From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1LCHOP-0002iV-LJ for qemu-devel@nongnu.org; Mon, 15 Dec 2008 12:38:05 -0500 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1LCHON-0002fv-Ey for qemu-devel@nongnu.org; Mon, 15 Dec 2008 12:38:04 -0500 Received: from [199.232.76.173] (port=58581 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1LCHON-0002fg-9C for qemu-devel@nongnu.org; Mon, 15 Dec 2008 12:38:03 -0500 Received: from mx20.gnu.org ([199.232.41.8]:52728) by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1LCHOM-0001SW-UH for qemu-devel@nongnu.org; Mon, 15 Dec 2008 12:38:03 -0500 Received: from mail.codesourcery.com ([65.74.133.4]) by mx20.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1LCHOL-0005Nw-RA for qemu-devel@nongnu.org; Mon, 15 Dec 2008 12:38:02 -0500 Date: Mon, 15 Dec 2008 09:38:00 -0800 From: Nathan Froyd Subject: Re: [Qemu-devel] [PATCH] target-ppc: add Altivec instructions, patch-bomb version Message-ID: <20081215173800.GA25983@codesourcery.com> References: <1229307315-16807-1-git-send-email-froydnj@codesourcery.com> <20081215064059.GB22069@hall.aurel32.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20081215064059.GB22069@hall.aurel32.net> Reply-To: qemu-devel@nongnu.org List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org On Mon, Dec 15, 2008 at 07:40:59AM +0100, Aurelien Jarno wrote: > And a question for the whole series: how have you tested those > instructions? The instructions from the original (dyngen) patch were tested using a modified version of: * test-powerpc.cpp - PowerPC regression testing * * Kheperix (C) 2003-2005 Gwenole Beauchesne that didn't require poking around in the guts of QEMU. I turned it into a standalone application that could be run in Linux. I then ran it on a 74xx machine and under usermode emulation with QEMU and compared the results. I assumed that porting it over would show no difference in functionality, but I see from running it this morning that m{f,t}vscr, vmul{e,o}*, and vs{l,r} are broken. Everything else works OK. I will poke at the failing instructions and submit updated patches for them. -Nathan