Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-15 Thread Mark Salter
On Wed, 2014-01-15 at 16:32 -0800, Laura Abbott wrote: > On 1/9/2014 7:50 PM, Mark Salter wrote: > > > diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c > > index 987a7f5..038fb75 100644 > > --- a/arch/arm/kernel/setup.c > > +++ b/arch/arm/kernel/setup.c > > @@ -36,6 +36,7 @@ > >

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-15 Thread Russell King - ARM Linux
On Wed, Jan 15, 2014 at 04:32:03PM -0800, Laura Abbott wrote: > Neither x86 nor arm64 get rid of existing mapping in paging_init. Is > there a reason this needs to be here for arm? Are there concerns about > alias mappings? The early mappings we setup in the assembly code are imperfect - they'

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-15 Thread Laura Abbott
On 1/9/2014 7:50 PM, Mark Salter wrote: diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c index 987a7f5..038fb75 100644 --- a/arch/arm/kernel/setup.c +++ b/arch/arm/kernel/setup.c @@ -36,6 +36,7 @@ #include #include #include +#include #include #include #include

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Mark Salter
On Fri, 2014-01-10 at 12:40 -0800, Stephen Boyd wrote: > Just some minor nitpicks > > On 01/09/14 19:50, Mark Salter wrote: > > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig > > index c1f1a7e..78a79a6a 100644 > > --- a/arch/arm/Kconfig > > +++ b/arch/arm/Kconfig > > @@ -1842,6 +1842,17 @@ confi

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Mark Salter
On Fri, 2014-01-10 at 10:11 -0600, Rob Herring wrote: > On Thu, Jan 9, 2014 at 9:50 PM, Mark Salter wrote: > > +config EARLY_IOREMAP > > + depends on MMU > > Is it possible to implement a !MMU version of early_ioremap that > simply returns the phys address rather than have this dependency?

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Stephen Boyd
Just some minor nitpicks On 01/09/14 19:50, Mark Salter wrote: > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig > index c1f1a7e..78a79a6a 100644 > --- a/arch/arm/Kconfig > +++ b/arch/arm/Kconfig > @@ -1842,6 +1842,17 @@ config UACCESS_WITH_MEMCPY > However, if the CPU data cache is using

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Rob Herring
On Fri, Jan 10, 2014 at 10:58 AM, Catalin Marinas wrote: > On Fri, Jan 10, 2014 at 04:11:03PM +, Rob Herring wrote: >> On Thu, Jan 9, 2014 at 9:50 PM, Mark Salter wrote: >> > This patch uses the generic early_ioremap code to implement >> > early_ioremap for ARM. The ARM-specific bits come mos

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Rob Herring
On Fri, Jan 10, 2014 at 10:58 AM, Catalin Marinas wrote: > On Fri, Jan 10, 2014 at 04:11:03PM +, Rob Herring wrote: >> On Thu, Jan 9, 2014 at 9:50 PM, Mark Salter wrote: >> > This patch uses the generic early_ioremap code to implement >> > early_ioremap for ARM. The ARM-specific bits come mos

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Catalin Marinas
On Fri, Jan 10, 2014 at 04:11:03PM +, Rob Herring wrote: > On Thu, Jan 9, 2014 at 9:50 PM, Mark Salter wrote: > > This patch uses the generic early_ioremap code to implement > > early_ioremap for ARM. The ARM-specific bits come mostly from > > an earlier patch from Leif Lindholm > > here: > >

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Rob Herring
On Thu, Jan 9, 2014 at 9:50 PM, Mark Salter wrote: > This patch uses the generic early_ioremap code to implement > early_ioremap for ARM. The ARM-specific bits come mostly from > an earlier patch from Leif Lindholm > here: > > https://lkml.org/lkml/2013/10/3/279 I think this will be useful to

Re: [PATCH v3 4/6] arm: add early_ioremap support

2014-01-10 Thread Russell King - ARM Linux
On Thu, Jan 09, 2014 at 10:50:31PM -0500, Mark Salter wrote: > +#define FIXMAP_PAGE_NORMAL (L_PTE_MT_WRITEBACK | L_PTE_YOUNG | L_PTE_PRESENT) > +#define FIXMAP_PAGE_IO (L_PTE_MT_DEV_NONSHARED | L_PTE_YOUNG | L_PTE_PRESENT) Doesn't this also want L_PTE_XN for at least IO mappings - maybe for the no