Fix Arm Ltd board node name collisions

The node name changes in commits 25b892b583cc ("ARM: dts: arm: Update register-bit-led nodes 'reg' and node names") and 2d3de197a818 ("ARM:
dts: arm: Update ICST clock nodes 'reg' and node names") caused name collisions with the kernel's handling of ICST clock names and platform device names. This series fixes the kernel side to handle the DT changes.

This does break DT ABI compatibility which is partially mitigated if these changes are backported to stable. The alternative is reverting the referenced DT changes and coming up with different node names without unit-addresses. Ultimately, ABI issues are platform maintainers' decision, not mine.

Rob

Rob Herring (2):
of: Support using 'mask' in making device bus id
clk: versatile: clk-icst: Ensure clock names are unique

 drivers/clk/versatile/clk-icst.c |  6 ++++--
 drivers/of/platform.c            | 10 ++++++++--
 2 files changed, 12 insertions(+), 4 deletions(-)

2.32.0

Commit 25b892b583cc ("ARM: dts: arm: Update register-bit-led nodes 'reg' and node names") added a 'reg' property to nodes. This change has the side effect of changing how the kernel generates the device name. The assumption was a translatable 'reg' address is unique. However, in the case of the register-bit-led binding (and a few others) that is not the case. The 'mask' property must also be used in this case to make a unique device name.

---
This should be applied to stable to minimize DT ABI breakage.

---
 drivers/of/platform.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

index 07813fb1ef37..b3faf89744aa 100644
@@ -76,6 +76,7 @@ static void of_device_make_bus_id(struct device *dev)
 	struct device_node *node = dev->of_node;
 	const __be32 *reg;
 	u64 addr;
+	u32 mask;
 
 	/* Construct the name, using parent nodes if necessary to ensure uniqueness */
 	while (node->parent) {

@@ -85,8 +86,13 @@ static void of_device_make_bus_id(struct device *dev)
 		 */
 		reg = of_get_property(node, "reg", NULL);
 		if (reg && (addr = of_translate_address(node, reg)) != OF_BAD_ADDR) {
-			dev_set_name(dev, dev_name(dev) ? "%llx.%pOFn:%s" : "%llx.%pOFn",
-				     addr, node, dev_name(dev));
+			if (!of_property_read_u32(node, "mask", &mask))
+				dev_set_name(dev, dev_name(dev) ? "%llx.%x.%pOFn:%s" : "%llx.%x.%pOFn",
+					     addr, ffs(mask) - 1, node, dev_name(dev));
+
+			else
+				dev_set_name(dev, dev_name(dev) ? "%llx.%pOFn:%s" : "%llx.%pOFn",
+					     addr, node, dev_name(dev));
 			return;
 		}
 

2.32.0

Yours,
Linus Walleij

Commit 2d3de197a818 ("ARM: dts: arm: Update ICST clock nodes 'reg' and node names") moved to using generic node names. That results in trying to register multiple clocks with the same name. Fix this by including the unit-address in the clock name.

---
This should be applied to stable to minimize DT ABI breakage.

---
 drivers/clk/versatile/clk-icst.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

index 77fd0ecaf155..d52f976dc875 100644
@@ -484,7 +484,7 @@ static void __init of_syscon_icst_setup(struct device_node *np)
 	struct device_node *parent;
 	struct regmap *map;
 	struct clk_icst_desc icst_desc;
-	const char *name = np->name;
+	const char *name;
 	const char *parent_name;
 	struct clk *regclk;
 	enum icst_control_type ctype;

@@ -533,15 +533,17 @@ static void __init of_syscon_icst_setup(struct device_node *np)
 		icst_desc.params = &icst525_apcp_cm_params;
 		ctype = ICST_INTEGRATOR_CP_CM_MEM;
 	} else {
-		pr_err("unknown ICST clock %s\n", name);
+		pr_err("unknown ICST clock %pOF\n", np);
 		return;
 	}
 
 	/* Parent clock name is not the same as node parent */
 	parent_name = of_clk_get_parent_name(np, 0);
+	name = kasprintf(GFP_KERNEL, "%pOFP", np);
 
 	regclk = icst_clk_setup(NULL, &icst_desc, name, parent_name, map, ctype);
 	if (IS_ERR(regclk)) {
+		kfree(name);
 		pr_err("error setting up syscon ICST clock %s\n", name);
 		return;
 	}

2.32.0

Yours,
Linus Walleij

Quoting Rob Herring (2021-11-09 08:46:50)

I don't mind breaking the ABI for these boards since everyone I know rebuild the DT with the kernel and boot. I don't know if other users flash it though, so check what Sudeep says.

Yours,
Linus Walleij