From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 55A732C011E for ; Mon, 9 Sep 2013 20:09:30 +1000 (EST) Message-ID: <1378721359.11525.3.camel@pasglop> Subject: Re: [PATCH] powerpc: OE=1 Form Instructions Not Decoded Correctly From: Benjamin Herrenschmidt To: Paul Mackerras Date: Mon, 09 Sep 2013 20:09:19 +1000 In-Reply-To: <20130908223559.GA495@iris.ozlabs.ibm.com> References: <20130908223559.GA495@iris.ozlabs.ibm.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Cc: Tom Musta , linuxppc-dev@lists.ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Mon, 2013-09-09 at 08:35 +1000, Paul Mackerras wrote: > On Fri, Sep 06, 2013 at 10:13:00AM -0500, Tom Musta wrote: > > To: linuxppc-dev@lists.ozlabs.org > > Subject: [PATCH] powerpc: OE=1 Form Instructions Not Decoded Correctly > > From: Tom Musta > > > > PowerISA uses instruction bit 21 to indicate that the overflow (OV) bit > > of the XER is to be set, as well as its corresponding sticky bit (SO). > > This patch addresses two defects in the implementation of the PowerISA > > single step code for this category of instructions: (a) the OE=1 case > > is not correctly accounted for in the case statement for the extended > > opcode handling. (b) the implementation is not setting XER[OV] and > > XER[SO]. > > Are you seeing any actual problems arising from the OE=1 instructions > not being emulated? This code was designed primarily for emulating > instructions in the kernel, which is written in C, and the C compiler > doesn't emit OE=1 instructions -- or at least it didn't in the past. > So, does the impetus for this change come because the C compiler is > now emitting these instructions, or because this code is being used on > non-kernel instructions, or just for completeness? Your patch > description needs to include answers to these kinds of questions. Isn't that code occasionally used with uprobes too nowadays ? Cheers, Ben. > Also, you need to indent your code correctly according to > Documentation/CodingStyle. > > Paul. > _______________________________________________ > Linuxppc-dev mailing list > Linuxppc-dev@lists.ozlabs.org > https://lists.ozlabs.org/listinfo/linuxppc-dev