[PATCH net-next v2 0/2] net: axienet: allow instantiation without a device tree

[PATCH net-next v2 0/2] net: axienet: allow instantiation without a device tree

[Sagi Maimon]

The ADVA TimeCard X2 is a PCIe timing card driven by ptp_ocp.  Its FPGA
contains an AXI Ethernet MAC with an AXI DMA engine.  The plan discussed
on netdev [1] is for ptp_ocp to register the MAC as a child platform
device described by a software node, so that this driver runs it rather
than a copy in ptp_ocp; Vadim Fedorenko agreed that a platform device is
the way to go.

These two patches are what axienet needs for that, apart from one more
change - letting the DMA device differ from the platform device - which
has no user without ptp_ocp and will be posted together with the ptp_ocp
patches:

1/2 reads the probe-time configuration through the device_property and
    fwnode APIs, so it can come from a software node.  No change for
    device tree users.
2/2 adds a MODULE_ALIAS, so a device registered by name autoloads the
    module.

Both were tested on the X2 with a local ptp_ocp change.

[1] https://lore.kernel.org/netdev/CAMuE1bHr84n5LKcj02HmMR5XefPtvQ-UEJAx7+UKfYvRgS0eDg@mail.gmail.com/

Changes in v2:
- 1/2: convert only the four lookups a software node provides:
  xlnx,rxmem, phy-mode, the MAC address and the PHY connection.
  xlnx,switch-x-sgmii and the deprecated xlnx,phy-type stay OF only, as
  Andrew asked, and so do xlnx,txcsum, xlnx,rxcsum and the "dmas" test,
  which the X2 does not use either.
- 1/2: v1's hardware test was not re-run.  The X2's software node sets
  none of the lookups that went back to OF, so each finds nothing with
  either version.
- 2/2: unchanged.
- v1: https://lore.kernel.org/netdev/20260927073843.346910-1-maimon.sagi@gmail.com/

Sagi Maimon (2):
  net: axienet: use device_property and fwnode APIs for probe-time
    config
  net: axienet: add a MODULE_ALIAS for platform instantiation

 .../net/ethernet/xilinx/xilinx_axienet_main.c | 22 ++++++++++++++-----
 1 file changed, 16 insertions(+), 6 deletions(-)


base-commit: 014d795c73837ea2339a4ea8e8f82c6e959b845d
-- 
2.47.0

[PATCH net-next v2 1/2] net: axienet: use device_property and fwnode APIs for probe-time config

[Sagi Maimon]

axienet_probe() reads its configuration exclusively through the of_*
API, so the driver can only be instantiated from a device tree node.

The ADVA TimeCard X2 is a PCIe timing card, driven by ptp_ocp, whose
FPGA contains an AXI Ethernet MAC with an AXI DMA engine.  ptp_ocp is
going to register the MAC as a child platform device described by a
software node, so that this driver runs it rather than a copy in
ptp_ocp.  That needs the probe-time configuration to be readable from a
software node.

Convert only the lookups such a caller needs: "xlnx,rxmem", which probe
requires; "phy-mode"; the MAC address; and the PHY connection, which is
where phylink finds a "fixed-link" child.  On a device tree this is a
no-op: device_property_read_u32() dispatches to the of_* implementation
when dev->fwnode is an OF node, phylink_of_phy_connect() is a wrapper
around phylink_fwnode_phy_connect(), and device_get_phy_mode() and
device_get_mac_address() are thin fwnode wrappers around the same
lookups.  device_get_mac_address() also keeps the "mac-address" nvmem
cell fallback that of_get_mac_address() has.

Note that device_get_phy_mode() returns the mode as a positive value
rather than through an out parameter, so the error test changes from
"if (ret)" to "if (ret < 0)".

Everything else stays on the OF API.  "xlnx,txcsum", "xlnx,rxcsum" and
"xlnx,switch-x-sgmii" are optional, and "xlnx,phy-type" is deprecated in
favour of "phy-mode".  Without an OF node each of those lookups finds
nothing, exactly as for a device tree node that leaves them out: no
checksum offload, no runtime SGMII/1000BASE-X switching and a
fall-through to "phy-mode".  The "dmas" test likewise selects the
built-in AXI DMA engine.  The "axistream-connected", "pcs-handle" and
"phy-handle" phandle lookups stay too: a caller without an OF node takes
neither branch, getting its DMA registers from its own platform
resources, and it cannot use the SGMII and 1000BASE-X modes, which need
a PCS.

No functional change intended.

Tested on the X2 with a local ptp_ocp change, with CONFIG_OF disabled
and enabled: the interface probes with its configuration read from the
software node and passes traffic.  No device tree board was available,
so the claim that this is a no-op for device tree users rests on the
dispatch described above.

Assisted-by: LLM sparse
Signed-off-by: Sagi Maimon <maimon.sagi@gmail.com>
---
 drivers/net/ethernet/xilinx/xilinx_axienet_main.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
index 782f903d318f..0f2170af9b10 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
@@ -33,6 +33,7 @@
 #include <linux/of_irq.h>
 #include <linux/of_address.h>
 #include <linux/platform_device.h>
+#include <linux/property.h>
 #include <linux/skbuff.h>
 #include <linux/math64.h>
 #include <linux/phy.h>
@@ -1671,7 +1672,7 @@ static int axienet_open(struct net_device *ndev)
 	ret = axienet_device_reset(ndev);
 	axienet_unlock_mii(lp);
 
-	ret = phylink_of_phy_connect(lp->phylink, lp->dev->of_node, 0);
+	ret = phylink_fwnode_phy_connect(lp->phylink, dev_fwnode(lp->dev), 0);
 	if (ret) {
 		dev_err(lp->dev, "phylink_of_phy_connect() failed: %d\n", ret);
 		return ret;
@@ -2898,7 +2899,7 @@ static int axienet_probe(struct platform_device *pdev)
 	 * Here we check for memory allocated for Rx/Tx in the hardware from
 	 * the device-tree and accordingly set flags.
 	 */
-	ret = of_property_read_u32(pdev->dev.of_node, "xlnx,rxmem", &lp->rxmem);
+	ret = device_property_read_u32(&pdev->dev, "xlnx,rxmem", &lp->rxmem);
 	if (ret)
 		return dev_err_probe(&pdev->dev, ret,
 				     "failed to read xlnx,rxmem property\n");
@@ -2930,9 +2931,10 @@ static int axienet_probe(struct platform_device *pdev)
 			return -EINVAL;
 		}
 	} else {
-		ret = of_get_phy_mode(pdev->dev.of_node, &lp->phy_mode);
-		if (ret)
+		ret = device_get_phy_mode(&pdev->dev);
+		if (ret < 0)
 			return ret;
+		lp->phy_mode = ret;
 	}
 	if (lp->switch_x_sgmii && lp->phy_mode != PHY_INTERFACE_MODE_SGMII &&
 	    lp->phy_mode != PHY_INTERFACE_MODE_1000BASEX) {
@@ -3060,7 +3062,7 @@ static int axienet_probe(struct platform_device *pdev)
 		dev_info(&pdev->dev, "Ethernet core IRQ not defined\n");
 
 	/* Retrieve the MAC address */
-	ret = of_get_mac_address(pdev->dev.of_node, mac_addr);
+	ret = device_get_mac_address(&pdev->dev, mac_addr);
 	if (!ret) {
 		axienet_set_mac_address(ndev, mac_addr);
 	} else {
-- 
2.47.0

[PATCH net-next v2 2/2] net: axienet: add a MODULE_ALIAS for platform instantiation

[Sagi Maimon]

The module is built as xilinx_emac.ko while the platform driver it
registers is named xilinx_axienet.  For a device tree instantiation that
mismatch does not matter, because the OF match table supplies the
modalias udev uses to load the module.

A device registered by name from another driver has no OF node and so no
OF modalias.  It advertises platform:xilinx_axienet, which matches
nothing, and the device is created but never bound.  ptp_ocp is going to
register the AXI Ethernet MAC of the ADVA TimeCard X2 this way, as a
child platform device described by a software node.

Add the corresponding MODULE_ALIAS so name-based instantiation can
autoload the driver, and give the name a define now that it is used in
more than one place.

Tested on the X2 with a local ptp_ocp change: with this patch the module
is loaded and binds when ptp_ocp registers the device; without it the
device stays unbound until xilinx_emac is loaded by hand.

Assisted-by: LLM sparse
Signed-off-by: Sagi Maimon <maimon.sagi@gmail.com>
---
 drivers/net/ethernet/xilinx/xilinx_axienet_main.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
index 0f2170af9b10..ae8ed4495ea0 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
@@ -48,6 +48,7 @@
 #include "xilinx_axienet.h"
 
 /* Descriptors defines for Tx and Rx DMA */
+#define AXIENET_DRIVER_NAME		"xilinx_axienet"
 #define TX_BD_NUM_DEFAULT		128
 #define RX_BD_NUM_DEFAULT		1024
 #define TX_BD_NUM_MIN			(MAX_SKB_FRAGS + 1)
@@ -3223,7 +3224,7 @@ static struct platform_driver axienet_driver = {
 	.remove = axienet_remove,
 	.shutdown = axienet_shutdown,
 	.driver = {
-		 .name = "xilinx_axienet",
+		 .name = AXIENET_DRIVER_NAME,
 		 .pm = &axienet_pm_ops,
 		 .of_match_table = axienet_of_match,
 	},
@@ -3231,6 +3232,13 @@ static struct platform_driver axienet_driver = {
 
 module_platform_driver(axienet_driver);
 
+/* The module is named xilinx_emac, the platform driver xilinx_axienet.  A
+ * device registered by name rather than from firmware advertises a
+ * platform:xilinx_axienet modalias, which without this matches no module:
+ * udev cannot autoload the driver and the device stays unbound.
+ */
+MODULE_ALIAS("platform:" AXIENET_DRIVER_NAME);
+
 MODULE_DESCRIPTION("Xilinx Axi Ethernet driver");
 MODULE_AUTHOR("Xilinx");
 MODULE_LICENSE("GPL");
-- 
2.47.0

Re: [PATCH net-next v2 1/2] net: axienet: use device_property and fwnode APIs for probe-time config

[Andrew Lunn]

On Mon, Sep 28, 2026 at 10:42:55AM +0300, Sagi Maimon wrote:
> axienet_probe() reads its configuration exclusively through the of_*
> API, so the driver can only be instantiated from a device tree node.
> 
> The ADVA TimeCard X2 is a PCIe timing card, driven by ptp_ocp, whose
> FPGA contains an AXI Ethernet MAC with an AXI DMA engine.  ptp_ocp is
> going to register the MAC as a child platform device described by a
> software node, so that this driver runs it rather than a copy in
> ptp_ocp.  That needs the probe-time configuration to be readable from a
> software node.
> 
> Convert only the lookups such a caller needs: "xlnx,rxmem", which probe
> requires; "phy-mode"; the MAC address; and the PHY connection, which is
> where phylink finds a "fixed-link" child.  On a device tree this is a
> no-op: device_property_read_u32() dispatches to the of_* implementation
> when dev->fwnode is an OF node, phylink_of_phy_connect() is a wrapper
> around phylink_fwnode_phy_connect(), and device_get_phy_mode() and
> device_get_mac_address() are thin fwnode wrappers around the same
> lookups.  device_get_mac_address() also keeps the "mac-address" nvmem
> cell fallback that of_get_mac_address() has.
> 
> Note that device_get_phy_mode() returns the mode as a positive value
> rather than through an out parameter, so the error test changes from
> "if (ret)" to "if (ret < 0)".
> 
> Everything else stays on the OF API.  "xlnx,txcsum", "xlnx,rxcsum" and
> "xlnx,switch-x-sgmii" are optional, and "xlnx,phy-type" is deprecated in
> favour of "phy-mode".  Without an OF node each of those lookups finds
> nothing, exactly as for a device tree node that leaves them out: no
> checksum offload, no runtime SGMII/1000BASE-X switching and a
> fall-through to "phy-mode".  The "dmas" test likewise selects the
> built-in AXI DMA engine.  The "axistream-connected", "pcs-handle" and
> "phy-handle" phandle lookups stay too: a caller without an OF node takes
> neither branch, getting its DMA registers from its own platform
> resources, and it cannot use the SGMII and 1000BASE-X modes, which need
> a PCS.
> 
> No functional change intended.
> 
> Tested on the X2 with a local ptp_ocp change, with CONFIG_OF disabled
> and enabled: the interface probes with its configuration read from the
> software node and passes traffic.  No device tree board was available,
> so the claim that this is a no-op for device tree users rests on the
> dispatch described above.
> 
> Assisted-by: LLM sparse
> Signed-off-by: Sagi Maimon <maimon.sagi@gmail.com>

Lets see what the AI reviewer says about this, but it looks O.K. to me:

Reviewed-by: Andrew Lunn <andrew@lunn.ch>

    Andrew