| Message ID | 1364563104.1345.74.camel@x61.thuisdomein (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Paul, On Fri, Mar 29, 2013 at 01:18:24PM +0000, Paul Bolle wrote: > Signed-off-by: Paul Bolle <pebolle@tiscali.nl> > --- > 0) Untested. > > 1) This typo was first introduced in v3.4: commit > e5ab85800820edd907d3f43f285e1232f84d5a41 ("ARM: 7382/1: mm: truncate > memory banks to fit in 4GB space for classic MMU"). The second instance > was added in v3.9-rc2: commit d61947a164760ac520cb416768afdf38c33d60e7 > (" ARM: 7657/1: head: fix swapper and idmap population with LPAE and > big-endian"). So testing is really needed here. > > 2) If this passes (review and) testing, this probably needs to go to > stable too. > > arch/arm/kernel/head.S | 2 +- > arch/arm/kernel/setup.c | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S > index e0eb9a1..8bac553 100644 > --- a/arch/arm/kernel/head.S > +++ b/arch/arm/kernel/head.S > @@ -267,7 +267,7 @@ __create_page_tables: > addne r6, r6, #1 << SECTION_SHIFT > strne r6, [r3] > > -#if defined(CONFIG_LPAE) && defined(CONFIG_CPU_ENDIAN_BE8) > +#if defined(CONFIG_ARM_LPAE) && defined(CONFIG_CPU_ENDIAN_BE8) > sub r4, r4, #4 @ Fixup page table pointer > @ for 64-bit descriptors > #endif The problem with this code is that there are very few people with platforms on which they can test it. You need an LPAE-capable CPU (Cortex-A7 or Cortex-A15) sitting on a big-endian platform. The only platform that I have for that is KVM, which doesn't have earlyprintk support. Now, without the offending patch, such a configuration didn't even boot, but I can't verify that the mapping actually works. Of course, what you have written was the intended behaviour, so perhaps we should just go with that for the time being. Will
diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S index e0eb9a1..8bac553 100644 --- a/arch/arm/kernel/head.S +++ b/arch/arm/kernel/head.S @@ -267,7 +267,7 @@ __create_page_tables: addne r6, r6, #1 << SECTION_SHIFT strne r6, [r3] -#if defined(CONFIG_LPAE) && defined(CONFIG_CPU_ENDIAN_BE8) +#if defined(CONFIG_ARM_LPAE) && defined(CONFIG_CPU_ENDIAN_BE8) sub r4, r4, #4 @ Fixup page table pointer @ for 64-bit descriptors #endif diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c index 3f6cbb2..62ead98 100644 --- a/arch/arm/kernel/setup.c +++ b/arch/arm/kernel/setup.c @@ -524,7 +524,7 @@ int __init arm_add_memory(phys_addr_t start, phys_addr_t size) size -= start & ~PAGE_MASK; bank->start = PAGE_ALIGN(start); -#ifndef CONFIG_LPAE +#ifndef CONFIG_ARM_LPAE if (bank->start + size < bank->start) { printk(KERN_CRIT "Truncating memory at 0x%08llx to fit in " "32-bit physical address space\n", (long long)start);
Signed-off-by: Paul Bolle <pebolle@tiscali.nl> --- 0) Untested. 1) This typo was first introduced in v3.4: commit e5ab85800820edd907d3f43f285e1232f84d5a41 ("ARM: 7382/1: mm: truncate memory banks to fit in 4GB space for classic MMU"). The second instance was added in v3.9-rc2: commit d61947a164760ac520cb416768afdf38c33d60e7 (" ARM: 7657/1: head: fix swapper and idmap population with LPAE and big-endian"). So testing is really needed here. 2) If this passes (review and) testing, this probably needs to go to stable too. arch/arm/kernel/head.S | 2 +- arch/arm/kernel/setup.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)