On Mon, Apr 8, 2024 at 6:40 PM Saravana Kannan <saravanak@xxxxxxxxxx> wrote:
>
> On Mon, Apr 8, 2024 at 4:13 PM Saravana Kannan <saravanak@xxxxxxxxxx> wrote:
> >
> > Get fw_devlink to work well with overlay devices.
> >
> > Signed-off-by: Saravana Kannan <saravanak@xxxxxxxxxx>
> > ---
> > drivers/base/core.c | 10 ++++++++++
> > drivers/of/dynamic.c | 8 ++++++++
> > include/linux/fwnode.h | 2 ++
> > 3 files changed, 20 insertions(+)
> >
> > diff --git a/drivers/base/core.c b/drivers/base/core.c
> > index 5f4e03336e68..d856f9c5d601 100644
> > --- a/drivers/base/core.c
> > +++ b/drivers/base/core.c
> > @@ -237,6 +237,16 @@ static void __fw_devlink_pickup_dangling_consumers(struct fwnode_handle *fwnode,
> > __fw_devlink_pickup_dangling_consumers(child, new_sup);
> > }
> >
> > +
> > +void fw_devlink_pickup_dangling_consumers(struct fwnode_handle *child,
> > + struct fwnode_handle *parent)
> > +{
> > + mutex_lock(&fwnode_link_lock);
> > + __fw_devlink_pickup_dangling_consumers(child, parent);
> > + __fw_devlink_link_to_consumers(parent->dev);
> > + mutex_unlock(&fwnode_link_lock);
> > +}
> > +
> > static DEFINE_MUTEX(device_links_lock);
> > DEFINE_STATIC_SRCU(device_links_srcu);
> >
> > diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c
> > index 19a1a38554f2..0a936f46820e 100644
> > --- a/drivers/of/dynamic.c
> > +++ b/drivers/of/dynamic.c
> > @@ -237,6 +237,7 @@ static void __of_attach_node(struct device_node *np)
> > int of_attach_node(struct device_node *np)
> > {
> > struct of_reconfig_data rd;
> > + struct fwnode_handle *fwnode, *parent;
> >
> > memset(&rd, 0, sizeof(rd));
> > rd.dn = np;
> > @@ -246,6 +247,13 @@ int of_attach_node(struct device_node *np)
> > mutex_unlock(&of_mutex);
> >
> > of_reconfig_notify(OF_RECONFIG_ATTACH_NODE, &rd);
> > + fwnode = of_fwnode_handle(np);
> > + fwnode_for_each_parent_node(fwnode, parent)
> > + if (parent->dev) {
> > + fw_devlink_pickup_dangling_consumers(fwnode, parent);
> > + fwnode_handle_put(parent);
> > + break;
> > + }
>
> I'm clearly calling this in the wrong location. Please move this logic
> to __of_changeset_entry_notify() and for the case
> OF_RECONFIG_ATTACH_NODE. Haven't fully thought through the DETACH
> case, but it should work correctly for that case too. If not, I'll
> take care of that next.
>
I'll send out a RFC v2 with the code fixed up in a few minutes.
> -Saravana
>
> >
> > return 0;
> > }
> > diff --git a/include/linux/fwnode.h b/include/linux/fwnode.h
> > index 0d79070c5a70..4b3f697a90e8 100644
> > --- a/include/linux/fwnode.h
> > +++ b/include/linux/fwnode.h
> > @@ -220,6 +220,8 @@ int fwnode_link_add(struct fwnode_handle *con, struct fwnode_handle *sup,
> > u8 flags);
> > void fwnode_links_purge(struct fwnode_handle *fwnode);
> > void fw_devlink_purge_absent_suppliers(struct fwnode_handle *fwnode);
> > +void fw_devlink_pickup_dangling_consumers(struct fwnode_handle *child,
> > + struct fwnode_handle *parent);
> > bool fw_devlink_is_strict(void);
> >
> > #endif
> > --
> > 2.44.0.478.gd926399ef9-goog
> >
