On 10/2/2020 4:05 AM, Grant Likely wrote:

On 30/09/2020 17:04, Calvin Johnson wrote:

quoted

Extract phy_id from compatible string. This will be used by
fwnode_mdiobus_register_phy() to create phy device using the
phy_id.

Signed-off-by: Calvin Johnson <redacted>
---

  drivers/net/phy/phy_device.c | 32 +++++++++++++++++++++++++++++++-
  include/linux/phy.h          |  5 +++++
  2 files changed, 36 insertions(+), 1 deletion(-)

diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index c4aec56d0a95..162abde6223d 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c

@@ -9,6 +9,7 @@

  #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+#include <linux/acpi.h>
  #include <linux/bitmap.h>
  #include <linux/delay.h>
  #include <linux/errno.h>

@@ -845,6 +846,27 @@ static int get_phy_c22_id(struct mii_bus *bus, 

int addr, u32 *phy_id)
      return 0;
  }
+/* Extract the phy ID from the compatible string of the form
+ * ethernet-phy-idAAAA.BBBB.
+ */
+int fwnode_get_phy_id(struct fwnode_handle *fwnode, u32 *phy_id)
+{
+    unsigned int upper, lower;
+    const char *cp;
+    int ret;
+
+    ret = fwnode_property_read_string(fwnode, "compatible", &cp);
+    if (ret)
+        return ret;
+
+    if (sscanf(cp, "ethernet-phy-id%4x.%4x", &upper, &lower) == 2) {
+        *phy_id = ((upper & 0xFFFF) << 16) | (lower & 0xFFFF);
+        return 0;
+    }
+    return -EINVAL;
+}
+EXPORT_SYMBOL(fwnode_get_phy_id);

This block, and the changes in patch 4 duplicate functions from 
drivers/of/of_mdio.c, but it doesn't refactor anything in 
drivers/of/of_mdio.c to use the new path. Is your intent to bring all of 
the parsing in these functions of "compatible" into the ACPI code path?

If so, then the existing code path needs to be refactored to work with 
fwnode_handle instead of device_node.

If not, then the DT path in these functions should call out to of_mdio, 
while the ACPI path only does what is necessary.

Rob has been asking before to have drivers/of/of_mdio.c be merged or at 
least relocated within drivers/net/phy where it would naturally belong. 
As a preliminary step towards ACPI support that would seem reasonable to do.

Then, as Grant suggests you can start re-factoring as much as possible 
with using fwnode_handle.
-- 
Florian