Message ID | 20190202025048.14772-1-mdf@kernel.org |
---|---|
State | Superseded, archived |
Headers | show |
Series | [net-next,RFC/PATCH] net: nixge: Make mdio child node optional | expand |
Context | Check | Description |
---|---|---|
robh/checkpatch | warning | "total: 0 errors, 1 warnings, 0 checks, 81 lines checked" |
On Fri, Feb 01, 2019 at 06:50:48PM -0800, Moritz Fischer wrote: > Make MDIO child optional and only instantiate the > MDIO bus if the child is actually present. > > There are currently no (in-tree) users of this > binding; all (out-of-tree) users use overlays that > get shipped together with the FPGA images that contain > the IP. > > This will significantly increase maintainabilty > of future revisions of this IP. > > Signed-off-by: Moritz Fischer <mdf@kernel.org> > Cc: Andrew Lunn <andrew@lunn.ch> > Cc: Rob Herring <robh+dt@kernel.org> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
On Mon, Feb 04, 2019 at 03:58:54PM +0100, Andrew Lunn wrote: > On Fri, Feb 01, 2019 at 06:50:48PM -0800, Moritz Fischer wrote: > > Make MDIO child optional and only instantiate the > > MDIO bus if the child is actually present. > > > > There are currently no (in-tree) users of this > > binding; all (out-of-tree) users use overlays that > > get shipped together with the FPGA images that contain > > the IP. > > > > This will significantly increase maintainabilty > > of future revisions of this IP. > > > > Signed-off-by: Moritz Fischer <mdf@kernel.org> > > Cc: Andrew Lunn <andrew@lunn.ch> > > Cc: Rob Herring <robh+dt@kernel.org> > > Reviewed-by: Andrew Lunn <andrew@lunn.ch> Thanks. > > Andrew I'll resubmit as a series together with the fixed-link change. Since Dave has applied Alex' patches I'll need to rebase anyways, otherwise it doesn't apply anymore to net-next. Cheers, Moritz
From: Andrew Lunn <andrew@lunn.ch> Date: Mon, 4 Feb 2019 15:58:54 +0100 > On Fri, Feb 01, 2019 at 06:50:48PM -0800, Moritz Fischer wrote: >> Make MDIO child optional and only instantiate the >> MDIO bus if the child is actually present. >> >> There are currently no (in-tree) users of this >> binding; all (out-of-tree) users use overlays that >> get shipped together with the FPGA images that contain >> the IP. >> >> This will significantly increase maintainabilty >> of future revisions of this IP. >> >> Signed-off-by: Moritz Fischer <mdf@kernel.org> >> Cc: Andrew Lunn <andrew@lunn.ch> >> Cc: Rob Herring <robh+dt@kernel.org> > > Reviewed-by: Andrew Lunn <andrew@lunn.ch> Moritz, please repost without the RFC tag and with Andrew's tag added. Thanks.
diff --git a/Documentation/devicetree/bindings/net/nixge.txt b/Documentation/devicetree/bindings/net/nixge.txt index e55af7f0881a..5a0072b69ce7 100644 --- a/Documentation/devicetree/bindings/net/nixge.txt +++ b/Documentation/devicetree/bindings/net/nixge.txt @@ -10,6 +10,9 @@ Required properties: - nvmem-cells: Phandle of nvmem cell containing the MAC address - nvmem-cell-names: Should be "address" +Optional properties: +- mdio subnode to indicate presence of MDIO controller + Examples (10G generic PHY): nixge0: ethernet@40000000 { compatible = "ni,xge-enet-2.00"; @@ -25,8 +28,26 @@ Examples (10G generic PHY): phy-mode = "xgmii"; phy-handle = <ðernet_phy1>; - ethernet_phy1: ethernet-phy@4 { - compatible = "ethernet-phy-ieee802.3-c45"; - reg = <4>; + mdio { + ethernet_phy1: ethernet-phy@4 { + compatible = "ethernet-phy-ieee802.3-c45"; + reg = <4>; + }; }; }; + +Examples (10G generic PHY, no MDIO): + nixge0: ethernet@40000000 { + compatible = "ni,xge-enet-2.00"; + reg = <0x40000000 0x6000>; + + nvmem-cells = <ð1_addr>; + nvmem-cell-names = "address"; + + interrupts = <0 29 IRQ_TYPE_LEVEL_HIGH>, <0 30 IRQ_TYPE_LEVEL_HIGH>; + interrupt-names = "rx", "tx"; + interrupt-parent = <&intc>; + + phy-mode = "xgmii"; + phy-handle = <ðernet_phy1>; + }; diff --git a/drivers/net/ethernet/ni/nixge.c b/drivers/net/ethernet/ni/nixge.c index 1e408d1a9b5f..93d5263cc57e 100644 --- a/drivers/net/ethernet/ni/nixge.c +++ b/drivers/net/ethernet/ni/nixge.c @@ -1230,6 +1230,7 @@ static int nixge_probe(struct platform_device *pdev) struct nixge_priv *priv; struct net_device *ndev; struct resource *dmares; + struct device_node *mn; const u8 *mac_addr; int err; @@ -1286,10 +1287,14 @@ static int nixge_probe(struct platform_device *pdev) priv->coalesce_count_rx = XAXIDMA_DFT_RX_THRESHOLD; priv->coalesce_count_tx = XAXIDMA_DFT_TX_THRESHOLD; - err = nixge_mdio_setup(priv, pdev->dev.of_node); - if (err) { - netdev_err(ndev, "error registering mdio bus"); - goto free_netdev; + mn = of_get_child_by_name(pdev->dev.of_node, "mdio"); + if (mn) { + err = nixge_mdio_setup(priv, mn); + of_node_put(mn); + if (err) { + netdev_err(ndev, "error registering mdio bus"); + goto free_netdev; + } } priv->phy_mode = of_get_phy_mode(pdev->dev.of_node); @@ -1315,7 +1320,8 @@ static int nixge_probe(struct platform_device *pdev) return 0; unregister_mdio: - mdiobus_unregister(priv->mii_bus); + if (priv->mii_bus) + mdiobus_unregister(priv->mii_bus); free_netdev: free_netdev(ndev); @@ -1330,7 +1336,8 @@ static int nixge_remove(struct platform_device *pdev) unregister_netdev(ndev); - mdiobus_unregister(priv->mii_bus); + if (priv->mii_bus) + mdiobus_unregister(priv->mii_bus); free_netdev(ndev);
Make MDIO child optional and only instantiate the MDIO bus if the child is actually present. There are currently no (in-tree) users of this binding; all (out-of-tree) users use overlays that get shipped together with the FPGA images that contain the IP. This will significantly increase maintainabilty of future revisions of this IP. Signed-off-by: Moritz Fischer <mdf@kernel.org> Cc: Andrew Lunn <andrew@lunn.ch> Cc: Rob Herring <robh+dt@kernel.org> --- Hi Rob, Andrew, I know generally changing bindings is a no-no. Ultimately I'm working on adding fixed-link support to this driver, during the review of that Andrew suggested to drop the wrong binding if I don't need legacy behavior. In my case I can pretty much guarantee there's no users, even out of tree due to how we ship the software with our devices (overlays & fpga images always go together). I think in the long this change will make the code easier to maintain and allow to easier support the different variants. Thanks, Moritz --- .../devicetree/bindings/net/nixge.txt | 27 ++++++++++++++++--- drivers/net/ethernet/ni/nixge.c | 19 ++++++++----- 2 files changed, 37 insertions(+), 9 deletions(-)