Re: [Ct173-kernel-internal] [PATCH 0/4] ARM: tegra: Support for big-endian (version 3)

[Dmitry Osipenko <digetx-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

20.01.2015 14:45, Ben Dooks пишет:
> On 20/01/15 11:36, Dmitry Osipenko wrote:
>> 20.01.2015 13:31, Ben Dooks пишет:
>>> On 09/01/15 15:27, Bob Mottram wrote:
>>>> During testing of big-endian on a Jetson TK1 with baserock, the
>>>> following
>>>> issues were found and fixed. Note that to get the system running with
>>>> big-endian you will need to disable ARCH_SUSPEND_POSSIBLE within
>>>> arch/arm/Kconfig.
>>>>
>>>> Bob Mottram (4):
>>>>     ARM: tegra: Add a separate big-endian configuration
>>>>     ARM: gpio: tegra: Maintain endianness
>>>>     ARM: tegra: enable big-endian support
>>>>     ARM: tegra: fix big-endian issue with irq code
>>>
>>> Is it possible to get the tegra core patches applied? These seem to work
>>> enough to get a system running and once the PM fixes are in there should
>>> be no problems running big-endian tegra?
>>>
>> I don't think it's enough. Upstream kernel doesn't accept half working
>> solutions. It's not only PM that needs fixing, but i2c, timer, reset
>> handler and possibly more. You are enabling big-endian across all tegra
>> generations, but tegra2 won't boot at all for example. Patches must be
>> tested on all hw before applying. And nit: big-endian should be enabled
>> after all dependencies resolved.
>
> Ok, we hadn't found any issue with the reset. I thought we had the
> timer fix in this series. Most of the other drivers didn't use
> any __raw code so we're reasonably confident they will work.
>
> I didn't see any issues with the i2c from visual inspection, what's
> wrong with it in big endian?
>
I've sent patch.

-- 
Dmitry