From mboxrd@z Thu Jan  1 00:00:00 1970
From: segher@kernel.crashing.org (Segher Boessenkool)
Date: Thu, 16 Nov 2017 11:01:44 -0600
Subject: [PATCH v2 11/18] arm64: make mrs_s and msr_s macros work with LTO
In-Reply-To: <20171116164608.GG94341@samitolvanen.mtv.corp.google.com>
References: <20171115213428.22559-1-samitolvanen@google.com>
 <20171115213428.22559-12-samitolvanen@google.com>
 <20171116115433.GG9361@arm.com>
 <20171116130742.wijnukjetoupi4hc@yury-thinkpad>
 <20171116135650.GQ10515@gate.crashing.org>
 <20171116164608.GG94341@samitolvanen.mtv.corp.google.com>
Message-ID: <20171116170144.GR10515@gate.crashing.org>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Thu, Nov 16, 2017 at 08:46:08AM -0800, Sami Tolvanen wrote:
> On Thu, Nov 16, 2017 at 07:56:50AM -0600, Segher Boessenkool wrote:
> > The compiler is fine, the assembler is fine (and the linker has
> > nothing to do with it).  Your code is not fine.
> 
> Would you care to elaborate? The current code assumes that macros are
> visible in other inline assembly blocks, and LLVM developers seem to
> feel this isn't correct behavior. This patch fixes the current code so
> it works with both assemblers.

If you say e.g.

void f(void)
{
	asm(".macro something\n\t.endm");
}

there is nothing that prevents the compiler from emitting this more
than once.  Expecting things to be emitted in whatever order is a bad
idea, too.

The thing with .purgem can work.  Inelegant, sure, but it can work :-)
Just make sure you do the macro define, the code that uses it, and the
undefine, all in the same inline asm statement.


Segher