powerpc/mm/hash64: Map all the kernel regions in the same 0xc range

This patch maps vmalloc, IO and vmemap regions in the 0xc address range
instead of the current 0xd and 0xf range. This brings the mapping closer
to radix translation mode.

With hash 64K page size each of this region is 512TB whereas with 4K config
we are limited by the max page table range of 64TB and hence there regions
are of 16TB size.

The kernel mapping is now:

 On 4K hash

     kernel_region_map_size = 16TB
     kernel vmalloc start   = 0xc000100000000000
     kernel IO start        = 0xc000200000000000
     kernel vmemmap start   = 0xc000300000000000

64K hash, 64K radix and 4k radix:

     kernel_region_map_size = 512TB
     kernel vmalloc start   = 0xc008000000000000
     kernel IO start        = 0xc00a000000000000
     kernel vmemmap start   = 0xc00c000000000000

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

1 files changed, 61 insertions, 34 deletions

diff --git a/arch/powerpc/include/asm/book3s/64/hash.h b/arch/powerpc/include/asm/book3s/64/hash.h
index 8cbc4106d449..76741a221910 100644
--- a/arch/powerpc/include/asm/book3s/64/hash.h
+++ b/arch/powerpc/include/asm/book3s/64/hash.h
@@ -29,6 +29,10 @@
 #define H_PGTABLE_EADDR_SIZE	(H_PTE_INDEX_SIZE + H_PMD_INDEX_SIZE + \
 				 H_PUD_INDEX_SIZE + H_PGD_INDEX_SIZE + PAGE_SHIFT)
 #define H_PGTABLE_RANGE		(ASM_CONST(1) << H_PGTABLE_EADDR_SIZE)
+/*
+ * Top 2 bits are ignored in page table walk.
+ */
+#define EA_MASK			(~(0xcUL << 60))
 
 /*
  * We store the slot details in the second half of page table.
@@ -42,53 +46,56 @@
 #endif
 
 /*
- * Define the address range of the kernel non-linear virtual area. In contrast
- * to the linear mapping, this is managed using the kernel page tables and then
- * inserted into the hash page table to actually take effect, similarly to user
- * mappings.
+ * +------------------------------+
+ * |                              |
+ * |                              |
+ * |                              |
+ * +------------------------------+  Kernel virtual map end (0xc00e000000000000)
+ * |                              |
+ * |                              |
+ * |      512TB/16TB of vmemmap   |
+ * |                              |
+ * |                              |
+ * +------------------------------+  Kernel vmemmap  start
+ * |                              |
+ * |      512TB/16TB of IO map    |
+ * |                              |
+ * +------------------------------+  Kernel IO map start
+ * |                              |
+ * |      512TB/16TB of vmap      |
+ * |                              |
+ * +------------------------------+  Kernel virt start (0xc008000000000000)
+ * |                              |
+ * |                              |
+ * |                              |
+ * +------------------------------+  Kernel linear (0xc.....)
  */
-#define H_KERN_VIRT_START ASM_CONST(0xD000000000000000)
 
-/*
- * Allow virtual mapping of one context size.
- * 512TB for 64K page size
- * 64TB for 4K page size
- */
-#define H_KERN_VIRT_SIZE (1UL << MAX_EA_BITS_PER_CONTEXT)
+#define H_VMALLOC_START		H_KERN_VIRT_START
+#define H_VMALLOC_SIZE		H_KERN_MAP_SIZE
+#define H_VMALLOC_END		(H_VMALLOC_START + H_VMALLOC_SIZE)
 
-/*
- * 8TB IO mapping size
- */
-#define H_KERN_IO_SIZE ASM_CONST(0x80000000000) /* 8T */
-
-/*
- * The vmalloc space starts at the beginning of the kernel non-linear virtual
- * region, and occupies 504T (64K) or 56T (4K)
- */
-#define H_VMALLOC_START H_KERN_VIRT_START
-#define H_VMALLOC_SIZE (H_KERN_VIRT_SIZE - H_KERN_IO_SIZE)
-#define H_VMALLOC_END  (H_VMALLOC_START + H_VMALLOC_SIZE)
+#define H_KERN_IO_START		H_VMALLOC_END
+#define H_KERN_IO_SIZE		H_KERN_MAP_SIZE
+#define H_KERN_IO_END		(H_KERN_IO_START + H_KERN_IO_SIZE)
 
-#define H_KERN_IO_START	H_VMALLOC_END
-#define H_KERN_IO_END	(H_KERN_VIRT_START + H_KERN_VIRT_SIZE)
+#define H_VMEMMAP_START		H_KERN_IO_END
+#define H_VMEMMAP_SIZE		H_KERN_MAP_SIZE
+#define H_VMEMMAP_END		(H_VMEMMAP_START + H_VMEMMAP_SIZE)
 
 /*
  * Region IDs
  */
-#define REGION_SHIFT		60UL
-#define REGION_MASK		(0xfUL << REGION_SHIFT)
-#define REGION_ID(ea)		(((unsigned long)(ea)) >> REGION_SHIFT)
-
-#define VMALLOC_REGION_ID	(REGION_ID(H_VMALLOC_START))
-#define KERNEL_REGION_ID	(REGION_ID(PAGE_OFFSET))
-#define VMEMMAP_REGION_ID	(0xfUL)	/* Server only */
-#define USER_REGION_ID		(0UL)
+#define USER_REGION_ID		1
+#define KERNEL_REGION_ID	2
+#define VMALLOC_REGION_ID	3
+#define IO_REGION_ID		4
+#define VMEMMAP_REGION_ID	5
 
 /*
  * Defines the address of the vmemap area, in its own region on
  * hash table CPUs.
  */
-#define H_VMEMMAP_BASE		(VMEMMAP_REGION_ID << REGION_SHIFT)
 
 #ifdef CONFIG_PPC_MM_SLICES
 #define HAVE_ARCH_UNMAPPED_AREA
@@ -104,6 +111,26 @@
 #define H_PUD_BAD_BITS		(PMD_TABLE_SIZE-1)
 
 #ifndef __ASSEMBLY__
+static inline int get_region_id(unsigned long ea)
+{
+	int id = (ea >> 60UL);
+
+	if (id == 0)
+		return USER_REGION_ID;
+
+	VM_BUG_ON(id != 0xc);
+	VM_BUG_ON(ea >= H_VMEMMAP_END);
+
+	if (ea >= H_VMEMMAP_START)
+		return VMEMMAP_REGION_ID;
+	else if (ea >= H_KERN_IO_START)
+		return IO_REGION_ID;
+	else if (ea >= H_VMALLOC_START)
+		return VMALLOC_REGION_ID;
+
+	return KERNEL_REGION_ID;
+}
+
 #define	hash__pmd_bad(pmd)		(pmd_val(pmd) & H_PMD_BAD_BITS)
 #define	hash__pud_bad(pud)		(pud_val(pud) & H_PUD_BAD_BITS)
 static inline int hash__pgd_bad(pgd_t pgd)