On 14/03/2019 10.29, Uwe Kleine-König wrote: > Hello, > > On Wed, Mar 13, 2019 at 09:26:12PM +0100, Rasmus Villemoes wrote: >> --- a/drivers/leds/trigger/ledtrig-netdev.c >> +++ b/drivers/leds/trigger/ledtrig-netdev.c >> @@ -122,7 +122,8 @@ static ssize_t device_name_store(struct device *dev, >> trigger_data->net_dev = NULL; >> } >> >> - strncpy(trigger_data->device_name, buf, size); >> + memcpy(trigger_data->device_name, buf, size); >> + trigger_data->device_name[size] = '\0'; > > This is open-coding > > strlcpy(trigger_data->device_name, buf, size);
No. size here is the number of bytes userspace wrote, which never (well, almost never, they could do something odd) contain a nul byte. Passing that as size to strlcpy would guarantee that we chopped off the last character from the user input. And while I do think the generic sysfs layer guarantees that the PAGE_SIZE buffer is zeroed before reading from userspace, I don't really want to rely on buf being a nul-terminated string (which it must be if using strlcpy - remember that that does a full strlen() to compute its return value). If anything, one could do strlcpy(, buf, size+1), but IMO copying the actual characters the user wrote, then explicitly making that into a nul-terminated string is much more straight-forward. Rasmus
