So I've been annoyed lately with having a bunch of devices such as i2c
eeproms (for use by VPDs, server world !) and other bits and pieces that
I want to be able to identify from userspace, and possibly provide
additional data about from FW.
Basically, it boils down to correlating the sysfs device with the OF
tree device node, so that user space can use device-tree info such as
additional "location" or "label" (or whatever else we can come up with)
properties to identify a given device, or get some attributes of use
about it, etc...
Now, so far, we've done that in some subsystem in a fairly ad-hoc basis
using "devspec" properties. For example, PCI creates them if it can
correlate the probed device with a DT node. Some powerpc specific busses
do that too.
However, i2c doesn't and it would be nice to have something more generic
since technically any device can have a corresponding device tree node
these days.
This patch achieves this by adding an "of_node" symlink to devices that
have a non-NULL dev->of_node pointer.
Signed-off-by: Benjamin Herrenschmidt <[email protected]>
---
drivers/base/core.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/drivers/base/core.c b/drivers/base/core.c
index 97e2baf..61ef574 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -469,6 +469,7 @@ static int device_add_attrs(struct device *dev)
{
struct class *class = dev->class;
const struct device_type *type = dev->type;
+ struct device_node *of_node;
int error;
if (class) {
@@ -493,6 +494,13 @@ static int device_add_attrs(struct device *dev)
goto err_remove_dev_groups;
}
+ of_node = dev_of_node(dev);
+ if (of_node) {
+ error = sysfs_create_link(&dev->kobj, &of_node->kobj, "of_node");
+ if (error)
+ dev_warn(dev, "Error %d creating of_node link\n", error);
+ }
+
return 0;
err_remove_dev_groups:
On Monday 16 February 2015 14:59:16 Benjamin Herrenschmidt wrote:
> So I've been annoyed lately with having a bunch of devices such as i2c
> eeproms (for use by VPDs, server world !) and other bits and pieces that
> I want to be able to identify from userspace, and possibly provide
> additional data about from FW.
>
> Basically, it boils down to correlating the sysfs device with the OF
> tree device node, so that user space can use device-tree info such as
> additional "location" or "label" (or whatever else we can come up with)
> properties to identify a given device, or get some attributes of use
> about it, etc...
>
> Now, so far, we've done that in some subsystem in a fairly ad-hoc basis
> using "devspec" properties. For example, PCI creates them if it can
> correlate the probed device with a DT node. Some powerpc specific busses
> do that too.
>
> However, i2c doesn't and it would be nice to have something more generic
> since technically any device can have a corresponding device tree node
> these days.
>
> This patch achieves this by adding an "of_node" symlink to devices that
> have a non-NULL dev->of_node pointer.
>
> Signed-off-by: Benjamin Herrenschmidt <[email protected]>
I like the idea, this seems very useful.
Acked-by: Arnd Bergmann <[email protected]>
+GregKH
On Sun, Feb 15, 2015 at 9:59 PM, Benjamin Herrenschmidt
<[email protected]> wrote:
> So I've been annoyed lately with having a bunch of devices such as i2c
> eeproms (for use by VPDs, server world !) and other bits and pieces that
> I want to be able to identify from userspace, and possibly provide
> additional data about from FW.
>
> Basically, it boils down to correlating the sysfs device with the OF
> tree device node, so that user space can use device-tree info such as
> additional "location" or "label" (or whatever else we can come up with)
> properties to identify a given device, or get some attributes of use
> about it, etc...
>
> Now, so far, we've done that in some subsystem in a fairly ad-hoc basis
> using "devspec" properties. For example, PCI creates them if it can
> correlate the probed device with a DT node. Some powerpc specific busses
> do that too.
>
> However, i2c doesn't and it would be nice to have something more generic
> since technically any device can have a corresponding device tree node
> these days.
>
> This patch achieves this by adding an "of_node" symlink to devices that
> have a non-NULL dev->of_node pointer.
Doesn't this need sysfs documentation?
> Signed-off-by: Benjamin Herrenschmidt <[email protected]>
I don't mind applying, but this is Greg's code.
Rob
> ---
> drivers/base/core.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/drivers/base/core.c b/drivers/base/core.c
> index 97e2baf..61ef574 100644
> --- a/drivers/base/core.c
> +++ b/drivers/base/core.c
> @@ -469,6 +469,7 @@ static int device_add_attrs(struct device *dev)
> {
> struct class *class = dev->class;
> const struct device_type *type = dev->type;
> + struct device_node *of_node;
> int error;
>
> if (class) {
> @@ -493,6 +494,13 @@ static int device_add_attrs(struct device *dev)
> goto err_remove_dev_groups;
> }
>
> + of_node = dev_of_node(dev);
> + if (of_node) {
> + error = sysfs_create_link(&dev->kobj, &of_node->kobj, "of_node");
> + if (error)
> + dev_warn(dev, "Error %d creating of_node link\n", error);
> + }
> +
> return 0;
>
> err_remove_dev_groups:
>
>
>
On Tue, Feb 17, 2015 at 09:00:47AM -0600, Rob Herring wrote:
> +GregKH
>
> On Sun, Feb 15, 2015 at 9:59 PM, Benjamin Herrenschmidt
> <[email protected]> wrote:
> > So I've been annoyed lately with having a bunch of devices such as i2c
> > eeproms (for use by VPDs, server world !) and other bits and pieces that
> > I want to be able to identify from userspace, and possibly provide
> > additional data about from FW.
> >
> > Basically, it boils down to correlating the sysfs device with the OF
> > tree device node, so that user space can use device-tree info such as
> > additional "location" or "label" (or whatever else we can come up with)
> > properties to identify a given device, or get some attributes of use
> > about it, etc...
> >
> > Now, so far, we've done that in some subsystem in a fairly ad-hoc basis
> > using "devspec" properties. For example, PCI creates them if it can
> > correlate the probed device with a DT node. Some powerpc specific busses
> > do that too.
> >
> > However, i2c doesn't and it would be nice to have something more generic
> > since technically any device can have a corresponding device tree node
> > these days.
> >
> > This patch achieves this by adding an "of_node" symlink to devices that
> > have a non-NULL dev->of_node pointer.
>
> Doesn't this need sysfs documentation?
Yes it does.
> > Signed-off-by: Benjamin Herrenschmidt <[email protected]>
>
> I don't mind applying, but this is Greg's code.
>
> Rob
>
> > ---
> > drivers/base/core.c | 8 ++++++++
> > 1 file changed, 8 insertions(+)
> >
> > diff --git a/drivers/base/core.c b/drivers/base/core.c
> > index 97e2baf..61ef574 100644
> > --- a/drivers/base/core.c
> > +++ b/drivers/base/core.c
> > @@ -469,6 +469,7 @@ static int device_add_attrs(struct device *dev)
> > {
> > struct class *class = dev->class;
> > const struct device_type *type = dev->type;
> > + struct device_node *of_node;
> > int error;
> >
> > if (class) {
> > @@ -493,6 +494,13 @@ static int device_add_attrs(struct device *dev)
> > goto err_remove_dev_groups;
> > }
> >
> > + of_node = dev_of_node(dev);
> > + if (of_node) {
> > + error = sysfs_create_link(&dev->kobj, &of_node->kobj, "of_node");
> > + if (error)
> > + dev_warn(dev, "Error %d creating of_node link\n", error);
> > + }
You don't clean up the link when the device is removed?
And why do it here and not in device_add_class_symlinks() like all other
devices do?
thanks,
greg k-h
On Tue, 2015-02-17 at 07:09 -0800, Greg Kroah-Hartman wrote:
> >
> > Doesn't this need sysfs documentation?
>
> Yes it does.
Ok, will do.
> > > Signed-off-by: Benjamin Herrenschmidt <[email protected]>
> >
> > I don't mind applying, but this is Greg's code.
> >
> > Rob
> >
> > > ---
> > > drivers/base/core.c | 8 ++++++++
> > > 1 file changed, 8 insertions(+)
> > >
> > > diff --git a/drivers/base/core.c b/drivers/base/core.c
> > > index 97e2baf..61ef574 100644
> > > --- a/drivers/base/core.c
> > > +++ b/drivers/base/core.c
> > > @@ -469,6 +469,7 @@ static int device_add_attrs(struct device *dev)
> > > {
> > > struct class *class = dev->class;
> > > const struct device_type *type = dev->type;
> > > + struct device_node *of_node;
> > > int error;
> > >
> > > if (class) {
> > > @@ -493,6 +494,13 @@ static int device_add_attrs(struct device *dev)
> > > goto err_remove_dev_groups;
> > > }
> > >
> > > + of_node = dev_of_node(dev);
> > > + if (of_node) {
> > > + error = sysfs_create_link(&dev->kobj, &of_node->kobj, "of_node");
> > > + if (error)
> > > + dev_warn(dev, "Error %d creating of_node link\n", error);
> > > + }
>
> You don't clean up the link when the device is removed?
Hrm, I suppose I do :-)
> And why do it here and not in device_add_class_symlinks() like all other
> devices do?
Well, device_add_class_symlinks() had "class" in the name so I
(incorrectly) assumed that was for class related stuff, but I see
that the parent pointer is there too. Also originally my patch was
adding an attribute and got turned into a symlink later. I'll move
it.
Cheers,
Ben.
> thanks,
>
> greg k-h
> --
> To unsubscribe from this list: send the line "unsubscribe devicetree" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, Feb 15, 2015 at 7:59 PM, Benjamin Herrenschmidt
<[email protected]> wrote:
> So I've been annoyed lately with having a bunch of devices such as i2c
> eeproms (for use by VPDs, server world !) and other bits and pieces that
> I want to be able to identify from userspace, and possibly provide
> additional data about from FW.
>
> Basically, it boils down to correlating the sysfs device with the OF
> tree device node, so that user space can use device-tree info such as
> additional "location" or "label" (or whatever else we can come up with)
> properties to identify a given device, or get some attributes of use
> about it, etc...
>
> Now, so far, we've done that in some subsystem in a fairly ad-hoc basis
> using "devspec" properties. For example, PCI creates them if it can
> correlate the probed device with a DT node. Some powerpc specific busses
> do that too.
>
> However, i2c doesn't and it would be nice to have something more generic
> since technically any device can have a corresponding device tree node
> these days.
>
> This patch achieves this by adding an "of_node" symlink to devices that
> have a non-NULL dev->of_node pointer.
>
> Signed-off-by: Benjamin Herrenschmidt <[email protected]>
Concept:
Acked-by: Olof Johansson <[email protected]>
We've come across similar needs on product, and always had to do some
less-than-optimal alternate solution. This seems like a good idea to
me.
-Olof