On 12/27/2016 05:31 PM, Kishon Vijay Abraham I wrote: > Hi, > > On Thursday 22 December 2016 03:12 PM, Jaehoon Chung wrote: >> If some ops-> callback function are not assigend, then it should do the >> unexpect behavior. >> To prevent the potential NULL pointer dereference, check the each >> callback functions before doing operation. > > The call backs checks are done after the mutex. Moreover even if the call > backs > are not assigned, the user can call the phy ops for doing pm_runtime.
Yes. I found this patch also is wrong. Thanks for pointing out. Best Regards, Jaehoon Chung > > Thanks > Kishon > >> >> Signed-off-by: Jaehoon Chung <[email protected]> >> --- >> drivers/phy/phy-core.c | 8 ++++---- >> 1 file changed, 4 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/phy/phy-core.c b/drivers/phy/phy-core.c >> index a268f4d6f3e9..e4eb4431c8a4 100644 >> --- a/drivers/phy/phy-core.c >> +++ b/drivers/phy/phy-core.c >> @@ -219,7 +219,7 @@ int phy_init(struct phy *phy) >> { >> int ret; >> >> - if (!phy) >> + if (!phy || !phy->ops->init) >> return 0; >> >> ret = phy_pm_runtime_get_sync(phy); >> @@ -248,7 +248,7 @@ int phy_exit(struct phy *phy) >> { >> int ret; >> >> - if (!phy) >> + if (!phy || !phy->ops->exit) >> return 0; >> >> ret = phy_pm_runtime_get_sync(phy); >> @@ -277,7 +277,7 @@ int phy_power_on(struct phy *phy) >> { >> int ret = 0; >> >> - if (!phy) >> + if (!phy || !phy->ops->power_on) >> goto out; >> >> if (phy->pwr) { >> @@ -319,7 +319,7 @@ int phy_power_off(struct phy *phy) >> { >> int ret; >> >> - if (!phy) >> + if (!phy || !phy->ops->power_off) >> return 0; >> >> mutex_lock(&phy->mutex); >> > > >
