From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from na01-by2-obe.outbound.protection.outlook.com (mail-by2lp0242.outbound.protection.outlook.com [207.46.163.242]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 8E9602C0089 for ; Wed, 18 Dec 2013 09:38:36 +1100 (EST) Message-ID: <1387319904.10013.482.camel@snotra.buserror.net> Subject: Re: [PATCH v2] powerpc 8xx: Loading kernels over 8Mbytes without CONFIG_PIN_TLB From: Scott Wood To: leroy christophe Date: Tue, 17 Dec 2013 16:38:24 -0600 In-Reply-To: <52AFE70D.70407@c-s.fr> References: <20131210112945.E4E311A2BF3@localhost.localdomain> <1386714253.10013.123.camel@snotra.buserror.net> <52A79E31.9070304@c-s.fr> <1386717537.10013.150.camel@snotra.buserror.net> <52A7A58F.8090705@c-s.fr> <1387234621.10013.408.camel@snotra.buserror.net> <52AFE70D.70407@c-s.fr> Content-Type: text/plain; charset="UTF-8" MIME-Version: 1.0 Cc: linuxppc-dev@lists.ozlabs.org, Paul Mackerras , linux-kernel@vger.kernel.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 2013-12-17 at 06:54 +0100, leroy christophe wrote: > Le 16/12/2013 23:57, Scott Wood a écrit : > > On Wed, 2013-12-11 at 00:36 +0100, leroy christophe wrote: > >> Le 11/12/2013 00:18, Scott Wood a écrit : > >>> There wasn't previously an ifdef specifically around the setting of > >>> SPRN_MD_CTR. That's new. There was an ifdef around the entire block, > >>> which has gone away because you are now trying to map more than 8M > >>> regardless of CONFIG_PIN_TLB, but that has nothing to do with whether > >>> there should be an ifdef around SPRN_MD_CTR. > >>> > >>> > >> Euh, ok, but then we have to fix it in the whole function, not only in > >> this block. Do you think it is worth doing it ? > > Fix what in the whole function? I was asking what harm there would be > > if you just remove all the CONFIG_PIN_TLB ifdefs except around the > > actual RSV4I setting -- do we really care what value goes in MD_CTR for > > the non-pinned case, as long as it's different for each entry? > > > > > MD_CTR is decremented after each entry added. > However, the function populates entry 28, then 29, then 30, then 31. At > the end MD_CTR has then value 30, ready to overide entry 30 then 29 then > 28 then 27 ..... > > So I will remove all the CONFIG_PIN_TLB, but I'll also have to fix the > value set in MD_CTR to start from 31, won't I ? OK, so the answer is that we rely on autodecrement avoiding entries over 28 in the CONFIG_PIN_TLB case. Leave the ifdefs, then. > Do you have any comment/recommendation on my tentative v3 patch where I > have tried to implement based on the use of r7 as you recommended ? I haven't reviewed it yet. -Scott