[RFC PATCH 1/2] powerpc/mm/hash: Rename get_ea_context to get_user_context

@@ -208,7 +208,7 @@ extern void radix_init_pseries(void);
 static inline void radix_init_pseries(void) { };
 #endif
 
-static inline int get_ea_context(mm_context_t *ctx, unsigned long ea)
+static inline int get_user_context(mm_context_t *ctx, unsigned long ea)
 {
 	int index = ea >> MAX_EA_BITS_PER_CONTEXT;
 

@@ -223,7 +223,7 @@ static inline int get_ea_context(mm_context_t *ctx, unsigned long ea)
 static inline unsigned long get_user_vsid(mm_context_t *ctx,
 					  unsigned long ea, int ssize)
 {
-	unsigned long context = get_ea_context(ctx, ea);
+	unsigned long context = get_user_context(ctx, ea);
 
 	return get_vsid(context, ea, ssize);
 }

@@ -81,7 +81,7 @@ static inline bool need_extra_context(struct mm_struct *mm, unsigned long ea)
 {
 	int context_id;
 
-	context_id = get_ea_context(&mm->context, ea);
+	context_id = get_user_context(&mm->context, ea);
 	if (!context_id)
 		return true;
 	return false;

@@ -464,7 +464,7 @@ static long slb_allocate_user(struct mm_struct *mm, unsigned long ea)
 	if (ea >= mm->context.slb_addr_limit)
 		return -EFAULT;
 
-	context = get_ea_context(&mm->context, ea);
+	context = get_user_context(&mm->context, ea);
 	if (!context)
 		return -EFAULT;
 

@@ -558,6 +558,26 @@ extern void slb_set_size(u16 size);
 #define ESID_BITS_MASK		((1 << ESID_BITS) - 1)
 #define ESID_BITS_1T_MASK	((1 << ESID_BITS_1T) - 1)
 
+/*
+ * Now certain config support MAX_PHYSMEM more than 512TB. Hence we will need
+ * to use more than one context for linear mapping the kernel.
+ * For vmalloc and memmap, we use just one context with 512TB. With 64 byte
+ * struct page size, we need ony 32 TB in memmap for 2PB (51 bits (MAX_PHYSMEM_BITS)).
+ */
+#if (MAX_PHYSMEM_BITS > MAX_EA_BITS_PER_CONTEXT)
+#define MAX_KERNEL_CTX_CNT	(1UL << (MAX_PHYSMEM_BITS - MAX_EA_BITS_PER_CONTEXT))
+#else
+#define MAX_KERNEL_CTX_CNT	1
+#endif
+
+#define MAX_VMALLOC_CTX_CNT	1
+#define MAX_MEMMAP_CTX_CNT	1
+
+/*
+ * Would be nice to use KERNEL_REGION_ID here
+ */
+#define KERNEL_REGION_CONTEXT_OFFSET	(0xc - 1)
+
 /*
  * 256MB segment
  * The proto-VSID space has 2^(CONTEX_BITS + ESID_BITS) - 1 segments

@@ -568,12 +588,13 @@ extern void slb_set_size(u16 size);
  * We also need to avoid the last segment of the last context, because that
  * would give a protovsid of 0x1fffffffff. That will result in a VSID 0
  * because of the modulo operation in vsid scramble.
+ *
+ * We add one extra context to MIN_USER_CONTEXT so that we can map kernel
+ * context easily. The +1 is to map the unused 0xe region mapping.
  */
 #define MAX_USER_CONTEXT	((ASM_CONST(1) << CONTEXT_BITS) - 2)
-#define MIN_USER_CONTEXT	(5)
-
-/* Would be nice to use KERNEL_REGION_ID here */
-#define KERNEL_REGION_CONTEXT_OFFSET	(0xc - 1)
+#define MIN_USER_CONTEXT	(MAX_KERNEL_CTX_CNT + MAX_VMALLOC_CTX_CNT + \
+				 MAX_MEMMAP_CTX_CNT + 2)
 
 /*
  * For platforms that support on 65bit VA we limit the context bits

@@ -733,6 +754,23 @@ static inline unsigned long get_vsid(unsigned long context, unsigned long ea,
 	return vsid_scramble(protovsid, VSID_MULTIPLIER_1T, vsid_bits);
 }
 
+static inline unsigned long get_kernel_context(unsigned long ea)
+{
+	unsigned long region_id = REGION_ID(ea);
+	unsigned long ctx = region_id - KERNEL_REGION_CONTEXT_OFFSET;
+	/*
+	 * For linear mapping we do support multiple context
+	 */
+	if (region_id == KERNEL_REGION_ID) {
+		/*
+		 * We already verified ea to be not beyond the addr limit.
+		 */
+		ctx +=  ((ea & ~REGION_MASK) >> MAX_EA_BITS_PER_CONTEXT);
+	}
+
+	return ctx;
+}
+
 /*
  * This is only valid for addresses >= PAGE_OFFSET
  */

@@ -755,8 +793,7 @@ static inline unsigned long get_kernel_vsid(unsigned long ea, int ssize)
 	 * So we can compute the context from the region (top nibble) by
 	 * subtracting 11, or 0xc - 1.
 	 */
-	context = (ea >> 60) - KERNEL_REGION_CONTEXT_OFFSET;
-
+	context = get_kernel_context(ea);
 	return get_vsid(context, ea, ssize);
 }
 

@@ -309,6 +309,21 @@ static inline u16 get_mm_addr_key(struct mm_struct *mm, unsigned long address)
  */
 #define MMU_PAGE_COUNT	16
 
+/*
+ * If we store section details in page->flags we can't increase the MAX_PHYSMEM_BITS
+ * if we increase SECTIONS_WIDTH we will not store node details in page->flags and
+ * page_to_nid does a page->section->node lookup
+ * Hence only increase for VMEMMAP. Further depending on SPARSEMEM_EXTREME reduce
+ * memory requirements with large number of sections.
+ * 51 bits is the max physical real address on POWER9
+ */
+#if defined(CONFIG_SPARSEMEM_VMEMMAP) && defined(CONFIG_SPARSEMEM_EXTREME) &&	\
+	defined (CONFIG_PPC_64K_PAGES)
+#define MAX_PHYSMEM_BITS        51
+#else
+#define MAX_PHYSMEM_BITS        46
+#endif
+
 #ifdef CONFIG_PPC_BOOK3S_64
 #include <asm/book3s/64/mmu.h>
 #else /* CONFIG_PPC_BOOK3S_64 */

@@ -9,17 +9,6 @@
  * MAX_PHYSMEM_BITS		2^N: how much memory we can have in that space
  */
 #define SECTION_SIZE_BITS       24
-/*
- * If we store section details in page->flags we can't increase the MAX_PHYSMEM_BITS
- * if we increase SECTIONS_WIDTH we will not store node details in page->flags and
- * page_to_nid does a page->section->node lookup
- * Hence only increase for VMEMMAP.
- */
-#ifdef CONFIG_SPARSEMEM_VMEMMAP
-#define MAX_PHYSMEM_BITS        47
-#else
-#define MAX_PHYSMEM_BITS        46
-#endif
 
 #endif /* CONFIG_SPARSEMEM */
 

@@ -423,16 +423,27 @@ static long slb_allocate_kernel(unsigned long ea, unsigned long id)
 	unsigned long flags;
 	int ssize;
 
-	if ((ea & ~REGION_MASK) >= (1ULL << MAX_EA_BITS_PER_CONTEXT))
-		return -EFAULT;
-
 	if (id == KERNEL_REGION_ID) {
+
+		/* We only support upto MAX_PHYSMEM_BITS */
+		if ((ea & ~REGION_MASK) > (1UL << MAX_PHYSMEM_BITS))
+			return -EFAULT;
+
 		flags = SLB_VSID_KERNEL | mmu_psize_defs[mmu_linear_psize].sllp;
+
 #ifdef CONFIG_SPARSEMEM_VMEMMAP
 	} else if (id == VMEMMAP_REGION_ID) {
+
+		if ((ea & ~REGION_MASK) >= (1ULL << MAX_EA_BITS_PER_CONTEXT))
+			return -EFAULT;
+
 		flags = SLB_VSID_KERNEL | mmu_psize_defs[mmu_vmemmap_psize].sllp;
 #endif
 	} else if (id == VMALLOC_REGION_ID) {
+
+		if ((ea & ~REGION_MASK) >= (1ULL << MAX_EA_BITS_PER_CONTEXT))
+			return -EFAULT;
+
 		if (ea < H_VMALLOC_END)
 			flags = get_paca()->vmalloc_sllp;
 		else

@@ -445,8 +456,7 @@ static long slb_allocate_kernel(unsigned long ea, unsigned long id)
 	if (!mmu_has_feature(MMU_FTR_1T_SEGMENT))
 		ssize = MMU_SEGSIZE_256M;
 
-	context = id - KERNEL_REGION_CONTEXT_OFFSET;
-
+	context = get_kernel_context(ea);
 	return slb_insert_entry(ea, context, flags, ssize, true);
 }