From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1M8v6g-0005A1-Iy for qemu-devel@nongnu.org; Tue, 26 May 2009 07:46:10 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1M8v6f-00059E-IX for qemu-devel@nongnu.org; Tue, 26 May 2009 07:46:09 -0400 Received: from [199.232.76.173] (port=45809 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1M8v6e-00058m-GY for qemu-devel@nongnu.org; Tue, 26 May 2009 07:46:08 -0400 Received: from fg-out-1718.google.com ([72.14.220.159]:62937) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1M8v6d-0003wS-Fq for qemu-devel@nongnu.org; Tue, 26 May 2009 07:46:08 -0400 Received: by fg-out-1718.google.com with SMTP id l27so727569fgb.8 for ; Tue, 26 May 2009 04:46:06 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20090526112915.GA3224@shareable.org> References: <20090423171503.GC4629@derik> <200904301708.01385.paul@codesourcery.com> <20090523165033.GC8037@derik> <200905241931.19368.paul@codesourcery.com> <20090526093944.GB4784@derik> <761ea48b0905260242v4290a835rd8b958c7544e667c@mail.gmail.com> <20090526095652.GA32352@shareable.org> <761ea48b0905260308m5ba590cbl31781a20bd3d001f@mail.gmail.com> <20090526112915.GA3224@shareable.org> Date: Tue, 26 May 2009 13:46:06 +0200 Message-ID: <761ea48b0905260446n20fbc791ofb17404f08880e7b@mail.gmail.com> Subject: Re: [Qemu-devel] [PATCH 1/16] ARM Add ARM 920T identifiers From: Laurent Desnogues Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jamie Lokier Cc: Vincent Sanders , qemu-devel@nongnu.org On Tue, May 26, 2009 at 1:29 PM, Jamie Lokier wrote: > Laurent Desnogues wrote: [...] >> It shouldn't. =A0I was thinking about how most people are misusing their >> toolchain and think that if QEMU runs correctly some code, it should >> run on the target. =A0In that particular case, that will fail :-) =A0For= instance >> LDM with PC used as procedure return does not behave the same >> in v4T and v5 IIRC. > > I agree. =A0Especially for testing toolchains and architecture-specific > code, it would be quite nice :-) > > I gather the x86 target is similarly lax with checking some things > that a real x86 checks. The existing ARM target is lax too, but quite close. The main problem is that the ARM architecture defines what is IMPLEMENTATION DEFINED and also contains many UNPREDICTABLE behaviors, and there are programs that rely on both, which basically means that if one wanted an accurate simulation of all that, the ARM target would be even more unreadable than it is now :) Anyway architectural compliance would be enough and I think it's highly desirable. Laurent