Skip to contentSkip to conversation

of_devlink: Minor fixes and new properties

PATCH In review Saravana Kannan 3 participants 9 messages Nov 5, 2019 – Nov 7, 2019 Device tree

lore
Share & tools
lore
#0 lore raw

cover of_devlink: Minor fixes and new properties

Sending again since the cover letter missed everyone/mailing lists.

First two patches are just code clean up and logging with no functional difference. The 3rd patch adds support for the following DT properties:
iommus, mboxes and io-channels.

These patches are on top of driver-core-next since that's where the rest of the of_devlink patches are.

Greg and Rob,
On a side note, I was wondering if I should rename the of_devlink kernel command line to fw_devlink and move it out of drivers/of/property.c and into drivers/base/core.c. This feature isn't really limited to devicetree, so I don't see a need to have devicetree specific kernel command line option. Please let me know if that sounds okay to you.

Thanks,
Saravana

Saravana Kannan (3):
of: property: Minor style clean up of of_link_to_phandle()
of: property: Make it easy to add device links from DT properties
of: property: Add device link support for iommus, mboxes and
io-channels

 drivers/of/property.c | 78 +++++++++++++++++++++++++++++++++----------
 1 file changed, 61 insertions(+), 17 deletions(-)
Signature · 1 line
2.24.0.rc1.363.gb1bccd3e3d-goog
#1 lore raw

PATCH 1/3 of: property: Minor style clean up of of_link_to_phandle()

Adding a debug log instead of silently ignoring a phandle for an early device. Also, return the right error code instead of 0 even though the actual execution flow won't change.

---
 drivers/of/property.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/of/property.c +8 -2
index e225ab17f598..fbc201330ba0 100644
@@ -1051,8 +1051,14 @@ static int of_link_to_phandle(struct device *dev, struct device_node *sup_np,
 	sup_dev = get_dev_from_fwnode(&sup_np->fwnode);
 	is_populated = of_node_check_flag(sup_np, OF_POPULATED);
 	of_node_put(sup_np);
-	if (!sup_dev)
-		return is_populated ? 0 : -EAGAIN;
+	if (!sup_dev && is_populated) {
+		/* Early device without struct device. */
+		dev_dbg(dev, "Not linking to %pOFP - No struct device\n",
+			sup_np);
+		return -ENODEV;
+	} else if (!sup_dev) {
+		return -EAGAIN;
+	}
 	if (!device_link_add(dev, sup_dev, dl_flags))
 		ret = -EAGAIN;
 	put_device(sup_dev);
Signature · 1 line
2.24.0.rc1.363.gb1bccd3e3d-goog
maintainer Reviewed-by on v1 1/3
#2 lore raw

Re: [PATCH v1 1/3] of: property: Minor style clean up of of_link_to_phandle()

Quoted 9 lines · Saravana Kannan

Adding a debug log instead of silently ignoring a phandle for an early device. Also, return the right error code instead of 0 even though the actual execution flow won't change.

---
 drivers/of/property.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
#3 lore raw

PATCH 2/3 of: property: Make it easy to add device links from DT properties

In reply to Saravana Kannan

Add a DEFINE_SIMPLE_PROP macro to make it easy to add support for simple properties with fixed names that just list phandles and phandle args.

Add a DEFINE_SUFFIX_PROP macro to make it easy to add support for properties with fixes suffix that just list phandles and phandle args.

---
 drivers/of/property.c | 62 ++++++++++++++++++++++++++++++++-----------
 1 file changed, 47 insertions(+), 15 deletions(-)
drivers/of/property.c +47 -15
index fbc201330ba0..812b69a029d1 100644
@@ -1101,17 +1101,11 @@ static struct device_node *parse_prop_cells(struct device_node *np,
 	return sup_args.np;
 }
 
-static struct device_node *parse_clocks(struct device_node *np,
-					const char *prop_name, int index)
-{
-	return parse_prop_cells(np, prop_name, index, "clocks", "#clock-cells");
-}
-
-static struct device_node *parse_interconnects(struct device_node *np,
-					       const char *prop_name, int index)
-{
-	return parse_prop_cells(np, prop_name, index, "interconnects",
-				"#interconnect-cells");
+#define DEFINE_SIMPLE_PROP(fname, name, cells)				  \
+static struct device_node *parse_##fname(struct device_node *np,	  \
+					const char *prop_name, int index) \
+{									  \
+	return parse_prop_cells(np, prop_name, index, name, cells);	  \
 }
 
 static int strcmp_suffix(const char *str, const char *suffix)
@@ -1125,13 +1119,47 @@ static int strcmp_suffix(const char *str, const char *suffix)
 	return strcmp(str + len - suffix_len, suffix);
 }
 
-static struct device_node *parse_regulators(struct device_node *np,
-					    const char *prop_name, int index)
+/**
+ * parse_suffix_prop_cells - Suffix property parsing function for suppliers
+ *
+ * @np:		Pointer to device tree node containing a list
+ * @prop_name:	Name of property to be parsed. Expected to hold phandle values
+ * @index:	For properties holding a list of phandles, this is the index
+ *		into the list.
+ * @suffix:	Property suffix that is known to contain list of phandle(s) to
+ *		supplier(s)
+ * @cells_name:	property name that specifies phandles' arguments count
+ *
+ * This is a helper function to parse properties that have a known fixed suffix
+ * and are a list of phandles and phandle arguments.
+ *
+ * Returns:
+ * - phandle node pointer with refcount incremented. Caller must of_node_put()
+ *   on it when done.
+ * - NULL if no phandle found at index
+ */
+static struct device_node *parse_suffix_prop_cells(struct device_node *np,
+					    const char *prop_name, int index,
+					    const char *suffix,
+					    const char *cells_name)
 {
-	if (index || strcmp_suffix(prop_name, "-supply"))
+	struct of_phandle_args sup_args;
+
+	if (strcmp_suffix(prop_name, suffix))
 		return NULL;
 
-	return of_parse_phandle(np, prop_name, 0);
+	if (of_parse_phandle_with_args(np, prop_name, cells_name, index,
+				       &sup_args))
+		return NULL;
+
+	return sup_args.np;
+}
+
+#define DEFINE_SUFFIX_PROP(fname, suffix, cells)			     \
+static struct device_node *parse_##fname(struct device_node *np,	     \
+					const char *prop_name, int index)    \
+{									     \
+	return parse_suffix_prop_cells(np, prop_name, index, suffix, cells); \
 }
 
 /**
@@ -1155,6 +1183,10 @@ struct supplier_bindings {
 					  const char *prop_name, int index);
 };
 
+DEFINE_SIMPLE_PROP(clocks, "clocks", "#clock-cells")
+DEFINE_SIMPLE_PROP(interconnects, "interconnects", "#interconnect-cells")
+DEFINE_SUFFIX_PROP(regulators, "-supply", NULL)
+
 static const struct supplier_bindings of_supplier_bindings[] = {
 	{ .parse_prop = parse_clocks, },
 	{ .parse_prop = parse_interconnects, },
Signature · 1 line
2.24.0.rc1.363.gb1bccd3e3d-goog
maintainer Reviewed-by on v1 2/3
#4 lore raw

Re: [PATCH v1 2/3] of: property: Make it easy to add device links from DT properties

Quoted 11 lines · Saravana Kannan

Add a DEFINE_SIMPLE_PROP macro to make it easy to add support for simple properties with fixed names that just list phandles and phandle args.

Add a DEFINE_SUFFIX_PROP macro to make it easy to add support for properties with fixes suffix that just list phandles and phandle args.

---
 drivers/of/property.c | 62 ++++++++++++++++++++++++++++++++-----------
 1 file changed, 47 insertions(+), 15 deletions(-)
#5 lore raw

PATCH 3/3 of: property: Add device link support for iommus, mboxes and io-channels

In reply to Saravana Kannan

Add support for creating device links out of more DT properties.

---
 drivers/of/property.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/of/property.c +6 -0
index 812b69a029d1..0fa04692e3cc 100644
@@ -1185,11 +1185,17 @@ struct supplier_bindings {
 
 DEFINE_SIMPLE_PROP(clocks, "clocks", "#clock-cells")
 DEFINE_SIMPLE_PROP(interconnects, "interconnects", "#interconnect-cells")
+DEFINE_SIMPLE_PROP(iommus, "iommus", "#iommu-cells")
+DEFINE_SIMPLE_PROP(mboxes, "mboxes", "#mbox-cells")
+DEFINE_SIMPLE_PROP(io_channels, "io-channel", "#io-channel-cells")
 DEFINE_SUFFIX_PROP(regulators, "-supply", NULL)
 
 static const struct supplier_bindings of_supplier_bindings[] = {
 	{ .parse_prop = parse_clocks, },
 	{ .parse_prop = parse_interconnects, },
+	{ .parse_prop = parse_iommus, },
+	{ .parse_prop = parse_mboxes, },
+	{ .parse_prop = parse_io_channels, },
 	{ .parse_prop = parse_regulators, },
 	{}
 };
Signature · 1 line
2.24.0.rc1.363.gb1bccd3e3d-goog
#7 lore raw
In reply to Saravana Kannan
Quoted 10 lines · Saravana Kannan

Sending again since the cover letter missed everyone/mailing lists.

First two patches are just code clean up and logging with no functional difference. The 3rd patch adds support for the following DT properties:
iommus, mboxes and io-channels.

These patches are on top of driver-core-next since that's where the rest of the of_devlink patches are.

Quoted 6 lines

Greg and Rob,
On a side note, I was wondering if I should rename the of_devlink kernel command line to fw_devlink and move it out of drivers/of/property.c and into drivers/base/core.c. This feature isn't really limited to devicetree, so I don't see a need to have devicetree specific kernel command line option. Please let me know if that sounds okay to you.

Hi Rob,

Thanks for the reviews. Can you let me know what you think of this too?

Rob/Greg,

If I rename of_devlink to fw_devlink, I might also make it a setting like fw_devlink=none/permissive/enforce
- none would be same as disabled completely.
- permissive would use SYNC_STATE_ONLY for all device links created by firmware. So it won't block any probes even with cycles but sync_state() will still work correctly.
- enforce would be the current "of_devlinkg=1" behavior where direct dependencies would block probing and the child dependencies would use SYNC_STATE_ONLY.

-Saravana

#8 lore raw
Quoted 24 lines · Saravana Kannan
Quoted 10 lines · Saravana Kannan

Sending again since the cover letter missed everyone/mailing lists.

First two patches are just code clean up and logging with no functional difference. The 3rd patch adds support for the following DT properties:
iommus, mboxes and io-channels.

These patches are on top of driver-core-next since that's where the rest of the of_devlink patches are.

Quoted 6 lines

Greg and Rob,
On a side note, I was wondering if I should rename the of_devlink kernel command line to fw_devlink and move it out of drivers/of/property.c and into drivers/base/core.c. This feature isn't really limited to devicetree, so I don't see a need to have devicetree specific kernel command line option. Please let me know if that sounds okay to you.

Hi Rob,

Thanks for the reviews. Can you let me know what you think of this too?

Rob/Greg,

Quoted 9 lines

If I rename of_devlink to fw_devlink, I might also make it a setting like fw_devlink=none/permissive/enforce
- none would be same as disabled completely.
- permissive would use SYNC_STATE_ONLY for all device links created by firmware. So it won't block any probes even with cycles but sync_state() will still work correctly.
- enforce would be the current "of_devlinkg=1" behavior where direct dependencies would block probing and the child dependencies would use SYNC_STATE_ONLY.

Renaming makes sense to me, and all of the above is fine as well.

thanks,

greg k-h