[S390] zero page cache synonyms

If the zero page is mapped to virtual user space addresses that differ
only in bit 2^12 or 2^13 we get L1 cache synonyms which can affect
performance. Follow the mips model and use multiple zero pages to avoid
the synonyms.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

arch/s390/include/asm/pgtable.h

@@ -46,11 +46,27 @@ extern void vmem_map_init(void);
 #define update_mmu_cache(vma, address, ptep)     do { } while (0)
 
 /*
- * ZERO_PAGE is a global shared page that is always zero: used
+ * ZERO_PAGE is a global shared page that is always zero; used
  * for zero-mapped memory areas etc..
  */
-extern char empty_zero_page[PAGE_SIZE];
-#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
+
+extern unsigned long empty_zero_page;
+extern unsigned long zero_page_mask;
+
+#define ZERO_PAGE(vaddr) \
+	(virt_to_page((void *)(empty_zero_page + \
+	 (((unsigned long)(vaddr)) &zero_page_mask))))
+
+#define is_zero_pfn is_zero_pfn
+static inline int is_zero_pfn(unsigned long pfn)
+{
+	extern unsigned long zero_pfn;
+	unsigned long offset_from_zero_pfn = pfn - zero_pfn;
+	return offset_from_zero_pfn <= (zero_page_mask >> PAGE_SHIFT);
+}
+
+#define my_zero_pfn(addr)	page_to_pfn(ZERO_PAGE(addr))
+
 #endif /* !__ASSEMBLY__ */
 
 /*

arch/s390/mm/init.c

@@ -42,9 +42,52 @@ DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 
 pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__((__aligned__(PAGE_SIZE)));
 
-char  empty_zero_page[PAGE_SIZE] __attribute__((__aligned__(PAGE_SIZE)));
+unsigned long empty_zero_page, zero_page_mask;
 EXPORT_SYMBOL(empty_zero_page);
 
+static unsigned long setup_zero_pages(void)
+{
+	struct cpuid cpu_id;
+	unsigned int order;
+	unsigned long size;
+	struct page *page;
+	int i;
+
+	get_cpu_id(&cpu_id);
+	switch (cpu_id.machine) {
+	case 0x9672:	/* g5 */
+	case 0x2064:	/* z900 */
+	case 0x2066:	/* z900 */
+	case 0x2084:	/* z990 */
+	case 0x2086:	/* z990 */
+	case 0x2094:	/* z9-109 */
+	case 0x2096:	/* z9-109 */
+		order = 0;
+		break;
+	case 0x2097:	/* z10 */
+	case 0x2098:	/* z10 */
+	default:
+		order = 2;
+		break;
+	}
+
+	empty_zero_page = __get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
+	if (!empty_zero_page)
+		panic("Out of memory in setup_zero_pages");
+
+	page = virt_to_page((void *) empty_zero_page);
+	split_page(page, order);
+	for (i = 1 << order; i > 0; i--) {
+		SetPageReserved(page);
+		page++;
+	}
+
+	size = PAGE_SIZE << order;
+	zero_page_mask = (size - 1) & PAGE_MASK;
+
+	return 1UL << order;
+}
+
 /*
  * paging_init() sets up the page tables
  */
@@ -92,14 +135,12 @@ void __init mem_init(void)
         max_mapnr = num_physpages = max_low_pfn;
         high_memory = (void *) __va(max_low_pfn * PAGE_SIZE);
 
-        /* clear the zero-page */
-        memset(empty_zero_page, 0, PAGE_SIZE);
-
 	/* Setup guest page hinting */
 	cmma_init();
 
 	/* this will put all low memory onto the freelists */
 	totalram_pages += free_all_bootmem();
+	totalram_pages -= setup_zero_pages();	/* Setup zeroed pages. */
 
 	reservedpages = 0;