From mboxrd@z Thu Jan 1 00:00:00 1970 From: ben.dooks@codethink.co.uk (Ben Dooks) Date: Tue, 15 May 2012 16:01:22 +0100 Subject: [PATCH 6/8] arm: mach-armada: add support for Armada XP board with device tree In-Reply-To: <4FB26E03.2030401@gmail.com> References: <1337072084-21967-1-git-send-email-thomas.petazzoni@free-electrons.com> <4FB25FE9.3060209@codethink.co.uk> <20120515161655.504f1403@skate> <201205151425.30706.arnd@arndb.de> <4FB26E03.2030401@gmail.com> Message-ID: <4FB26FC2.9050805@codethink.co.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 15/05/12 15:53, Rob Herring wrote: > On 05/15/2012 09:25 AM, Arnd Bergmann wrote: >> On Tuesday 15 May 2012, Thomas Petazzoni wrote: >>> Le Tue, 15 May 2012 14:53:45 +0100, >>> Ben Dooks a ?crit : >>> >>>> Since the two board support files are identical, except for the names >>>> they print, I'd say this is the job of one file. It can always be >>>> split later. >>> >>> The initial motivation for keeping two files here is that the two SoC >>> have a different number of PCIe memory areas, and those areas are >>> typically mapped in ->map_io(). However: >>> >>> * Maybe those mappings can be done using a normal ioremap() rather >>> than in ->map_io(), according to DT informations (but most other ARM >>> SoC support at the moment seem to do PCI mappings using static >>> mappings in ->map_io) >> >> I'm pretty sure we can use ioremap for new PCI implementations now. >> Also, you can scan the device tree in map_io() if necesary, to see which >> PCIe ports are enabled. > > For the i/o windows, they should be static because we want a fixed > virtual address across platforms. 0xfef00000 is the planned address. > Memory windows should be ioremap. It might be worth looking at changing ioremap() to work nicely with large page entries if it hasn't already been done. -- Ben Dooks http://www.codethink.co.uk/ Senior Engineer Codethink - Providing Genius