On Wed, Jul 1, 2020 at 9:13 AM Calvin Johnson
<calvin.johnson@xxxxxxxxxxx> wrote:
>
> Modify dpaa2_mac_connect() to support ACPI along with DT.
> Modify dpaa2_mac_get_node() to get the dpmac fwnode from either
> DT or ACPI.
> Replace of_get_phy_mode with fwnode_get_phy_mode to get
> phy-mode for a dpmac_node.
> Define and use helper function find_phy_device() to find phy_dev
> that is later connected to mac->phylink.
...
> #include "dpaa2-eth.h"
> #include "dpaa2-mac.h"
> +#include <linux/acpi.h>
> +#include <linux/platform_device.h>
Can we put (more) generic headers atop of (more) private ones?
...
> + struct fwnode_handle *fsl_mc_fwnode = dev->parent->parent->fwnode;
dev_fwnode() please.
> + struct fwnode_handle *dpmacs, *dpmac = NULL;
> + struct device *fsl_mc = dev->parent->parent;
So. something like
struct device *fsl_mc = dev->parent->parent;
struct fwnode_handle *fsl_mc_fwnode = dev_fwnode(fsl_mc);
...
> + dpmacs = device_get_named_child_node(fsl_mc, "dpmacs");
If you have fwnode, why to use device_* API?
dpmacs = fwnode_get_named_child_node(fsl_mc_fwnode, "dpmacs");
> + if (!dpmacs)
> + return NULL;
> +
> + while ((dpmac = fwnode_get_next_child_node(dpmacs, dpmac))) {
> + err = fwnode_property_read_u32(dpmac, "reg", &id);
> + if (err)
> + continue;
> + if (id == dpmac_id)
> + return dpmac;
> + }
...
> + } else if (is_acpi_node(fsl_mc_fwnode)) {
is_acpi_device_node() ?
> + adev = acpi_find_child_device(ACPI_COMPANION(dev->parent),
> + dpmac_id, false);
> + if (adev)
> + return (&adev->fwnode);
No need to have parentheses. Don't we have some special macro to get
fwnode out of ACPI device?
...
> + err = fwnode_get_phy_mode(dpmac_node);
> + if (err > 0)
> + return err;
Positive?! Why? What's going on here?
...
> + if (is_of_node(dpmac_node))
> + err = phylink_of_phy_connect(mac->phylink,
> + to_of_node(dpmac_node), 0);
> + else if (is_acpi_node(dpmac_node)) {
> + phy_dev = find_phy_device(dpmac_node);
> + if (IS_ERR(phy_dev))
> + goto err_phylink_destroy;
> + err = phylink_connect_phy(mac->phylink, phy_dev);
Can't you rather provide phylink_fwnode_connect_phy API and drop this
conditional tree entirely?
...
> + if (is_of_node(dpmac_node))
Redundant.
> + of_node_put(to_of_node(dpmac_node));
Honestly, looking at this code, I think one needs a bit more time to
get into fwnode paradigm and APIs.
...
> + if (is_of_node(dpmac_node))
Ditto.
> + of_node_put(to_of_node(dpmac_node));
--
With Best Regards,
Andy Shevchenko
