[PATCH 0/3] of: generic infrastructure fixes

[PATCH 0/3] of: generic infrastructure fixes

[Pantelis Antoniou]

The first patch renames the *_node_sysfs methods to _node_post since
this is more accurate when more work takes place besides sysfs
tweaking when the next patch is using a hashtable for phandles.
This is a win for very large blobs with a large number of phandles.
Current we use an exhaustive search which is not optimal at all.

Finally, a long standing request of applied overlays populating
the symbol list is implemented; this allows staggered overlay
application with a symbol from a previous overlay being used
by another after it.

Pantelis Antoniou (3):
  of: rename *_node_sysfs to _node_post
  of: Support hashtable lookups for phandles
  of: overlay: Pick up label symbols from overlays.

 drivers/of/base.c       |  39 +++++++++++++++---
 drivers/of/dynamic.c    |  18 ++++++---
 drivers/of/of_private.h |  35 ++++++++++++++++-
 drivers/of/overlay.c    | 102 ++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/of/unittest.c   |   4 +-
 include/linux/of.h      |   2 +
 6 files changed, 186 insertions(+), 14 deletions(-)

-- 
1.7.12

[PATCH 1/3] of: rename *_node_sysfs to _node_post

[Pantelis Antoniou]

Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
---
 drivers/of/base.c       |  4 ++--
 drivers/of/dynamic.c    | 10 +++++-----
 drivers/of/of_private.h |  4 ++--
 drivers/of/unittest.c   |  4 ++--
 4 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 9bbca56..20bbc2f 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -156,7 +156,7 @@ int __of_add_property_sysfs(struct device_node *np, struct property *pp)
 	return rc;
 }
 
-int __of_attach_node_sysfs(struct device_node *np)
+int __of_attach_node_post(struct device_node *np)
 {
 	const char *name;
 	struct property *pp;
@@ -203,7 +203,7 @@ void __init of_core_init(void)
 		return;
 	}
 	for_each_of_allnodes(np)
-		__of_attach_node_sysfs(np);
+		__of_attach_node_post(np);
 	mutex_unlock(&of_mutex);
 
 	/* Symlink in /proc as required by userspace ABI */
diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
index 350b1cf..eb31a55 100644
--- a/drivers/of/dynamic.c
+++ b/drivers/of/dynamic.c
@@ -41,7 +41,7 @@ void of_node_put(struct device_node *node)
 }
 EXPORT_SYMBOL(of_node_put);
 
-void __of_detach_node_sysfs(struct device_node *np)
+void __of_detach_node_post(struct device_node *np)
 {
 	struct property *pp;
 
@@ -251,7 +251,7 @@ int of_attach_node(struct device_node *np)
 	__of_attach_node(np);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 
-	__of_attach_node_sysfs(np);
+	__of_attach_node_post(np);
 	mutex_unlock(&of_mutex);
 
 	of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, &rd);
@@ -304,7 +304,7 @@ int of_detach_node(struct device_node *np)
 	__of_detach_node(np);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 
-	__of_detach_node_sysfs(np);
+	__of_detach_node_post(np);
 	mutex_unlock(&of_mutex);
 
 	of_reconfig_notify(OF_RECONFIG_DETACH_NODE, &rd);
@@ -628,10 +628,10 @@ static int __of_changeset_entry_apply(struct of_changeset_entry *ce)
 
 	switch (ce->action) {
 	case OF_RECONFIG_ATTACH_NODE:
-		__of_attach_node_sysfs(ce->np);
+		__of_attach_node_post(ce->np);
 		break;
 	case OF_RECONFIG_DETACH_NODE:
-		__of_detach_node_sysfs(ce->np);
+		__of_detach_node_post(ce->np);
 		break;
 	case OF_RECONFIG_ADD_PROPERTY:
 		/* ignore duplicate names */
diff --git a/drivers/of/of_private.h b/drivers/of/of_private.h
index 092cba7..10b0342 100644
--- a/drivers/of/of_private.h
+++ b/drivers/of/of_private.h
@@ -79,9 +79,9 @@ extern void __of_update_property_sysfs(struct device_node *np,
 		struct property *newprop, struct property *oldprop);
 
 extern void __of_attach_node(struct device_node *np);
-extern int __of_attach_node_sysfs(struct device_node *np);
+extern int __of_attach_node_post(struct device_node *np);
 extern void __of_detach_node(struct device_node *np);
-extern void __of_detach_node_sysfs(struct device_node *np);
+extern void __of_detach_node_post(struct device_node *np);
 
 /* iterators for transactions, used for overlays */
 /* forward iterator */
diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
index e5a5ec0..7ea3689 100644
--- a/drivers/of/unittest.c
+++ b/drivers/of/unittest.c
@@ -931,7 +931,7 @@ static int attach_node_and_children(struct device_node *np)
 	of_node_clear_flag(np, OF_DETACHED);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 
-	__of_attach_node_sysfs(np);
+	__of_attach_node_post(np);
 	mutex_unlock(&of_mutex);
 
 	while (child) {
@@ -989,7 +989,7 @@ static int __init unittest_data_add(void)
 	if (!of_root) {
 		of_root = unittest_data_node;
 		for_each_of_allnodes(np)
-			__of_attach_node_sysfs(np);
+			__of_attach_node_post(np);
 		of_aliases = of_find_node_by_path("/aliases");
 		of_chosen = of_find_node_by_path("/chosen");
 		return 0;
-- 
1.7.12

Re: [PATCH 1/3] of: rename *_node_sysfs to _node_post

[Rob Herring]

On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
<pantelis.antoniou@konsulko.com> wrote:

The "why" goes here.

> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
> ---
>  drivers/of/base.c       |  4 ++--
>  drivers/of/dynamic.c    | 10 +++++-----
>  drivers/of/of_private.h |  4 ++--
>  drivers/of/unittest.c   |  4 ++--
>  4 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index 9bbca56..20bbc2f 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -156,7 +156,7 @@ int __of_add_property_sysfs(struct device_node *np, struct property *pp)
>         return rc;
>  }
>
> -int __of_attach_node_sysfs(struct device_node *np)
> +int __of_attach_node_post(struct device_node *np)
>  {
>         const char *name;
>         struct property *pp;
> @@ -203,7 +203,7 @@ void __init of_core_init(void)
>                 return;
>         }
>         for_each_of_allnodes(np)
> -               __of_attach_node_sysfs(np);
> +               __of_attach_node_post(np);
>         mutex_unlock(&of_mutex);
>
>         /* Symlink in /proc as required by userspace ABI */
> diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
> index 350b1cf..eb31a55 100644
> --- a/drivers/of/dynamic.c
> +++ b/drivers/of/dynamic.c
> @@ -41,7 +41,7 @@ void of_node_put(struct device_node *node)
>  }
>  EXPORT_SYMBOL(of_node_put);
>
> -void __of_detach_node_sysfs(struct device_node *np)
> +void __of_detach_node_post(struct device_node *np)
>  {
>         struct property *pp;
>
> @@ -251,7 +251,7 @@ int of_attach_node(struct device_node *np)
>         __of_attach_node(np);
>         raw_spin_unlock_irqrestore(&devtree_lock, flags);
>
> -       __of_attach_node_sysfs(np);
> +       __of_attach_node_post(np);
>         mutex_unlock(&of_mutex);
>
>         of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, &rd);
> @@ -304,7 +304,7 @@ int of_detach_node(struct device_node *np)
>         __of_detach_node(np);
>         raw_spin_unlock_irqrestore(&devtree_lock, flags);
>
> -       __of_detach_node_sysfs(np);
> +       __of_detach_node_post(np);
>         mutex_unlock(&of_mutex);
>
>         of_reconfig_notify(OF_RECONFIG_DETACH_NODE, &rd);
> @@ -628,10 +628,10 @@ static int __of_changeset_entry_apply(struct of_changeset_entry *ce)
>
>         switch (ce->action) {
>         case OF_RECONFIG_ATTACH_NODE:
> -               __of_attach_node_sysfs(ce->np);
> +               __of_attach_node_post(ce->np);
>                 break;
>         case OF_RECONFIG_DETACH_NODE:
> -               __of_detach_node_sysfs(ce->np);
> +               __of_detach_node_post(ce->np);
>                 break;
>         case OF_RECONFIG_ADD_PROPERTY:
>                 /* ignore duplicate names */
> diff --git a/drivers/of/of_private.h b/drivers/of/of_private.h
> index 092cba7..10b0342 100644
> --- a/drivers/of/of_private.h
> +++ b/drivers/of/of_private.h
> @@ -79,9 +79,9 @@ extern void __of_update_property_sysfs(struct device_node *np,
>                 struct property *newprop, struct property *oldprop);
>
>  extern void __of_attach_node(struct device_node *np);
> -extern int __of_attach_node_sysfs(struct device_node *np);
> +extern int __of_attach_node_post(struct device_node *np);
>  extern void __of_detach_node(struct device_node *np);
> -extern void __of_detach_node_sysfs(struct device_node *np);
> +extern void __of_detach_node_post(struct device_node *np);
>
>  /* iterators for transactions, used for overlays */
>  /* forward iterator */
> diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
> index e5a5ec0..7ea3689 100644
> --- a/drivers/of/unittest.c
> +++ b/drivers/of/unittest.c
> @@ -931,7 +931,7 @@ static int attach_node_and_children(struct device_node *np)
>         of_node_clear_flag(np, OF_DETACHED);
>         raw_spin_unlock_irqrestore(&devtree_lock, flags);
>
> -       __of_attach_node_sysfs(np);
> +       __of_attach_node_post(np);
>         mutex_unlock(&of_mutex);
>
>         while (child) {
> @@ -989,7 +989,7 @@ static int __init unittest_data_add(void)
>         if (!of_root) {
>                 of_root = unittest_data_node;
>                 for_each_of_allnodes(np)
> -                       __of_attach_node_sysfs(np);
> +                       __of_attach_node_post(np);
>                 of_aliases = of_find_node_by_path("/aliases");
>                 of_chosen = of_find_node_by_path("/chosen");
>                 return 0;
> --
> 1.7.12
>

Re: [PATCH 1/3] of: rename *_node_sysfs to _node_post

[Pantelis Antoniou]

Hi Rob,

> On May 10, 2016, at 00:49 , Rob Herring <robherring2@gmail.com> wrote:
> 
> On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
> <pantelis.antoniou@konsulko.com> wrote:
> 
> The "why" goes here.
> 

OK.

>> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
>> ---
>> drivers/of/base.c       |  4 ++--
>> drivers/of/dynamic.c    | 10 +++++-----
>> drivers/of/of_private.h |  4 ++--
>> drivers/of/unittest.c   |  4 ++--
>> 4 files changed, 11 insertions(+), 11 deletions(-)
>> 
>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>> index 9bbca56..20bbc2f 100644
>> --- a/drivers/of/base.c
>> +++ b/drivers/of/base.c
>> @@ -156,7 +156,7 @@ int __of_add_property_sysfs(struct device_node *np, struct property *pp)
>>        return rc;
>> }
>> 
>> -int __of_attach_node_sysfs(struct device_node *np)
>> +int __of_attach_node_post(struct device_node *np)
>> {
>>        const char *name;
>>        struct property *pp;
>> @@ -203,7 +203,7 @@ void __init of_core_init(void)
>>                return;
>>        }
>>        for_each_of_allnodes(np)
>> -               __of_attach_node_sysfs(np);
>> +               __of_attach_node_post(np);
>>        mutex_unlock(&of_mutex);
>> 
>>        /* Symlink in /proc as required by userspace ABI */
>> diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
>> index 350b1cf..eb31a55 100644
>> --- a/drivers/of/dynamic.c
>> +++ b/drivers/of/dynamic.c
>> @@ -41,7 +41,7 @@ void of_node_put(struct device_node *node)
>> }
>> EXPORT_SYMBOL(of_node_put);
>> 
>> -void __of_detach_node_sysfs(struct device_node *np)
>> +void __of_detach_node_post(struct device_node *np)
>> {
>>        struct property *pp;
>> 
>> @@ -251,7 +251,7 @@ int of_attach_node(struct device_node *np)
>>        __of_attach_node(np);
>>        raw_spin_unlock_irqrestore(&devtree_lock, flags);
>> 
>> -       __of_attach_node_sysfs(np);
>> +       __of_attach_node_post(np);
>>        mutex_unlock(&of_mutex);
>> 
>>        of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, &rd);
>> @@ -304,7 +304,7 @@ int of_detach_node(struct device_node *np)
>>        __of_detach_node(np);
>>        raw_spin_unlock_irqrestore(&devtree_lock, flags);
>> 
>> -       __of_detach_node_sysfs(np);
>> +       __of_detach_node_post(np);
>>        mutex_unlock(&of_mutex);
>> 
>>        of_reconfig_notify(OF_RECONFIG_DETACH_NODE, &rd);
>> @@ -628,10 +628,10 @@ static int __of_changeset_entry_apply(struct of_changeset_entry *ce)
>> 
>>        switch (ce->action) {
>>        case OF_RECONFIG_ATTACH_NODE:
>> -               __of_attach_node_sysfs(ce->np);
>> +               __of_attach_node_post(ce->np);
>>                break;
>>        case OF_RECONFIG_DETACH_NODE:
>> -               __of_detach_node_sysfs(ce->np);
>> +               __of_detach_node_post(ce->np);
>>                break;
>>        case OF_RECONFIG_ADD_PROPERTY:
>>                /* ignore duplicate names */
>> diff --git a/drivers/of/of_private.h b/drivers/of/of_private.h
>> index 092cba7..10b0342 100644
>> --- a/drivers/of/of_private.h
>> +++ b/drivers/of/of_private.h
>> @@ -79,9 +79,9 @@ extern void __of_update_property_sysfs(struct device_node *np,
>>                struct property *newprop, struct property *oldprop);
>> 
>> extern void __of_attach_node(struct device_node *np);
>> -extern int __of_attach_node_sysfs(struct device_node *np);
>> +extern int __of_attach_node_post(struct device_node *np);
>> extern void __of_detach_node(struct device_node *np);
>> -extern void __of_detach_node_sysfs(struct device_node *np);
>> +extern void __of_detach_node_post(struct device_node *np);
>> 
>> /* iterators for transactions, used for overlays */
>> /* forward iterator */
>> diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
>> index e5a5ec0..7ea3689 100644
>> --- a/drivers/of/unittest.c
>> +++ b/drivers/of/unittest.c
>> @@ -931,7 +931,7 @@ static int attach_node_and_children(struct device_node *np)
>>        of_node_clear_flag(np, OF_DETACHED);
>>        raw_spin_unlock_irqrestore(&devtree_lock, flags);
>> 
>> -       __of_attach_node_sysfs(np);
>> +       __of_attach_node_post(np);
>>        mutex_unlock(&of_mutex);
>> 
>>        while (child) {
>> @@ -989,7 +989,7 @@ static int __init unittest_data_add(void)
>>        if (!of_root) {
>>                of_root = unittest_data_node;
>>                for_each_of_allnodes(np)
>> -                       __of_attach_node_sysfs(np);
>> +                       __of_attach_node_post(np);
>>                of_aliases = of_find_node_by_path("/aliases");
>>                of_chosen = of_find_node_by_path("/chosen");
>>                return 0;
>> --
>> 1.7.12
>> 

[PATCH 2/3] of: Support hashtable lookups for phandles

[Pantelis Antoniou]

Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
---
 drivers/of/base.c       | 35 ++++++++++++++++++++++++++++++++---
 drivers/of/dynamic.c    |  8 ++++++++
 drivers/of/of_private.h | 31 +++++++++++++++++++++++++++++++
 include/linux/of.h      |  2 ++
 4 files changed, 73 insertions(+), 3 deletions(-)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 20bbc2f..770cb95 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -27,6 +27,7 @@
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/proc_fs.h>
+#include <linux/rhashtable.h>
 
 #include "of_private.h"
 
@@ -41,6 +42,16 @@ static const char *of_stdout_options;
 
 struct kset *of_kset;
 
+const struct rhashtable_params of_phandle_ht_params = {
+	.key_offset = offsetof(struct device_node, phandle), /* base offset */
+	.key_len = sizeof(phandle),
+	.head_offset = offsetof(struct device_node, ht_node),
+	.automatic_shrinking = true,
+};
+
+struct rhashtable of_phandle_ht;
+bool of_phandle_ht_initialized;
+
 /*
  * Used to protect the of_aliases, to hold off addition of nodes to sysfs.
  * This mutex must be held whenever modifications are being made to the
@@ -162,6 +173,12 @@ int __of_attach_node_post(struct device_node *np)
 	struct property *pp;
 	int rc;
 
+	if (of_phandle_ht_available()) {
+		rc = of_phandle_ht_insert(np);
+		WARN(rc, "insert to phandle hash fail @%s\n",
+				of_node_full_name(np));
+	}
+
 	if (!IS_ENABLED(CONFIG_SYSFS))
 		return 0;
 
@@ -194,6 +211,13 @@ void __init of_core_init(void)
 	struct device_node *np;
 	int ret;
 
+	ret = rhashtable_init(&of_phandle_ht, &of_phandle_ht_params);
+	if (ret) {
+		pr_warn("devicetree: Failed to initialize hashtable\n");
+		return;
+	}
+	of_phandle_ht_initialized = 1;
+
 	/* Create the kset, and register existing nodes */
 	mutex_lock(&of_mutex);
 	of_kset = kset_create_and_add("devicetree", NULL, firmware_kobj);
@@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
 		return NULL;
 
 	raw_spin_lock_irqsave(&devtree_lock, flags);
-	for_each_of_allnodes(np)
-		if (np->phandle == handle)
-			break;
+	/* when we're ready use the hash table */
+	if (of_phandle_ht_available() && !in_interrupt())
+		np = of_phandle_ht_lookup(handle);
+	else { /* fallback */
+		for_each_of_allnodes(np)
+			if (np->phandle == handle)
+				break;
+	}
 	of_node_get(np);
 	raw_spin_unlock_irqrestore(&devtree_lock, flags);
 	return np;
diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
index eb31a55..bdebdf5 100644
--- a/drivers/of/dynamic.c
+++ b/drivers/of/dynamic.c
@@ -11,6 +11,7 @@
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/proc_fs.h>
+#include <linux/rhashtable.h>
 
 #include "of_private.h"
 
@@ -44,6 +45,13 @@ EXPORT_SYMBOL(of_node_put);
 void __of_detach_node_post(struct device_node *np)
 {
 	struct property *pp;
+	int rc;
+
+	if (of_phandle_ht_available()) {
+		rc = of_phandle_ht_remove(np);
+		WARN(rc, "remove from phandle hash fail @%s\n",
+				of_node_full_name(np));
+	}
 
 	if (!IS_ENABLED(CONFIG_SYSFS))
 		return;
diff --git a/drivers/of/of_private.h b/drivers/of/of_private.h
index 10b0342..88b3b8f 100644
--- a/drivers/of/of_private.h
+++ b/drivers/of/of_private.h
@@ -101,4 +101,35 @@ static inline int of_overlay_init(void)
 }
 #endif
 
+extern const struct rhashtable_params of_phandle_ht_params;
+extern struct rhashtable of_phandle_ht;
+extern bool of_phandle_ht_initialized;
+
+static inline bool of_phandle_ht_available(void)
+{
+	return of_phandle_ht_initialized;
+}
+
+static inline int of_phandle_ht_insert(struct device_node *np)
+{
+	if (!np || !np->phandle)
+		return 0;
+	return rhashtable_insert_fast(&of_phandle_ht,
+		&np->ht_node, of_phandle_ht_params);
+}
+
+static inline int of_phandle_ht_remove(struct device_node *np)
+{
+	if (!np || !np->phandle)
+		return 0;
+	return rhashtable_remove_fast(&of_phandle_ht,
+		&np->ht_node, of_phandle_ht_params);
+}
+
+static inline struct device_node *of_phandle_ht_lookup(phandle handle)
+{
+	return rhashtable_lookup_fast(&of_phandle_ht,
+			&handle, of_phandle_ht_params);
+}
+
 #endif /* _LINUX_OF_PRIVATE_H */
diff --git a/include/linux/of.h b/include/linux/of.h
index 582bc45..56bb62d 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -25,6 +25,7 @@
 #include <linux/notifier.h>
 #include <linux/property.h>
 #include <linux/list.h>
+#include <linux/rhashtable.h>
 
 #include <asm/byteorder.h>
 #include <asm/errno.h>
@@ -52,6 +53,7 @@ struct device_node {
 	phandle phandle;
 	const char *full_name;
 	struct fwnode_handle fwnode;
+	struct rhash_head ht_node;
 
 	struct	property *properties;
 	struct	property *deadprops;	/* removed properties */
-- 
1.7.12

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Geert Uytterhoeven]

Hi Pantelis,

On Mon, May 9, 2016 at 8:11 PM, Pantelis Antoniou
<pantelis.antoniou@konsulko.com> wrote:
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c

> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>                 return NULL;
>
>         raw_spin_lock_irqsave(&devtree_lock, flags);
> -       for_each_of_allnodes(np)
> -               if (np->phandle == handle)
> -                       break;
> +       /* when we're ready use the hash table */
> +       if (of_phandle_ht_available() && !in_interrupt())

I guess the !in_interrupt() test is because of the locking inside
rhashtable_lookup_fast()?

> +               np = of_phandle_ht_lookup(handle);
> +       else { /* fallback */
> +               for_each_of_allnodes(np)
> +                       if (np->phandle == handle)
> +                               break;
> +       }

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Rob Herring]

On Mon, May 9, 2016 at 3:38 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> Hi Pantelis,
>
> On Mon, May 9, 2016 at 8:11 PM, Pantelis Antoniou
> <pantelis.antoniou@konsulko.com> wrote:
>> --- a/drivers/of/base.c
>> +++ b/drivers/of/base.c
>
>> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>>                 return NULL;
>>
>>         raw_spin_lock_irqsave(&devtree_lock, flags);
>> -       for_each_of_allnodes(np)
>> -               if (np->phandle == handle)
>> -                       break;
>> +       /* when we're ready use the hash table */
>> +       if (of_phandle_ht_available() && !in_interrupt())
>
> I guess the !in_interrupt() test is because of the locking inside
> rhashtable_lookup_fast()?

Not a use we should support. Just warn for anyone parsing DT in
interrupt context.

Rob

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Pantelis Antoniou]

Hi Rob,

> On May 10, 2016, at 00:11 , Rob Herring <robherring2@gmail.com> wrote:
> 
> On Mon, May 9, 2016 at 3:38 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>> Hi Pantelis,
>> 
>> On Mon, May 9, 2016 at 8:11 PM, Pantelis Antoniou
>> <pantelis.antoniou@konsulko.com> wrote:
>>> --- a/drivers/of/base.c
>>> +++ b/drivers/of/base.c
>> 
>>> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>>>                return NULL;
>>> 
>>>        raw_spin_lock_irqsave(&devtree_lock, flags);
>>> -       for_each_of_allnodes(np)
>>> -               if (np->phandle == handle)
>>> -                       break;
>>> +       /* when we're ready use the hash table */
>>> +       if (of_phandle_ht_available() && !in_interrupt())
>> 
>> I guess the !in_interrupt() test is because of the locking inside
>> rhashtable_lookup_fast()?
> 
> Not a use we should support. Just warn for anyone parsing DT in
> interrupt context.
> 

That’s not about users calling in interrupt context. It’s when we’re
very early in the boot sequence we’re under interrupt context and
calls to the hash methods cannot be made.

> Rob

Regards

— Pantelis

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Rob Herring]

On Tue, May 10, 2016 at 8:45 AM, Pantelis Antoniou
<pantelis.antoniou@konsulko.com> wrote:
> Hi Rob,
>
>> On May 10, 2016, at 00:11 , Rob Herring <robherring2@gmail.com> wrote:
>>
>> On Mon, May 9, 2016 at 3:38 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>>> Hi Pantelis,
>>>
>>> On Mon, May 9, 2016 at 8:11 PM, Pantelis Antoniou
>>> <pantelis.antoniou@konsulko.com> wrote:
>>>> --- a/drivers/of/base.c
>>>> +++ b/drivers/of/base.c
>>>
>>>> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>>>>                return NULL;
>>>>
>>>>        raw_spin_lock_irqsave(&devtree_lock, flags);
>>>> -       for_each_of_allnodes(np)
>>>> -               if (np->phandle == handle)
>>>> -                       break;
>>>> +       /* when we're ready use the hash table */
>>>> +       if (of_phandle_ht_available() && !in_interrupt())
>>>
>>> I guess the !in_interrupt() test is because of the locking inside
>>> rhashtable_lookup_fast()?
>>
>> Not a use we should support. Just warn for anyone parsing DT in
>> interrupt context.
>>
>
> That’s not about users calling in interrupt context. It’s when we’re
> very early in the boot sequence we’re under interrupt context and
> calls to the hash methods cannot be made.

I don't understand. When exactly are we in interrupt context?

Rob

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Rob Herring]

On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
<pantelis.antoniou@konsulko.com> wrote:

Why this is needed goes here.

Any data on how much time the current code takes?

> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
> ---
>  drivers/of/base.c       | 35 ++++++++++++++++++++++++++++++++---
>  drivers/of/dynamic.c    |  8 ++++++++
>  drivers/of/of_private.h | 31 +++++++++++++++++++++++++++++++
>  include/linux/of.h      |  2 ++
>  4 files changed, 73 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index 20bbc2f..770cb95 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -27,6 +27,7 @@
>  #include <linux/slab.h>
>  #include <linux/string.h>
>  #include <linux/proc_fs.h>
> +#include <linux/rhashtable.h>
>
>  #include "of_private.h"
>
> @@ -41,6 +42,16 @@ static const char *of_stdout_options;
>
>  struct kset *of_kset;
>
> +const struct rhashtable_params of_phandle_ht_params = {
> +       .key_offset = offsetof(struct device_node, phandle), /* base offset */
> +       .key_len = sizeof(phandle),
> +       .head_offset = offsetof(struct device_node, ht_node),
> +       .automatic_shrinking = true,
> +};
> +
> +struct rhashtable of_phandle_ht;
> +bool of_phandle_ht_initialized;

Get rid of this and of_phandle_ht_available. It should always be initialized.

> +
>  /*
>   * Used to protect the of_aliases, to hold off addition of nodes to sysfs.
>   * This mutex must be held whenever modifications are being made to the
> @@ -162,6 +173,12 @@ int __of_attach_node_post(struct device_node *np)
>         struct property *pp;
>         int rc;
>
> +       if (of_phandle_ht_available()) {
> +               rc = of_phandle_ht_insert(np);
> +               WARN(rc, "insert to phandle hash fail @%s\n",
> +                               of_node_full_name(np));
> +       }
> +
>         if (!IS_ENABLED(CONFIG_SYSFS))
>                 return 0;
>
> @@ -194,6 +211,13 @@ void __init of_core_init(void)
>         struct device_node *np;
>         int ret;
>
> +       ret = rhashtable_init(&of_phandle_ht, &of_phandle_ht_params);
> +       if (ret) {
> +               pr_warn("devicetree: Failed to initialize hashtable\n");
> +               return;
> +       }
> +       of_phandle_ht_initialized = 1;
> +
>         /* Create the kset, and register existing nodes */
>         mutex_lock(&of_mutex);
>         of_kset = kset_create_and_add("devicetree", NULL, firmware_kobj);
> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>                 return NULL;
>
>         raw_spin_lock_irqsave(&devtree_lock, flags);
> -       for_each_of_allnodes(np)
> -               if (np->phandle == handle)
> -                       break;
> +       /* when we're ready use the hash table */
> +       if (of_phandle_ht_available() && !in_interrupt())
> +               np = of_phandle_ht_lookup(handle);
> +       else { /* fallback */
> +               for_each_of_allnodes(np)
> +                       if (np->phandle == handle)
> +                               break;
> +       }
>         of_node_get(np);
>         raw_spin_unlock_irqrestore(&devtree_lock, flags);
>         return np;
> diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
> index eb31a55..bdebdf5 100644
> --- a/drivers/of/dynamic.c
> +++ b/drivers/of/dynamic.c
> @@ -11,6 +11,7 @@
>  #include <linux/slab.h>
>  #include <linux/string.h>
>  #include <linux/proc_fs.h>
> +#include <linux/rhashtable.h>
>
>  #include "of_private.h"
>
> @@ -44,6 +45,13 @@ EXPORT_SYMBOL(of_node_put);
>  void __of_detach_node_post(struct device_node *np)
>  {
>         struct property *pp;
> +       int rc;
> +
> +       if (of_phandle_ht_available()) {
> +               rc = of_phandle_ht_remove(np);
> +               WARN(rc, "remove from phandle hash fail @%s\n",
> +                               of_node_full_name(np));

Can this really fail?

Rob

Re: [PATCH 2/3] of: Support hashtable lookups for phandles

[Pantelis Antoniou]

Hi Rob,

> On May 10, 2016, at 00:21 , Rob Herring <robherring2@gmail.com> wrote:
> 
> On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
> <pantelis.antoniou@konsulko.com> wrote:
> 
> Why this is needed goes here.
> 
> Any data on how much time the current code takes?
> 

I’ll get some numbers. The current algorithm is exhaustive search of
all nodes (not even nodes that have phandles). So it’s kind of obvious
bad on large blobs.

>> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
>> ---
>> drivers/of/base.c       | 35 ++++++++++++++++++++++++++++++++---
>> drivers/of/dynamic.c    |  8 ++++++++
>> drivers/of/of_private.h | 31 +++++++++++++++++++++++++++++++
>> include/linux/of.h      |  2 ++
>> 4 files changed, 73 insertions(+), 3 deletions(-)
>> 
>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>> index 20bbc2f..770cb95 100644
>> --- a/drivers/of/base.c
>> +++ b/drivers/of/base.c
>> @@ -27,6 +27,7 @@
>> #include <linux/slab.h>
>> #include <linux/string.h>
>> #include <linux/proc_fs.h>
>> +#include <linux/rhashtable.h>
>> 
>> #include "of_private.h"
>> 
>> @@ -41,6 +42,16 @@ static const char *of_stdout_options;
>> 
>> struct kset *of_kset;
>> 
>> +const struct rhashtable_params of_phandle_ht_params = {
>> +       .key_offset = offsetof(struct device_node, phandle), /* base offset */
>> +       .key_len = sizeof(phandle),
>> +       .head_offset = offsetof(struct device_node, ht_node),
>> +       .automatic_shrinking = true,
>> +};
>> +
>> +struct rhashtable of_phandle_ht;
>> +bool of_phandle_ht_initialized;
> 
> Get rid of this and of_phandle_ht_available. It should always be initialized.
> 

Not that simple. We have to support the case where calls are made very early
in the boot sequence before the initialization.

>> +
>> /*
>>  * Used to protect the of_aliases, to hold off addition of nodes to sysfs.
>>  * This mutex must be held whenever modifications are being made to the
>> @@ -162,6 +173,12 @@ int __of_attach_node_post(struct device_node *np)
>>        struct property *pp;
>>        int rc;
>> 
>> +       if (of_phandle_ht_available()) {
>> +               rc = of_phandle_ht_insert(np);
>> +               WARN(rc, "insert to phandle hash fail @%s\n",
>> +                               of_node_full_name(np));
>> +       }
>> +
>>        if (!IS_ENABLED(CONFIG_SYSFS))
>>                return 0;
>> 
>> @@ -194,6 +211,13 @@ void __init of_core_init(void)
>>        struct device_node *np;
>>        int ret;
>> 
>> +       ret = rhashtable_init(&of_phandle_ht, &of_phandle_ht_params);
>> +       if (ret) {
>> +               pr_warn("devicetree: Failed to initialize hashtable\n");
>> +               return;
>> +       }
>> +       of_phandle_ht_initialized = 1;
>> +
>>        /* Create the kset, and register existing nodes */
>>        mutex_lock(&of_mutex);
>>        of_kset = kset_create_and_add("devicetree", NULL, firmware_kobj);
>> @@ -1073,9 +1097,14 @@ struct device_node *of_find_node_by_phandle(phandle handle)
>>                return NULL;
>> 
>>        raw_spin_lock_irqsave(&devtree_lock, flags);
>> -       for_each_of_allnodes(np)
>> -               if (np->phandle == handle)
>> -                       break;
>> +       /* when we're ready use the hash table */
>> +       if (of_phandle_ht_available() && !in_interrupt())
>> +               np = of_phandle_ht_lookup(handle);
>> +       else { /* fallback */
>> +               for_each_of_allnodes(np)
>> +                       if (np->phandle == handle)
>> +                               break;
>> +       }
>>        of_node_get(np);
>>        raw_spin_unlock_irqrestore(&devtree_lock, flags);
>>        return np;
>> diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
>> index eb31a55..bdebdf5 100644
>> --- a/drivers/of/dynamic.c
>> +++ b/drivers/of/dynamic.c
>> @@ -11,6 +11,7 @@
>> #include <linux/slab.h>
>> #include <linux/string.h>
>> #include <linux/proc_fs.h>
>> +#include <linux/rhashtable.h>
>> 
>> #include "of_private.h"
>> 
>> @@ -44,6 +45,13 @@ EXPORT_SYMBOL(of_node_put);
>> void __of_detach_node_post(struct device_node *np)
>> {
>>        struct property *pp;
>> +       int rc;
>> +
>> +       if (of_phandle_ht_available()) {
>> +               rc = of_phandle_ht_remove(np);
>> +               WARN(rc, "remove from phandle hash fail @%s\n",
>> +                               of_node_full_name(np));
> 
> Can this really fail?
> 

No, unless the structure is corrupted. I can move it under DEBUG.

> Rob
> —

Regards

— Pantelis

[PATCH 3/3] of: overlay: Pick up label symbols from overlays.

[Pantelis Antoniou]

Insert overlay symbols to the base tree when applied.
This makes it possible to apply an overlay that references a label
that a previously inserted overlay had.

Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>

---
 drivers/of/overlay.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 102 insertions(+)

diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
index a274d65..a7956a2 100644
--- a/drivers/of/overlay.c
+++ b/drivers/of/overlay.c
@@ -514,6 +514,101 @@ static int of_free_overlay_info(struct of_overlay *ov)
 	return 0;
 }
 
+static int of_overlay_add_symbols(
+		struct device_node *tree,
+		struct of_overlay *ov)
+{
+	struct of_overlay_info *ovinfo;
+	struct device_node *root_sym = NULL;
+	struct device_node *child = NULL;
+	struct property *prop;
+	const char *path, *s;
+	char *new_path;
+	int i, len, err;
+
+	/* this may fail (if no fixups are required) */
+	root_sym = of_find_node_by_path("/__symbols__");
+
+	/* do nothing if no root symbols */
+	if (!root_sym)
+		return 0;
+
+	/* locate the symbols & fixups nodes on resolve */
+	for_each_child_of_node(tree, child) {
+		if (of_node_cmp(child->name, "__symbols__") == 0)
+			goto found;
+	}
+	/* no symbols, no problem */
+	of_node_put(root_sym);
+	return 0;
+
+found:
+	err = -EINVAL;
+	for_each_property_of_node(child, prop) {
+
+		/* skip properties added automatically */
+		if (of_prop_cmp(prop->name, "name") == 0)
+			continue;
+
+		err = of_property_read_string(child,
+				prop->name, &path);
+		if (err != 0) {
+			pr_err("%s: Could not find symbol '%s'\n",
+					__func__, prop->name);
+			continue;
+		}
+
+
+		/* now find fragment index */
+		s = path;
+
+		/* compare paths to find fragment index */
+		ovinfo = NULL;
+		len = -1;
+		for (i = 0; i < ov->count; i++) {
+			ovinfo = &ov->ovinfo_tab[i];
+
+			pr_debug("%s: #%d: overlay->name=%s target->name=%s\n",
+					__func__, i, ovinfo->overlay->full_name,
+					ovinfo->target->full_name);
+
+			len = strlen(ovinfo->overlay->full_name);
+			if (strncasecmp(path, ovinfo->overlay->full_name,
+						len) == 0 && path[len] == '/')
+				break;
+		}
+
+		if (i >= ov->count)
+			continue;
+
+		pr_debug("%s: found target at #%d\n", __func__, i);
+		new_path = kasprintf(GFP_KERNEL, "%s%s",
+				ovinfo->target->full_name,
+				path + len);
+		if (!new_path) {
+			pr_err("%s: Failed to allocate propname for \"%s\"\n",
+					__func__, prop->name);
+			continue;
+		}
+
+		err = of_changeset_add_property_string(&ov->cset, root_sym,
+				prop->name, new_path);
+
+		/* free always */
+		kfree(new_path);
+
+		if (err) {
+			pr_err("%s: Failed to add property for \"%s\"\n",
+					__func__, prop->name);
+		}
+	}
+
+	of_node_put(child);
+	of_node_put(root_sym);
+
+	return 0;
+}
+
 static LIST_HEAD(ov_list);
 static DEFINE_IDR(ov_idr);
 
@@ -642,6 +737,13 @@ static int __of_overlay_create(struct device_node *tree,
 		goto err_abort_trans;
 	}
 
+	err = of_overlay_add_symbols(tree, ov);
+	if (err) {
+		pr_err("%s: of_overlay_add_symbols() failed for tree@%s\n",
+				__func__, tree->full_name);
+		goto err_abort_trans;
+	}
+
 	/* apply the changeset */
 	err = __of_changeset_apply(&ov->cset);
 	if (err) {
-- 
1.7.12

Re: [PATCH 3/3] of: overlay: Pick up label symbols from overlays.

[Marek Vasut]

On 05/09/2016 08:11 PM, Pantelis Antoniou wrote:
> Insert overlay symbols to the base tree when applied.
> This makes it possible to apply an overlay that references a label
> that a previously inserted overlay had.
> 
> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> 
> ---
>  drivers/of/overlay.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 102 insertions(+)
> 
> diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
> index a274d65..a7956a2 100644
> --- a/drivers/of/overlay.c
> +++ b/drivers/of/overlay.c
> @@ -514,6 +514,101 @@ static int of_free_overlay_info(struct of_overlay *ov)
>  	return 0;
>  }
>  
> +static int of_overlay_add_symbols(
> +		struct device_node *tree,
> +		struct of_overlay *ov)
> +{
> +	struct of_overlay_info *ovinfo;
> +	struct device_node *root_sym = NULL;
> +	struct device_node *child = NULL;
> +	struct property *prop;
> +	const char *path, *s;
> +	char *new_path;
> +	int i, len, err;
> +
> +	/* this may fail (if no fixups are required) */
> +	root_sym = of_find_node_by_path("/__symbols__");
> +
> +	/* do nothing if no root symbols */
> +	if (!root_sym)
> +		return 0;
> +
> +	/* locate the symbols & fixups nodes on resolve */
> +	for_each_child_of_node(tree, child) {
> +		if (of_node_cmp(child->name, "__symbols__") == 0)
> +			goto found;
> +	}
> +	/* no symbols, no problem */
> +	of_node_put(root_sym);
> +	return 0;
> +
> +found:

You might want to factor this loop into a separate function instead of
the elaborate use of goto .

> +	err = -EINVAL;
> +	for_each_property_of_node(child, prop) {
> +
> +		/* skip properties added automatically */
> +		if (of_prop_cmp(prop->name, "name") == 0)
> +			continue;
> +
> +		err = of_property_read_string(child,
> +				prop->name, &path);
> +		if (err != 0) {
> +			pr_err("%s: Could not find symbol '%s'\n",
> +					__func__, prop->name);

It is useful to print the errno here I think.

> +			continue;
> +		}
> +
> +
> +		/* now find fragment index */
> +		s = path;
> +
> +		/* compare paths to find fragment index */
> +		ovinfo = NULL;
> +		len = -1;
> +		for (i = 0; i < ov->count; i++) {
> +			ovinfo = &ov->ovinfo_tab[i];
> +
> +			pr_debug("%s: #%d: overlay->name=%s target->name=%s\n",
> +					__func__, i, ovinfo->overlay->full_name,
> +					ovinfo->target->full_name);
> +
> +			len = strlen(ovinfo->overlay->full_name);
> +			if (strncasecmp(path, ovinfo->overlay->full_name,
> +						len) == 0 && path[len] == '/')
> +				break;
> +		}
> +
> +		if (i >= ov->count)
> +			continue;
> +
> +		pr_debug("%s: found target at #%d\n", __func__, i);
> +		new_path = kasprintf(GFP_KERNEL, "%s%s",
> +				ovinfo->target->full_name,
> +				path + len);
> +		if (!new_path) {
> +			pr_err("%s: Failed to allocate propname for \"%s\"\n",
> +					__func__, prop->name);

You ran out of memory here, so the pr_err() will probably also fail.

> +			continue;
> +		}
> +
> +		err = of_changeset_add_property_string(&ov->cset, root_sym,
> +				prop->name, new_path);
> +
> +		/* free always */
> +		kfree(new_path);
> +
> +		if (err) {
> +			pr_err("%s: Failed to add property for \"%s\"\n",
> +					__func__, prop->name);
> +		}
> +	}
> +
> +	of_node_put(child);
> +	of_node_put(root_sym);
> +
> +	return 0;
> +}
> +
>  static LIST_HEAD(ov_list);
>  static DEFINE_IDR(ov_idr);
>  
> @@ -642,6 +737,13 @@ static int __of_overlay_create(struct device_node *tree,
>  		goto err_abort_trans;
>  	}
>  
> +	err = of_overlay_add_symbols(tree, ov);
> +	if (err) {
> +		pr_err("%s: of_overlay_add_symbols() failed for tree@%s\n",
> +				__func__, tree->full_name);
> +		goto err_abort_trans;
> +	}
> +
>  	/* apply the changeset */
>  	err = __of_changeset_apply(&ov->cset);
>  	if (err) {
> 


-- 
Best regards,
Marek Vasut

Re: [PATCH 3/3] of: overlay: Pick up label symbols from overlays.

[Rob Herring]

On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
<pantelis.antoniou@konsulko.com> wrote:
> Insert overlay symbols to the base tree when applied.
> This makes it possible to apply an overlay that references a label
> that a previously inserted overlay had.
>
> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
>
> ---
>  drivers/of/overlay.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 102 insertions(+)
>
> diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
> index a274d65..a7956a2 100644
> --- a/drivers/of/overlay.c
> +++ b/drivers/of/overlay.c
> @@ -514,6 +514,101 @@ static int of_free_overlay_info(struct of_overlay *ov)
>         return 0;
>  }
>
> +static int of_overlay_add_symbols(
> +               struct device_node *tree,
> +               struct of_overlay *ov)
> +{
> +       struct of_overlay_info *ovinfo;
> +       struct device_node *root_sym = NULL;
> +       struct device_node *child = NULL;
> +       struct property *prop;
> +       const char *path, *s;
> +       char *new_path;
> +       int i, len, err;
> +
> +       /* this may fail (if no fixups are required) */
> +       root_sym = of_find_node_by_path("/__symbols__");
> +
> +       /* do nothing if no root symbols */
> +       if (!root_sym)
> +               return 0;
> +
> +       /* locate the symbols & fixups nodes on resolve */
> +       for_each_child_of_node(tree, child) {
> +               if (of_node_cmp(child->name, "__symbols__") == 0)
> +                       goto found;

Doesn't of_get_child_by_name work here?

> +       }
> +       /* no symbols, no problem */
> +       of_node_put(root_sym);
> +       return 0;
> +
> +found:

As Marek said, get rid of the goto.

> +       err = -EINVAL;
> +       for_each_property_of_node(child, prop) {
> +
> +               /* skip properties added automatically */
> +               if (of_prop_cmp(prop->name, "name") == 0)
> +                       continue;
> +
> +               err = of_property_read_string(child,
> +                               prop->name, &path);
> +               if (err != 0) {
> +                       pr_err("%s: Could not find symbol '%s'\n",
> +                                       __func__, prop->name);

Can you introduce and use a common pr_fmt prefix here and elsewhere
(just what you are adding).

> +                       continue;
> +               }
> +
> +

Extra blank line.

> +               /* now find fragment index */
> +               s = path;
> +
> +               /* compare paths to find fragment index */
> +               ovinfo = NULL;
> +               len = -1;

Move these into the for init.

> +               for (i = 0; i < ov->count; i++) {
> +                       ovinfo = &ov->ovinfo_tab[i];
> +
> +                       pr_debug("%s: #%d: overlay->name=%s target->name=%s\n",
> +                                       __func__, i, ovinfo->overlay->full_name,
> +                                       ovinfo->target->full_name);
> +
> +                       len = strlen(ovinfo->overlay->full_name);
> +                       if (strncasecmp(path, ovinfo->overlay->full_name,
> +                                               len) == 0 && path[len] == '/')
> +                               break;
> +               }
> +
> +               if (i >= ov->count)
> +                       continue;
> +
> +               pr_debug("%s: found target at #%d\n", __func__, i);
> +               new_path = kasprintf(GFP_KERNEL, "%s%s",
> +                               ovinfo->target->full_name,
> +                               path + len);
> +               if (!new_path) {
> +                       pr_err("%s: Failed to allocate propname for \"%s\"\n",
> +                                       __func__, prop->name);
> +                       continue;

If this fails, is there much point in continuing?

> +               }
> +
> +               err = of_changeset_add_property_string(&ov->cset, root_sym,
> +                               prop->name, new_path);
> +
> +               /* free always */
> +               kfree(new_path);
> +
> +               if (err) {

No need for brackets.

> +                       pr_err("%s: Failed to add property for \"%s\"\n",
> +                                       __func__, prop->name);
> +               }
> +       }
> +
> +       of_node_put(child);
> +       of_node_put(root_sym);
> +
> +       return 0;
> +}
> +
>  static LIST_HEAD(ov_list);
>  static DEFINE_IDR(ov_idr);
>
> @@ -642,6 +737,13 @@ static int __of_overlay_create(struct device_node *tree,
>                 goto err_abort_trans;
>         }
>
> +       err = of_overlay_add_symbols(tree, ov);
> +       if (err) {
> +               pr_err("%s: of_overlay_add_symbols() failed for tree@%s\n",
> +                               __func__, tree->full_name);
> +               goto err_abort_trans;
> +       }
> +
>         /* apply the changeset */
>         err = __of_changeset_apply(&ov->cset);
>         if (err) {
> --
> 1.7.12
>

Re: [PATCH 3/3] of: overlay: Pick up label symbols from overlays.

[Pantelis Antoniou]

Hi Rob,

> On May 10, 2016, at 00:47 , Rob Herring <robherring2@gmail.com> wrote:
> 
> On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
> <pantelis.antoniou@konsulko.com> wrote:
>> Insert overlay symbols to the base tree when applied.
>> This makes it possible to apply an overlay that references a label
>> that a previously inserted overlay had.
>> 
>> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
>> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
>> 
>> ---
>> drivers/of/overlay.c | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++
>> 1 file changed, 102 insertions(+)
>> 
>> diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
>> index a274d65..a7956a2 100644
>> --- a/drivers/of/overlay.c
>> +++ b/drivers/of/overlay.c
>> @@ -514,6 +514,101 @@ static int of_free_overlay_info(struct of_overlay *ov)
>>        return 0;
>> }
>> 
>> +static int of_overlay_add_symbols(
>> +               struct device_node *tree,
>> +               struct of_overlay *ov)
>> +{
>> +       struct of_overlay_info *ovinfo;
>> +       struct device_node *root_sym = NULL;
>> +       struct device_node *child = NULL;
>> +       struct property *prop;
>> +       const char *path, *s;
>> +       char *new_path;
>> +       int i, len, err;
>> +
>> +       /* this may fail (if no fixups are required) */
>> +       root_sym = of_find_node_by_path("/__symbols__");
>> +
>> +       /* do nothing if no root symbols */
>> +       if (!root_sym)
>> +               return 0;
>> +
>> +       /* locate the symbols & fixups nodes on resolve */
>> +       for_each_child_of_node(tree, child) {
>> +               if (of_node_cmp(child->name, "__symbols__") == 0)
>> +                       goto found;
> 
> Doesn't of_get_child_by_name work here?
> 

No, we’re holding of_mutex.

>> +       }
>> +       /* no symbols, no problem */
>> +       of_node_put(root_sym);
>> +       return 0;
>> +
>> +found:
> 
> As Marek said, get rid of the goto.
> 
>> +       err = -EINVAL;
>> +       for_each_property_of_node(child, prop) {
>> +
>> +               /* skip properties added automatically */
>> +               if (of_prop_cmp(prop->name, "name") == 0)
>> +                       continue;
>> +
>> +               err = of_property_read_string(child,
>> +                               prop->name, &path);
>> +               if (err != 0) {
>> +                       pr_err("%s: Could not find symbol '%s'\n",
>> +                                       __func__, prop->name);
> 
> Can you introduce and use a common pr_fmt prefix here and elsewhere
> (just what you are adding).

OK

> 
>> +                       continue;
>> +               }
>> +
>> +
> 
> Extra blank line.
> 
>> +               /* now find fragment index */
>> +               s = path;
>> +
>> +               /* compare paths to find fragment index */
>> +               ovinfo = NULL;
>> +               len = -1;
> 
> Move these into the for init.
> 
>> +               for (i = 0; i < ov->count; i++) {
>> +                       ovinfo = &ov->ovinfo_tab[i];
>> +
>> +                       pr_debug("%s: #%d: overlay->name=%s target->name=%s\n",
>> +                                       __func__, i, ovinfo->overlay->full_name,
>> +                                       ovinfo->target->full_name);
>> +
>> +                       len = strlen(ovinfo->overlay->full_name);
>> +                       if (strncasecmp(path, ovinfo->overlay->full_name,
>> +                                               len) == 0 && path[len] == '/')
>> +                               break;
>> +               }
>> +
>> +               if (i >= ov->count)
>> +                       continue;
>> +
>> +               pr_debug("%s: found target at #%d\n", __func__, i);
>> +               new_path = kasprintf(GFP_KERNEL, "%s%s",
>> +                               ovinfo->target->full_name,
>> +                               path + len);
>> +               if (!new_path) {
>> +                       pr_err("%s: Failed to allocate propname for \"%s\"\n",
>> +                                       __func__, prop->name);
>> +                       continue;
> 
> If this fails, is there much point in continuing?
> 

No

>> +               }
>> +
>> +               err = of_changeset_add_property_string(&ov->cset, root_sym,
>> +                               prop->name, new_path);
>> +
>> +               /* free always */
>> +               kfree(new_path);
>> +
>> +               if (err) {
> 
> No need for brackets.
> 
>> +                       pr_err("%s: Failed to add property for \"%s\"\n",
>> +                                       __func__, prop->name);
>> +               }
>> +       }
>> +
>> +       of_node_put(child);
>> +       of_node_put(root_sym);
>> +
>> +       return 0;
>> +}
>> +
>> static LIST_HEAD(ov_list);
>> static DEFINE_IDR(ov_idr);
>> 
>> @@ -642,6 +737,13 @@ static int __of_overlay_create(struct device_node *tree,
>>                goto err_abort_trans;
>>        }
>> 
>> +       err = of_overlay_add_symbols(tree, ov);
>> +       if (err) {
>> +               pr_err("%s: of_overlay_add_symbols() failed for tree@%s\n",
>> +                               __func__, tree->full_name);
>> +               goto err_abort_trans;
>> +       }
>> +
>>        /* apply the changeset */
>>        err = __of_changeset_apply(&ov->cset);
>>        if (err) {
>> --
>> 1.7.12
>> 

Regards

— Pantelis

Re: [PATCH 3/3] of: overlay: Pick up label symbols from overlays.

[Rob Herring]

On Tue, May 10, 2016 at 8:56 AM, Pantelis Antoniou
<panto@antoniou-consulting.com> wrote:
> Hi Rob,
>
>> On May 10, 2016, at 00:47 , Rob Herring <robherring2@gmail.com> wrote:
>>
>> On Mon, May 9, 2016 at 1:11 PM, Pantelis Antoniou
>> <pantelis.antoniou@konsulko.com> wrote:
>>> Insert overlay symbols to the base tree when applied.
>>> This makes it possible to apply an overlay that references a label
>>> that a previously inserted overlay had.

[...]

>>> +       /* locate the symbols & fixups nodes on resolve */
>>> +       for_each_child_of_node(tree, child) {
>>> +               if (of_node_cmp(child->name, "__symbols__") == 0)
>>> +                       goto found;
>>
>> Doesn't of_get_child_by_name work here?
>>
>
> No, we’re holding of_mutex.

So. Go look at the function.

Rob