA in DT ?

Who are you directing this question to and where's that patch posted?

Re: [PATCH v6] sun4i-emac.c: add dma support

[Corentin Labbe]

Le Mon, May 30, 2022 at 11:48:19AM -0700, Jakub Kicinski a écrit :
> On Mon, 30 May 2022 06:51:14 +0200 Corentin Labbe wrote:
> > Any news on patch which enable sun4i-emac DMA in DT ?
> 
> Who are you directing this question to and where's that patch posted?

I am sorry, I fail to set the right "to:"
My question was for Conley Lee.

This serie was applied but the DT part was never posted.
So sun4i-emac can handle DMA but is not enabled at all.

The DT patch is easy, so without answer, I will send it.

Regards

Re: [PATCH v6] sun4i-emac.c: add dma support

[Corentin Labbe]

Le Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com a écrit :
> From: Conley Lee <conleylee@foxmail.com>
> 
> Thanks for your review. Here is the new version for this patch.
> 
> This patch adds support for the emac rx dma present on sun4i. The emac
> is able to move packets from rx fifo to RAM by using dma.
> 
> Change since v4.
>   - rename sbk field to skb
>   - rename alloc_emac_dma_req to emac_alloc_dma_req
>   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
>     sleeping
>   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
>   - fix some code style issues 
> 
> Change since v5.
>   - fix some code style issue
> 

Hello

I just tested this on a sun4i-a10-olinuxino-lime

I got:
[    2.922812] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): get io resource from device: 0x1c0b000, size = 4096
[    2.934512] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): failed to request dma channel. dma is disabled
[    2.945740] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): configure dma failed. disable dma.
[    2.957887] sun4i-emac 1c0b000.ethernet: eth0: at (ptrval), IRQ 19 MAC: 02:49:09:40:ab:3d

On which board did you test it and how ?

Regards

Re: [PATCH v6] sun4i-emac.c: add dma support

[Corentin Labbe]

Le Sat, Jan 01, 2022 at 03:09:01PM +0800, Conley Lee a écrit :
> On 12/31/21 at 11:43上午, Corentin Labbe wrote:
> > Date: Fri, 31 Dec 2021 11:43:53 +0100
> > From: Corentin Labbe <clabbe.montjoie@gmail.com>
> > To: conleylee@foxmail.com
> > Cc: davem@davemloft.net, kuba@kernel.org, mripard@kernel.org,
> >  wens@csie.org, netdev@vger.kernel.org,
> >  linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
> > Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> > 
> > Le Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com a écrit :
> > > From: Conley Lee <conleylee@foxmail.com>
> > > 
> > > Thanks for your review. Here is the new version for this patch.
> > > 
> > > This patch adds support for the emac rx dma present on sun4i. The emac
> > > is able to move packets from rx fifo to RAM by using dma.
> > > 
> > > Change since v4.
> > >   - rename sbk field to skb
> > >   - rename alloc_emac_dma_req to emac_alloc_dma_req
> > >   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
> > >     sleeping
> > >   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
> > >   - fix some code style issues 
> > > 
> > > Change since v5.
> > >   - fix some code style issue
> > > 
> > 
> > Hello
> > 
> > I just tested this on a sun4i-a10-olinuxino-lime
> > 
> > I got:
> > [    2.922812] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): get io resource from device: 0x1c0b000, size = 4096
> > [    2.934512] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): failed to request dma channel. dma is disabled
> > [    2.945740] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): configure dma failed. disable dma.
> > [    2.957887] sun4i-emac 1c0b000.ethernet: eth0: at (ptrval), IRQ 19 MAC: 02:49:09:40:ab:3d
> > 
> > On which board did you test it and how ?
> > 
> > Regards
> 
> Sorry. I sent the email with text/html format. This email is an clean version.
> 
> In order to enable dma rx channel. `dmas` and `dma-names` properties
> should be added to emac section in dts:
> 
> emac: ethernet@1c0b000 {
> 	...
> 	dmas = <&dma SUN4I_DMA_DEDICATED 7>;
> 	dma-names = "rx";
> 	...
> }

Helo

Yes I figured that out. But you should have done a patch serie adding this.
Your patch is now applied but it is a useless change without the dtb change.
You should also probably update the driver binding (Documentation/devicetree/bindings/net/allwinner,sun4i-a10-emac.yaml) since you add new members to DT node.

Furthermore, why did you add RX only and not TX dma also ?

Probably it is too late since patch is applied but it is:
Tested-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Tested-on: sun4i-a10-olinuxino-lime

Regards

Re: [PATCH v6] sun4i-emac.c: add dma support

[Corentin Labbe]

Le Mon, Jan 03, 2022 at 10:55:04AM +0800, Conley Lee a écrit :
> On 01/02/22 at 06:38下午, Corentin Labbe wrote:
> > Date: Sun, 2 Jan 2022 18:38:51 +0100
> > From: Corentin Labbe <clabbe.montjoie@gmail.com>
> > To: Conley Lee <conleylee@foxmail.com>
> > Cc: davem@davemloft.net, kuba@kernel.org, mripard@kernel.org,
> >  wens@csie.org, netdev@vger.kernel.org,
> >  linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
> >  linux-sunxi@lists.linux.dev, jernej.skrabec@gmail.com
> > Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> > 
> > Le Sat, Jan 01, 2022 at 03:09:01PM +0800, Conley Lee a écrit :
> > > On 12/31/21 at 11:43上午, Corentin Labbe wrote:
> > > > Date: Fri, 31 Dec 2021 11:43:53 +0100
> > > > From: Corentin Labbe <clabbe.montjoie@gmail.com>
> > > > To: conleylee@foxmail.com
> > > > Cc: davem@davemloft.net, kuba@kernel.org, mripard@kernel.org,
> > > >  wens@csie.org, netdev@vger.kernel.org,
> > > >  linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
> > > > Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> > > > 
> > > > Le Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com a écrit :
> > > > > From: Conley Lee <conleylee@foxmail.com>
> > > > > 
> > > > > Thanks for your review. Here is the new version for this patch.
> > > > > 
> > > > > This patch adds support for the emac rx dma present on sun4i. The emac
> > > > > is able to move packets from rx fifo to RAM by using dma.
> > > > > 
> > > > > Change since v4.
> > > > >   - rename sbk field to skb
> > > > >   - rename alloc_emac_dma_req to emac_alloc_dma_req
> > > > >   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
> > > > >     sleeping
> > > > >   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
> > > > >   - fix some code style issues 
> > > > > 
> > > > > Change since v5.
> > > > >   - fix some code style issue
> > > > > 
> > > > 
> > > > Hello
> > > > 
> > > > I just tested this on a sun4i-a10-olinuxino-lime
> > > > 
> > > > I got:
> > > > [    2.922812] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): get io resource from device: 0x1c0b000, size = 4096
> > > > [    2.934512] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): failed to request dma channel. dma is disabled
> > > > [    2.945740] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): configure dma failed. disable dma.
> > > > [    2.957887] sun4i-emac 1c0b000.ethernet: eth0: at (ptrval), IRQ 19 MAC: 02:49:09:40:ab:3d
> > > > 
> > > > On which board did you test it and how ?
> > > > 
> > > > Regards
> > > 
> > > Sorry. I sent the email with text/html format. This email is an clean version.
> > > 
> > > In order to enable dma rx channel. `dmas` and `dma-names` properties
> > > should be added to emac section in dts:
> > > 
> > > emac: ethernet@1c0b000 {
> > > 	...
> > > 	dmas = <&dma SUN4I_DMA_DEDICATED 7>;
> > > 	dma-names = "rx";
> > > 	...
> > > }
> > 
> > Helo
> > 
> > Yes I figured that out. But you should have done a patch serie adding this.
> > Your patch is now applied but it is a useless change without the dtb change.
> > You should also probably update the driver binding (Documentation/devicetree/bindings/net/allwinner,sun4i-a10-emac.yaml) since you add new members to DT node.
> > 
> > Furthermore, why did you add RX only and not TX dma also ?
> > 
> > Probably it is too late since patch is applied but it is:
> > Tested-by: Corentin Labbe <clabbe.montjoie@gmail.com>
> > Tested-on: sun4i-a10-olinuxino-lime
> > 
> > Regards
> 
> Thanks for your suggestion. I will submit a patch to add those changes
> later. 
> 
> And the reason why I didn't add TX support is becasuse there is no any
> public page to describe sun4i emac TX DMA register map. So, I don't known
> how to enable TX DMA at hardware level. If you has any page or datasheet
> about EMAC TX DMA, can you share with me ? Thanks.

Hello

You can find TX DMA info on the R40 Use manual (8.10.5.2 Register Name: EMAC_TX_MODE)

You should keep all people in CC when you answer to someone.

Regards

Re: [PATCH v6] sun4i-emac.c: add dma support

[Conley Lee]

On 01/03/22 at 12:42下午, Corentin Labbe wrote:
> Date: Mon, 3 Jan 2022 12:42:58 +0100
> From: Corentin Labbe <clabbe.montjoie@gmail.com>
> To: Conley Lee <conleylee@foxmail.com>
> Cc: davem@davemloft.net, mripard@kernel.org, wens@csie.org,
>  jernej.skrabec@gmail.com, netdev@vger.kernel.org,
>  linux-arm-kernel@lists.infradead.org, linux-sunxi@lists.linux.dev,
>  linux-kernel@vger.kernel.org
> Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> 
> Le Mon, Jan 03, 2022 at 10:55:04AM +0800, Conley Lee a écrit :
> > On 01/02/22 at 06:38下午, Corentin Labbe wrote:
> > > Date: Sun, 2 Jan 2022 18:38:51 +0100
> > > From: Corentin Labbe <clabbe.montjoie@gmail.com>
> > > To: Conley Lee <conleylee@foxmail.com>
> > > Cc: davem@davemloft.net, kuba@kernel.org, mripard@kernel.org,
> > >  wens@csie.org, netdev@vger.kernel.org,
> > >  linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
> > >  linux-sunxi@lists.linux.dev, jernej.skrabec@gmail.com
> > > Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> > > 
> > > Le Sat, Jan 01, 2022 at 03:09:01PM +0800, Conley Lee a écrit :
> > > > On 12/31/21 at 11:43上午, Corentin Labbe wrote:
> > > > > Date: Fri, 31 Dec 2021 11:43:53 +0100
> > > > > From: Corentin Labbe <clabbe.montjoie@gmail.com>
> > > > > To: conleylee@foxmail.com
> > > > > Cc: davem@davemloft.net, kuba@kernel.org, mripard@kernel.org,
> > > > >  wens@csie.org, netdev@vger.kernel.org,
> > > > >  linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
> > > > > Subject: Re: [PATCH v6] sun4i-emac.c: add dma support
> > > > > 
> > > > > Le Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com a écrit :
> > > > > > From: Conley Lee <conleylee@foxmail.com>
> > > > > > 
> > > > > > Thanks for your review. Here is the new version for this patch.
> > > > > > 
> > > > > > This patch adds support for the emac rx dma present on sun4i. The emac
> > > > > > is able to move packets from rx fifo to RAM by using dma.
> > > > > > 
> > > > > > Change since v4.
> > > > > >   - rename sbk field to skb
> > > > > >   - rename alloc_emac_dma_req to emac_alloc_dma_req
> > > > > >   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
> > > > > >     sleeping
> > > > > >   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
> > > > > >   - fix some code style issues 
> > > > > > 
> > > > > > Change since v5.
> > > > > >   - fix some code style issue
> > > > > > 
> > > > > 
> > > > > Hello
> > > > > 
> > > > > I just tested this on a sun4i-a10-olinuxino-lime
> > > > > 
> > > > > I got:
> > > > > [    2.922812] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): get io resource from device: 0x1c0b000, size = 4096
> > > > > [    2.934512] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): failed to request dma channel. dma is disabled
> > > > > [    2.945740] sun4i-emac 1c0b000.ethernet (unnamed net_device) (uninitialized): configure dma failed. disable dma.
> > > > > [    2.957887] sun4i-emac 1c0b000.ethernet: eth0: at (ptrval), IRQ 19 MAC: 02:49:09:40:ab:3d
> > > > > 
> > > > > On which board did you test it and how ?
> > > > > 
> > > > > Regards
> > > > 
> > > > Sorry. I sent the email with text/html format. This email is an clean version.
> > > > 
> > > > In order to enable dma rx channel. `dmas` and `dma-names` properties
> > > > should be added to emac section in dts:
> > > > 
> > > > emac: ethernet@1c0b000 {
> > > > 	...
> > > > 	dmas = <&dma SUN4I_DMA_DEDICATED 7>;
> > > > 	dma-names = "rx";
> > > > 	...
> > > > }
> > > 
> > > Helo
> > > 
> > > Yes I figured that out. But you should have done a patch serie adding this.
> > > Your patch is now applied but it is a useless change without the dtb change.
> > > You should also probably update the driver binding (Documentation/devicetree/bindings/net/allwinner,sun4i-a10-emac.yaml) since you add new members to DT node.
> > > 
> > > Furthermore, why did you add RX only and not TX dma also ?
> > > 
> > > Probably it is too late since patch is applied but it is:
> > > Tested-by: Corentin Labbe <clabbe.montjoie@gmail.com>
> > > Tested-on: sun4i-a10-olinuxino-lime
> > > 
> > > Regards
> > 
> > Thanks for your suggestion. I will submit a patch to add those changes
> > later. 
> > 
> > And the reason why I didn't add TX support is becasuse there is no any
> > public page to describe sun4i emac TX DMA register map. So, I don't known
> > how to enable TX DMA at hardware level. If you has any page or datasheet
> > about EMAC TX DMA, can you share with me ? Thanks.
> 
> Hello
> 
> You can find TX DMA info on the R40 Use manual (8.10.5.2 Register Name: EMAC_TX_MODE)
> 
> You should keep all people in CC when you answer to someone.
> 
> Regards

Haha, got it! I have been looking for the docs about emac register map
for a long time. You really help me a lot. I will submit a new patch to
add driver bidding and enable both RX and TX channels. Thanks ~

Re: [PATCH v1] sun4i-emac.c: enable emac tx dma

[Corentin Labbe]

Le Sun, Jan 09, 2022 at 05:17:55PM +0800, conleylee@foxmail.com a écrit :
> From: conley <conleylee@foxmail.com>
> 
> Hello
> I am reading the R40 user manual and trying to create a new path to enable
> emac tx dma channel. According to the figure 8-21(TX Operation Diagram),
> I try to enable emac tx dma channel by the follow steps:
> 1. enable tx dma mode
> 2. set packet lengths
> 2. move data from skb to tx fifo by using dma in xmit function.
> 3. start transfer from tx fifo to phy in dma tx done callback
> 
> But it doesn't work. emac tx interrupt and dma finished interrupt are
> raised, but no packets are transmitted (I test it by tcpdump).
> Do you know how to configure the emac tx dma correctly? Thanks ~
> 

Hello

Here are my thoughts to help you:
- Your email is not a real patch, but an ask for help, so you should not use [ PATCH ] in the subject.
- If it was a patch, "v1" is not necessary
- Your patch below is doing too many unrelated different things, it is hard to see the DMA TX enable part
- I think you could first send a preliminary patch which adds all EMAC_INT_CTL_TX_xxx which are already used by the driver (to reduce the diff)
- Without the DTB change, it is hard to see the whole picture, did you correctly use the right dma number for an easy example.
- Knowing also the board (and so PHY, modes etc...) could help
- I think your priority should not to add TX, but to fix reported problems to your initial patch (build warnings/error https://marc.info/?l=linux-arm-kernel&m=164159846213585&w=2) since your work on TX will need to be applied after this.
- For the previous point, always build test with at least 2 different 32/64 arch. And if possible a total different arch (like x86_64).

Anyway, I will try to test your patch on my a10 board

Regards

Re: [PATCH v1] sun4i-emac.c: enable emac tx dma

[Conley Lee]

On 01/09/22 at 09:45下午, Corentin Labbe wrote:
> Date: Sun, 9 Jan 2022 21:45:01 +0100
> From: Corentin Labbe <clabbe.montjoie@gmail.com>
> To: conleylee@foxmail.com
> Cc: davem@davemloft.net, mripard@kernel.org, wens@csie.org,
>  jernej.skrabec@gmail.com, netdev@vger.kernel.org,
>  linux-arm-kernel@lists.infradead.org, linux-sunxi@lists.linux.dev,
>  linux-kernel@vger.kernel.org
> Subject: Re: [PATCH v1] sun4i-emac.c: enable emac tx dma
> 
> Le Sun, Jan 09, 2022 at 05:17:55PM +0800, conleylee@foxmail.com a écrit :
> > From: conley <conleylee@foxmail.com>
> > 
> > Hello
> > I am reading the R40 user manual and trying to create a new path to enable
> > emac tx dma channel. According to the figure 8-21(TX Operation Diagram),
> > I try to enable emac tx dma channel by the follow steps:
> > 1. enable tx dma mode
> > 2. set packet lengths
> > 2. move data from skb to tx fifo by using dma in xmit function.
> > 3. start transfer from tx fifo to phy in dma tx done callback
> > 
> > But it doesn't work. emac tx interrupt and dma finished interrupt are
> > raised, but no packets are transmitted (I test it by tcpdump).
> > Do you know how to configure the emac tx dma correctly? Thanks ~
> > 
> 
> Hello
> 
> Here are my thoughts to help you:
> - Your email is not a real patch, but an ask for help, so you should not use [ PATCH ] in the subject.
> - If it was a patch, "v1" is not necessary
> - Your patch below is doing too many unrelated different things, it is hard to see the DMA TX enable part
> - I think you could first send a preliminary patch which adds all EMAC_INT_CTL_TX_xxx which are already used by the driver (to reduce the diff)
> - Without the DTB change, it is hard to see the whole picture, did you correctly use the right dma number for an easy example.
> - Knowing also the board (and so PHY, modes etc...) could help
> - I think your priority should not to add TX, but to fix reported problems to your initial patch (build warnings/error https://marc.info/?l=linux-arm-kernel&m=164159846213585&w=2) since your work on TX will need to be applied after this.
> - For the previous point, always build test with at least 2 different 32/64 arch. And if possible a total different arch (like x86_64).
> 
> Anyway, I will try to test your patch on my a10 board
> 
> Regards
Thanks a lot for your help ~ I will submit an patch which add all
register related. By the way, the build warnings problems have been fixed
by Jakub Kicinski <kuba@kernel.org>.

Re: [PATCH v6] sun4i-emac.c: add dma support

[Kees Cook]

On Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com wrote:
> From: Conley Lee <conleylee@foxmail.com>
>
> Thanks for your review. Here is the new version for this patch.
>
> This patch adds support for the emac rx dma present on sun4i. The emac
> is able to move packets from rx fifo to RAM by using dma.
>
> Change since v4.
>   - rename sbk field to skb
>   - rename alloc_emac_dma_req to emac_alloc_dma_req
>   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
>     sleeping
>   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
>   - fix some code style issues
>
> Change since v5.
>   - fix some code style issue
>
> Signed-off-by: Conley Lee <conleylee@foxmail.com>
> ---
>  drivers/net/ethernet/allwinner/sun4i-emac.c | 200 ++++++++++++++++++++
>  1 file changed, 200 insertions(+)

This is causing build failures:

$ make CROSS_COMPILE=aarch64-linux-gnu- ARCH=arm64 allmodconfig
$ make CROSS_COMPILE=aarch64-linux-gnu- ARCH=arm64 -j... -k -s
drivers/net/ethernet/allwinner/sun4i-emac.c: In function 'emac_configure_dma':
drivers/net/ethernet/allwinner/sun4i-emac.c:922:60: error: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
  922 |         netdev_info(ndev, "get io resource from device: 0x%x, size = %u\n",
      |                                                           ~^
      |                                                            |                                      |                                                            unsigned int
      |                                                           %llx
  923 |                     regs->start, resource_size(regs));
      |                     ~~~~~~~~~~~
      |                         |
      |                         resource_size_t {aka long long unsigned int}
drivers/net/ethernet/allwinner/sun4i-emac.c:922:71: error: format '%u' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
  922 |         netdev_info(ndev, "get io resource from device: 0x%x, size = %u\n",
      |                                                                      ~^
      |                                                                       |
      |                                                                       unsigned int
      |                                                                      %llu
  923 |                     regs->start, resource_size(regs));
      |                                  ~~~~~~~~~~~~~~~~~~~
      |                                  |
      |                                  resource_size_t {aka long long unsigned int}


--
Kees Cook

Re: [PATCH v6] sun4i-emac.c: add dma support

[Jakub Kicinski]

On Fri, 7 Jan 2022 15:34:18 -0800 Kees Cook wrote:
> On Wed, Dec 29, 2021 at 09:43:51AM +0800, conleylee@foxmail.com wrote:
> > From: Conley Lee <conleylee@foxmail.com>
> >
> > Thanks for your review. Here is the new version for this patch.
> >
> > This patch adds support for the emac rx dma present on sun4i. The emac
> > is able to move packets from rx fifo to RAM by using dma.
> >
> > Change since v4.
> >   - rename sbk field to skb
> >   - rename alloc_emac_dma_req to emac_alloc_dma_req
> >   - using kzalloc(..., GPF_ATOMIC) in interrupt context to avoid
> >     sleeping
> >   - retry by using emac_inblk_32bit when emac_dma_inblk_32bit fails
> >   - fix some code style issues
> >
> > Change since v5.
> >   - fix some code style issue
> >
> > Signed-off-by: Conley Lee <conleylee@foxmail.com>
>
> This is causing build failures:
> 
> $ make CROSS_COMPILE=aarch64-linux-gnu- ARCH=arm64 allmodconfig
> $ make CROSS_COMPILE=aarch64-linux-gnu- ARCH=arm64 -j... -k -s
> drivers/net/ethernet/allwinner/sun4i-emac.c: In function 'emac_configure_dma':
> drivers/net/ethernet/allwinner/sun4i-emac.c:922:60: error: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
>   922 |         netdev_info(ndev, "get io resource from device: 0x%x, size = %u\n",
>       |                                                           ~^
>       |                                                            |                                      |                                                            unsigned int
>       |                                                           %llx
>   923 |                     regs->start, resource_size(regs));
>       |                     ~~~~~~~~~~~
>       |                         |
>       |                         resource_size_t {aka long long unsigned int}
> drivers/net/ethernet/allwinner/sun4i-emac.c:922:71: error: format '%u' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
>   922 |         netdev_info(ndev, "get io resource from device: 0x%x, size = %u\n",
>       |                                                                      ~^
>       |                                                                       |
>       |                                                                       unsigned int
>       |                                                                      %llu
>   923 |                     regs->start, resource_size(regs));
>       |                                  ~~~~~~~~~~~~~~~~~~~
>       |                                  |
>       |                                  resource_size_t {aka long long unsigned int}
> 

Ugh, I saw this and somehow it didn't enter my brain that it's new.
%pa right? Let me test that and send a fix before we close net-next..