[PATCH 0/2] i2c: prevent id collisions in the DT case

[PATCH 0/2] i2c: prevent id collisions in the DT case

[Wolfram Sang]

This series tries to fix the issue reported here[1]. In short: A DT entry wants
a fixed id but its probe was deferred. When it was probed again, its id was
taken by a dynamically assigned muxed bus. So, the probe failed.

Let's start all dynamically assigned ids after the highest fixed one. We do
this for legacy platform devices already.

Please let me know what you think. If you can accept the new of-helper
function, I'd like an ack so I can take the patches via my i2c tree.

Thanks,

   Wolfram


[1] http://thread.gmane.org/gmane.linux.drivers.i2c/22166

Wolfram Sang (2):
  of: base: add function to get highest id of an alias stem
  i2c: busses with dynamic ids should start after fixed ids for DT

 drivers/i2c/i2c-core.c |  8 ++++++++
 drivers/of/base.c      | 26 ++++++++++++++++++++++++++
 include/linux/of.h     |  6 ++++++
 3 files changed, 40 insertions(+)

-- 
2.1.4

[PATCH 1/2] of: base: add function to get highest id of an alias stem

[Wolfram Sang]

I2C supports adding adapters using either a dynamic or fixed id. The
latter is provided by aliases in the DT case. To prevent id collisions
of those two types, install this function which gives us the highest
fixed id, so we can then let the dynamically created ones come after
this highest number.

Signed-off-by: Wolfram Sang <wsa-z923LK4zBo2bacvFa/9K2g@public.gmane.org>
---

Because this function is so similar to of_alias_get_id(), I thought of merging
them into __of_alias_get_id(np, stem, bool get_highest) and have two call
wrappers but then decided the decreased readability is not worth the hazzle.

 drivers/of/base.c  | 26 ++++++++++++++++++++++++++
 include/linux/of.h |  6 ++++++
 2 files changed, 32 insertions(+)

diff --git a/drivers/of/base.c b/drivers/of/base.c
index 0a8aeb8523fe7d..63cba04aacf686 100644
--- a/drivers/of/base.c
+++ b/drivers/of/base.c
@@ -1958,6 +1958,32 @@ int of_alias_get_id(struct device_node *np, const char *stem)
 }
 EXPORT_SYMBOL_GPL(of_alias_get_id);
 
+/**
+ * of_alias_get_highest_id - Get highest alias id for the given stem
+ * @stem:	Alias stem to be examined
+ *
+ * The function travels the lookup table to get the highest alias id for the
+ * given alias stem.  It returns the alias id if found.
+ */
+int of_alias_get_highest_id(const char *stem)
+{
+	struct alias_prop *app;
+	int id = -ENODEV;
+
+	mutex_lock(&of_mutex);
+	list_for_each_entry(app, &aliases_lookup, link) {
+		if (strcmp(app->stem, stem) != 0)
+			continue;
+
+		if (app->id > id)
+			id = app->id;
+	}
+	mutex_unlock(&of_mutex);
+
+	return id;
+}
+EXPORT_SYMBOL_GPL(of_alias_get_highest_id);
+
 const __be32 *of_prop_next_u32(struct property *prop, const __be32 *cur,
 			       u32 *pu)
 {
diff --git a/include/linux/of.h b/include/linux/of.h
index dfde07e77a632b..9bfcc18ceab3bf 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -332,6 +332,7 @@ extern int of_count_phandle_with_args(const struct device_node *np,
 
 extern void of_alias_scan(void * (*dt_alloc)(u64 size, u64 align));
 extern int of_alias_get_id(struct device_node *np, const char *stem);
+extern int of_alias_get_highest_id(const char *stem);
 
 extern int of_machine_is_compatible(const char *compat);
 
@@ -594,6 +595,11 @@ static inline int of_alias_get_id(struct device_node *np, const char *stem)
 	return -ENOSYS;
 }
 
+static inline int of_alias_get_highest_id(const char *stem)
+{
+	return -ENOSYS;
+}
+
 static inline int of_machine_is_compatible(const char *compat)
 {
 	return 0;
-- 
2.1.4

Re: [PATCH 1/2] of: base: add function to get highest id of an alias stem

[Rob Herring]

On Thu, Mar 12, 2015 at 11:17 AM, Wolfram Sang <wsa@the-dreams.de> wrote:
> I2C supports adding adapters using either a dynamic or fixed id. The
> latter is provided by aliases in the DT case. To prevent id collisions
> of those two types, install this function which gives us the highest
> fixed id, so we can then let the dynamically created ones come after
> this highest number.
>
> Signed-off-by: Wolfram Sang <wsa@the-dreams.de>

Acked-by: Rob Herring <robh@kernel.org>

> ---
>
> Because this function is so similar to of_alias_get_id(), I thought of merging
> them into __of_alias_get_id(np, stem, bool get_highest) and have two call
> wrappers but then decided the decreased readability is not worth the hazzle.
>
>  drivers/of/base.c  | 26 ++++++++++++++++++++++++++
>  include/linux/of.h |  6 ++++++
>  2 files changed, 32 insertions(+)
>
> diff --git a/drivers/of/base.c b/drivers/of/base.c
> index 0a8aeb8523fe7d..63cba04aacf686 100644
> --- a/drivers/of/base.c
> +++ b/drivers/of/base.c
> @@ -1958,6 +1958,32 @@ int of_alias_get_id(struct device_node *np, const char *stem)
>  }
>  EXPORT_SYMBOL_GPL(of_alias_get_id);
>
> +/**
> + * of_alias_get_highest_id - Get highest alias id for the given stem
> + * @stem:      Alias stem to be examined
> + *
> + * The function travels the lookup table to get the highest alias id for the
> + * given alias stem.  It returns the alias id if found.
> + */
> +int of_alias_get_highest_id(const char *stem)
> +{
> +       struct alias_prop *app;
> +       int id = -ENODEV;
> +
> +       mutex_lock(&of_mutex);
> +       list_for_each_entry(app, &aliases_lookup, link) {
> +               if (strcmp(app->stem, stem) != 0)
> +                       continue;
> +
> +               if (app->id > id)
> +                       id = app->id;
> +       }
> +       mutex_unlock(&of_mutex);
> +
> +       return id;
> +}
> +EXPORT_SYMBOL_GPL(of_alias_get_highest_id);
> +
>  const __be32 *of_prop_next_u32(struct property *prop, const __be32 *cur,
>                                u32 *pu)
>  {
> diff --git a/include/linux/of.h b/include/linux/of.h
> index dfde07e77a632b..9bfcc18ceab3bf 100644
> --- a/include/linux/of.h
> +++ b/include/linux/of.h
> @@ -332,6 +332,7 @@ extern int of_count_phandle_with_args(const struct device_node *np,
>
>  extern void of_alias_scan(void * (*dt_alloc)(u64 size, u64 align));
>  extern int of_alias_get_id(struct device_node *np, const char *stem);
> +extern int of_alias_get_highest_id(const char *stem);
>
>  extern int of_machine_is_compatible(const char *compat);
>
> @@ -594,6 +595,11 @@ static inline int of_alias_get_id(struct device_node *np, const char *stem)
>         return -ENOSYS;
>  }
>
> +static inline int of_alias_get_highest_id(const char *stem)
> +{
> +       return -ENOSYS;
> +}
> +
>  static inline int of_machine_is_compatible(const char *compat)
>  {
>         return 0;
> --
> 2.1.4
>

Re: [PATCH 1/2] of: base: add function to get highest id of an alias stem

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 492 bytes --]

On Thu, Mar 12, 2015 at 05:17:58PM +0100, Wolfram Sang wrote:
> I2C supports adding adapters using either a dynamic or fixed id. The
> latter is provided by aliases in the DT case. To prevent id collisions
> of those two types, install this function which gives us the highest
> fixed id, so we can then let the dynamically created ones come after
> this highest number.
> 
> Signed-off-by: Wolfram Sang <wsa-z923LK4zBo2bacvFa/9K2g@public.gmane.org>

Applied to for-next, thanks!


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH 1/2] of: base: add function to get highest id of an alias stem

[Grant Likely]

On Wed, 18 Mar 2015 13:46:31 +0100
, Wolfram Sang <wsa@the-dreams.de>
 wrote:
> On Thu, Mar 12, 2015 at 05:17:58PM +0100, Wolfram Sang wrote:
> > I2C supports adding adapters using either a dynamic or fixed id. The
> > latter is provided by aliases in the DT case. To prevent id collisions
> > of those two types, install this function which gives us the highest
> > fixed id, so we can then let the dynamically created ones come after
> > this highest number.
> > 
> > Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
> 
> Applied to for-next, thanks!

Looks good to me to. Thanks for doing this. It should also be useful for
other devices.

g.

[PATCH 2/2] i2c: busses with dynamic ids should start after fixed ids for DT

[Wolfram Sang]

Make sure dynamic ids do not interfere with fixed ones and let them
start after the highest fixed id. This patch might cause different
bus-numbers for dynamic ids, however it fixes a bug. Assume:

- fixed id0 defers probe
- fixed id1 succeeds and registers a muxed bus with dynamic id
- muxed bus gets id0
- fixed id0 wants to probe again, but its fixed id is gone now
- fixed id0 probe fails

With this patch, the fixed ids are always reserved in the DT case.
For legacy board init, we already have a mechanism like this in
i2c_register_board_info().

Reported-by: Bob Feretich <bob.feretich@rafresearch.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
---
 drivers/i2c/i2c-core.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/i2c/i2c-core.c b/drivers/i2c/i2c-core.c
index 210cf4874cb7ea..4dda9a529f4464 100644
--- a/drivers/i2c/i2c-core.c
+++ b/drivers/i2c/i2c-core.c
@@ -1878,6 +1878,14 @@ static int __init i2c_init(void)
 {
 	int retval;
 
+	retval = of_alias_get_highest_id("i2c");
+
+	down_write(&__i2c_board_lock);
+	if (retval >= __i2c_first_dynamic_bus_num) {
+		__i2c_first_dynamic_bus_num = retval + 1;
+	}
+	up_write(&__i2c_board_lock);
+
 	retval = bus_register(&i2c_bus_type);
 	if (retval)
 		return retval;
-- 
2.1.4

Re: [PATCH 2/2] i2c: busses with dynamic ids should start after fixed ids for DT

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 886 bytes --]

On Thu, Mar 12, 2015 at 05:17:59PM +0100, Wolfram Sang wrote:
> Make sure dynamic ids do not interfere with fixed ones and let them
> start after the highest fixed id. This patch might cause different
> bus-numbers for dynamic ids, however it fixes a bug. Assume:
> 
> - fixed id0 defers probe
> - fixed id1 succeeds and registers a muxed bus with dynamic id
> - muxed bus gets id0
> - fixed id0 wants to probe again, but its fixed id is gone now
> - fixed id0 probe fails
> 
> With this patch, the fixed ids are always reserved in the DT case.
> For legacy board init, we already have a mechanism like this in
> i2c_register_board_info().
> 
> Reported-by: Bob Feretich <bob.feretich-8wbKi1faPaosQv5ZqcSHkQ@public.gmane.org>
> Signed-off-by: Wolfram Sang <wsa-z923LK4zBo2bacvFa/9K2g@public.gmane.org>

Fixed the checkpatch warning and applied to for-next, thanks!


[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]