Path: csiph.com!aioe.org!bofh.it!news.nic.it!robomod
From: Russell King - ARM Linux <linux@armlinux.org.uk>
Newsgroups: linux.kernel
Subject: Re: [PATCH 1/3] ARM: Generalize fncpy implementation
Date: Fri, 16 Jun 2017 14:00:01 +0200
Message-ID: <tSY4V-7Xg-1@gated-at.bofh.it>
References: <tSMn7-tC-3@gated-at.bofh.it> <tSMn7-tC-5@gated-at.bofh.it>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.5.23 (2014-03-12)
Sender: robomod@news.nic.it
Approved: robomod@news.nic.it
Lines: 21
Organization: linux.* mail to news gateway
Xref: csiph.com linux.kernel:1667663

On Thu, Jun 15, 2017 at 04:21:15PM -0700, Florian Fainelli wrote:
> ARM's fncpy implementation is actually suitable for a large number of
> platforms since the only assumption it makes is just the function's
> alignment (8 bytes) which also happens to fulfil other architectures,
> including but not limited to ARM64.

NAK.  This is really not "generic" because the whole point of this is
that it encapsulates architecture specific knowledge - in the case of
ARM, the fact that bit 1 is used to indicate whether the code is to be
run in Thumb mode or ARM mode.

That clearly does not belong in an asm-generic version of this.

I'm not saying "don't provide an asm-generic" version, I'm saying don't
use the ARM version as an asm-generic implementation, because it is
nothing of the sort.

-- 
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line: currently at 9.6Mbps down 400kbps up
according to speedtest.net.