Hi Alexander,
alex.aring@xxxxxxxxx wrote on Sun, 23 Jan 2022 15:34:14 -0500:
> Hi,
>
> On Thu, 20 Jan 2022 at 06:21, Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote:
> >
> > Drivers are expected to set the PHY current_channel and current_page
> > according to their default state. The hwsim driver is advertising being
> > configured on channel 13 by default but that is not reflected in its own
> > internal pib structure. In order to ensure that this driver consider the
> > current channel as being 13 internally, we can call hwsim_hw_channel()
> > instead of creating an empty pib structure.
> >
> > We assume here that kvfree_rcu(NULL) is a valid call.
> >
> > Fixes: f25da51fdc38 ("ieee802154: hwsim: add replacement for fakelb")
> > Signed-off-by: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
> > ---
> > drivers/net/ieee802154/mac802154_hwsim.c | 10 +---------
> > 1 file changed, 1 insertion(+), 9 deletions(-)
> >
> > diff --git a/drivers/net/ieee802154/mac802154_hwsim.c b/drivers/net/ieee802154/mac802154_hwsim.c
> > index 8caa61ec718f..795f8eb5387b 100644
> > --- a/drivers/net/ieee802154/mac802154_hwsim.c
> > +++ b/drivers/net/ieee802154/mac802154_hwsim.c
> > @@ -732,7 +732,6 @@ static int hwsim_add_one(struct genl_info *info, struct device *dev,
> > {
> > struct ieee802154_hw *hw;
> > struct hwsim_phy *phy;
> > - struct hwsim_pib *pib;
> > int idx;
> > int err;
> >
> > @@ -780,13 +779,8 @@ static int hwsim_add_one(struct genl_info *info, struct device *dev,
> >
> > /* hwsim phy channel 13 as default */
> > hw->phy->current_channel = 13;
> > - pib = kzalloc(sizeof(*pib), GFP_KERNEL);
> > - if (!pib) {
> > - err = -ENOMEM;
> > - goto err_pib;
> > - }
> > + hwsim_hw_channel(hw, hw->phy->current_page, hw->phy->current_channel);
>
> Probably you saw it already; this will end in a
> "suspicious_RCU_usage", that's because of an additional lock check in
> hwsim_hw_channel() which checks if rtnl is held. However, in this
> situation it's not necessary to hold the rtnl lock because we know the
> phy is not being registered yet.
yes, indeed!
>
> Either we change it to rcu_derefence() but then we would reduce the
> check if rtnl lock is being held or just simply initial the default
> pib->channel here to 13 which makes that whole patch a one line fix.
In general I like to drop more lines than I add, hence my first choice
but just setting pib->channel to 13 also makes sense here and avoids
oversimplifying the rtnl check in hwsim_hw_channel(), so let's go for
it.
Thanks,
Miquèl
