On Tue, Sep 03, 2019 at 02:28:00PM +0300, Andy Shevchenko wrote: > On Tue, Sep 03, 2019 at 11:28:16AM +0200, Petr Mladek wrote: > > On Tue 2019-09-03 10:52:33, Petr Mladek wrote: > > > On Mon 2019-09-02 16:57:30, Sakari Ailus wrote: > > > > Factor out static kobject_string() function that simply calls > > > > device_node_string(), and thus remove references to kobjects (as these are > > > > struct device_node). > > > > > - return kobject_string(buf, end, ptr, spec, fmt); > > > > + return device_node_string(buf, end, ptr, spec, fmt + 1); > > > > > > I know that this come from from kobject_string(). But please, modify > > > it to follow the style used by other %p modifiers. I mean to pass > > > "fmt" as is and then use: > > > > > > if (fmt[1] != 'F') > > > > Ah, I see that it would need more changes in device_node_string(). > > OK, let's leave the patch as is. I am sorry for the noise. > > I came to the same conclusions, though can we consider to drop this patch? It's a cleanup. I'd prefer to keep the patch. Albeit fmt++; in device_node_string() would do the trick of avoiding fmt + 1 in the caller. That said, I'd prefer to keep the original patch as-is. -- Sakari Ailus sakari.ailus@xxxxxxxxxxxxxxx
