RE: [PATCH] powerpc: Avoid clang warnings around setjmp and longjmp

From: David Laight <David.Laight@ACULAB.COM>
To: 'Nathan Chancellor' <natechancellor@gmail.com>,
	Nick Desaulniers <ndesaulniers@google.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
	"# 3.4.x" <stable@vger.kernel.org>,
	clang-built-linux <clang-built-linux@googlegroups.com>,
	Paul Mackerras <paulus@samba.org>,
	linuxppc-dev <linuxppc-dev@lists.ozlabs.org>
Subject: RE: [PATCH] powerpc: Avoid clang warnings around setjmp and longjmp
Date: Thu, 29 Aug 2019 09:59:48 +0000	[thread overview]
Message-ID: <6801a83ed6d54d95b87a41c57ef6e6b0@AcuMS.aculab.com> (raw)
In-Reply-To: <20190828184529.GC127646@archlinux-threadripper>

From: Nathan Chancellor
> Sent: 28 August 2019 19:45
...
> However, I think that -fno-builtin-* would be appropriate here because
> we are providing our own setjmp implementation, meaning clang should not
> be trying to do anything with the builtin implementation like building a
> declaration for it.

Isn't implementing setjmp impossible unless you tell the compiler that
you function is 'setjmp-like' ?

For instance I think it all goes horribly wrong if the generated code
looks like:
	push local_variable
	// setup arguments to setjmp
	call setjmp
	pop local_variable
	// check return value of setjmp

With a naive compiler and simple ABI setjmp just has to save the
return address, stack pointer and caller saved registers.
With modern compilers and ABI I doubt you can implement setjmp
without some help from the compiler.

It is probably best to avoid setjmp/longjmp completely.

	David

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