On Tue, 25 Oct 2022 01:59:50 +0200, Jason A. Donenfeld wrote: > > Hi Al, > > On Mon, Oct 24, 2022 at 11:11 PM Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote: > > > > On Mon, Oct 24, 2022 at 06:29:29PM +0200, Jason A. Donenfeld wrote: > > > With char becoming unsigned by default, and with `char` alone being > > > ambiguous and based on architecture, signed chars need to be marked > > > explicitly as such. This fixes warnings like: > > > > It might make sparse to STFU, but it does *not* resolve the underlying > > issue: > > > > vortex_adb_checkinout() returns a number in range of 0..31 on success > > and -ENOMEM on failure. Quite a few callers don't bother to check... > > Yea, I saw that. I assume that the places that don't check don't > *need* to check. But maybe this driver is junk and other bugs lurk. > I'm not sure. Either way, I think this change is certainly an > improvement on the status quo. I don't intend to develop further on > it, but feel free to send patches atop once this lands. Yes, the driver is likely broken in some other way, too. Such a corner case doesn't hit usually, and it's unlikely worth to spend too much time on it. In anyway, I applied the patch now to for-next branch. Thanks! Takashi
