Re: K7/Athlon Optimization Causes Build Fail on Kernel 3.3 i686

From: Borislav Petkov <bp@alien8.de>
To: "H. Peter Anvin" <hpa@zytor.com>
Cc: Matt Fleming <matt.fleming@intel.com>,
	Trevor Turner <turn3r.tr3v0r@gmail.com>,
	Randy Dunlap <rdunlap@xenotime.net>,
	linux-kernel@vger.kernel.org, Al Viro <viro@zeniv.linux.org.uk>
Subject: Re: K7/Athlon Optimization Causes Build Fail on Kernel 3.3 i686
Date: Thu, 29 Mar 2012 22:46:23 +0200	[thread overview]
Message-ID: <20120329204622.GA8655@liondog.tnic> (raw)
In-Reply-To: <4F74A9D8.1040808@zytor.com>

On Thu, Mar 29, 2012 at 11:28:40AM -0700, H. Peter Anvin wrote:
> I like this one for urgent as a quick fix, but it doesn't solve the real
> problem which is that boot/compressed shouldn't pick up these macros at all.

But it does through:

<arch/x86/include/asm/string_32.h>:

...
static inline void *__memcpy3d(void *to, const void *from, size_t len)
{
	if (len < 512)
		return __memcpy(to, from, len);
	return _mmx_memcpy(to, from, len);
}

#define memcpy(t, f, n)				\
	(__builtin_constant_p((n))		\
	 ? __constant_memcpy3d((t), (f), (n))	\
	 : __memcpy3d((t), (f), (n)))

> It should be possible to build with MK7 (which is an optimization
> option) and EFI (a functionality option) since they don't inherently
> conflict.  It just doesn't necessarily have to happen *right now*.

I know, technically it is doable, but do you see a K7 box with an EFI
firmware? I, at least, don't think so... :-)

Now, if you think that those optimizations could be used on other boxes
which could sport EFI, then that's a whole another question...

-- 
Regards/Gruss,
    Boris.