linux-kernel - Re: [patches] [PATCH] [8/35] x86_64: a memcpy that tries to reducecache pressure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <4635B62D.76E4.0078.0@novell.com>
Date:	Mon, 30 Apr 2007 08:26:05 +0100
From:	"Jan Beulich" <jbeulich@...ell.com>
To:	"Bryan O'Sullivan" <bos@...hscale.com>, "Andi Kleen" <ak@...e.de>
Cc:	"Roland Dreier" <rolandd@...co.com>,
	<linux-kernel@...r.kernel.org>, <patches@...-64.org>
Subject: Re: [patches] [PATCH] [8/35] x86_64: a memcpy that tries to
	reducecache pressure

>+	testb	$8, %dxl	/* rdx is 3,5,6,7,9..15 */

Could you use the more conventional name %dl (or whatever is being meant)
here?

>+.L42:
>+	prefetchnta	128(%rsi)

I think I commented similarly on a previous version of the patch: This will 
- result in still bringing the first 128 bytes into the cache
- prevent caching of memory beyond the buffer being dealt with

>+	sfence

Similarly, I continue to believe this is wrong here (again I commented before
that this is likely unnecessary, or if at all, it should be an lfence).

>+	orl	%edx, %edx

Operations of this type generally should use 'test', not 'or' or 'and'.

Jan


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/