[PATCH net next] net: axienet: Use dedicated ethtool_ops for the dmaengine path

[PATCH net next] net: axienet: Use dedicated ethtool_ops for the dmaengine path

[Suraj Gupta]

The dmaengine path shares ethtool_ops with the legacy AXI DMA path,
including .get_coalesce/.set_coalesce that poke XAXIDMA_*_CR_OFFSET
directly. In dmaengine mode lp->dma_regs is not mapped by axienet, so
those ethtool calls touch unmapped/unrelated memory and report values
unrelated to the channel actually in use.

.get_ringparam/.set_ringparam only touch lp->rx_bd_num/lp->tx_bd_num,
fields used only by the legacy path for BD ring sizing. In dmaengine
mode the descriptor ring is owned by the dmaengine provider and these
fields are not consulted, so reporting them is misleading.

No dmaengine API exists today to query or program either coalescing
or ring size on behalf of the client, so neither can be exposed
meaningfully in dmaengine mode.

Add axienet_ethtool_dmaengine_ops without the coalesce and ringparam
hooks. Also move the ethtool_ops assignment from early probe into the
if/else alongside netdev_ops, so the legacy and dmaengine paths pick
their respective ops in one place. No functional change for the
legacy DMA path.

Signed-off-by: Suraj Gupta <suraj.gupta2@amd.com>
---
 .../net/ethernet/xilinx/xilinx_axienet_main.c | 27 ++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
index 263c4b67fd5a..fcf517069d16 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
@@ -2536,6 +2536,25 @@ static const struct ethtool_ops axienet_ethtool_ops = {
 	.get_rmon_stats = axienet_ethtool_get_rmon_stats,
 };
 
+static const struct ethtool_ops axienet_ethtool_dmaengine_ops = {
+	.get_drvinfo    = axienet_ethtools_get_drvinfo,
+	.get_regs_len   = axienet_ethtools_get_regs_len,
+	.get_regs       = axienet_ethtools_get_regs,
+	.get_link       = ethtool_op_get_link,
+	.get_pauseparam = axienet_ethtools_get_pauseparam,
+	.set_pauseparam = axienet_ethtools_set_pauseparam,
+	.get_link_ksettings = axienet_ethtools_get_link_ksettings,
+	.set_link_ksettings = axienet_ethtools_set_link_ksettings,
+	.nway_reset	= axienet_ethtools_nway_reset,
+	.get_ethtool_stats = axienet_ethtools_get_ethtool_stats,
+	.get_strings    = axienet_ethtools_get_strings,
+	.get_sset_count = axienet_ethtools_get_sset_count,
+	.get_pause_stats = axienet_ethtools_get_pause_stats,
+	.get_eth_mac_stats = axienet_ethtool_get_eth_mac_stats,
+	.get_eth_ctrl_stats = axienet_ethtool_get_eth_ctrl_stats,
+	.get_rmon_stats = axienet_ethtool_get_rmon_stats,
+};
+
 static struct axienet_local *pcs_to_axienet_local(struct phylink_pcs *pcs)
 {
 	return container_of(pcs, struct axienet_local, pcs);
@@ -2792,7 +2811,6 @@ static int axienet_probe(struct platform_device *pdev)
 
 	SET_NETDEV_DEV(ndev, &pdev->dev);
 	ndev->features = NETIF_F_SG;
-	ndev->ethtool_ops = &axienet_ethtool_ops;
 
 	/* MTU range: 64 - 9000 */
 	ndev->min_mtu = 64;
@@ -3021,10 +3039,13 @@ static int axienet_probe(struct platform_device *pdev)
 		lp->use_dmaengine = 1;
 	}
 
-	if (lp->use_dmaengine)
+	if (lp->use_dmaengine) {
 		ndev->netdev_ops = &axienet_netdev_dmaengine_ops;
-	else
+		ndev->ethtool_ops = &axienet_ethtool_dmaengine_ops;
+	} else {
 		ndev->netdev_ops = &axienet_netdev_ops;
+		ndev->ethtool_ops = &axienet_ethtool_ops;
+	}
 	/* Check for Ethernet core IRQ (optional) */
 	if (lp->eth_irq <= 0)
 		dev_info(&pdev->dev, "Ethernet core IRQ not defined\n");
-- 
2.25.1

Re: [PATCH net next] net: axienet: Use dedicated ethtool_ops for the dmaengine path

[Jacob Keller]

On 6/1/2026 5:44 AM, Suraj Gupta wrote:
> The dmaengine path shares ethtool_ops with the legacy AXI DMA path,
> including .get_coalesce/.set_coalesce that poke XAXIDMA_*_CR_OFFSET
> directly. In dmaengine mode lp->dma_regs is not mapped by axienet, so
> those ethtool calls touch unmapped/unrelated memory and report values
> unrelated to the channel actually in use.
> 
> .get_ringparam/.set_ringparam only touch lp->rx_bd_num/lp->tx_bd_num,
> fields used only by the legacy path for BD ring sizing. In dmaengine
> mode the descriptor ring is owned by the dmaengine provider and these
> fields are not consulted, so reporting them is misleading.
> 
> No dmaengine API exists today to query or program either coalescing
> or ring size on behalf of the client, so neither can be exposed
> meaningfully in dmaengine mode.
> 
> Add axienet_ethtool_dmaengine_ops without the coalesce and ringparam
> hooks. Also move the ethtool_ops assignment from early probe into the
> if/else alongside netdev_ops, so the legacy and dmaengine paths pick
> their respective ops in one place. No functional change for the
> legacy DMA path.
> 
> Signed-off-by: Suraj Gupta <suraj.gupta2@amd.com>
> ---

Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>

>  .../net/ethernet/xilinx/xilinx_axienet_main.c | 27 ++++++++++++++++---
>  1 file changed, 24 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
> index 263c4b67fd5a..fcf517069d16 100644
> --- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
> +++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
> @@ -2536,6 +2536,25 @@ static const struct ethtool_ops axienet_ethtool_ops = {
>  	.get_rmon_stats = axienet_ethtool_get_rmon_stats,
>  };
>  
> +static const struct ethtool_ops axienet_ethtool_dmaengine_ops = {
> +	.get_drvinfo    = axienet_ethtools_get_drvinfo,
> +	.get_regs_len   = axienet_ethtools_get_regs_len,
> +	.get_regs       = axienet_ethtools_get_regs,
> +	.get_link       = ethtool_op_get_link,
> +	.get_pauseparam = axienet_ethtools_get_pauseparam,
> +	.set_pauseparam = axienet_ethtools_set_pauseparam,
> +	.get_link_ksettings = axienet_ethtools_get_link_ksettings,
> +	.set_link_ksettings = axienet_ethtools_set_link_ksettings,
> +	.nway_reset	= axienet_ethtools_nway_reset,
> +	.get_ethtool_stats = axienet_ethtools_get_ethtool_stats,
> +	.get_strings    = axienet_ethtools_get_strings,
> +	.get_sset_count = axienet_ethtools_get_sset_count,
> +	.get_pause_stats = axienet_ethtools_get_pause_stats,
> +	.get_eth_mac_stats = axienet_ethtool_get_eth_mac_stats,
> +	.get_eth_ctrl_stats = axienet_ethtool_get_eth_ctrl_stats,
> +	.get_rmon_stats = axienet_ethtool_get_rmon_stats,
> +};
> +
>  static struct axienet_local *pcs_to_axienet_local(struct phylink_pcs *pcs)
>  {
>  	return container_of(pcs, struct axienet_local, pcs);
> @@ -2792,7 +2811,6 @@ static int axienet_probe(struct platform_device *pdev)
>  
>  	SET_NETDEV_DEV(ndev, &pdev->dev);
>  	ndev->features = NETIF_F_SG;
> -	ndev->ethtool_ops = &axienet_ethtool_ops;
>  
>  	/* MTU range: 64 - 9000 */
>  	ndev->min_mtu = 64;
> @@ -3021,10 +3039,13 @@ static int axienet_probe(struct platform_device *pdev)
>  		lp->use_dmaengine = 1;
>  	}
>  
> -	if (lp->use_dmaengine)
> +	if (lp->use_dmaengine) {
>  		ndev->netdev_ops = &axienet_netdev_dmaengine_ops;
> -	else
> +		ndev->ethtool_ops = &axienet_ethtool_dmaengine_ops;
> +	} else {
>  		ndev->netdev_ops = &axienet_netdev_ops;
> +		ndev->ethtool_ops = &axienet_ethtool_ops;
> +	}
>  	/* Check for Ethernet core IRQ (optional) */
>  	if (lp->eth_irq <= 0)
>  		dev_info(&pdev->dev, "Ethernet core IRQ not defined\n");

Re: [PATCH net next] net: axienet: Use dedicated ethtool_ops for the dmaengine path

[patchwork-bot+netdevbpf]

Hello:

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

On Mon, 1 Jun 2026 18:14:54 +0530 you wrote:
> The dmaengine path shares ethtool_ops with the legacy AXI DMA path,
> including .get_coalesce/.set_coalesce that poke XAXIDMA_*_CR_OFFSET
> directly. In dmaengine mode lp->dma_regs is not mapped by axienet, so
> those ethtool calls touch unmapped/unrelated memory and report values
> unrelated to the channel actually in use.
> 
> .get_ringparam/.set_ringparam only touch lp->rx_bd_num/lp->tx_bd_num,
> fields used only by the legacy path for BD ring sizing. In dmaengine
> mode the descriptor ring is owned by the dmaengine provider and these
> fields are not consulted, so reporting them is misleading.
> 
> [...]

Here is the summary with links:
  - [net,next] net: axienet: Use dedicated ethtool_ops for the dmaengine path
    https://git.kernel.org/netdev/net-next/c/c1c3d01e3a90

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