[PATCH v7 00/11] complete deferred page initialization

[PATCH v7 00/11] complete deferred page initialization

[Pavel Tatashin]

Changelog:
v8 - v7
- Added Acked-by's from Dave Miller for SPARC changes
- Fixed a minor compiling issue on tile architecture reported by kbuild

v7 - v6
- Addressed comments from Michal Hocko
- memblock_discard() patch was removed from this series and integrated
  separately
- Fixed bug reported by kbuild test robot new patch:
  mm: zero reserved and unavailable struct pages
- Removed patch
  x86/mm: reserve only exiting low pages
  As, it is not needed anymore, because of the previous fix
- Re-wrote deferred_init_memmap(), found and fixed an existing bug, where
  page variable is not reset when zone holes present.
- Merged several patches together per Michal request
- Added performance data including raw logs

v6 - v5
- Fixed ARM64 + kasan code, as reported by Ard Biesheuvel
- Tested ARM64 code in qemu and found few more issues, that I fixed in this
  iteration
- Added page roundup/rounddown to x86 and arm zeroing routines to zero the
  whole allocated range, instead of only provided address range.
- Addressed SPARC related comment from Sam Ravnborg
- Fixed section mismatch warnings related to memblock_discard().

v5 - v4
- Fixed build issues reported by kbuild on various configurations

v4 - v3
- Rewrote code to zero sturct pages in __init_single_page() as
  suggested by Michal Hocko
- Added code to handle issues related to accessing struct page
  memory before they are initialized.

v3 - v2
- Addressed David Miller comments about one change per patch:
    * Splited changes to platforms into 4 patches
    * Made "do not zero vmemmap_buf" as a separate patch

v2 - v1
- Per request, added s390 to deferred "struct page" zeroing
- Collected performance data on x86 which proofs the importance to
  keep memset() as prefetch (see below).

SMP machines can benefit from the DEFERRED_STRUCT_PAGE_INIT config option,
which defers initializing struct pages until all cpus have been started so
it can be done in parallel.

However, this feature is sub-optimal, because the deferred page
initialization code expects that the struct pages have already been zeroed,
and the zeroing is done early in boot with a single thread only.  Also, we
access that memory and set flags before struct pages are initialized. All
of this is fixed in this patchset.

In this work we do the following:
- Never read access struct page until it was initialized
- Never set any fields in struct pages before they are initialized
- Zero struct page at the beginning of struct page initialization


==========================================================================
Performance improvements on x86 machine with 8 nodes:
Intel(R) Xeon(R) CPU E7-8895 v3 @ 2.60GHz and 1T of memory:
                        TIME          SPEED UP
base no deferred:       95.796233s
fix no deferred:        79.978956s    19.77%

base deferred:          77.254713s
fix deferred:           55.050509s    40.34%
==========================================================================
SPARC M6 3600 MHz with 15T of memory
                        TIME          SPEED UP
base no deferred:       358.335727s
fix no deferred:        302.320936s   18.52%

base deferred:          237.534603s
fix deferred:           182.103003s   30.44%
==========================================================================
Raw dmesg output with timestamps:
x86 base no deferred:    https://hastebin.com/ofunepurit.scala
x86 base deferred:       https://hastebin.com/ifazegeyas.scala
x86 fix no deferred:     https://hastebin.com/pegocohevo.scala
x86 fix deferred:        https://hastebin.com/ofupevikuk.scala
sparc base no deferred:  https://hastebin.com/ibobeteken.go
sparc base deferred:     https://hastebin.com/fariqimiyu.go
sparc fix no deferred:   https://hastebin.com/muhegoheyi.go
sparc fix deferred:      https://hastebin.com/xadinobutu.go

Pavel Tatashin (11):
  x86/mm: setting fields in deferred pages
  sparc64/mm: setting fields in deferred pages
  mm: deferred_init_memmap improvements
  sparc64: simplify vmemmap_populate
  mm: defining memblock_virt_alloc_try_nid_raw
  mm: zero struct pages during initialization
  sparc64: optimized struct page zeroing
  mm: zero reserved and unavailable struct pages
  x86/kasan: explicitly zero kasan shadow memory
  arm64/kasan: explicitly zero kasan shadow memory
  mm: stop zeroing memory during allocation in vmemmap

 arch/arm64/mm/kasan_init.c          |  42 ++++++++
 arch/sparc/include/asm/pgtable_64.h |  30 ++++++
 arch/sparc/mm/init_64.c             |  31 +++---
 arch/x86/mm/init_64.c               |   9 +-
 arch/x86/mm/kasan_init_64.c         |  66 ++++++++++++
 include/linux/bootmem.h             |  27 +++++
 include/linux/memblock.h            |  16 +++
 include/linux/mm.h                  |  26 +++++
 mm/memblock.c                       |  60 +++++++++--
 mm/page_alloc.c                     | 207 ++++++++++++++++++++----------------
 mm/sparse-vmemmap.c                 |  15 ++-
 mm/sparse.c                         |   6 +-
 12 files changed, 406 insertions(+), 129 deletions(-)

-- 
2.14.1

[PATCH v8 01/11] x86/mm: setting fields in deferred pages

[Pavel Tatashin]

Without deferred struct page feature (CONFIG_DEFERRED_STRUCT_PAGE_INIT),
flags and other fields in "struct page"es are never changed prior to first
initializing struct pages by going through __init_single_page().

With deferred struct page feature enabled, however, we set fields in
register_page_bootmem_info that are subsequently clobbered right after in
free_all_bootmem:

        mem_init() {
                register_page_bootmem_info();
                free_all_bootmem();
                ...
        }

When register_page_bootmem_info() is called only non-deferred struct pages
are initialized. But, this function goes through some reserved pages which
might be part of the deferred, and thus are not yet initialized.

  mem_init
   register_page_bootmem_info
    register_page_bootmem_info_node
     get_page_bootmem
      .. setting fields here ..
      such as: page->freelist = (void *)type;

  free_all_bootmem()
   free_low_memory_core_early()
    for_each_reserved_mem_region()
     reserve_bootmem_region()
      init_reserved_page() <- Only if this is deferred reserved page
       __init_single_pfn()
        __init_single_page()
            memset(0) <-- Loose the set fields here

We end-up with issue where, currently we do not observe problem as memory
is explicitly zeroed. But, if flag asserts are changed we can start hitting
issues.

Also, because in this patch series we will stop zeroing struct page memory
during allocation, we must make sure that struct pages are properly
initialized prior to using them.

The deferred-reserved pages are initialized in free_all_bootmem().
Therefore, the fix is to switch the above calls.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 arch/x86/mm/init_64.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index 048fbe8fc274..42b4b7a585c2 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -1173,12 +1173,17 @@ void __init mem_init(void)
 
 	/* clear_bss() already clear the empty_zero_page */
 
-	register_page_bootmem_info();
-
 	/* this will put all memory onto the freelists */
 	free_all_bootmem();
 	after_bootmem = 1;
 
+	/* Must be done after boot memory is put on freelist, because here we
+	 * might set fields in deferred struct pages that have not yet been
+	 * initialized, and free_all_bootmem() initializes all the reserved
+	 * deferred pages for us.
+	 */
+	register_page_bootmem_info();
+
 	/* Register memory areas for /proc/kcore */
 	kclist_add(&kcore_vsyscall, (void *)VSYSCALL_ADDR,
 			 PAGE_SIZE, KCORE_OTHER);
-- 
2.14.1

[PATCH v8 02/11] sparc64/mm: setting fields in deferred pages

[Pavel Tatashin]

Without deferred struct page feature (CONFIG_DEFERRED_STRUCT_PAGE_INIT),
flags and other fields in "struct page"es are never changed prior to first
initializing struct pages by going through __init_single_page().

With deferred struct page feature enabled there is a case where we set some
fields prior to initializing:

mem_init() {
     register_page_bootmem_info();
     free_all_bootmem();
     ...
}

When register_page_bootmem_info() is called only non-deferred struct pages
are initialized. But, this function goes through some reserved pages which
might be part of the deferred, and thus are not yet initialized.

mem_init
register_page_bootmem_info
register_page_bootmem_info_node
 get_page_bootmem
  .. setting fields here ..
  such as: page->freelist = (void *)type;

free_all_bootmem()
free_low_memory_core_early()
 for_each_reserved_mem_region()
  reserve_bootmem_region()
   init_reserved_page() <- Only if this is deferred reserved page
    __init_single_pfn()
     __init_single_page()
      memset(0) <-- Loose the set fields here

We end-up with similar issue as in the previous patch, where currently we
do not observe problem as memory is zeroed. But, if flag asserts are
changed we can start hitting issues.

Also, because in this patch series we will stop zeroing struct page memory
during allocation, we must make sure that struct pages are properly
initialized prior to using them.

The deferred-reserved pages are initialized in free_all_bootmem().
Therefore, the fix is to switch the above calls.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Acked-by: David S. Miller <davem@davemloft.net>
---
 arch/sparc/mm/init_64.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/sparc/mm/init_64.c b/arch/sparc/mm/init_64.c
index b2ba410b26f4..078f1352736e 100644
--- a/arch/sparc/mm/init_64.c
+++ b/arch/sparc/mm/init_64.c
@@ -2539,9 +2539,15 @@ void __init mem_init(void)
 {
 	high_memory = __va(last_valid_pfn << PAGE_SHIFT);
 
-	register_page_bootmem_info();
 	free_all_bootmem();
 
+	/* Must be done after boot memory is put on freelist, because here we
+	 * might set fields in deferred struct pages that have not yet been
+	 * initialized, and free_all_bootmem() initializes all the reserved
+	 * deferred pages for us.
+	 */
+	register_page_bootmem_info();
+
 	/*
 	 * Set up the zero page, mark it reserved, so that page count
 	 * is not manipulated when freeing the page from user ptes.
-- 
2.14.1

[PATCH v8 03/11] mm: deferred_init_memmap improvements

[Pavel Tatashin]

This patch fixes two issues in deferred_init_memmap

=====
In deferred_init_memmap() where all deferred struct pages are initialized
we have a check like this:

if (page->flags) {
	VM_BUG_ON(page_zone(page) != zone);
	goto free_range;
}

This way we are checking if the current deferred page has already been
initialized. It works, because memory for struct pages has been zeroed, and
the only way flags are not zero if it went through __init_single_page()
before.  But, once we change the current behavior and won't zero the memory
in memblock allocator, we cannot trust anything inside "struct page"es
until they are initialized. This patch fixes this.

The deferred_init_memmap() is re-written to loop through only free memory
ranges provided by memblock.

=====
This patch fixes another existing issue on systems that have holes in
zones i.e CONFIG_HOLES_IN_ZONE is defined.

In for_each_mem_pfn_range() we have code like this:

if (!pfn_valid_within(pfn)
	goto free_range;

Note: 'page' is not set to NULL and is not incremented but 'pfn' advances.
Thus means if deferred struct pages are enabled on systems with these kind
of holes, linux would get memory corruptions. I have fixed this issue by
defining a new macro that performs all the necessary operations when we
free the current set of pages.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 mm/page_alloc.c | 161 +++++++++++++++++++++++++++-----------------------------
 1 file changed, 78 insertions(+), 83 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index c841af88836a..d132c801d2c1 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1410,14 +1410,17 @@ void clear_zone_contiguous(struct zone *zone)
 }
 
 #ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
-static void __init deferred_free_range(struct page *page,
-					unsigned long pfn, int nr_pages)
+static void __init deferred_free_range(unsigned long pfn,
+				       unsigned long nr_pages)
 {
-	int i;
+	struct page *page;
+	unsigned long i;
 
-	if (!page)
+	if (!nr_pages)
 		return;
 
+	page = pfn_to_page(pfn);
+
 	/* Free a large naturally-aligned chunk if possible */
 	if (nr_pages == pageblock_nr_pages &&
 	    (pfn & (pageblock_nr_pages - 1)) == 0) {
@@ -1443,19 +1446,82 @@ static inline void __init pgdat_init_report_one_done(void)
 		complete(&pgdat_init_all_done_comp);
 }
 
+#define DEFERRED_FREE(nr_free, free_base_pfn, page)			\
+({									\
+	unsigned long nr = (nr_free);					\
+									\
+	deferred_free_range((free_base_pfn), (nr));			\
+	(free_base_pfn) = 0;						\
+	(nr_free) = 0;							\
+	page = NULL;							\
+	nr;								\
+})
+
+static unsigned long deferred_init_range(int nid, int zid, unsigned long pfn,
+					 unsigned long end_pfn)
+{
+	struct mminit_pfnnid_cache nid_init_state = { };
+	unsigned long nr_pgmask = pageblock_nr_pages - 1;
+	unsigned long free_base_pfn = 0;
+	unsigned long nr_pages = 0;
+	unsigned long nr_free = 0;
+	struct page *page = NULL;
+
+	for (; pfn < end_pfn; pfn++) {
+		/*
+		 * First we check if pfn is valid on architectures where it is
+		 * possible to have holes within pageblock_nr_pages. On systems
+		 * where it is not possible, this function is optimized out.
+		 *
+		 * Then, we check if a current large page is valid by only
+		 * checking the validity of the head pfn.
+		 *
+		 * meminit_pfn_in_nid is checked on systems where pfns can
+		 * interleave within a node: a pfn is between start and end
+		 * of a node, but does not belong to this memory node.
+		 *
+		 * Finally, we minimize pfn page lookups and scheduler checks by
+		 * performing it only once every pageblock_nr_pages.
+		 */
+		if (!pfn_valid_within(pfn)) {
+			nr_pages += DEFERRED_FREE(nr_free, free_base_pfn, page);
+		} else if (!(pfn & nr_pgmask) && !pfn_valid(pfn)) {
+			nr_pages += DEFERRED_FREE(nr_free, free_base_pfn, page);
+		} else if (!meminit_pfn_in_nid(pfn, nid, &nid_init_state)) {
+			nr_pages += DEFERRED_FREE(nr_free, free_base_pfn, page);
+		} else if (page && (pfn & nr_pgmask)) {
+			page++;
+			__init_single_page(page, pfn, zid, nid);
+			nr_free++;
+		} else {
+			nr_pages += DEFERRED_FREE(nr_free, free_base_pfn, page);
+			page = pfn_to_page(pfn);
+			__init_single_page(page, pfn, zid, nid);
+			free_base_pfn = pfn;
+			nr_free = 1;
+			cond_resched();
+		}
+	}
+	/* Free the last block of pages to allocator */
+	nr_pages += DEFERRED_FREE(nr_free, free_base_pfn, page);
+
+	return nr_pages;
+}
+
 /* Initialise remaining memory on a node */
 static int __init deferred_init_memmap(void *data)
 {
 	pg_data_t *pgdat = data;
 	int nid = pgdat->node_id;
-	struct mminit_pfnnid_cache nid_init_state = { };
 	unsigned long start = jiffies;
 	unsigned long nr_pages = 0;
-	unsigned long walk_start, walk_end;
-	int i, zid;
+	unsigned long spfn, epfn;
+	phys_addr_t spa, epa;
+	int zid;
 	struct zone *zone;
 	unsigned long first_init_pfn = pgdat->first_deferred_pfn;
 	const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id);
+	u64 i;
 
 	if (first_init_pfn == ULONG_MAX) {
 		pgdat_init_report_one_done();
@@ -1477,83 +1543,12 @@ static int __init deferred_init_memmap(void *data)
 		if (first_init_pfn < zone_end_pfn(zone))
 			break;
 	}
+	first_init_pfn = max(zone->zone_start_pfn, first_init_pfn);
 
-	for_each_mem_pfn_range(i, nid, &walk_start, &walk_end, NULL) {
-		unsigned long pfn, end_pfn;
-		struct page *page = NULL;
-		struct page *free_base_page = NULL;
-		unsigned long free_base_pfn = 0;
-		int nr_to_free = 0;
-
-		end_pfn = min(walk_end, zone_end_pfn(zone));
-		pfn = first_init_pfn;
-		if (pfn < walk_start)
-			pfn = walk_start;
-		if (pfn < zone->zone_start_pfn)
-			pfn = zone->zone_start_pfn;
-
-		for (; pfn < end_pfn; pfn++) {
-			if (!pfn_valid_within(pfn))
-				goto free_range;
-
-			/*
-			 * Ensure pfn_valid is checked every
-			 * pageblock_nr_pages for memory holes
-			 */
-			if ((pfn & (pageblock_nr_pages - 1)) == 0) {
-				if (!pfn_valid(pfn)) {
-					page = NULL;
-					goto free_range;
-				}
-			}
-
-			if (!meminit_pfn_in_nid(pfn, nid, &nid_init_state)) {
-				page = NULL;
-				goto free_range;
-			}
-
-			/* Minimise pfn page lookups and scheduler checks */
-			if (page && (pfn & (pageblock_nr_pages - 1)) != 0) {
-				page++;
-			} else {
-				nr_pages += nr_to_free;
-				deferred_free_range(free_base_page,
-						free_base_pfn, nr_to_free);
-				free_base_page = NULL;
-				free_base_pfn = nr_to_free = 0;
-
-				page = pfn_to_page(pfn);
-				cond_resched();
-			}
-
-			if (page->flags) {
-				VM_BUG_ON(page_zone(page) != zone);
-				goto free_range;
-			}
-
-			__init_single_page(page, pfn, zid, nid);
-			if (!free_base_page) {
-				free_base_page = page;
-				free_base_pfn = pfn;
-				nr_to_free = 0;
-			}
-			nr_to_free++;
-
-			/* Where possible, batch up pages for a single free */
-			continue;
-free_range:
-			/* Free the current block of pages to allocator */
-			nr_pages += nr_to_free;
-			deferred_free_range(free_base_page, free_base_pfn,
-								nr_to_free);
-			free_base_page = NULL;
-			free_base_pfn = nr_to_free = 0;
-		}
-		/* Free the last block of pages to allocator */
-		nr_pages += nr_to_free;
-		deferred_free_range(free_base_page, free_base_pfn, nr_to_free);
-
-		first_init_pfn = max(end_pfn, first_init_pfn);
+	for_each_free_mem_range(i, nid, MEMBLOCK_NONE, &spa, &epa, NULL) {
+		spfn = max_t(unsigned long, first_init_pfn, PFN_UP(spa));
+		epfn = min_t(unsigned long, zone_end_pfn(zone), PFN_DOWN(epa));
+		nr_pages += deferred_init_range(nid, zid, spfn, epfn);
 	}
 
 	/* Sanity check that the next zone really is unpopulated */
-- 
2.14.1

[PATCH v8 04/11] sparc64: simplify vmemmap_populate

[Pavel Tatashin]

Remove duplicating code by using common functions
vmemmap_pud_populate and vmemmap_pgd_populate.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Acked-by: David S. Miller <davem@davemloft.net>
---
 arch/sparc/mm/init_64.c | 23 ++++++-----------------
 1 file changed, 6 insertions(+), 17 deletions(-)

diff --git a/arch/sparc/mm/init_64.c b/arch/sparc/mm/init_64.c
index 078f1352736e..fc47afa518f5 100644
--- a/arch/sparc/mm/init_64.c
+++ b/arch/sparc/mm/init_64.c
@@ -2642,30 +2642,19 @@ int __meminit vmemmap_populate(unsigned long vstart, unsigned long vend,
 	vstart = vstart & PMD_MASK;
 	vend = ALIGN(vend, PMD_SIZE);
 	for (; vstart < vend; vstart += PMD_SIZE) {
-		pgd_t *pgd = pgd_offset_k(vstart);
+		pgd_t *pgd = vmemmap_pgd_populate(vstart, node);
 		unsigned long pte;
 		pud_t *pud;
 		pmd_t *pmd;
 
-		if (pgd_none(*pgd)) {
-			pud_t *new = vmemmap_alloc_block(PAGE_SIZE, node);
+		if (!pgd)
+			return -ENOMEM;
 
-			if (!new)
-				return -ENOMEM;
-			pgd_populate(&init_mm, pgd, new);
-		}
-
-		pud = pud_offset(pgd, vstart);
-		if (pud_none(*pud)) {
-			pmd_t *new = vmemmap_alloc_block(PAGE_SIZE, node);
-
-			if (!new)
-				return -ENOMEM;
-			pud_populate(&init_mm, pud, new);
-		}
+		pud = vmemmap_pud_populate(pgd, vstart, node);
+		if (!pud)
+			return -ENOMEM;
 
 		pmd = pmd_offset(pud, vstart);
-
 		pte = pmd_val(*pmd);
 		if (!(pte & _PAGE_VALID)) {
 			void *block = vmemmap_alloc_block(PMD_SIZE, node);
-- 
2.14.1

[PATCH v8 05/11] mm: defining memblock_virt_alloc_try_nid_raw

[Pavel Tatashin]

* A new variant of memblock_virt_alloc_* allocations:
memblock_virt_alloc_try_nid_raw()
    - Does not zero the allocated memory
    - Does not panic if request cannot be satisfied

* optimize early system hash allocations

Clients can call alloc_large_system_hash() with flag: HASH_ZERO to specify
that memory that was allocated for system hash needs to be zeroed,
otherwise the memory does not need to be zeroed, and client will initialize
it.

If memory does not need to be zero'd, call the new
memblock_virt_alloc_raw() interface, and thus improve the boot performance.

* debug for raw alloctor

When CONFIG_DEBUG_VM is enabled, this patch sets all the memory that is
returned by memblock_virt_alloc_try_nid_raw() to ones to ensure that no
places excpect zeroed memory.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Acked-by: Michal Hocko <mhocko@suse.com>
---
 include/linux/bootmem.h | 27 ++++++++++++++++++++++
 mm/memblock.c           | 60 +++++++++++++++++++++++++++++++++++++++++++------
 mm/page_alloc.c         | 15 ++++++-------
 3 files changed, 87 insertions(+), 15 deletions(-)

diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h
index e223d91b6439..ea30b3987282 100644
--- a/include/linux/bootmem.h
+++ b/include/linux/bootmem.h
@@ -160,6 +160,9 @@ extern void *__alloc_bootmem_low_node(pg_data_t *pgdat,
 #define BOOTMEM_ALLOC_ANYWHERE		(~(phys_addr_t)0)
 
 /* FIXME: Move to memblock.h at a point where we remove nobootmem.c */
+void *memblock_virt_alloc_try_nid_raw(phys_addr_t size, phys_addr_t align,
+				      phys_addr_t min_addr,
+				      phys_addr_t max_addr, int nid);
 void *memblock_virt_alloc_try_nid_nopanic(phys_addr_t size,
 		phys_addr_t align, phys_addr_t min_addr,
 		phys_addr_t max_addr, int nid);
@@ -176,6 +179,14 @@ static inline void * __init memblock_virt_alloc(
 					    NUMA_NO_NODE);
 }
 
+static inline void * __init memblock_virt_alloc_raw(
+					phys_addr_t size,  phys_addr_t align)
+{
+	return memblock_virt_alloc_try_nid_raw(size, align, BOOTMEM_LOW_LIMIT,
+					    BOOTMEM_ALLOC_ACCESSIBLE,
+					    NUMA_NO_NODE);
+}
+
 static inline void * __init memblock_virt_alloc_nopanic(
 					phys_addr_t size, phys_addr_t align)
 {
@@ -257,6 +268,14 @@ static inline void * __init memblock_virt_alloc(
 	return __alloc_bootmem(size, align, BOOTMEM_LOW_LIMIT);
 }
 
+static inline void * __init memblock_virt_alloc_raw(
+					phys_addr_t size,  phys_addr_t align)
+{
+	if (!align)
+		align = SMP_CACHE_BYTES;
+	return __alloc_bootmem_nopanic(size, align, BOOTMEM_LOW_LIMIT);
+}
+
 static inline void * __init memblock_virt_alloc_nopanic(
 					phys_addr_t size, phys_addr_t align)
 {
@@ -309,6 +328,14 @@ static inline void * __init memblock_virt_alloc_try_nid(phys_addr_t size,
 					  min_addr);
 }
 
+static inline void * __init memblock_virt_alloc_try_nid_raw(
+			phys_addr_t size, phys_addr_t align,
+			phys_addr_t min_addr, phys_addr_t max_addr, int nid)
+{
+	return ___alloc_bootmem_node_nopanic(NODE_DATA(nid), size, align,
+				min_addr, max_addr);
+}
+
 static inline void * __init memblock_virt_alloc_try_nid_nopanic(
 			phys_addr_t size, phys_addr_t align,
 			phys_addr_t min_addr, phys_addr_t max_addr, int nid)
diff --git a/mm/memblock.c b/mm/memblock.c
index 91205780e6b1..1f299fb1eb08 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -1327,7 +1327,6 @@ static void * __init memblock_virt_alloc_internal(
 	return NULL;
 done:
 	ptr = phys_to_virt(alloc);
-	memset(ptr, 0, size);
 
 	/*
 	 * The min_count is set to 0 so that bootmem allocated blocks
@@ -1340,6 +1339,45 @@ static void * __init memblock_virt_alloc_internal(
 	return ptr;
 }
 
+/**
+ * memblock_virt_alloc_try_nid_raw - allocate boot memory block without zeroing
+ * memory and without panicking
+ * @size: size of memory block to be allocated in bytes
+ * @align: alignment of the region and block's size
+ * @min_addr: the lower bound of the memory region from where the allocation
+ *	  is preferred (phys address)
+ * @max_addr: the upper bound of the memory region from where the allocation
+ *	      is preferred (phys address), or %BOOTMEM_ALLOC_ACCESSIBLE to
+ *	      allocate only from memory limited by memblock.current_limit value
+ * @nid: nid of the free area to find, %NUMA_NO_NODE for any node
+ *
+ * Public function, provides additional debug information (including caller
+ * info), if enabled. Does not zero allocated memory, does not panic if request
+ * cannot be satisfied.
+ *
+ * RETURNS:
+ * Virtual address of allocated memory block on success, NULL on failure.
+ */
+void * __init memblock_virt_alloc_try_nid_raw(
+			phys_addr_t size, phys_addr_t align,
+			phys_addr_t min_addr, phys_addr_t max_addr,
+			int nid)
+{
+	void *ptr;
+
+	memblock_dbg("%s: %llu bytes align=0x%llx nid=%d from=0x%llx max_addr=0x%llx %pF\n",
+		     __func__, (u64)size, (u64)align, nid, (u64)min_addr,
+		     (u64)max_addr, (void *)_RET_IP_);
+
+	ptr = memblock_virt_alloc_internal(size, align,
+					   min_addr, max_addr, nid);
+#ifdef CONFIG_DEBUG_VM
+	if (ptr && size > 0)
+		memset(ptr, 0xff, size);
+#endif
+	return ptr;
+}
+
 /**
  * memblock_virt_alloc_try_nid_nopanic - allocate boot memory block
  * @size: size of memory block to be allocated in bytes
@@ -1351,8 +1389,8 @@ static void * __init memblock_virt_alloc_internal(
  *	      allocate only from memory limited by memblock.current_limit value
  * @nid: nid of the free area to find, %NUMA_NO_NODE for any node
  *
- * Public version of _memblock_virt_alloc_try_nid_nopanic() which provides
- * additional debug information (including caller info), if enabled.
+ * Public function, provides additional debug information (including caller
+ * info), if enabled. This function zeroes the allocated memory.
  *
  * RETURNS:
  * Virtual address of allocated memory block on success, NULL on failure.
@@ -1362,11 +1400,17 @@ void * __init memblock_virt_alloc_try_nid_nopanic(
 				phys_addr_t min_addr, phys_addr_t max_addr,
 				int nid)
 {
+	void *ptr;
+
 	memblock_dbg("%s: %llu bytes align=0x%llx nid=%d from=0x%llx max_addr=0x%llx %pF\n",
 		     __func__, (u64)size, (u64)align, nid, (u64)min_addr,
 		     (u64)max_addr, (void *)_RET_IP_);
-	return memblock_virt_alloc_internal(size, align, min_addr,
-					     max_addr, nid);
+
+	ptr = memblock_virt_alloc_internal(size, align,
+					   min_addr, max_addr, nid);
+	if (ptr)
+		memset(ptr, 0, size);
+	return ptr;
 }
 
 /**
@@ -1380,7 +1424,7 @@ void * __init memblock_virt_alloc_try_nid_nopanic(
  *	      allocate only from memory limited by memblock.current_limit value
  * @nid: nid of the free area to find, %NUMA_NO_NODE for any node
  *
- * Public panicking version of _memblock_virt_alloc_try_nid_nopanic()
+ * Public panicking version of memblock_virt_alloc_try_nid_nopanic()
  * which provides debug information (including caller info), if enabled,
  * and panics if the request can not be satisfied.
  *
@@ -1399,8 +1443,10 @@ void * __init memblock_virt_alloc_try_nid(
 		     (u64)max_addr, (void *)_RET_IP_);
 	ptr = memblock_virt_alloc_internal(size, align,
 					   min_addr, max_addr, nid);
-	if (ptr)
+	if (ptr) {
+		memset(ptr, 0, size);
 		return ptr;
+	}
 
 	panic("%s: Failed to allocate %llu bytes align=0x%llx nid=%d from=0x%llx max_addr=0x%llx\n",
 	      __func__, (u64)size, (u64)align, nid, (u64)min_addr,
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index d132c801d2c1..a8dbd405ed94 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -7299,18 +7299,17 @@ void *__init alloc_large_system_hash(const char *tablename,
 
 	log2qty = ilog2(numentries);
 
-	/*
-	 * memblock allocator returns zeroed memory already, so HASH_ZERO is
-	 * currently not used when HASH_EARLY is specified.
-	 */
 	gfp_flags = (flags & HASH_ZERO) ? GFP_ATOMIC | __GFP_ZERO : GFP_ATOMIC;
 	do {
 		size = bucketsize << log2qty;
-		if (flags & HASH_EARLY)
-			table = memblock_virt_alloc_nopanic(size, 0);
-		else if (hashdist)
+		if (flags & HASH_EARLY) {
+			if (flags & HASH_ZERO)
+				table = memblock_virt_alloc_nopanic(size, 0);
+			else
+				table = memblock_virt_alloc_raw(size, 0);
+		} else if (hashdist) {
 			table = __vmalloc(size, gfp_flags, PAGE_KERNEL);
-		else {
+		} else {
 			/*
 			 * If bucketsize is not a power-of-two, we may free
 			 * some pages at the end of hash table which
-- 
2.14.1

[PATCH v8 06/11] mm: zero struct pages during initialization

[Pavel Tatashin]

Add struct page zeroing as a part of initialization of other fields in
__init_single_page().

This single thread performance collected on: Intel(R) Xeon(R) CPU E7-8895
v3 @ 2.60GHz with 1T of memory (268400646 pages in 8 nodes):

                        BASE            FIX
sparse_init     11.244671836s   0.007199623s
zone_sizes_init  4.879775891s   8.355182299s
                  --------------------------
Total           16.124447727s   8.362381922s

sparse_init is where memory for struct pages is zeroed, and the zeroing
part is moved later in this patch into __init_single_page(), which is
called from zone_sizes_init().

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Acked-by: Michal Hocko <mhocko@suse.com>
---
 include/linux/mm.h | 9 +++++++++
 mm/page_alloc.c    | 1 +
 2 files changed, 10 insertions(+)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index f8c10d336e42..50b74d628243 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -94,6 +94,15 @@ extern int mmap_rnd_compat_bits __read_mostly;
 #define mm_forbids_zeropage(X)	(0)
 #endif
 
+/*
+ * On some architectures it is expensive to call memset() for small sizes.
+ * Those architectures should provide their own implementation of "struct page"
+ * zeroing by defining this macro in <asm/pgtable.h>.
+ */
+#ifndef mm_zero_struct_page
+#define mm_zero_struct_page(pp)  ((void)memset((pp), 0, sizeof(struct page)))
+#endif
+
 /*
  * Default maximum number of active map areas, this limits the number of vmas
  * per mm struct. Users can overwrite this number by sysctl but there is a
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index a8dbd405ed94..4b630ee91430 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1170,6 +1170,7 @@ static void free_one_page(struct zone *zone,
 static void __meminit __init_single_page(struct page *page, unsigned long pfn,
 				unsigned long zone, int nid)
 {
+	mm_zero_struct_page(page);
 	set_page_links(page, zone, nid, pfn);
 	init_page_count(page);
 	page_mapcount_reset(page);
-- 
2.14.1

[PATCH v8 07/11] sparc64: optimized struct page zeroing

[Pavel Tatashin]

Add an optimized mm_zero_struct_page(), so struct page's are zeroed without
calling memset(). We do eight to ten regular stores based on the size of
struct page. Compiler optimizes out the conditions of switch() statement.

SPARC-M6 with 15T of memory, single thread performance:

                               BASE            FIX  OPTIMIZED_FIX
        bootmem_init   28.440467985s   2.305674818s   2.305161615s
free_area_init_nodes  202.845901673s 225.343084508s 172.556506560s
                      --------------------------------------------
Total                 231.286369658s 227.648759326s 174.861668175s

BASE:  current linux
FIX:   This patch series without "optimized struct page zeroing"
OPTIMIZED_FIX: This patch series including the current patch.

bootmem_init() is where memory for struct pages is zeroed during
allocation. Note, about two seconds in this function is a fixed time: it
does not increase as memory is increased.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Acked-by: David S. Miller <davem@davemloft.net>
---
 arch/sparc/include/asm/pgtable_64.h | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

diff --git a/arch/sparc/include/asm/pgtable_64.h b/arch/sparc/include/asm/pgtable_64.h
index 4fefe3762083..8ed478abc630 100644
--- a/arch/sparc/include/asm/pgtable_64.h
+++ b/arch/sparc/include/asm/pgtable_64.h
@@ -230,6 +230,36 @@ extern unsigned long _PAGE_ALL_SZ_BITS;
 extern struct page *mem_map_zero;
 #define ZERO_PAGE(vaddr)	(mem_map_zero)
 
+/* This macro must be updated when the size of struct page grows above 80
+ * or reduces below 64.
+ * The idea that compiler optimizes out switch() statement, and only
+ * leaves clrx instructions
+ */
+#define	mm_zero_struct_page(pp) do {					\
+	unsigned long *_pp = (void *)(pp);				\
+									\
+	 /* Check that struct page is either 64, 72, or 80 bytes */	\
+	BUILD_BUG_ON(sizeof(struct page) & 7);				\
+	BUILD_BUG_ON(sizeof(struct page) < 64);				\
+	BUILD_BUG_ON(sizeof(struct page) > 80);				\
+									\
+	switch (sizeof(struct page)) {					\
+	case 80:							\
+		_pp[9] = 0;	/* fallthrough */			\
+	case 72:							\
+		_pp[8] = 0;	/* fallthrough */			\
+	default:							\
+		_pp[7] = 0;						\
+		_pp[6] = 0;						\
+		_pp[5] = 0;						\
+		_pp[4] = 0;						\
+		_pp[3] = 0;						\
+		_pp[2] = 0;						\
+		_pp[1] = 0;						\
+		_pp[0] = 0;						\
+	}								\
+} while (0)
+
 /* PFNs are real physical page numbers.  However, mem_map only begins to record
  * per-page information starting at pfn_base.  This is to handle systems where
  * the first physical page in the machine is at some huge physical address,
-- 
2.14.1

[PATCH v8 08/11] mm: zero reserved and unavailable struct pages

[Pavel Tatashin]

Some memory is reserved but unavailable: not present in memblock.memory
(because not backed by physical pages), but present in memblock.reserved.
Such memory has backing struct pages, but they are not initialized by going
through __init_single_page().

In some cases these struct pages are accessed even if they do not contain
any data. One example is page_to_pfn() might access page->flags if this is
where section information is stored (CONFIG_SPARSEMEM,
SECTION_IN_PAGE_FLAGS).

Since, struct pages are zeroed in __init_single_page(), and not during
allocation time, we must zero such struct pages explicitly.

The patch involves adding a new memblock iterator:
	for_each_resv_unavail_range(i, p_start, p_end)

Which iterates through reserved && !memory lists, and we zero struct pages
explicitly by calling mm_zero_struct_page().

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 include/linux/memblock.h | 16 ++++++++++++++++
 include/linux/mm.h       |  6 ++++++
 mm/page_alloc.c          | 30 ++++++++++++++++++++++++++++++
 3 files changed, 52 insertions(+)

diff --git a/include/linux/memblock.h b/include/linux/memblock.h
index bae11c7e7bf3..bdd4268f9323 100644
--- a/include/linux/memblock.h
+++ b/include/linux/memblock.h
@@ -237,6 +237,22 @@ unsigned long memblock_next_valid_pfn(unsigned long pfn, unsigned long max_pfn);
 	for_each_mem_range_rev(i, &memblock.memory, &memblock.reserved,	\
 			       nid, flags, p_start, p_end, p_nid)
 
+/**
+ * for_each_resv_unavail_range - iterate through reserved and unavailable memory
+ * @i: u64 used as loop variable
+ * @flags: pick from blocks based on memory attributes
+ * @p_start: ptr to phys_addr_t for start address of the range, can be %NULL
+ * @p_end: ptr to phys_addr_t for end address of the range, can be %NULL
+ *
+ * Walks over unavailabled but reserved (reserved && !memory) areas of memblock.
+ * Available as soon as memblock is initialized.
+ * Note: because this memory does not belong to any physical node, flags and
+ * nid arguments do not make sense and thus not exported as arguments.
+ */
+#define for_each_resv_unavail_range(i, p_start, p_end)			\
+	for_each_mem_range(i, &memblock.reserved, &memblock.memory,	\
+			   NUMA_NO_NODE, MEMBLOCK_NONE, p_start, p_end, NULL)
+
 static inline void memblock_set_region_flags(struct memblock_region *r,
 					     unsigned long flags)
 {
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 50b74d628243..a7bba4ce79ba 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2010,6 +2010,12 @@ extern int __meminit __early_pfn_to_nid(unsigned long pfn,
 					struct mminit_pfnnid_cache *state);
 #endif
 
+#ifdef CONFIG_HAVE_MEMBLOCK
+void zero_resv_unavail(void);
+#else
+static inline void zero_resv_unavail(void) {}
+#endif
+
 extern void set_dma_reserve(unsigned long new_dma_reserve);
 extern void memmap_init_zone(unsigned long, int, unsigned long,
 				unsigned long, enum memmap_context);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4b630ee91430..1d38d391dffd 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6202,6 +6202,34 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
 	free_area_init_core(pgdat);
 }
 
+#ifdef CONFIG_HAVE_MEMBLOCK
+/*
+ * Only struct pages that are backed by physical memory are zeroed and
+ * initialized by going through __init_single_page(). But, there are some
+ * struct pages which are reserved in memblock allocator and their fields
+ * may be accessed (for example page_to_pfn() on some configuration accesses
+ * flags). We must explicitly zero those struct pages.
+ */
+void __paginginit zero_resv_unavail(void)
+{
+	phys_addr_t start, end;
+	unsigned long pfn;
+	u64 i, pgcnt;
+
+	/* Loop through ranges that are reserved, but do not have reported
+	 * physical memory backing.
+	 */
+	pgcnt = 0;
+	for_each_resv_unavail_range(i, &start, &end) {
+		for (pfn = PFN_DOWN(start); pfn < PFN_UP(end); pfn++) {
+			mm_zero_struct_page(pfn_to_page(pfn));
+			pgcnt++;
+		}
+	}
+	pr_info("Reserved but unavailable: %lld pages", pgcnt);
+}
+#endif /* CONFIG_HAVE_MEMBLOCK */
+
 #ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
 
 #if MAX_NUMNODES > 1
@@ -6625,6 +6653,7 @@ void __init free_area_init_nodes(unsigned long *max_zone_pfn)
 			node_set_state(nid, N_MEMORY);
 		check_for_memory(pgdat, nid);
 	}
+	zero_resv_unavail();
 }
 
 static int __init cmdline_parse_core(char *p, unsigned long *core)
@@ -6788,6 +6817,7 @@ void __init free_area_init(unsigned long *zones_size)
 {
 	free_area_init_node(0, zones_size,
 			__pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL);
+	zero_resv_unavail();
 }
 
 static int page_alloc_cpu_dead(unsigned int cpu)
-- 
2.14.1

[PATCH v8 09/11] x86/kasan: explicitly zero kasan shadow memory

[Pavel Tatashin]

To optimize the performance of struct page initialization,
vmemmap_populate() will no longer zero memory.

We must explicitly zero the memory that is allocated by vmemmap_populate()
for kasan, as this memory does not go through struct page initialization
path.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 arch/x86/mm/kasan_init_64.c | 66 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 66 insertions(+)

diff --git a/arch/x86/mm/kasan_init_64.c b/arch/x86/mm/kasan_init_64.c
index bc84b73684b7..cc0399032673 100644
--- a/arch/x86/mm/kasan_init_64.c
+++ b/arch/x86/mm/kasan_init_64.c
@@ -84,6 +84,66 @@ static struct notifier_block kasan_die_notifier = {
 };
 #endif
 
+/*
+ * x86 variant of vmemmap_populate() uses either PMD_SIZE pages or base pages
+ * to map allocated memory.  This routine determines the page size for the given
+ * address from vmemmap.
+ */
+static u64 get_vmemmap_pgsz(u64 addr)
+{
+	pgd_t *pgd;
+	p4d_t *p4d;
+	pud_t *pud;
+	pmd_t *pmd;
+
+	pgd = pgd_offset_k(addr);
+	BUG_ON(pgd_none(*pgd) || pgd_large(*pgd));
+
+	p4d = p4d_offset(pgd, addr);
+	BUG_ON(p4d_none(*p4d) || p4d_large(*p4d));
+
+	pud = pud_offset(p4d, addr);
+	BUG_ON(pud_none(*pud) || pud_large(*pud));
+
+	pmd = pmd_offset(pud, addr);
+	BUG_ON(pmd_none(*pmd));
+
+	if (pmd_large(*pmd))
+		return PMD_SIZE;
+	return PAGE_SIZE;
+}
+
+/*
+ * Memory that was allocated by vmemmap_populate is not zeroed, so we must
+ * zero it here explicitly.
+ */
+static void
+zero_vmemmap_populated_memory(void)
+{
+	u64 i, start, end;
+
+	for (i = 0; i < E820_MAX_ENTRIES && pfn_mapped[i].end; i++) {
+		void *kaddr_start = pfn_to_kaddr(pfn_mapped[i].start);
+		void *kaddr_end = pfn_to_kaddr(pfn_mapped[i].end);
+
+		start = (u64)kasan_mem_to_shadow(kaddr_start);
+		end = (u64)kasan_mem_to_shadow(kaddr_end);
+
+		/* Round to the start end of the mapped pages */
+		start = rounddown(start, get_vmemmap_pgsz(start));
+		end = roundup(end, get_vmemmap_pgsz(start));
+		memset((void *)start, 0, end - start);
+	}
+
+	start = (u64)kasan_mem_to_shadow(_stext);
+	end = (u64)kasan_mem_to_shadow(_end);
+
+	/* Round to the start end of the mapped pages */
+	start = rounddown(start, get_vmemmap_pgsz(start));
+	end = roundup(end, get_vmemmap_pgsz(start));
+	memset((void *)start, 0, end - start);
+}
+
 void __init kasan_early_init(void)
 {
 	int i;
@@ -146,6 +206,12 @@ void __init kasan_init(void)
 	load_cr3(init_top_pgt);
 	__flush_tlb_all();
 
+	/*
+	 * vmemmap_populate does not zero the memory, so we need to zero it
+	 * explicitly
+	 */
+	zero_vmemmap_populated_memory();
+
 	/*
 	 * kasan_zero_page has been used as early shadow memory, thus it may
 	 * contain some garbage. Now we can clear and write protect it, since
-- 
2.14.1

[PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Pavel Tatashin]

To optimize the performance of struct page initialization,
vmemmap_populate() will no longer zero memory.

We must explicitly zero the memory that is allocated by vmemmap_populate()
for kasan, as this memory does not go through struct page initialization
path.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 arch/arm64/mm/kasan_init.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

diff --git a/arch/arm64/mm/kasan_init.c b/arch/arm64/mm/kasan_init.c
index 81f03959a4ab..e78a9ecbb687 100644
--- a/arch/arm64/mm/kasan_init.c
+++ b/arch/arm64/mm/kasan_init.c
@@ -135,6 +135,41 @@ static void __init clear_pgds(unsigned long start,
 		set_pgd(pgd_offset_k(start), __pgd(0));
 }
 
+/*
+ * Memory that was allocated by vmemmap_populate is not zeroed, so we must
+ * zero it here explicitly.
+ */
+static void
+zero_vmemmap_populated_memory(void)
+{
+	struct memblock_region *reg;
+	u64 start, end;
+
+	for_each_memblock(memory, reg) {
+		start = __phys_to_virt(reg->base);
+		end = __phys_to_virt(reg->base + reg->size);
+
+		if (start >= end)
+			break;
+
+		start = (u64)kasan_mem_to_shadow((void *)start);
+		end = (u64)kasan_mem_to_shadow((void *)end);
+
+		/* Round to the start end of the mapped pages */
+		start = round_down(start, SWAPPER_BLOCK_SIZE);
+		end = round_up(end, SWAPPER_BLOCK_SIZE);
+		memset((void *)start, 0, end - start);
+	}
+
+	start = (u64)kasan_mem_to_shadow(_text);
+	end = (u64)kasan_mem_to_shadow(_end);
+
+	/* Round to the start end of the mapped pages */
+	start = round_down(start, SWAPPER_BLOCK_SIZE);
+	end = round_up(end, SWAPPER_BLOCK_SIZE);
+	memset((void *)start, 0, end - start);
+}
+
 void __init kasan_init(void)
 {
 	u64 kimg_shadow_start, kimg_shadow_end;
@@ -205,8 +240,15 @@ void __init kasan_init(void)
 			pfn_pte(sym_to_pfn(kasan_zero_page), PAGE_KERNEL_RO));
 
 	memset(kasan_zero_page, 0, PAGE_SIZE);
+
 	cpu_replace_ttbr1(lm_alias(swapper_pg_dir));
 
+	/*
+	 * vmemmap_populate does not zero the memory, so we need to zero it
+	 * explicitly
+	 */
+	zero_vmemmap_populated_memory();
+
 	/* At this point kasan is fully initialized. Enable error messages */
 	init_task.kasan_depth = 0;
 	pr_info("KernelAddressSanitizer initialized\n");
-- 
2.14.1

[PATCH v8 11/11] mm: stop zeroing memory during allocation in vmemmap

[Pavel Tatashin]

vmemmap_alloc_block() will no longer zero the block, so zero memory
at its call sites for everything except struct pages.  Struct page memory
is zero'd by struct page initialization.

Replace allocators in sprase-vmemmap to use the non-zeroing version. So,
we will get the performance improvement by zeroing the memory in parallel
when struct pages are zeroed.

Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
---
 include/linux/mm.h  | 11 +++++++++++
 mm/sparse-vmemmap.c | 15 +++++++--------
 mm/sparse.c         |  6 +++---
 3 files changed, 21 insertions(+), 11 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index a7bba4ce79ba..25848764570f 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2501,6 +2501,17 @@ static inline void *vmemmap_alloc_block_buf(unsigned long size, int node)
 	return __vmemmap_alloc_block_buf(size, node, NULL);
 }
 
+static inline void *vmemmap_alloc_block_zero(unsigned long size, int node)
+{
+	void *p = vmemmap_alloc_block(size, node);
+
+	if (!p)
+		return NULL;
+	memset(p, 0, size);
+
+	return p;
+}
+
 void vmemmap_verify(pte_t *, int, unsigned long, unsigned long);
 int vmemmap_populate_basepages(unsigned long start, unsigned long end,
 			       int node);
diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c
index d1a39b8051e0..c2f5654e7c9d 100644
--- a/mm/sparse-vmemmap.c
+++ b/mm/sparse-vmemmap.c
@@ -41,7 +41,7 @@ static void * __ref __earlyonly_bootmem_alloc(int node,
 				unsigned long align,
 				unsigned long goal)
 {
-	return memblock_virt_alloc_try_nid(size, align, goal,
+	return memblock_virt_alloc_try_nid_raw(size, align, goal,
 					    BOOTMEM_ALLOC_ACCESSIBLE, node);
 }
 
@@ -54,9 +54,8 @@ void * __meminit vmemmap_alloc_block(unsigned long size, int node)
 	if (slab_is_available()) {
 		struct page *page;
 
-		page = alloc_pages_node(node,
-			GFP_KERNEL | __GFP_ZERO | __GFP_RETRY_MAYFAIL,
-			get_order(size));
+		page = alloc_pages_node(node, GFP_KERNEL | __GFP_RETRY_MAYFAIL,
+					get_order(size));
 		if (page)
 			return page_address(page);
 		return NULL;
@@ -183,7 +182,7 @@ pmd_t * __meminit vmemmap_pmd_populate(pud_t *pud, unsigned long addr, int node)
 {
 	pmd_t *pmd = pmd_offset(pud, addr);
 	if (pmd_none(*pmd)) {
-		void *p = vmemmap_alloc_block(PAGE_SIZE, node);
+		void *p = vmemmap_alloc_block_zero(PAGE_SIZE, node);
 		if (!p)
 			return NULL;
 		pmd_populate_kernel(&init_mm, pmd, p);
@@ -195,7 +194,7 @@ pud_t * __meminit vmemmap_pud_populate(p4d_t *p4d, unsigned long addr, int node)
 {
 	pud_t *pud = pud_offset(p4d, addr);
 	if (pud_none(*pud)) {
-		void *p = vmemmap_alloc_block(PAGE_SIZE, node);
+		void *p = vmemmap_alloc_block_zero(PAGE_SIZE, node);
 		if (!p)
 			return NULL;
 		pud_populate(&init_mm, pud, p);
@@ -207,7 +206,7 @@ p4d_t * __meminit vmemmap_p4d_populate(pgd_t *pgd, unsigned long addr, int node)
 {
 	p4d_t *p4d = p4d_offset(pgd, addr);
 	if (p4d_none(*p4d)) {
-		void *p = vmemmap_alloc_block(PAGE_SIZE, node);
+		void *p = vmemmap_alloc_block_zero(PAGE_SIZE, node);
 		if (!p)
 			return NULL;
 		p4d_populate(&init_mm, p4d, p);
@@ -219,7 +218,7 @@ pgd_t * __meminit vmemmap_pgd_populate(unsigned long addr, int node)
 {
 	pgd_t *pgd = pgd_offset_k(addr);
 	if (pgd_none(*pgd)) {
-		void *p = vmemmap_alloc_block(PAGE_SIZE, node);
+		void *p = vmemmap_alloc_block_zero(PAGE_SIZE, node);
 		if (!p)
 			return NULL;
 		pgd_populate(&init_mm, pgd, p);
diff --git a/mm/sparse.c b/mm/sparse.c
index 83b3bf6461af..d22f51bb7c79 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -437,9 +437,9 @@ void __init sparse_mem_maps_populate_node(struct page **map_map,
 	}
 
 	size = PAGE_ALIGN(size);
-	map = memblock_virt_alloc_try_nid(size * map_count,
-					  PAGE_SIZE, __pa(MAX_DMA_ADDRESS),
-					  BOOTMEM_ALLOC_ACCESSIBLE, nodeid);
+	map = memblock_virt_alloc_try_nid_raw(size * map_count,
+					      PAGE_SIZE, __pa(MAX_DMA_ADDRESS),
+					      BOOTMEM_ALLOC_ACCESSIBLE, nodeid);
 	if (map) {
 		for (pnum = pnum_begin; pnum < pnum_end; pnum++) {
 			if (!present_section_nr(pnum))
-- 
2.14.1

[PATCH v8 00/11] complete deferred page initialization

[Pavel Tatashin]

Copy paste error, changing the subject for the header to v8 from v7.

On 09/14/2017 06:35 PM, Pavel Tatashin wrote:
> Changelog:
> v8 - v7
> - Added Acked-by's from Dave Miller for SPARC changes
> - Fixed a minor compiling issue on tile architecture reported by kbuild
>
> v7 - v6
> - Addressed comments from Michal Hocko
> - memblock_discard() patch was removed from this series and integrated
>    separately
> - Fixed bug reported by kbuild test robot new patch:
>    mm: zero reserved and unavailable struct pages
> - Removed patch
>    x86/mm: reserve only exiting low pages
>    As, it is not needed anymore, because of the previous fix
> - Re-wrote deferred_init_memmap(), found and fixed an existing bug, where
>    page variable is not reset when zone holes present.
> - Merged several patches together per Michal request
> - Added performance data including raw logs
>
> v6 - v5
> - Fixed ARM64 + kasan code, as reported by Ard Biesheuvel
> - Tested ARM64 code in qemu and found few more issues, that I fixed in this
>    iteration
> - Added page roundup/rounddown to x86 and arm zeroing routines to zero the
>    whole allocated range, instead of only provided address range.
> - Addressed SPARC related comment from Sam Ravnborg
> - Fixed section mismatch warnings related to memblock_discard().
>
> v5 - v4
> - Fixed build issues reported by kbuild on various configurations
>
> v4 - v3
> - Rewrote code to zero sturct pages in __init_single_page() as
>    suggested by Michal Hocko
> - Added code to handle issues related to accessing struct page
>    memory before they are initialized.
>
> v3 - v2
> - Addressed David Miller comments about one change per patch:
>      * Splited changes to platforms into 4 patches
>      * Made "do not zero vmemmap_buf" as a separate patch
>
> v2 - v1
> - Per request, added s390 to deferred "struct page" zeroing
> - Collected performance data on x86 which proofs the importance to
>    keep memset() as prefetch (see below).
>
> SMP machines can benefit from the DEFERRED_STRUCT_PAGE_INIT config option,
> which defers initializing struct pages until all cpus have been started so
> it can be done in parallel.
>
> However, this feature is sub-optimal, because the deferred page
> initialization code expects that the struct pages have already been zeroed,
> and the zeroing is done early in boot with a single thread only.  Also, we
> access that memory and set flags before struct pages are initialized. All
> of this is fixed in this patchset.
>
> In this work we do the following:
> - Never read access struct page until it was initialized
> - Never set any fields in struct pages before they are initialized
> - Zero struct page at the beginning of struct page initialization
>
>
> ==========================================================================
> Performance improvements on x86 machine with 8 nodes:
> Intel(R) Xeon(R) CPU E7-8895 v3 @ 2.60GHz and 1T of memory:
>                          TIME          SPEED UP
> base no deferred:       95.796233s
> fix no deferred:        79.978956s    19.77%
>
> base deferred:          77.254713s
> fix deferred:           55.050509s    40.34%
> ==========================================================================
> SPARC M6 3600 MHz with 15T of memory
>                          TIME          SPEED UP
> base no deferred:       358.335727s
> fix no deferred:        302.320936s   18.52%
>
> base deferred:          237.534603s
> fix deferred:           182.103003s   30.44%
> ==========================================================================
> Raw dmesg output with timestamps:
> x86 base no deferred:    https://hastebin.com/ofunepurit.scala
> x86 base deferred:       https://hastebin.com/ifazegeyas.scala
> x86 fix no deferred:     https://hastebin.com/pegocohevo.scala
> x86 fix deferred:        https://hastebin.com/ofupevikuk.scala
> sparc base no deferred:  https://hastebin.com/ibobeteken.go
> sparc base deferred:     https://hastebin.com/fariqimiyu.go
> sparc fix no deferred:   https://hastebin.com/muhegoheyi.go
> sparc fix deferred:      https://hastebin.com/xadinobutu.go
>
> Pavel Tatashin (11):
>    x86/mm: setting fields in deferred pages
>    sparc64/mm: setting fields in deferred pages
>    mm: deferred_init_memmap improvements
>    sparc64: simplify vmemmap_populate
>    mm: defining memblock_virt_alloc_try_nid_raw
>    mm: zero struct pages during initialization
>    sparc64: optimized struct page zeroing
>    mm: zero reserved and unavailable struct pages
>    x86/kasan: explicitly zero kasan shadow memory
>    arm64/kasan: explicitly zero kasan shadow memory
>    mm: stop zeroing memory during allocation in vmemmap
>
>   arch/arm64/mm/kasan_init.c          |  42 ++++++++
>   arch/sparc/include/asm/pgtable_64.h |  30 ++++++
>   arch/sparc/mm/init_64.c             |  31 +++---
>   arch/x86/mm/init_64.c               |   9 +-
>   arch/x86/mm/kasan_init_64.c         |  66 ++++++++++++
>   include/linux/bootmem.h             |  27 +++++
>   include/linux/memblock.h            |  16 +++
>   include/linux/mm.h                  |  26 +++++
>   mm/memblock.c                       |  60 +++++++++--
>   mm/page_alloc.c                     | 207 ++++++++++++++++++++----------------
>   mm/sparse-vmemmap.c                 |  15 ++-
>   mm/sparse.c                         |   6 +-
>   12 files changed, 406 insertions(+), 129 deletions(-)
>

Re: [PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Mark Rutland]

On Thu, Sep 14, 2017 at 06:35:16PM -0400, Pavel Tatashin wrote:
> To optimize the performance of struct page initialization,
> vmemmap_populate() will no longer zero memory.
> 
> We must explicitly zero the memory that is allocated by vmemmap_populate()
> for kasan, as this memory does not go through struct page initialization
> path.
> 
> Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
> Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
> Reviewed-by: Daniel Jordan <daniel.m.jordan@oracle.com>
> Reviewed-by: Bob Picco <bob.picco@oracle.com>
> ---
>  arch/arm64/mm/kasan_init.c | 42 ++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 42 insertions(+)
> 
> diff --git a/arch/arm64/mm/kasan_init.c b/arch/arm64/mm/kasan_init.c
> index 81f03959a4ab..e78a9ecbb687 100644
> --- a/arch/arm64/mm/kasan_init.c
> +++ b/arch/arm64/mm/kasan_init.c
> @@ -135,6 +135,41 @@ static void __init clear_pgds(unsigned long start,
>  		set_pgd(pgd_offset_k(start), __pgd(0));
>  }
>  
> +/*
> + * Memory that was allocated by vmemmap_populate is not zeroed, so we must
> + * zero it here explicitly.
> + */
> +static void
> +zero_vmemmap_populated_memory(void)
> +{
> +	struct memblock_region *reg;
> +	u64 start, end;
> +
> +	for_each_memblock(memory, reg) {
> +		start = __phys_to_virt(reg->base);
> +		end = __phys_to_virt(reg->base + reg->size);
> +
> +		if (start >= end)
> +			break;
> +
> +		start = (u64)kasan_mem_to_shadow((void *)start);
> +		end = (u64)kasan_mem_to_shadow((void *)end);
> +
> +		/* Round to the start end of the mapped pages */
> +		start = round_down(start, SWAPPER_BLOCK_SIZE);
> +		end = round_up(end, SWAPPER_BLOCK_SIZE);
> +		memset((void *)start, 0, end - start);
> +	}
> +
> +	start = (u64)kasan_mem_to_shadow(_text);
> +	end = (u64)kasan_mem_to_shadow(_end);
> +
> +	/* Round to the start end of the mapped pages */
> +	start = round_down(start, SWAPPER_BLOCK_SIZE);
> +	end = round_up(end, SWAPPER_BLOCK_SIZE);
> +	memset((void *)start, 0, end - start);
> +}

I really don't see the need to duplicate the existing logic to iterate over
memblocks, calculate the addresses, etc.

Why can't we just have a zeroing wrapper? e.g. something like the below.

I really don't see why we couldn't have a generic function in core code to do
this, even if vmemmap_populate() doesn't.

Thanks,
Mark.

---->8----
diff --git a/arch/arm64/mm/kasan_init.c b/arch/arm64/mm/kasan_init.c
index 81f0395..698d065 100644
--- a/arch/arm64/mm/kasan_init.c
+++ b/arch/arm64/mm/kasan_init.c
@@ -135,6 +135,17 @@ static void __init clear_pgds(unsigned long start,
                set_pgd(pgd_offset_k(start), __pgd(0));
 }
 
+void kasan_populate_shadow(unsigned long shadow_start, unsigned long shadow_end,
+                          nid_t nid)
+{
+       shadow_start = round_down(shadow_start, SWAPPER_BLOCK_SIZE);
+       shadow_end = round_up(shadow_end, SWAPPER_BLOCK_SIZE);
+
+       vmemmap_populate(shadow_start, shadow_end, nid);
+
+       memset((void *)shadow_start, 0, shadow_end - shadow_start);
+}
+
 void __init kasan_init(void)
 {
        u64 kimg_shadow_start, kimg_shadow_end;
@@ -161,8 +172,8 @@ void __init kasan_init(void)
 
        clear_pgds(KASAN_SHADOW_START, KASAN_SHADOW_END);
 
-       vmemmap_populate(kimg_shadow_start, kimg_shadow_end,
-                        pfn_to_nid(virt_to_pfn(lm_alias(_text))));
+       kasah_populate_shadow(kimg_shadow_start, kimg_shadow_end,
+                             pfn_to_nid(virt_to_pfn(lm_alias(_text))));
 
        /*
         * vmemmap_populate() has populated the shadow region that covers the
@@ -191,9 +202,9 @@ void __init kasan_init(void)
                if (start >= end)
                        break;
 
-               vmemmap_populate((unsigned long)kasan_mem_to_shadow(start),
-                               (unsigned long)kasan_mem_to_shadow(end),
-                               pfn_to_nid(virt_to_pfn(start)));
+               kasan_populate_shadow((unsigned long)kasan_mem_to_shadow(start),
+                                     (unsigned long)kasan_mem_to_shadow(end),
+                                     pfn_to_nid(virt_to_pfn(start)));
        }
 
        /*

Re: [PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Pavel Tatashin]

Hi Mark,

Thank you for looking at this. We can't do this because page table is 
not set until cpu_replace_ttbr1() is called. So, we can't do memset() on 
this memory until then.

Pasha

Re: [PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Mark Rutland]

On Thu, Sep 14, 2017 at 09:30:28PM -0400, Pavel Tatashin wrote:
> Hi Mark,
> 
> Thank you for looking at this. We can't do this because page table is not
> set until cpu_replace_ttbr1() is called. So, we can't do memset() on this
> memory until then.

I see. Sorry, I had missed that we were on the temporary tables at this point
in time.

I'm still not keen on duplicating the iteration. Can we split the vmemmap code
so that we have a variant that takes a GFP? 

That way we could explicitly pass __GFP_ZERO for those cases where we want a
zeroed page, and are happy to pay the cost of initialization.

Thanks
Mark.

Re: [PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Pavel Tatashin]

Hi Mark,

I had this option  back upto version 3, where zero flag was passed into 
vmemmap_alloc_block(), but I was asked to remove it, because it required 
too many changes in other places. So, the current approach is cleaner, 
but the idea is that kasan should use its own version of 
vmemmap_populate() for both x86 and ARM, but I think it is outside of 
the scope of this work.

See this comment from Ard Biesheuvel:
https://lkml.org/lkml/2017/8/3/948

"
KASAN uses vmemmap_populate as a convenience: kasan has nothing to do
with vmemmap, but the function already existed and happened to do what
KASAN requires.

Given that that will no longer be the case, it would be far better to
stop using vmemmap_populate altogether, and clone it into a KASAN
specific version (with an appropriate name) with the zeroing folded
into it.
"

If you think I should add these function in this project, than sure I 
can send a new version with kasanmap_populate() functions.

Thank you,
Pasha

On 09/15/2017 04:38 PM, Mark Rutland wrote:
> On Thu, Sep 14, 2017 at 09:30:28PM -0400, Pavel Tatashin wrote:
>> Hi Mark, Thank you for looking at this. We can't do this because page 
>> table is not set until cpu_replace_ttbr1() is called. So, we can't do 
>> memset() on this memory until then. 
> I see. Sorry, I had missed that we were on the temporary tables at 
> this point in time. I'm still not keen on duplicating the iteration. 
> Can we split the vmemmap code so that we have a variant that takes a 
> GFP? That way we could explicitly pass __GFP_ZERO for those cases 
> where we want a zeroed page, and are happy to pay the cost of 
> initialization. Thanks Mark.

Re: [PATCH v8 10/11] arm64/kasan: explicitly zero kasan shadow memory

[Mark Rutland]

On Fri, Sep 15, 2017 at 05:20:59PM -0400, Pavel Tatashin wrote:
> Hi Mark,
> 
> I had this option  back upto version 3, where zero flag was passed into
> vmemmap_alloc_block(), but I was asked to remove it, because it required too
> many changes in other places.

Ok. Sorry for bringing back a point that had already been covered.

> So, the current approach is cleaner, but the idea is that kasan should use
> its own version of vmemmap_populate() for both x86 and ARM, but I think it is
> outside of the scope of this work.

I appreciate that this is unrelated to your ultimate goal, and that this is
somewhat frustrating given the KASAN code is arguably abusing the
vmemmap_populate() interface.

However, I do think we need to migrate the KASAN code to a proper interface
immediately, rather than making it worse in the interim.

> If you think I should add these function in this project, than sure I can
> send a new version with kasanmap_populate() functions.

I would very much appreciate if you could send a version with a
kasan_map_populate() interface. I'm more than happy to review/test that portion
of the series, or to help if there's some problem which makes that difficult.

Thanks,
Mark.