On Thu, 20 Nov 2014, Oliver Neukum wrote: > On Thu, 2014-11-20 at 15:36 +0100, Bjørn Mork wrote: > > Oliver Neukum <oneukum@xxxxxxx> writes: > > > On Thu, 2014-11-20 at 15:09 +0100, Bjørn Mork wrote: > > >> Oliver Neukum <oneukum@xxxxxxx> writes: > > >> > On Thu, 2014-11-20 at 14:27 +0100, Johan Hovold wrote: > > >> > > > >> >> It looks like it's never freed in the alloc_fail8 error path of probe, > > >> >> though. > > >> > > > >> > Indeed. A case of the coder being blind to his own code. > > >> > Patch is forthcoming. > > >> > > >> Why does the driver have to allocate anything at all for this purpose? > > >> Couldn't it just save a pointer to the country functional descriptor, > > >> and let the "show" functions decode that when necessary? > > > > > > The show method would have to deal with a race against disconnect(). > > > > How is that different with the additional allocation? > > Good question. That was the original motivation. Sysfs attributes are > just a pain. Let me think. The sysfs filesystem itself insures that sysfs attributes never race with device removal. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
