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 AB39ADE055 for ; Wed, 8 Apr 2009 03:57:01 +1000 (EST) Message-Id: <606D70D7-6F85-48DF-854F-18DAE01ACA95@kernel.crashing.org> From: Kumar Gala To: Peter Bergner In-Reply-To: <1239125792.7557.7.camel@otta> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Mime-Version: 1.0 (Apple Message framework v930.3) Subject: Re: [PATCH] powerpc: Fix tlbilx opcode Date: Tue, 7 Apr 2009 12:54:53 -0500 References: <1239050219-4477-1-git-send-email-galak@kernel.crashing.org> <1239125792.7557.7.camel@otta> Cc: linuxppc-dev@ozlabs.org, Edmar Wienskoski-RA8797 List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Apr 7, 2009, at 12:36 PM, Peter Bergner wrote: > On Mon, 2009-04-06 at 15:36 -0500, Kumar Gala wrote: >> The tlbilx opcode was not matching the Power ISA 2.06 arch spec. >> The old opcode was an early suggested opcode that changed during the >> 2.06 architecture process. > [snip] >> #define PPC_INST_STSWI 0x7c0005aa >> #define PPC_INST_STSWX 0x7c00052a >> -#define PPC_INST_TLBILX 0x7c000626 >> +#define PPC_INST_TLBILX 0x7c000024 >> #define PPC_INST_WAIT 0x7c00007c > > How do you want to handle the current binutils code that implements > tlbilx > and its extended mnemonics? Should they be changed to use secondary > opcode > 18 instead of 787 as per the ISA 2.06 doc? YES ! :) > {"tlbilx", X(31,787), X_MASK, E500MC, > PPCNONE, {T, RA0, RB}}, > {"tlbilxlpid", XTO(31,787,0), XTO_MASK, E500MC, > PPCNONE, {0}}, > {"tlbilxpid", XTO(31,787,1), XTO_MASK, E500MC, > PPCNONE, {0}}, > {"tlbilxva", XTO(31,787,3), XTO_MASK, E500MC, > PPCNONE, {RA0, RB}}, > > If you want to change the tlbilx entry to use 18 rather than 787, > let me > know whether you want me to make the change or whether you (Edmar?) > will > handle it. > > I'll note the extended mnemonics should be moved before tlbilx so > that the > disassembler will use them (if appropriate) instead of the base > instruction. I'll check w/Edmar. - k