From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tom Rini Date: Tue, 3 Feb 2015 16:49:41 -0500 Subject: [U-Boot] [RFC PATCH 04/21] ARM: tegra: collect SoC sources into mach-tegra In-Reply-To: <20150128111146.EDD3.AA925319@jp.panasonic.com> References: <1422166283-20822-5-git-send-email-yamada.m@jp.panasonic.com> <54C67C29.3040401@wwwdotorg.org> <20150128111146.EDD3.AA925319@jp.panasonic.com> Message-ID: <20150203214941.GS10826@bill-the-cat> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On Wed, Jan 28, 2015 at 11:11:46AM +0900, Masahiro Yamada wrote: > Hi Stephen, > > > On Mon, 26 Jan 2015 10:40:57 -0700 > Stephen Warren wrote: > > > On 01/24/2015 11:11 PM, Masahiro Yamada wrote: > > > This commit moves files as follows: > > > > > > arch/arm/cpu/arm720t/tegra20/* -> arch/arm/mach-tegra/tegra20/* > > > arch/arm/cpu/arm720t/tegra30/* -> arch/arm/mach-tegra/tegra30/* > > ... > > > > Bikeshed: I know that this matches what the Linux kernel has for > > 32-bit ARM, but I've always disliked using the word "machine" to > > describe an SoC. Would just "tegra" or "soc-tegra" be better than > > "mach-tegra"? Feel free to entirely ignore this though; I don't feel > > too strongly. > > In hindsight, arch/arm/soc-* would have been clearer than arch/arm/mach-*. > However, we are already familiar with this directory name convention, > so "machine" looks OK to me. > > > Oh and one more thing: We're starting to work on 64-bit Tegra > > support in the Linux kernel. A fair amount of the code moved by this > > patch is likely to apply on both 32-bit and 64-bit Tegra. Will > > arch/arm support 64-bit within U-Boot, or will there be a separate > > directory for 32- and 64-bit ARM? If so, should this code all be > > moved to something more like soc/tegra/... or drivers/tegra/... or > > drivers/soc/tegra/... or ... so it can be shared between the > > architectures? > > We had a hot discussion when aarch64 support was introduced to U-Boot. > > Finally, the community chose arch/arm/cpu/armv8/ rather than arch/arm64/, > i.e. single-arch-directory. And I suspect that at some point we'll also end up moving things from arch/ into drivers/soc/ to mirror the kernel as well. But I still think arch/arm for both is the right direction :) -- Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: