Re: [PATCH v2 7/8] powerpc/fsl_booke: make sure PAGE_OFFSET map to memstart_addr for relocatable kernel

On Mon, Aug 05, 2013 at 07:14:00PM -0500, Scott Wood wrote:

On Thu, 2013-07-04 at 20:54 +0800, Kevin Hao wrote:

quoted

@@ -1222,6 +1266,9 @@ _GLOBAL(switch_to_as1)
 /*
  * Restore to the address space 0 and also invalidate the tlb entry created
  * by switch_to_as1.
+ * r3 - the tlb entry which should be invalidated
+ * r4 - __pa(PAGE_OFFSET in AS0) - pa(PAGE_OFFSET in AS1)
+ * r5 - device tree virtual address
 */
 _GLOBAL(restore_to_as0)

[snip]

quoted

diff --git a/arch/powerpc/mm/mmu_decl.h b/arch/powerpc/mm/mmu_decl.h
index 3a65644..8280dbb 100644
--- a/arch/powerpc/mm/mmu_decl.h
+++ b/arch/powerpc/mm/mmu_decl.h

@@ -149,7 +149,7 @@ extern void MMU_init_hw(void);
 extern unsigned long mmu_mapin_ram(unsigned long top);
 extern void adjust_total_lowmem(void);
 extern int switch_to_as1(void);
-extern void restore_to_as0(int);
+extern void restore_to_as0(int, int, void *);

"int" seems wrong for the second argument. Shouldn't it be phys_addr_t?

This is the offset between __pa(PAGE_OFFSET in AS0) and __pa(PAGE_OFFSET in AS1).
And the max offset should never exceed 768M. So a int type is safe here and
using this type also avoid the ugly #ifdef CONFIG_PHYS_64BIT.

Also, please use parameter names.

Sure.

Thanks,
Kevin

-Scott