public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: David Brown <davidb@codeaurora.org>
To: Will Deacon <will.deacon@arm.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>,
	linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH 1/4] msm: scm: Mark inline asm as volatile
Date: Sun, 27 Feb 2011 09:38:43 -0800	[thread overview]
Message-ID: <8yalj112x98.fsf@huya.qualcomm.com> (raw)
In-Reply-To: <1298805034.4626.4.camel@jazzbox> (Will Deacon's message of "Sun, 27 Feb 2011 11:10:33 +0000")

On Sun, Feb 27 2011, Will Deacon wrote:

> On Sat, 2011-02-26 at 18:12 +0000, David Brown wrote:
>> On Fri, Feb 25 2011, Will Deacon wrote:
>
>> > These asm blocks all have sensible looking output constraints. Why
>> > do they need to be marked volatile?
>> 
>> Without the volatile, the compiler is free to assume the only side
>> effects of the asm are to modify the output registers.  The volatile is
>> needed to indicate to the compiler that the asm has other side effects.
>
> As far as I know, volatile asm does two things:
>
> (1) It stops the compiler from reordering the asm block with respect to
> other volatile statements.
>
> (2) It prevents the compiler from optimising the block away when
> dataflow analysis indicates it's not required.
>
> If side-effects need to be indicated, won't a memory clobber do the
> trick?

Per the gcc manual:

    If your assembler instructions access memory in an unpredictable
   fashion, add `memory' to the list of clobbered registers.  This will
   cause GCC to not keep memory values cached in registers across the
   assembler instruction and not optimize stores or loads to that
   memory.  You will also want to add the `volatile' keyword if the
   memory affected is not listed in the inputs or outputs of the `asm',
   as the `memory' clobber does not count as a side-effect of the `asm'.
   If you know how large the accessed memory is, you can add it as input
   or output but if this is not known, you should add `memory'.  As an
   example, if you access ten bytes of a string, you can use a memory
   input like:

The smc instruction is similar to a syscall.  When in the secure world,
the processor is making state changes.  It's not quite correct to
declare this as memory, because the memory used when secure isn't even
accessible to us.  As far as I can tell, the volatile is the only way to
tell the compiler this.

David

-- 
Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.

  reply	other threads:[~2011-02-27 17:38 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-24 18:44 [PATCH 0/4] SCM fixes and updates Stephen Boyd
2011-02-24 18:44 ` [PATCH 1/4] msm: scm: Mark inline asm as volatile Stephen Boyd
2011-02-25 11:56   ` Will Deacon
2011-02-25 19:05     ` Stephen Boyd
2011-02-26 18:12     ` David Brown
2011-02-26 19:43       ` Nicolas Pitre
2011-02-27 17:41         ` David Brown
2011-02-28  2:21           ` Nicolas Pitre
2011-02-27 11:10       ` Will Deacon
2011-02-27 17:38         ` David Brown [this message]
2011-03-01 10:30           ` Will Deacon
2011-02-24 18:44 ` [PATCH 2/4] msm: scm: Fix improper register assignment Stephen Boyd
2011-02-25 13:23   ` Will Deacon
2011-02-25 19:22     ` Stephen Boyd
2011-02-26  5:09     ` Saravana Kannan
2011-02-26  8:47       ` Russell King - ARM Linux
2011-02-26 17:58         ` David Brown
2011-02-26 20:04           ` Nicolas Pitre
2011-03-01 10:37             ` Will Deacon
2011-03-01 21:29               ` Saravana Kannan
2011-03-02  0:02                 ` Nicolas Pitre
2011-03-01 13:54             ` Will Deacon
2011-02-24 18:44 ` [PATCH 3/4] msm: scm: Check for interruption immediately Stephen Boyd
2011-02-24 18:44 ` [PATCH 4/4] msm: scm: Get cacheline size from CTR Stephen Boyd
2011-02-24 19:01   ` Thomas Gleixner
2011-02-24 19:44     ` Stephen Boyd
2011-02-24 19:56       ` Thomas Gleixner
2011-03-01  4:21         ` Stephen Boyd
2011-02-24 19:32   ` Sergei Shtylyov
2011-02-24 19:50     ` Stephen Boyd
2011-02-24 19:55     ` Russell King - ARM Linux
2011-03-09 19:29 ` [PATCH 0/4] SCM fixes and updates Stephen Boyd
2011-03-10 20:06   ` David Brown

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=8yalj112x98.fsf@huya.qualcomm.com \
    --to=davidb@codeaurora.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sboyd@codeaurora.org \
    --cc=will.deacon@arm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox