x86, mm: introduce vmem_altmap to augment vmemmap_populate()

In support of providing struct page for large persistent memory
capacities, use struct vmem_altmap to change the default policy for
allocating memory for the memmap array.  The default vmemmap_populate()
allocates page table storage area from the page allocator.  Given
persistent memory capacities relative to DRAM it may not be feasible to
store the memmap in 'System Memory'.  Instead vmem_altmap represents
pre-allocated "device pages" to satisfy vmemmap_alloc_block_buf()
requests.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: kbuild test robot <lkp@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 92f9595..4af58a3 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -17,6 +17,7 @@
 #include <linux/sysctl.h>
 #include <linux/cpu.h>
 #include <linux/memory.h>
+#include <linux/memremap.h>
 #include <linux/memory_hotplug.h>
 #include <linux/highmem.h>
 #include <linux/vmalloc.h>
@@ -506,10 +507,25 @@
 	unsigned long i;
 	int err = 0;
 	int start_sec, end_sec;
+	struct vmem_altmap *altmap;
+
 	/* during initialize mem_map, align hot-added range to section */
 	start_sec = pfn_to_section_nr(phys_start_pfn);
 	end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
 
+	altmap = to_vmem_altmap((unsigned long) pfn_to_page(phys_start_pfn));
+	if (altmap) {
+		/*
+		 * Validate altmap is within bounds of the total request
+		 */
+		if (altmap->base_pfn != phys_start_pfn
+				|| vmem_altmap_offset(altmap) > nr_pages) {
+			pr_warn_once("memory add fail, invalid altmap\n");
+			return -EINVAL;
+		}
+		altmap->alloc = 0;
+	}
+
 	for (i = start_sec; i <= end_sec; i++) {
 		err = __add_section(nid, zone, section_nr_to_pfn(i));
 
@@ -731,7 +747,8 @@
 	pgdat_resize_unlock(zone->zone_pgdat, &flags);
 }
 
-static int __remove_section(struct zone *zone, struct mem_section *ms)
+static int __remove_section(struct zone *zone, struct mem_section *ms,
+		unsigned long map_offset)
 {
 	unsigned long start_pfn;
 	int scn_nr;
@@ -748,7 +765,7 @@
 	start_pfn = section_nr_to_pfn(scn_nr);
 	__remove_zone(zone, start_pfn);
 
-	sparse_remove_one_section(zone, ms);
+	sparse_remove_one_section(zone, ms, map_offset);
 	return 0;
 }
 
@@ -767,9 +784,32 @@
 		 unsigned long nr_pages)
 {
 	unsigned long i;
-	int sections_to_remove;
-	resource_size_t start, size;
-	int ret = 0;
+	unsigned long map_offset = 0;
+	int sections_to_remove, ret = 0;
+
+	/* In the ZONE_DEVICE case device driver owns the memory region */
+	if (is_dev_zone(zone)) {
+		struct page *page = pfn_to_page(phys_start_pfn);
+		struct vmem_altmap *altmap;
+
+		altmap = to_vmem_altmap((unsigned long) page);
+		if (altmap)
+			map_offset = vmem_altmap_offset(altmap);
+	} else {
+		resource_size_t start, size;
+
+		start = phys_start_pfn << PAGE_SHIFT;
+		size = nr_pages * PAGE_SIZE;
+
+		ret = release_mem_region_adjustable(&iomem_resource, start,
+					size);
+		if (ret) {
+			resource_size_t endres = start + size - 1;
+
+			pr_warn("Unable to release resource <%pa-%pa> (%d)\n",
+					&start, &endres, ret);
+		}
+	}
 
 	/*
 	 * We can only remove entire sections
@@ -777,23 +817,12 @@
 	BUG_ON(phys_start_pfn & ~PAGE_SECTION_MASK);
 	BUG_ON(nr_pages % PAGES_PER_SECTION);
 
-	start = phys_start_pfn << PAGE_SHIFT;
-	size = nr_pages * PAGE_SIZE;
-
-	/* in the ZONE_DEVICE case device driver owns the memory region */
-	if (!is_dev_zone(zone))
-		ret = release_mem_region_adjustable(&iomem_resource, start, size);
-	if (ret) {
-		resource_size_t endres = start + size - 1;
-
-		pr_warn("Unable to release resource <%pa-%pa> (%d)\n",
-				&start, &endres, ret);
-	}
-
 	sections_to_remove = nr_pages / PAGES_PER_SECTION;
 	for (i = 0; i < sections_to_remove; i++) {
 		unsigned long pfn = phys_start_pfn + i*PAGES_PER_SECTION;
-		ret = __remove_section(zone, __pfn_to_section(pfn));
+
+		ret = __remove_section(zone, __pfn_to_section(pfn), map_offset);
+		map_offset = 0;
 		if (ret)
 			break;
 	}
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 2540971..7ca3fe6 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -43,6 +43,7 @@
 #include <linux/vmalloc.h>
 #include <linux/vmstat.h>
 #include <linux/mempolicy.h>
+#include <linux/memremap.h>
 #include <linux/stop_machine.h>
 #include <linux/sort.h>
 #include <linux/pfn.h>
@@ -4485,8 +4486,9 @@
 void __meminit memmap_init_zone(unsigned long size, int nid, unsigned long zone,
 		unsigned long start_pfn, enum memmap_context context)
 {
-	pg_data_t *pgdat = NODE_DATA(nid);
+	struct vmem_altmap *altmap = to_vmem_altmap(__pfn_to_phys(start_pfn));
 	unsigned long end_pfn = start_pfn + size;
+	pg_data_t *pgdat = NODE_DATA(nid);
 	unsigned long pfn;
 	struct zone *z;
 	unsigned long nr_initialised = 0;
@@ -4494,6 +4496,13 @@
 	if (highest_memmap_pfn < end_pfn - 1)
 		highest_memmap_pfn = end_pfn - 1;
 
+	/*
+	 * Honor reservation requested by the driver for this ZONE_DEVICE
+	 * memory
+	 */
+	if (altmap && start_pfn == altmap->base_pfn)
+		start_pfn += altmap->reserve;
+
 	z = &pgdat->node_zones[zone];
 	for (pfn = start_pfn; pfn < end_pfn; pfn++) {
 		/*
diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c
index 4cba9c27..b60802b 100644
--- a/mm/sparse-vmemmap.c
+++ b/mm/sparse-vmemmap.c
@@ -20,6 +20,7 @@
 #include <linux/mm.h>
 #include <linux/mmzone.h>
 #include <linux/bootmem.h>
+#include <linux/memremap.h>
 #include <linux/highmem.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
@@ -70,7 +71,7 @@
 }
 
 /* need to make sure size is all the same during early stage */
-void * __meminit vmemmap_alloc_block_buf(unsigned long size, int node)
+static void * __meminit alloc_block_buf(unsigned long size, int node)
 {
 	void *ptr;
 
@@ -87,6 +88,77 @@
 	return ptr;
 }
 
+static unsigned long __meminit vmem_altmap_next_pfn(struct vmem_altmap *altmap)
+{
+	return altmap->base_pfn + altmap->reserve + altmap->alloc
+		+ altmap->align;
+}
+
+static unsigned long __meminit vmem_altmap_nr_free(struct vmem_altmap *altmap)
+{
+	unsigned long allocated = altmap->alloc + altmap->align;
+
+	if (altmap->free > allocated)
+		return altmap->free - allocated;
+	return 0;
+}
+
+/**
+ * vmem_altmap_alloc - allocate pages from the vmem_altmap reservation
+ * @altmap - reserved page pool for the allocation
+ * @nr_pfns - size (in pages) of the allocation
+ *
+ * Allocations are aligned to the size of the request
+ */
+static unsigned long __meminit vmem_altmap_alloc(struct vmem_altmap *altmap,
+		unsigned long nr_pfns)
+{
+	unsigned long pfn = vmem_altmap_next_pfn(altmap);
+	unsigned long nr_align;
+
+	nr_align = 1UL << find_first_bit(&nr_pfns, BITS_PER_LONG);
+	nr_align = ALIGN(pfn, nr_align) - pfn;
+
+	if (nr_pfns + nr_align > vmem_altmap_nr_free(altmap))
+		return ULONG_MAX;
+	altmap->alloc += nr_pfns;
+	altmap->align += nr_align;
+	return pfn + nr_align;
+}
+
+static void * __meminit altmap_alloc_block_buf(unsigned long size,
+		struct vmem_altmap *altmap)
+{
+	unsigned long pfn, nr_pfns;
+	void *ptr;
+
+	if (size & ~PAGE_MASK) {
+		pr_warn_once("%s: allocations must be multiple of PAGE_SIZE (%ld)\n",
+				__func__, size);
+		return NULL;
+	}
+
+	nr_pfns = size >> PAGE_SHIFT;
+	pfn = vmem_altmap_alloc(altmap, nr_pfns);
+	if (pfn < ULONG_MAX)
+		ptr = __va(__pfn_to_phys(pfn));
+	else
+		ptr = NULL;
+	pr_debug("%s: pfn: %#lx alloc: %ld align: %ld nr: %#lx\n",
+			__func__, pfn, altmap->alloc, altmap->align, nr_pfns);
+
+	return ptr;
+}
+
+/* need to make sure size is all the same during early stage */
+void * __meminit __vmemmap_alloc_block_buf(unsigned long size, int node,
+		struct vmem_altmap *altmap)
+{
+	if (altmap)
+		return altmap_alloc_block_buf(size, altmap);
+	return alloc_block_buf(size, node);
+}
+
 void __meminit vmemmap_verify(pte_t *pte, int node,
 				unsigned long start, unsigned long end)
 {
@@ -103,7 +175,7 @@
 	pte_t *pte = pte_offset_kernel(pmd, addr);
 	if (pte_none(*pte)) {
 		pte_t entry;
-		void *p = vmemmap_alloc_block_buf(PAGE_SIZE, node);
+		void *p = alloc_block_buf(PAGE_SIZE, node);
 		if (!p)
 			return NULL;
 		entry = pfn_pte(__pa(p) >> PAGE_SHIFT, PAGE_KERNEL);
diff --git a/mm/sparse.c b/mm/sparse.c
index d1b48b6..3717cee 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -748,7 +748,7 @@
 	if (!memmap)
 		return;
 
-	for (i = 0; i < PAGES_PER_SECTION; i++) {
+	for (i = 0; i < nr_pages; i++) {
 		if (PageHWPoison(&memmap[i])) {
 			atomic_long_sub(1, &num_poisoned_pages);
 			ClearPageHWPoison(&memmap[i]);
@@ -788,7 +788,8 @@
 		free_map_bootmem(memmap);
 }
 
-void sparse_remove_one_section(struct zone *zone, struct mem_section *ms)
+void sparse_remove_one_section(struct zone *zone, struct mem_section *ms,
+		unsigned long map_offset)
 {
 	struct page *memmap = NULL;
 	unsigned long *usemap = NULL, flags;
@@ -804,7 +805,8 @@
 	}
 	pgdat_resize_unlock(pgdat, &flags);
 
-	clear_hwpoisoned_pages(memmap, PAGES_PER_SECTION);
+	clear_hwpoisoned_pages(memmap + map_offset,
+			PAGES_PER_SECTION - map_offset);
 	free_section_usemap(memmap, usemap);
 }
 #endif /* CONFIG_MEMORY_HOTREMOVE */