On Wed, Oct 09, 2019 at 10:28:36AM -0700, Doug Anderson wrote: > On Wed, Oct 9, 2019 at 2:30 AM Daniel Thompson > <daniel.thomp...@linaro.org> wrote: > > > > @@ -741,7 +732,7 @@ int vkdb_printf(enum kdb_msgsrc src, const char > > > > *fmt, va_list ap) > > > > > > > > /* check for having reached the LINES number of printed lines */ > > > > if (kdb_nextline >= linecount) { > > > > - char buf1[16] = ""; > > > > + char ch; > > > > > > The type of "ch" should be the same as returned by kdb_getchar()? > > > Either "int" if you're keeping it "int" or "unsigned char"? > > > > Probably... although the assumption that kdb strings are char * is burnt > > in a lot of places so there will still be further tidy up needed. > > True. It doesn't matter a whole lot so if you think it's easier to > keep it as char that's OK too.
After looking at it from a number of angles I think we can have this match the return type of kdb_getchar()... but the best way to achieve this is to make kdb_getchar() return a unqualified char. That ends up consistent across the sub-system and shouldn't do any narrowing that wouldn't already have been happening inside kdb_read(). Daniel.