Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 10 Jun 2020 15:20:03 +0200 (CEST)
From:   Miroslav Benes <mbenes@suse.cz>
To:     Julien Thierry <jthierry@redhat.com>
cc:     linux-kernel@vger.kernel.org, jpoimboe@redhat.com,
        peterz@infradead.org, raphael.gault@arm.com
Subject: Re: [RFC PATCH 7/7] objtool: Make unwind_hints available for all
 architectures
In-Reply-To: <20200608152754.2483-8-jthierry@redhat.com>
Message-ID: <alpine.LSU.2.21.2006101516310.26666@pobox.suse.cz>
References: <20200608152754.2483-1-jthierry@redhat.com> <20200608152754.2483-8-jthierry@redhat.com>
User-Agent: Alpine 2.21 (LSU 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Julien,

On Mon, 8 Jun 2020, Julien Thierry wrote:

> Unwind hints are useful to give some information about the call frame
> or stack states in non-standard code.
> 
> Despite unwind hints being used in arch-independent code, the
> unwind_hint structure type itself is define in x86 kernel headers.
> 
> This is because what an unwind hint will describe is very architecture
> specific, both regarding the state and the affected registers.
> 
> To get to share this concept, expose the unwind_hint structure across
> architecutres. However, the hint types remain defined by the
> architecture code. Objtool then needs it's arch specific code to
> "decode" the unwind hint into a cfi_state.

I think it would be nice to split the patch. Something like.

1. current include/linux/frame.h mixes assembly and non-assembly 
definitions, so introduce ASSEMBLY ifdef first seems like a good idea to 
me.

2. move the relevant definitions to frame.h and add the file to 
sync-check

3. the rest of the patch

Would it make sense?

Otherwise, it looks good to me.

Miroslav