| Message ID | 20210408203611.544005-1-hdegoede@redhat.com |
|---|---|
| State | New |
| Headers | show |
| Series | usb: roles: Cakk try_module_get() from usb_role_switch_find_by_fwnode() | expand |
On Thu, Apr 08, 2021 at 10:36:11PM +0200, Hans de Goede wrote: > usb_role_switch_find_by_fwnode() returns a reference to the role-switch > which must be put by calling usb_role_switch_put(). > > usb_role_switch_put() calls module_put(sw->dev.parent->driver->owner), > add a matching try_module_get() to usb_role_switch_find_by_fwnode(), > making it behave the same as the other usb_role_switch functions > which return a reference. > > This avoids a WARN_ON being hit at kernel/module.c:1158 due to the > module-refcount going below 0. > Took me a while to figure out what the subject line is supposed to mean. s/Cakk/Call/ Otherwise Reviewed-by: Guenter Roeck <linux@roeck-us.net> It might be useful though to explain the difference between fwnode_usb_role_switch_get() and usb_role_switch_find_by_fwnode(), and why two different functions are needed, both passing fwnode as parameter and returning a pointer to usb_role_switch. Thanks, Guenter > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > --- > drivers/usb/roles/class.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/usb/roles/class.c b/drivers/usb/roles/class.c > index 97f37077b7f9..33b637d0d8d9 100644 > --- a/drivers/usb/roles/class.c > +++ b/drivers/usb/roles/class.c > @@ -189,6 +189,8 @@ usb_role_switch_find_by_fwnode(const struct fwnode_handle *fwnode) > return NULL; > > dev = class_find_device_by_fwnode(role_class, fwnode); > + if (dev) > + WARN_ON(!try_module_get(dev->parent->driver->owner)); > > return dev ? to_role_switch(dev) : NULL; > } > -- > 2.30.2 >
On Thu, Apr 08, 2021 at 04:09:04PM -0700, Guenter Roeck wrote: > On Thu, Apr 08, 2021 at 10:36:11PM +0200, Hans de Goede wrote: > > usb_role_switch_find_by_fwnode() returns a reference to the role-switch > > which must be put by calling usb_role_switch_put(). > > > > usb_role_switch_put() calls module_put(sw->dev.parent->driver->owner), > > add a matching try_module_get() to usb_role_switch_find_by_fwnode(), > > making it behave the same as the other usb_role_switch functions > > which return a reference. > > > > This avoids a WARN_ON being hit at kernel/module.c:1158 due to the > > module-refcount going below 0. > > > > Took me a while to figure out what the subject line is supposed > to mean. > > s/Cakk/Call/ > > Otherwise > > Reviewed-by: Guenter Roeck <linux@roeck-us.net> > > It might be useful though to explain the difference between > fwnode_usb_role_switch_get() and usb_role_switch_find_by_fwnode(), > and why two different functions are needed, both passing fwnode > as parameter and returning a pointer to usb_role_switch. Yes, the function names are confusing indeed. My proposal is to rename usb_role_switch_find_by_fwnode() to fwnode_to_usb_role_switch(). I can prepare a patch for that if you guys are OK with it, or Hans, would you prefer to send that together with this one? Actually, shouldn't this be marked as a fix? thanks, -- heikki
Hi, On 4/9/21 11:30 AM, Heikki Krogerus wrote: > On Thu, Apr 08, 2021 at 04:09:04PM -0700, Guenter Roeck wrote: >> On Thu, Apr 08, 2021 at 10:36:11PM +0200, Hans de Goede wrote: >>> usb_role_switch_find_by_fwnode() returns a reference to the role-switch >>> which must be put by calling usb_role_switch_put(). >>> >>> usb_role_switch_put() calls module_put(sw->dev.parent->driver->owner), >>> add a matching try_module_get() to usb_role_switch_find_by_fwnode(), >>> making it behave the same as the other usb_role_switch functions >>> which return a reference. >>> >>> This avoids a WARN_ON being hit at kernel/module.c:1158 due to the >>> module-refcount going below 0. >>> >> >> Took me a while to figure out what the subject line is supposed >> to mean. >> >> s/Cakk/Call/ >> >> Otherwise >> >> Reviewed-by: Guenter Roeck <linux@roeck-us.net> >> >> It might be useful though to explain the difference between >> fwnode_usb_role_switch_get() and usb_role_switch_find_by_fwnode(), >> and why two different functions are needed, both passing fwnode >> as parameter and returning a pointer to usb_role_switch. Sorry about thetypo, I completely missed that while preparing the patch, fixed for v2. > Yes, the function names are confusing indeed. My proposal is to rename > usb_role_switch_find_by_fwnode() to fwnode_to_usb_role_switch(). > > I can prepare a patch for that if you guys are OK with it, or Hans, > would you prefer to send that together with this one? If you can send a patch to apply on top of my v2 of this patch then that would be great. > Actually, shouldn't this be marked as a fix? That is a good point I've added a fixes tag for v2. Regards, Hans
diff --git a/drivers/usb/roles/class.c b/drivers/usb/roles/class.c index 97f37077b7f9..33b637d0d8d9 100644 --- a/drivers/usb/roles/class.c +++ b/drivers/usb/roles/class.c @@ -189,6 +189,8 @@ usb_role_switch_find_by_fwnode(const struct fwnode_handle *fwnode) return NULL; dev = class_find_device_by_fwnode(role_class, fwnode); + if (dev) + WARN_ON(!try_module_get(dev->parent->driver->owner)); return dev ? to_role_switch(dev) : NULL; }
usb_role_switch_find_by_fwnode() returns a reference to the role-switch which must be put by calling usb_role_switch_put(). usb_role_switch_put() calls module_put(sw->dev.parent->driver->owner), add a matching try_module_get() to usb_role_switch_find_by_fwnode(), making it behave the same as the other usb_role_switch functions which return a reference. This avoids a WARN_ON being hit at kernel/module.c:1158 due to the module-refcount going below 0. Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- drivers/usb/roles/class.c | 2 ++ 1 file changed, 2 insertions(+)