RE: mainline build failure for arm64 allmodconfig with clang

From: David Laight <David.Laight@ACULAB.COM>
To: 'Linus Torvalds' <torvalds@linux-foundation.org>,
	Nick Desaulniers <ndesaulniers@google.com>
Cc: Joe Perches <joe@perches.com>,
	Nathan Chancellor <nathan@kernel.org>,
	"Sudip Mukherjee (Codethink)" <sudipm.mukherjee@gmail.com>,
	Masahiro Yamada <masahiroy@kernel.org>,
	Michal Marek <michal.lkml@markovi.net>,
	"linux-kbuild@vger.kernel.org" <linux-kbuild@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	clang-built-linux <llvm@lists.linux.dev>,
	Justin Stitt <justinstitt@google.com>
Subject: RE: mainline build failure for arm64 allmodconfig with clang
Date: Mon, 15 Aug 2022 10:37:42 +0000	[thread overview]
Message-ID: <de9ccf7da0aa441b855fdfbd7761115a@AcuMS.aculab.com> (raw)
In-Reply-To: <CAHk-=wgJA=e-CLcvU5LRKu0bMLeAewXtOM6as1hFVeQAVkMPbg@mail.gmail.com>

...
> And often that char isn't actually a 'char', it is actually an 'int',
> either because you have situations like 'getch()', or you have simply
> just the usual C expression rules, ie you have something like
> 
>         isprint(c) ? c : '.'
> 
> where even if 'c' is of type 'char', the end result is 'int'.

Which is, of course, invalid C :-)
The domain of the isxxxx() functions is 'the values of char
cast to unsigned int' and EOF.
So typically [-1 .. 255].
Passing a signed char variable to isprint() is likely to
generate undesired behaviour (like a code dump).

It has to be said that I did a full trawl through the NetBSD
userspace code base to fix all the isxxx() calls and failed
to find a single one that could actually pass EOF.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)