[RFC PATCH 1/5] mm: large system hash use vmalloc for size > MAX_ORDER when !hashdist

[RFC PATCH 1/5] mm: large system hash use vmalloc for size > MAX_ORDER when !hashdist

[Nicholas Piggin]

The kernel currently clamps large system hashes to MAX_ORDER when
hashdist is not set, which is rather arbitrary.

vmalloc space is limited on 32-bit machines, but this shouldn't
result in much more used because of small physical memory.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 mm/page_alloc.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 59661106da16..1683d54d6405 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -7978,7 +7978,7 @@ void *__init alloc_large_system_hash(const char *tablename,
 			else
 				table = memblock_alloc_raw(size,
 							   SMP_CACHE_BYTES);
-		} else if (hashdist) {
+		} else if (get_order(size) >= MAX_ORDER || hashdist) {
 			table = __vmalloc(size, gfp_flags, PAGE_KERNEL);
 		} else {
 			/*
@@ -7986,10 +7986,8 @@ void *__init alloc_large_system_hash(const char *tablename,
 			 * some pages at the end of hash table which
 			 * alloc_pages_exact() automatically does
 			 */
-			if (get_order(size) < MAX_ORDER) {
-				table = alloc_pages_exact(size, gfp_flags);
-				kmemleak_alloc(table, size, 1, gfp_flags);
-			}
+			table = alloc_pages_exact(size, gfp_flags);
+			kmemleak_alloc(table, size, 1, gfp_flags);
 		}
 	} while (!table && size > PAGE_SIZE && --log2qty);
 
-- 
2.20.1

[RFC PATCH 2/5] mm: large system hash avoid vmap for non-NUMA machines when hashdist

[Nicholas Piggin]

hashdist currently always uses vmalloc when hashdist is true. When
there is only 1 online node and size <= MAX_ORDER, vmalloc can be
avoided.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 mm/page_alloc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 1683d54d6405..1312d4db5602 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -7978,7 +7978,8 @@ void *__init alloc_large_system_hash(const char *tablename,
 			else
 				table = memblock_alloc_raw(size,
 							   SMP_CACHE_BYTES);
-		} else if (get_order(size) >= MAX_ORDER || hashdist) {
+		} else if (get_order(size) >= MAX_ORDER ||
+				(hashdist && num_online_nodes() > 1)) {
 			table = __vmalloc(size, gfp_flags, PAGE_KERNEL);
 		} else {
 			/*
-- 
2.20.1

[RFC PATCH 3/5] mm/vmalloc: Hugepage vmalloc mappings

[Nicholas Piggin]

This appears to help cached git diff performance by about 5% on a
POWER9 (with 32MB dentry cache hash).

  Profiling git diff dTLB misses with a vanilla kernel:

  81.75%  git      [kernel.vmlinux]    [k] __d_lookup_rcu
   7.21%  git      [kernel.vmlinux]    [k] strncpy_from_user
   1.77%  git      [kernel.vmlinux]    [k] find_get_entry
   1.59%  git      [kernel.vmlinux]    [k] kmem_cache_free

            40,168      dTLB-miss
       0.100342754 seconds time elapsed

After this patch (and the subsequent powerpc HUGE_VMAP patches), the
dentry cache hash gets mapped with 2MB pages:

             2,987      dTLB-miss
       0.095933138 seconds time elapsed

elapsed time improvement isn't too scientific but seems consistent,
TLB misses certainly improves an order of magnitude. My laptop
takes a lot of misses here too, so x86 would be interesting to test,
I think it should just work there.

---
 include/linux/vmalloc.h |  1 +
 mm/vmalloc.c            | 87 +++++++++++++++++++++++++++--------------
 2 files changed, 59 insertions(+), 29 deletions(-)

diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h
index c6eebb839552..029635560306 100644
--- a/include/linux/vmalloc.h
+++ b/include/linux/vmalloc.h
@@ -42,6 +42,7 @@ struct vm_struct {
 	unsigned long		size;
 	unsigned long		flags;
 	struct page		**pages;
+	unsigned int		page_shift;
 	unsigned int		nr_pages;
 	phys_addr_t		phys_addr;
 	const void		*caller;
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index e5e9e1fcac01..c9ba88768bca 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -216,32 +216,34 @@ static int vmap_p4d_range(pgd_t *pgd, unsigned long addr,
  * Ie. pte at addr+N*PAGE_SIZE shall point to pfn corresponding to pages[N]
  */
 static int vmap_page_range_noflush(unsigned long start, unsigned long end,
-				   pgprot_t prot, struct page **pages)
+				   pgprot_t prot, struct page **pages,
+				   unsigned int page_shift)
 {
-	pgd_t *pgd;
-	unsigned long next;
 	unsigned long addr = start;
-	int err = 0;
-	int nr = 0;
+	unsigned int i, nr = (end - start) >> (PAGE_SHIFT + page_shift);
 
-	BUG_ON(addr >= end);
-	pgd = pgd_offset_k(addr);
-	do {
-		next = pgd_addr_end(addr, end);
-		err = vmap_p4d_range(pgd, addr, next, prot, pages, &nr);
+	for (i = 0; i < nr; i++) {
+		int err;
+
+		err = ioremap_page_range(addr,
+					addr + (PAGE_SIZE << page_shift),
+					__pa(page_address(pages[i])), prot);
 		if (err)
 			return err;
-	} while (pgd++, addr = next, addr != end);
+
+		addr += PAGE_SIZE << page_shift;
+	}
 
 	return nr;
 }
 
 static int vmap_page_range(unsigned long start, unsigned long end,
-			   pgprot_t prot, struct page **pages)
+			   pgprot_t prot, struct page **pages,
+			   unsigned int page_shift)
 {
 	int ret;
 
-	ret = vmap_page_range_noflush(start, end, prot, pages);
+	ret = vmap_page_range_noflush(start, end, prot, pages, page_shift);
 	flush_cache_vmap(start, end);
 	return ret;
 }
@@ -1189,7 +1191,7 @@ void *vm_map_ram(struct page **pages, unsigned int count, int node, pgprot_t pro
 		addr = va->va_start;
 		mem = (void *)addr;
 	}
-	if (vmap_page_range(addr, addr + size, prot, pages) < 0) {
+	if (vmap_page_range(addr, addr + size, prot, pages, 0) < 0) {
 		vm_unmap_ram(mem, count);
 		return NULL;
 	}
@@ -1305,7 +1307,7 @@ void __init vmalloc_init(void)
 int map_kernel_range_noflush(unsigned long addr, unsigned long size,
 			     pgprot_t prot, struct page **pages)
 {
-	return vmap_page_range_noflush(addr, addr + size, prot, pages);
+	return vmap_page_range_noflush(addr, addr + size, prot, pages, 0);
 }
 
 /**
@@ -1352,7 +1354,7 @@ int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page **pages)
 	unsigned long end = addr + get_vm_area_size(area);
 	int err;
 
-	err = vmap_page_range(addr, end, prot, pages);
+	err = vmap_page_range(addr, end, prot, pages, 0);
 
 	return err > 0 ? 0 : err;
 }
@@ -1395,8 +1397,9 @@ static struct vm_struct *__get_vm_area_node(unsigned long size,
 		return NULL;
 
 	if (flags & VM_IOREMAP)
-		align = 1ul << clamp_t(int, get_count_order_long(size),
-				       PAGE_SHIFT, IOREMAP_MAX_ORDER);
+		align = max(align,
+				1ul << clamp_t(int, get_count_order_long(size),
+				       PAGE_SHIFT, IOREMAP_MAX_ORDER));
 
 	area = kzalloc_node(sizeof(*area), gfp_mask & GFP_RECLAIM_MASK, node);
 	if (unlikely(!area))
@@ -1608,7 +1611,7 @@ static void __vunmap(const void *addr, int deallocate_pages)
 			struct page *page = area->pages[i];
 
 			BUG_ON(!page);
-			__free_pages(page, 0);
+			__free_pages(page, area->page_shift);
 		}
 
 		kvfree(area->pages);
@@ -1751,14 +1754,17 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 				 pgprot_t prot, int node)
 {
 	struct page **pages;
+	unsigned long addr = (unsigned long)area->addr;
+	unsigned long size = get_vm_area_size(area);
+	unsigned int page_shift = area->page_shift;
+	unsigned int shift = page_shift + PAGE_SHIFT;
 	unsigned int nr_pages, array_size, i;
 	const gfp_t nested_gfp = (gfp_mask & GFP_RECLAIM_MASK) | __GFP_ZERO;
 	const gfp_t alloc_mask = gfp_mask | __GFP_NOWARN;
 	const gfp_t highmem_mask = (gfp_mask & (GFP_DMA | GFP_DMA32)) ?
-					0 :
-					__GFP_HIGHMEM;
+					0 : __GFP_HIGHMEM;
 
-	nr_pages = get_vm_area_size(area) >> PAGE_SHIFT;
+	nr_pages = size >> shift;
 	array_size = (nr_pages * sizeof(struct page *));
 
 	area->nr_pages = nr_pages;
@@ -1779,10 +1785,8 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 	for (i = 0; i < area->nr_pages; i++) {
 		struct page *page;
 
-		if (node == NUMA_NO_NODE)
-			page = alloc_page(alloc_mask|highmem_mask);
-		else
-			page = alloc_pages_node(node, alloc_mask|highmem_mask, 0);
+		page = alloc_pages_node(node,
+				alloc_mask|highmem_mask, page_shift);
 
 		if (unlikely(!page)) {
 			/* Successfully allocated i pages, free them in __vunmap() */
@@ -1794,8 +1798,9 @@ static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
 			cond_resched();
 	}
 
-	if (map_vm_area(area, prot, pages))
+	if (vmap_page_range(addr, addr + size, prot, pages, page_shift) < 0)
 		goto fail;
+
 	return area->addr;
 
 fail:
@@ -1832,19 +1837,35 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align,
 	struct vm_struct *area;
 	void *addr;
 	unsigned long real_size = size;
+	unsigned long real_align = align;
+	unsigned long size_per_node;
+	unsigned int shift;
 
 	size = PAGE_ALIGN(size);
 	if (!size || (size >> PAGE_SHIFT) > totalram_pages())
 		goto fail;
 
+	size_per_node = size;
+	if (node == NUMA_NO_NODE)
+		size_per_node /= num_online_nodes();
+	if (size_per_node >= PMD_SIZE)
+		shift = PMD_SHIFT;
+	else
+		shift = PAGE_SHIFT;
+again:
+	align = max(real_align, 1UL << shift);
+	size = ALIGN(real_size, align);
+
 	area = __get_vm_area_node(size, align, VM_ALLOC | VM_UNINITIALIZED |
 				vm_flags, start, end, node, gfp_mask, caller);
 	if (!area)
 		goto fail;
 
+	area->page_shift = shift - PAGE_SHIFT;
+
 	addr = __vmalloc_area_node(area, gfp_mask, prot, node);
 	if (!addr)
-		return NULL;
+		goto fail;
 
 	/*
 	 * In this function, newly allocated vm_struct has VM_UNINITIALIZED
@@ -1858,8 +1879,16 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align,
 	return addr;
 
 fail:
-	warn_alloc(gfp_mask, NULL,
+	if (shift == PMD_SHIFT) {
+		shift = PAGE_SHIFT;
+		goto again;
+	}
+
+	if (!area) {
+		/* Warn for area allocation, page allocations already warn */
+		warn_alloc(gfp_mask, NULL,
 			  "vmalloc: allocation failure: %lu bytes", real_size);
+	}
 	return NULL;
 }
 
-- 
2.20.1

[RFC PATCH 4/5] powerpc/64s/radix: Enable HAVE_ARCH_HUGE_VMAP

[Nicholas Piggin]

This does not actually enable huge vmap mappings, because powerpc/64
ioremap does not call ioremap_page_range, but it is required before
implementing huge mappings in ioremap, because the generic vunmap code
needs to cope with them.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/Kconfig                     |  1 +
 arch/powerpc/mm/book3s64/radix_pgtable.c | 93 ++++++++++++++++++++++++
 2 files changed, 94 insertions(+)

diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index d7996cfaceca..ffac84600e0e 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -166,6 +166,7 @@ config PPC
 	select GENERIC_STRNLEN_USER
 	select GENERIC_TIME_VSYSCALL
 	select HAVE_ARCH_AUDITSYSCALL
+	select HAVE_ARCH_HUGE_VMAP		if PPC_BOOK3S_64 && PPC_RADIX_MMU
 	select HAVE_ARCH_JUMP_LABEL
 	select HAVE_ARCH_KASAN			if PPC32
 	select HAVE_ARCH_KGDB
diff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c
index c9bcf428dd2b..3bc9ade56277 100644
--- a/arch/powerpc/mm/book3s64/radix_pgtable.c
+++ b/arch/powerpc/mm/book3s64/radix_pgtable.c
@@ -1122,3 +1122,96 @@ void radix__ptep_modify_prot_commit(struct vm_area_struct *vma,
 
 	set_pte_at(mm, addr, ptep, pte);
 }
+
+int __init arch_ioremap_pud_supported(void)
+{
+	return radix_enabled();
+}
+
+int __init arch_ioremap_pmd_supported(void)
+{
+	return radix_enabled();
+}
+
+int p4d_free_pud_page(p4d_t *p4d, unsigned long addr)
+{
+	return 0;
+}
+
+int pud_set_huge(pud_t *pud, phys_addr_t addr, pgprot_t prot)
+{
+	pte_t *ptep = (pte_t *)pud;
+	pte_t new_pud = pfn_pte(__phys_to_pfn(addr), prot);
+
+	set_pte_at(&init_mm, 0 /* radix unused */, ptep, new_pud);
+
+	return 1;
+}
+
+int pud_clear_huge(pud_t *pud)
+{
+	if (pud_huge(*pud)) {
+		pud_clear(pud);
+		return 1;
+	}
+
+	return 0;
+}
+
+int pud_free_pmd_page(pud_t *pud, unsigned long addr)
+{
+	pmd_t *pmd;
+	int i;
+
+	pmd = (pmd_t *)pud_page_vaddr(*pud);
+	pud_clear(pud);
+
+	flush_tlb_kernel_range(addr, addr + PUD_SIZE);
+
+	for (i = 0; i < PTRS_PER_PMD; i++) {
+		if (!pmd_none(pmd[i])) {
+			pte_t *pte;
+			pte = (pte_t *)pmd_page_vaddr(pmd[i]);
+
+			pte_free_kernel(&init_mm, pte);
+		}
+	}
+
+	pmd_free(&init_mm, pmd);
+
+	return 1;
+}
+
+int pmd_set_huge(pmd_t *pmd, phys_addr_t addr, pgprot_t prot)
+{
+	pte_t *ptep = (pte_t *)pmd;
+	pte_t new_pmd = pfn_pte(__phys_to_pfn(addr), prot);
+
+	set_pte_at(&init_mm, 0 /* radix unused */, ptep, new_pmd);
+
+	return 1;
+}
+
+int pmd_clear_huge(pmd_t *pmd)
+{
+	if (pmd_huge(*pmd)) {
+		pmd_clear(pmd);
+		return 1;
+	}
+
+	return 0;
+}
+
+int pmd_free_pte_page(pmd_t *pmd, unsigned long addr)
+{
+	pte_t *pte;
+
+	pte = (pte_t *)pmd_page_vaddr(*pmd);
+	pmd_clear(pmd);
+
+	flush_tlb_kernel_range(addr, addr + PMD_SIZE);
+
+	pte_free_kernel(&init_mm, pte);
+
+	return 1;
+}
-- 
2.20.1

[RFC PATCH 5/5] powerpc/64s/radix: iomap use huge page mappings

[Nicholas Piggin]

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/book3s/64/pgtable.h |  8 +++
 arch/powerpc/mm/pgtable_64.c                 | 54 +++++++++++++++++---
 2 files changed, 56 insertions(+), 6 deletions(-)

diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h b/arch/powerpc/include/asm/book3s/64/pgtable.h
index 7dede2e34b70..93b8a99df88e 100644
--- a/arch/powerpc/include/asm/book3s/64/pgtable.h
+++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
@@ -274,6 +274,14 @@ extern unsigned long __vmalloc_end;
 #define VMALLOC_START	__vmalloc_start
 #define VMALLOC_END	__vmalloc_end
 
+static inline unsigned int ioremap_max_order(void)
+{
+	if (radix_enabled())
+		return PUD_SHIFT;
+	return 7 + PAGE_SHIFT; /* default from linux/vmalloc.h */
+}
+#define IOREMAP_MAX_ORDER ({ ioremap_max_order();})
+
 extern unsigned long __kernel_virt_start;
 extern unsigned long __kernel_virt_size;
 extern unsigned long __kernel_io_start;
diff --git a/arch/powerpc/mm/pgtable_64.c b/arch/powerpc/mm/pgtable_64.c
index d2d976ff8a0e..f660116251e6 100644
--- a/arch/powerpc/mm/pgtable_64.c
+++ b/arch/powerpc/mm/pgtable_64.c
@@ -112,7 +112,7 @@ unsigned long ioremap_bot = IOREMAP_BASE;
  * __ioremap_at - Low level function to establish the page tables
  *                for an IO mapping
  */
-void __iomem *__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_t prot)
+static void __iomem * hash__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_t prot)
 {
 	unsigned long i;
 
@@ -120,6 +120,50 @@ void __iomem *__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_
 	if (pgprot_val(prot) & H_PAGE_4K_PFN)
 		return NULL;
 
+	for (i = 0; i < size; i += PAGE_SIZE)
+		if (map_kernel_page((unsigned long)ea + i, pa + i, prot))
+			return NULL;
+
+	return (void __iomem *)ea;
+}
+
+static int radix__ioremap_page_range(unsigned long addr, unsigned long end,
+		       phys_addr_t phys_addr, pgprot_t prot)
+{
+	while (addr != end) {
+		if (!(addr & ~PUD_MASK) && !(phys_addr & ~PUD_MASK) &&
+				end - addr >= PUD_SIZE) {
+			if (radix__map_kernel_page(addr, phys_addr, prot, PUD_SIZE))
+				return -ENOMEM;
+			addr += PUD_SIZE;
+			phys_addr += PUD_SIZE;
+
+		} else if (!(addr & ~PMD_MASK) && !(phys_addr & ~PMD_MASK) &&
+				end - addr >= PMD_SIZE) {
+			if (radix__map_kernel_page(addr, phys_addr, prot, PMD_SIZE))
+				return -ENOMEM;
+			addr += PMD_SIZE;
+			phys_addr += PMD_SIZE;
+
+		} else {
+			if (radix__map_kernel_page(addr, phys_addr, prot, PAGE_SIZE))
+				return -ENOMEM;
+			addr += PAGE_SIZE;
+			phys_addr += PAGE_SIZE;
+		}
+	}
+	return 0;
+}
+
+static void __iomem * radix__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_t prot)
+{
+	if (radix__ioremap_page_range((unsigned long)ea, (unsigned long)ea + size, pa, prot))
+		return NULL;
+	return ea;
+}
+
+void __iomem *__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_t prot)
+{
 	if ((ea + size) >= (void *)IOREMAP_END) {
 		pr_warn("Outside the supported range\n");
 		return NULL;
@@ -129,11 +173,9 @@ void __iomem *__ioremap_at(phys_addr_t pa, void *ea, unsigned long size, pgprot_
 	WARN_ON(((unsigned long)ea) & ~PAGE_MASK);
 	WARN_ON(size & ~PAGE_MASK);
 
-	for (i = 0; i < size; i += PAGE_SIZE)
-		if (map_kernel_page((unsigned long)ea + i, pa + i, prot))
-			return NULL;
-
-	return (void __iomem *)ea;
+	if (radix_enabled())
+		return radix__ioremap_at(pa, ea, size, prot);
+	return hash__ioremap_at(pa, ea, size, prot);
 }
 
 /**
-- 
2.20.1