From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.pbcl.net ([88.198.119.4] helo=hetzner.pbcl.net) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1QmhWM-0002xo-SH for openembedded-core@lists.openembedded.org; Fri, 29 Jul 2011 09:30:10 +0200 Received: from blundell.swaffham-prior.co.uk ([91.216.112.25] helo=[192.168.114.3]) by hetzner.pbcl.net with esmtpsa (TLS1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.72) (envelope-from ) id 1QmhSG-0003gn-Kw for openembedded-core@lists.openembedded.org; Fri, 29 Jul 2011 09:25:56 +0200 From: Phil Blundell To: Patches and discussions about the oe-core layer In-Reply-To: <4E324C2D.2070201@gmail.com> References: <346abefc87d21d0cc111ef87a6e48f40c5b6cb0b.1311683981.git.richard.purdie@linuxfoundation.org> <4E324C2D.2070201@gmail.com> Date: Fri, 29 Jul 2011 08:25:50 +0100 Message-ID: <1311924350.3141.8.camel@lenovo.internal.reciva.com> Mime-Version: 1.0 X-Mailer: Evolution 2.32.3 Subject: Re: [PATCH 1/3] Add ARM tune file overhaul based largely on work from Mark Hatle X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: Patches and discussions about the oe-core layer List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 29 Jul 2011 07:30:10 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Thu, 2011-07-28 at 22:59 -0700, Khem Raj wrote: > here I guess we have to say AVAILTUNES += "arm920t arm920" > where arm920 is arm mode and 920t is thumb mode. but anyway I would > prefer that thumb is optionally added provided user asked for it > through DISTRO/MACHINE features then we should first make sure > that selected machine has thumb feature and if yes for both then we > should enable -mthumb compiler option which now it enable when thumb > appears in TUNE_FEATURES I don't think we want to go down the road of inventing CPU names for ARM-state variants of Thumb-capable cores. (In other words, we shouldn't start putting things like "arm920" in AVAILTUNES when there isn't actually a core named arm920.) Fundamentally I think you're right that the choice between ARM-state and Thumb-state (for t1 at least) ought to be under the control of the DISTRO and not selected purely on the basis of the hardware capabilities. For t2 it is a bit less of a clear-cut issue but, if there are genuinely people making ARMv7 cores with the Thumb decoder removed, I guess we need to have that same switch there as well. p.