diff mbox

[01/10] of: Keep track of populated platform devices

Message ID 1398707877-22596-2-git-send-email-pawel.moll@arm.com (mailing list archive)
State New, archived
Headers show

Commit Message

Pawel Moll April 28, 2014, 5:57 p.m. UTC
In "Device Tree powered" systems, platform devices are usually
massively populated with of_platform_populate() call, executed
at some level of initcalls, either by generic architecture
or by platform-specific code.

There are situations though where certain devices must be
created (and bound with drivers) before all the others.
This presents a challenge, as devices created explicitly
would be created again by of_platform_populate().

This patch tries to solve that issue in a generic way,
adding a "populated" flag for a DT node description.
Once set, this device will never be created again via
of_* API, so of_platform_populate() will skip such nodes
(and its children) in a similar way to the non-available
ones.

Cc: devicetree@vger.kernel.org
Signed-off-by: Pawel Moll <pawel.moll@arm.com>
---
 drivers/of/platform.c | 10 ++++++++--
 include/linux/of.h    |  1 +
 2 files changed, 9 insertions(+), 2 deletions(-)

Comments

Rob Herring April 28, 2014, 6:02 p.m. UTC | #1
On Mon, Apr 28, 2014 at 12:57 PM, Pawel Moll <pawel.moll@arm.com> wrote:
> In "Device Tree powered" systems, platform devices are usually
> massively populated with of_platform_populate() call, executed
> at some level of initcalls, either by generic architecture
> or by platform-specific code.
>
> There are situations though where certain devices must be
> created (and bound with drivers) before all the others.
> This presents a challenge, as devices created explicitly
> would be created again by of_platform_populate().
>
> This patch tries to solve that issue in a generic way,
> adding a "populated" flag for a DT node description.
> Once set, this device will never be created again via
> of_* API, so of_platform_populate() will skip such nodes
> (and its children) in a similar way to the non-available
> ones.
>
> Cc: devicetree@vger.kernel.org
> Signed-off-by: Pawel Moll <pawel.moll@arm.com>
> ---
>  drivers/of/platform.c | 10 ++++++++--
>  include/linux/of.h    |  1 +
>  2 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/of/platform.c b/drivers/of/platform.c
> index 404d1da..0ae757a 100644
> --- a/drivers/of/platform.c
> +++ b/drivers/of/platform.c
> @@ -204,7 +204,8 @@ static struct platform_device *of_platform_device_create_pdata(
>  {
>         struct platform_device *dev;
>
> -       if (!of_device_is_available(np))
> +       if (!of_device_is_available(np) ||
> +                       of_node_check_flag(np, OF_POPULATED))

This and the amba case should be a test_and_set operation to avoid a
race condition.

Rob
Grant Likely April 29, 2014, 12:56 p.m. UTC | #2
On Mon, 28 Apr 2014 13:02:28 -0500, Rob Herring <robherring2@gmail.com> wrote:
> On Mon, Apr 28, 2014 at 12:57 PM, Pawel Moll <pawel.moll@arm.com> wrote:
> > In "Device Tree powered" systems, platform devices are usually
> > massively populated with of_platform_populate() call, executed
> > at some level of initcalls, either by generic architecture
> > or by platform-specific code.
> >
> > There are situations though where certain devices must be
> > created (and bound with drivers) before all the others.
> > This presents a challenge, as devices created explicitly
> > would be created again by of_platform_populate().
> >
> > This patch tries to solve that issue in a generic way,
> > adding a "populated" flag for a DT node description.
> > Once set, this device will never be created again via
> > of_* API, so of_platform_populate() will skip such nodes
> > (and its children) in a similar way to the non-available
> > ones.
> >
> > Cc: devicetree@vger.kernel.org
> > Signed-off-by: Pawel Moll <pawel.moll@arm.com>
> > ---
> >  drivers/of/platform.c | 10 ++++++++--
> >  include/linux/of.h    |  1 +
> >  2 files changed, 9 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/of/platform.c b/drivers/of/platform.c
> > index 404d1da..0ae757a 100644
> > --- a/drivers/of/platform.c
> > +++ b/drivers/of/platform.c
> > @@ -204,7 +204,8 @@ static struct platform_device *of_platform_device_create_pdata(
> >  {
> >         struct platform_device *dev;
> >
> > -       if (!of_device_is_available(np))
> > +       if (!of_device_is_available(np) ||
> > +                       of_node_check_flag(np, OF_POPULATED))
> 
> This and the amba case should be a test_and_set operation to avoid a
> race condition.

Yes. It is an unlikely condition at the time that Pawel is handling, but
it should also work for later calls to of_platform_populate() which may
happen at runtime and in parallel.

g.
Pawel Moll April 30, 2014, 11:48 a.m. UTC | #3
On Tue, 2014-04-29 at 13:56 +0100, Grant Likely wrote:
> > > diff --git a/drivers/of/platform.c b/drivers/of/platform.c
> > > index 404d1da..0ae757a 100644
> > > --- a/drivers/of/platform.c
> > > +++ b/drivers/of/platform.c
> > > @@ -204,7 +204,8 @@ static struct platform_device *of_platform_device_create_pdata(
> > >  {
> > >         struct platform_device *dev;
> > >
> > > -       if (!of_device_is_available(np))
> > > +       if (!of_device_is_available(np) ||
> > > +                       of_node_check_flag(np, OF_POPULATED))
> > 
> > This and the amba case should be a test_and_set operation to avoid a
> > race condition.
> 
> Yes. It is an unlikely condition at the time that Pawel is handling, but
> it should also work for later calls to of_platform_populate() which may
> happen at runtime and in parallel.

Sure thing, will fix.

Thanks!

Pawel
diff mbox

Patch

diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index 404d1da..0ae757a 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -204,7 +204,8 @@  static struct platform_device *of_platform_device_create_pdata(
 {
 	struct platform_device *dev;
 
-	if (!of_device_is_available(np))
+	if (!of_device_is_available(np) ||
+			of_node_check_flag(np, OF_POPULATED))
 		return NULL;
 
 	dev = of_device_alloc(np, bus_id, parent);
@@ -230,6 +231,8 @@  static struct platform_device *of_platform_device_create_pdata(
 		return NULL;
 	}
 
+	of_node_set_flag(np, OF_POPULATED);
+
 	return dev;
 }
 
@@ -262,7 +265,8 @@  static struct amba_device *of_amba_device_create(struct device_node *node,
 
 	pr_debug("Creating amba device %s\n", node->full_name);
 
-	if (!of_device_is_available(node))
+	if (!of_device_is_available(node) ||
+			of_node_check_flag(node, OF_POPULATED))
 		return NULL;
 
 	dev = amba_device_alloc(NULL, 0, 0);
@@ -305,6 +309,8 @@  static struct amba_device *of_amba_device_create(struct device_node *node,
 		goto err_free;
 	}
 
+	of_node_set_flag(node, OF_POPULATED);
+
 	return dev;
 
 err_free:
diff --git a/include/linux/of.h b/include/linux/of.h
index 3bad8d1..fd3aaee 100644
--- a/include/linux/of.h
+++ b/include/linux/of.h
@@ -197,6 +197,7 @@  static inline unsigned long of_read_ulong(const __be32 *cell, int size)
 /* flag descriptions */
 #define OF_DYNAMIC	1 /* node and properties were allocated via kmalloc */
 #define OF_DETACHED	2 /* node has been detached from the device tree */
+#define OF_POPULATED	3 /* device already created for the node */
 
 #define OF_IS_DYNAMIC(x) test_bit(OF_DYNAMIC, &x->_flags)
 #define OF_MARK_DYNAMIC(x) set_bit(OF_DYNAMIC, &x->_flags)