[PATCH 01/11] memory hotplug prep: kill local_mapnr

[PATCH 01/11] memory hotplug prep: kill local_mapnr

[Dave Hansen]

This is unused, except for in an alpha header.  Keep the alpha
one, kill the rest.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/include/asm-i386/mmzone.h   |    6 ------
 memhotplug-dave/include/asm-m32r/mmzone.h   |    6 ------
 memhotplug-dave/include/asm-parisc/mmzone.h |    6 ------
 memhotplug-dave/include/asm-ppc64/mmzone.h  |    3 ---
 memhotplug-dave/include/asm-x86_64/mmzone.h |    2 --
 5 files changed, 23 deletions(-)

diff -puN include/asm-i386/mmzone.h~C0-kill-local_mapnr include/asm-i386/mmzone.h
--- memhotplug/include/asm-i386/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
+++ memhotplug-dave/include/asm-i386/mmzone.h	2005-08-18 14:59:43.000000000 -0700
@@ -88,12 +88,6 @@ static inline int pfn_to_nid(unsigned lo
 	__pgdat->node_start_pfn + __pgdat->node_spanned_pages;		\
 })
 
-#define local_mapnr(kvaddr)						\
-({									\
-	unsigned long __pfn = __pa(kvaddr) >> PAGE_SHIFT;		\
-	(__pfn - node_start_pfn(pfn_to_nid(__pfn)));			\
-})
-
 /* XXX: FIXME -- wli */
 #define kern_addr_valid(kaddr)	(0)
 
diff -puN include/asm-m32r/mmzone.h~C0-kill-local_mapnr include/asm-m32r/mmzone.h
--- memhotplug/include/asm-m32r/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
+++ memhotplug-dave/include/asm-m32r/mmzone.h	2005-08-18 14:59:43.000000000 -0700
@@ -21,12 +21,6 @@ extern struct pglist_data *node_data[];
 	__pgdat->node_start_pfn + __pgdat->node_spanned_pages - 1;	\
 })
 
-#define local_mapnr(kvaddr)						\
-({									\
-	unsigned long __pfn = __pa(kvaddr) >> PAGE_SHIFT;		\
-	(__pfn - node_start_pfn(pfn_to_nid(__pfn)));			\
-})
-
 #define pfn_to_page(pfn)						\
 ({									\
 	unsigned long __pfn = pfn;					\
diff -puN include/asm-parisc/mmzone.h~C0-kill-local_mapnr include/asm-parisc/mmzone.h
--- memhotplug/include/asm-parisc/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
+++ memhotplug-dave/include/asm-parisc/mmzone.h	2005-08-18 14:59:43.000000000 -0700
@@ -27,12 +27,6 @@ extern struct node_map_data node_data[];
 })
 #define node_localnr(pfn, nid)		((pfn) - node_start_pfn(nid))
 
-#define local_mapnr(kvaddr)						\
-({									\
-	unsigned long __pfn = __pa(kvaddr) >> PAGE_SHIFT;		\
-	(__pfn - node_start_pfn(pfn_to_nid(__pfn)));			\
-})
-
 #define pfn_to_page(pfn)						\
 ({									\
 	unsigned long __pfn = (pfn);					\
diff -puN include/asm-ppc64/mmzone.h~C0-kill-local_mapnr include/asm-ppc64/mmzone.h
--- memhotplug/include/asm-ppc64/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
+++ memhotplug-dave/include/asm-ppc64/mmzone.h	2005-08-18 14:59:43.000000000 -0700
@@ -67,9 +67,6 @@ static inline int pa_to_nid(unsigned lon
 #define node_start_pfn(nid)	(NODE_DATA(nid)->node_start_pfn)
 #define node_end_pfn(nid)	(NODE_DATA(nid)->node_end_pfn)
 
-#define local_mapnr(kvaddr) \
-	( (__pa(kvaddr) >> PAGE_SHIFT) - node_start_pfn(kvaddr_to_nid(kvaddr)) 
-
 #ifdef CONFIG_DISCONTIGMEM
 
 /*
diff -puN include/asm-x86_64/mmzone.h~C0-kill-local_mapnr include/asm-x86_64/mmzone.h
--- memhotplug/include/asm-x86_64/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
+++ memhotplug-dave/include/asm-x86_64/mmzone.h	2005-08-18 14:59:43.000000000 -0700
@@ -38,8 +38,6 @@ static inline __attribute__((pure)) int 
 
 #ifdef CONFIG_DISCONTIGMEM
 
-#define pfn_to_nid(pfn) phys_to_nid((unsigned long)(pfn) << PAGE_SHIFT)
-#define kvaddr_to_nid(kaddr)	phys_to_nid(__pa(kaddr))
 
 /* AK: this currently doesn't deal with invalid addresses. We'll see 
    if the 2.5 kernel doesn't pass them
_

[PATCH 02/11] memory hotplug prep: break out zone initialization

[Dave Hansen]

If a zone is empty at boot-time and then hot-added to later,
it needs to run the same init code that would have been run
on it at boot.

This patch breaks out zone table and per-cpu-pages functions
for use by the hotplug code.  You can almost see all of the
free_area_init_core() function on one page now. :)

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/mm/page_alloc.c |   98 +++++++++++++++++++++++-----------------
 1 files changed, 58 insertions(+), 40 deletions(-)

diff -puN mm/page_alloc.c~C1-pcp_zone_init mm/page_alloc.c
--- memhotplug/mm/page_alloc.c~C1-pcp_zone_init	2005-09-02 12:12:32.000000000 -0700
+++ memhotplug-dave/mm/page_alloc.c	2005-09-02 12:17:34.000000000 -0700
@@ -1865,6 +1865,60 @@ void __init setup_per_cpu_pageset()
 
 #endif
 
+static __devinit
+void zone_wait_table_init(struct zone *zone, unsigned long zone_size_pages)
+{
+	int i;
+	struct pglist_data *pgdat = zone->zone_pgdat;
+
+	/*
+	 * The per-page waitqueue mechanism uses hashed waitqueues
+	 * per zone.
+	 */
+	zone->wait_table_size = wait_table_size(zone_size_pages);
+	zone->wait_table_bits =	wait_table_bits(zone->wait_table_size);
+	zone->wait_table = (wait_queue_head_t *)
+		alloc_bootmem_node(pgdat, zone->wait_table_size
+					* sizeof(wait_queue_head_t));
+
+	for(i = 0; i < zone->wait_table_size; ++i)
+		init_waitqueue_head(zone->wait_table + i);
+}
+
+static __devinit void zone_pcp_init(struct zone *zone)
+{
+	int cpu;
+	unsigned long batch = zone_batchsize(zone);
+
+	for (cpu = 0; cpu < NR_CPUS; cpu++) {
+#ifdef CONFIG_NUMA
+		/* Early boot. Slab allocator not functional yet */
+		zone->pageset[cpu] = &boot_pageset[cpu];
+		setup_pageset(&boot_pageset[cpu],0);
+#else
+		setup_pageset(zone_pcp(zone,cpu), batch);
+#endif
+	}
+	printk(KERN_DEBUG "  %s zone: %lu pages, LIFO batch:%lu\n",
+		zone->name, zone->present_pages, batch);
+}
+
+static void init_currently_empty_zone(struct zone *zone,
+		unsigned long zone_start_pfn, unsigned long size)
+{
+	struct pglist_data *pgdat = zone->zone_pgdat;
+
+	zone_wait_table_init(zone, size);
+	pgdat->nr_zones = zone_idx(zone) + 1;
+
+	zone->zone_mem_map = pfn_to_page(zone_start_pfn);
+	zone->zone_start_pfn = zone_start_pfn;
+
+	memmap_init(size, pgdat->node_id, zone_idx(zone), zone_start_pfn);
+
+	zone_init_free_lists(pgdat, zone, zone->spanned_pages);
+}
+
 /*
  * Set up the zone data structures:
  *   - mark all pages reserved
@@ -1874,8 +1928,8 @@ void __init setup_per_cpu_pageset()
 static void __init free_area_init_core(struct pglist_data *pgdat,
 		unsigned long *zones_size, unsigned long *zholes_size)
 {
-	unsigned long i, j;
-	int cpu, nid = pgdat->node_id;
+	unsigned long j;
+	int nid = pgdat->node_id;
 	unsigned long zone_start_pfn = pgdat->node_start_pfn;
 
 	pgdat->nr_zones = 0;
@@ -1885,7 +1939,6 @@ static void __init free_area_init_core(s
 	for (j = 0; j < MAX_NR_ZONES; j++) {
 		struct zone *zone = pgdat->node_zones + j;
 		unsigned long size, realsize;
-		unsigned long batch;
 
 		realsize = size = zones_size[j];
 		if (zholes_size)
@@ -1905,19 +1958,7 @@ static void __init free_area_init_core(s
 
 		zone->temp_priority = zone->prev_priority = DEF_PRIORITY;
 
-		batch = zone_batchsize(zone);
-
-		for (cpu = 0; cpu < NR_CPUS; cpu++) {
-#ifdef CONFIG_NUMA
-			/* Early boot. Slab allocator not functional yet */
-			zone->pageset[cpu] = &boot_pageset[cpu];
-			setup_pageset(&boot_pageset[cpu],0);
-#else
-			setup_pageset(zone_pcp(zone,cpu), batch);
-#endif
-		}
-		printk(KERN_DEBUG "  %s zone: %lu pages, LIFO batch:%lu\n",
-				zone_names[j], realsize, batch);
+		zone_pcp_init(zone);
 		INIT_LIST_HEAD(&zone->active_list);
 		INIT_LIST_HEAD(&zone->inactive_list);
 		zone->nr_scan_active = 0;
@@ -1928,32 +1969,9 @@ static void __init free_area_init_core(s
 		if (!size)
 			continue;
 
-		/*
-		 * The per-page waitqueue mechanism uses hashed waitqueues
-		 * per zone.
-		 */
-		zone->wait_table_size = wait_table_size(size);
-		zone->wait_table_bits =
-			wait_table_bits(zone->wait_table_size);
-		zone->wait_table = (wait_queue_head_t *)
-			alloc_bootmem_node(pgdat, zone->wait_table_size
-						* sizeof(wait_queue_head_t));
-
-		for(i = 0; i < zone->wait_table_size; ++i)
-			init_waitqueue_head(zone->wait_table + i);
-
-		pgdat->nr_zones = j+1;
-
-		zone->zone_mem_map = pfn_to_page(zone_start_pfn);
-		zone->zone_start_pfn = zone_start_pfn;
-
-		memmap_init(size, nid, j, zone_start_pfn);
-
 		zonetable_add(zone, nid, j, zone_start_pfn, size);
-
+		init_currently_empty_zone(zone, zone_start_pfn, size);
 		zone_start_pfn += size;
-
-		zone_init_free_lists(pgdat, zone, zone->spanned_pages);
 	}
 }
 
_

[PATCH 03/11] memory hotplug prep: __section_nr helper

[Dave Hansen]

A little helper that we use in the hotplug code.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/include/linux/mmzone.h |   25 +++++++++++++++++++++++++
 1 files changed, 25 insertions(+)

diff -puN include/linux/mmzone.h~C3-__section_nr include/linux/mmzone.h
--- memhotplug/include/linux/mmzone.h~C3-__section_nr	2005-08-18 14:59:45.000000000 -0700
+++ memhotplug-dave/include/linux/mmzone.h	2005-08-18 14:59:45.000000000 -0700
@@ -511,6 +511,31 @@ static inline struct mem_section *__nr_t
 }
 
 /*
+ * Although written for the SPARSEMEM_EXTREME case, this happens
+ * to also work for the flat array case becase
+ * NR_SECTION_ROOTS==NR_MEM_SECTIONS.
+ */
+static inline int __section_nr(struct mem_section* ms)
+{
+	unsigned long root_nr;
+	struct mem_section* root;
+
+	for (root_nr = 0;
+	     root_nr < NR_MEM_SECTIONS;
+	     root_nr += SECTIONS_PER_ROOT) {
+		root = __nr_to_section(root_nr);
+
+		if (!root)
+			continue;
+
+		if ((ms >= root) && (ms < (root + SECTIONS_PER_ROOT)))
+		     break;
+	}
+
+	return (root_nr * SECTIONS_PER_ROOT) + (ms - root);
+}
+
+/*
  * We use the lower bits of the mem_map pointer to store
  * a little bit of information.  There should be at least
  * 3 bits here due to 32-bit alignment.
_

[PATCH 04/11] memory hotplug prep: fixup bad_range()

[Dave Hansen]

When doing memory hotplug operations, the size of existing zones can
obviously change.  This means that zone->zone_{start_pfn,spanned_pages}
can change.

There are currently no locks that protect these structure members.
However, they are rarely accessed at runtime.  Outside of swsusp, the
only place that I can find is bad_range().

So, split bad_range() up into two pieces: one that needs to be locked
and anther that doesn't.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/mm/page_alloc.c |   26 +++++++++++++++++++++-----
 1 files changed, 21 insertions(+), 5 deletions(-)

diff -puN mm/page_alloc.c~C5.1-bad_range-rework mm/page_alloc.c
--- memhotplug/mm/page_alloc.c~C5.1-bad_range-rework	2005-08-18 14:59:45.000000000 -0700
+++ memhotplug-dave/mm/page_alloc.c	2005-08-18 14:59:45.000000000 -0700
@@ -77,21 +77,37 @@ int min_free_kbytes = 1024;
 unsigned long __initdata nr_kernel_pages;
 unsigned long __initdata nr_all_pages;
 
-/*
- * Temporary debugging check for pages not lying within a given zone.
- */
-static int bad_range(struct zone *zone, struct page *page)
+static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
 {
 	if (page_to_pfn(page) >= zone->zone_start_pfn + zone->spanned_pages)
 		return 1;
 	if (page_to_pfn(page) < zone->zone_start_pfn)
 		return 1;
+
+	return 0;
+}
+
+static int page_is_consistent(struct zone *zone, struct page *page)
+{
 #ifdef CONFIG_HOLES_IN_ZONE
 	if (!pfn_valid(page_to_pfn(page)))
-		return 1;
+		return 0;
 #endif
 	if (zone != page_zone(page))
+		return 0;
+
+	return 1;
+}
+/*
+ * Temporary debugging check for pages not lying within a given zone.
+ */
+static int bad_range(struct zone *zone, struct page *page)
+{
+	if (page_outside_zone_boundaries(zone, page))
 		return 1;
+	if (!page_is_consistent(zone, page))
+		return 1;
+
 	return 0;
 }
 
_

[PATCH 05/11] memory hotplug locking: node_size_lock

[Dave Hansen]

pgdat->node_size_lock is basically only neeeded in one place in the
normal code: show_mem(), which is the arch-specific sysrq-m printing
function.

Strictly speaking, the architectures not doing memory hotplug do
no need this locking in show_mem().  However, they are all included
for completeness.  This should also make any future consolidation
of all of the implementations a little more straightforward.

This lock is also held in the sparsemem code during a memory removal,
as sections are invalidated.  This is the place there pfn_valid() is
made false for a memory area that's being removed.  The lock is
only required when doing pfn_valid() operations on memory which the
user does not already have a reference on the page, such as in
show_mem().

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/arch/alpha/mm/numa.c           |    3 ++
 memhotplug-dave/arch/i386/mm/pgtable.c         |    3 ++
 memhotplug-dave/arch/ia64/mm/discontig.c       |    7 ++++-
 memhotplug-dave/arch/m32r/mm/init.c            |    9 +++++-
 memhotplug-dave/arch/parisc/mm/init.c          |    3 ++
 memhotplug-dave/arch/ppc64/mm/init.c           |    6 ++++
 memhotplug-dave/include/linux/memory_hotplug.h |   34 +++++++++++++++++++++++++
 memhotplug-dave/include/linux/mmzone.h         |   12 ++++++++
 memhotplug-dave/mm/page_alloc.c                |    1 
 9 files changed, 76 insertions(+), 2 deletions(-)

diff -puN arch/alpha/mm/numa.c~C5.2-pgdat_size_lock arch/alpha/mm/numa.c
--- memhotplug/arch/alpha/mm/numa.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/alpha/mm/numa.c	2005-09-02 12:42:10.000000000 -0700
@@ -371,6 +371,8 @@ show_mem(void)
 	show_free_areas();
 	printk("Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
 	for_each_online_node(nid) {
+		unsigned long flags;
+		pgdat_resize_lock(NODE_DATA(nid), &flags);
 		i = node_spanned_pages(nid);
 		while (i-- > 0) {
 			struct page *page = nid_page_nr(nid, i);
@@ -384,6 +386,7 @@ show_mem(void)
 			else
 				shared += page_count(page) - 1;
 		}
+		pgdat_resize_unlock(NODE_DATA(nid), &flags);
 	}
 	printk("%ld pages of RAM\n",total);
 	printk("%ld free pages\n",free);
diff -puN arch/i386/mm/pgtable.c~C5.2-pgdat_size_lock arch/i386/mm/pgtable.c
--- memhotplug/arch/i386/mm/pgtable.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/i386/mm/pgtable.c	2005-09-02 12:42:10.000000000 -0700
@@ -31,11 +31,13 @@ void show_mem(void)
 	pg_data_t *pgdat;
 	unsigned long i;
 	struct page_state ps;
+	unsigned long flags;
 
 	printk(KERN_INFO "Mem-info:\n");
 	show_free_areas();
 	printk(KERN_INFO "Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
 	for_each_pgdat(pgdat) {
+		pgdat_resize_lock(pgdat, &flags);
 		for (i = 0; i < pgdat->node_spanned_pages; ++i) {
 			page = pgdat_page_nr(pgdat, i);
 			total++;
@@ -48,6 +50,7 @@ void show_mem(void)
 			else if (page_count(page))
 				shared += page_count(page) - 1;
 		}
+		pgdat_resize_unlock(pgdat, &flags);
 	}
 	printk(KERN_INFO "%d pages of RAM\n", total);
 	printk(KERN_INFO "%d pages of HIGHMEM\n", highmem);
diff -puN arch/ia64/mm/discontig.c~C5.2-pgdat_size_lock arch/ia64/mm/discontig.c
--- memhotplug/arch/ia64/mm/discontig.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/ia64/mm/discontig.c	2005-09-02 13:43:06.000000000 -0700
@@ -524,9 +524,13 @@ void show_mem(void)
 	show_free_areas();
 	printk("Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
 	for_each_pgdat(pgdat) {
-		unsigned long present = pgdat->node_present_pages;
+		unsigned long present;
+		unsigned long flags;
 		int shared = 0, cached = 0, reserved = 0;
+
 		printk("Node ID: %d\n", pgdat->node_id);
+		pgdat_resize_lock(pgdat, &flags);
+		present = pgdat->node_present_pages;
 		for(i = 0; i < pgdat->node_spanned_pages; i++) {
 			struct page *page = pgdat_page_nr(pgdat, i);
 			if (!ia64_pfn_valid(pgdat->node_start_pfn+i))
@@ -538,6 +542,7 @@ void show_mem(void)
 			else if (page_count(page))
 				shared += page_count(page)-1;
 		}
+		pgdat_resize_unlock(pgdat, &flags);
 		total_present += present;
 		total_reserved += reserved;
 		total_cached += cached;
diff -puN arch/m32r/mm/init.c~C5.2-pgdat_size_lock arch/m32r/mm/init.c
--- memhotplug/arch/m32r/mm/init.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/m32r/mm/init.c	2005-09-02 12:42:10.000000000 -0700
@@ -48,6 +48,8 @@ void show_mem(void)
 	show_free_areas();
 	printk("Free swap:       %6ldkB\n",nr_swap_pages<<(PAGE_SHIFT-10));
 	for_each_pgdat(pgdat) {
+		unsigned long flags;
+		pgdat_resize_lock(pgdat, &flags);
 		for (i = 0; i < pgdat->node_spanned_pages; ++i) {
 			page = pgdat_page_nr(pgdat, i);
 			total++;
@@ -60,6 +62,7 @@ void show_mem(void)
 			else if (page_count(page))
 				shared += page_count(page) - 1;
 		}
+		pgdat_resize_unlock(pgdat, &flags);
 	}
 	printk("%d pages of RAM\n", total);
 	printk("%d pages of HIGHMEM\n",highmem);
@@ -150,10 +153,14 @@ int __init reservedpages_count(void)
 	int reservedpages, nid, i;
 
 	reservedpages = 0;
-	for_each_online_node(nid)
+	for_each_online_node(nid) {
+		unsigned long flags;
+		pgdat_resize_lock(NODE_DATA(nid), &flags);
 		for (i = 0 ; i < MAX_LOW_PFN(nid) - START_PFN(nid) ; i++)
 			if (PageReserved(nid_page_nr(nid, i)))
 				reservedpages++;
+		pgdat_resize_unlock(NODE_DATA(nid), &flags);
+	}
 
 	return reservedpages;
 }
diff -puN arch/parisc/mm/init.c~C5.2-pgdat_size_lock arch/parisc/mm/init.c
--- memhotplug/arch/parisc/mm/init.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/parisc/mm/init.c	2005-09-02 12:42:10.000000000 -0700
@@ -505,7 +505,9 @@ void show_mem(void)
 
 		for (j = node_start_pfn(i); j < node_end_pfn(i); j++) {
 			struct page *p;
+			unsigned long flags;
 
+			pgdat_resize_lock(NODE_DATA(i), &flags);
 			p = nid_page_nr(i, j) - node_start_pfn(i);
 
 			total++;
@@ -517,6 +519,7 @@ void show_mem(void)
 				free++;
 			else
 				shared += page_count(p) - 1;
+			pgdat_resize_unlock(NODE_DATA(i), &flags);
         	}
 	}
 #endif
diff -puN arch/ppc64/mm/init.c~C5.2-pgdat_size_lock arch/ppc64/mm/init.c
--- memhotplug/arch/ppc64/mm/init.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/arch/ppc64/mm/init.c	2005-09-02 13:43:08.000000000 -0700
@@ -104,6 +104,8 @@ void show_mem(void)
 	show_free_areas();
 	printk("Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
 	for_each_pgdat(pgdat) {
+		unsigned long flags;
+		pgdat_resize_lock(pgdat, &flags);
 		for (i = 0; i < pgdat->node_spanned_pages; i++) {
 			page = pgdat_page_nr(pgdat, i);
 			total++;
@@ -114,6 +116,7 @@ void show_mem(void)
 			else if (page_count(page))
 				shared += page_count(page) - 1;
 		}
+		pgdat_resize_unlock(pgdat, &flags);
 	}
 	printk("%ld pages of RAM\n", total);
 	printk("%ld reserved pages\n", reserved);
@@ -648,11 +651,14 @@ void __init mem_init(void)
 #endif
 
 	for_each_pgdat(pgdat) {
+		unsigned long flags;
+		pgdat_resize_lock(pgdat, &flags);
 		for (i = 0; i < pgdat->node_spanned_pages; i++) {
 			page = pgdat_page_nr(pgdat, i);
 			if (PageReserved(page))
 				reservedpages++;
 		}
+		pgdat_resize_unlock(pgdat, &flags);
 	}
 
 	codesize = (unsigned long)&_etext - (unsigned long)&_stext;
diff -puN /dev/null include/linux/memory_hotplug.h
--- /dev/null	2005-03-30 22:36:15.000000000 -0800
+++ memhotplug-dave/include/linux/memory_hotplug.h	2005-09-02 13:43:13.000000000 -0700
@@ -0,0 +1,34 @@
+#ifndef __LINUX_MEMORY_HOTPLUG_H
+#define __LINUX_MEMORY_HOTPLUG_H
+
+#include <linux/mmzone.h>
+#include <linux/spinlock.h>
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+/*
+ * pgdat resizing functions
+ */
+static inline
+void pgdat_resize_lock(struct pglist_data *pgdat, unsigned long *flags)
+{
+	spin_lock_irqsave(&pgdat->node_size_lock, *flags);
+}
+static inline
+void pgdat_resize_unlock(struct pglist_data *pgdat, unsigned long *flags)
+{
+	spin_lock_irqrestore(&pgdat->node_size_lock, *flags);
+}
+static inline
+void pgdat_resize_init(struct pglist_data *pgdat)
+{
+	spin_lock_init(&pgdat->node_size_lock);
+}
+#else /* ! CONFIG_MEMORY_HOTPLUG */
+/*
+ * Stub functions for when hotplug is off
+ */
+static inline void pgdat_resize_lock(struct pglist_data *p, unsigned long *f) {}
+static inline void pgdat_resize_unlock(struct pglist_data *p, unsigned long *f) {}
+static inline void pgdat_resize_init(struct pglist_data *pgdat) {}
+#endif
+#endif /* __LINUX_MEMORY_HOTPLUG_H */
diff -puN include/linux/mmzone.h~C5.2-pgdat_size_lock include/linux/mmzone.h
--- memhotplug/include/linux/mmzone.h~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/include/linux/mmzone.h	2005-09-02 13:43:13.000000000 -0700
@@ -273,6 +273,16 @@ typedef struct pglist_data {
 	struct page *node_mem_map;
 #endif
 	struct bootmem_data *bdata;
+#ifdef CONFIG_MEMORY_HOTPLUG
+	/*
+	 * Must be held any time you expect node_start_pfn, node_present_pages
+	 * or node_spanned_pages stay constant.  Holding this will also
+	 * guarantee that any pfn_valid() stays that way.
+	 *
+	 * Nests above zone->lock and zone->size_seqlock.
+	 */
+	spinlock_t node_size_lock;
+#endif
 	unsigned long node_start_pfn;
 	unsigned long node_present_pages; /* total number of physical pages */
 	unsigned long node_spanned_pages; /* total size of physical page
@@ -293,6 +303,8 @@ typedef struct pglist_data {
 #endif
 #define nid_page_nr(nid, pagenr) 	pgdat_page_nr(NODE_DATA(nid),(pagenr))
 
+#include <linux/memory_hotplug.h>
+
 extern struct pglist_data *pgdat_list;
 
 void __get_zone_counts(unsigned long *active, unsigned long *inactive,
diff -puN mm/page_alloc.c~C5.2-pgdat_size_lock mm/page_alloc.c
--- memhotplug/mm/page_alloc.c~C5.2-pgdat_size_lock	2005-09-02 12:42:10.000000000 -0700
+++ memhotplug-dave/mm/page_alloc.c	2005-09-02 13:43:13.000000000 -0700
@@ -1948,6 +1948,7 @@ static void __init free_area_init_core(s
 	int nid = pgdat->node_id;
 	unsigned long zone_start_pfn = pgdat->node_start_pfn;
 
+	pgdat_resize_init(pgdat);
 	pgdat->nr_zones = 0;
 	init_waitqueue_head(&pgdat->kswapd_wait);
 	pgdat->kswapd_max_order = 0;
_

[PATCH 06/11] memory hotplug locking: zone span seqlock

[Dave Hansen]

See the "fixup bad_range()" patch for more information, but this
actually creates a the lock to protect things making assumptions
about a zone's size staying constant at runtime.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/include/linux/memory_hotplug.h |   39 +++++++++++++++++++++++--
 memhotplug-dave/include/linux/mmzone.h         |   15 +++++++++
 memhotplug-dave/mm/page_alloc.c                |   19 ++++++++----
 3 files changed, 66 insertions(+), 7 deletions(-)

diff -puN include/linux/memory_hotplug.h~C6-zone-span_seqlock include/linux/memory_hotplug.h
--- memhotplug/include/linux/memory_hotplug.h~C6-zone-span_seqlock	2005-09-02 12:42:11.000000000 -0700
+++ memhotplug-dave/include/linux/memory_hotplug.h	2005-09-02 13:43:10.000000000 -0700
@@ -16,13 +16,36 @@ void pgdat_resize_lock(struct pglist_dat
 static inline
 void pgdat_resize_unlock(struct pglist_data *pgdat, unsigned long *flags)
 {
-	spin_lock_irqrestore(&pgdat->node_size_lock, *flags);
+	spin_unlock_irqrestore(&pgdat->node_size_lock, *flags);
 }
 static inline
 void pgdat_resize_init(struct pglist_data *pgdat)
 {
 	spin_lock_init(&pgdat->node_size_lock);
 }
+/*
+ * Zone resizing functions
+ */
+static inline unsigned zone_span_seqbegin(struct zone *zone)
+{
+	return read_seqbegin(&zone->span_seqlock);
+}
+static inline int zone_span_seqretry(struct zone *zone, unsigned iv)
+{
+	return read_seqretry(&zone->span_seqlock, iv);
+}
+static inline void zone_span_writelock(struct zone *zone)
+{
+	write_seqlock(&zone->span_seqlock);
+}
+static inline void zone_span_writeunlock(struct zone *zone)
+{
+	write_sequnlock(&zone->span_seqlock);
+}
+static inline void zone_seqlock_init(struct zone *zone)
+{
+	seqlock_init(&zone->span_seqlock);
+}
 #else /* ! CONFIG_MEMORY_HOTPLUG */
 /*
  * Stub functions for when hotplug is off
@@ -30,5 +53,17 @@ void pgdat_resize_init(struct pglist_dat
 static inline void pgdat_resize_lock(struct pglist_data *p, unsigned long *f) {}
 static inline void pgdat_resize_unlock(struct pglist_data *p, unsigned long *f) {}
 static inline void pgdat_resize_init(struct pglist_data *pgdat) {}
-#endif
+
+static inline unsigned zone_span_seqbegin(struct zone *zone)
+{
+	return 0;
+}
+static inline int zone_span_seqretry(struct zone *zone, unsigned iv)
+{
+	return 0;
+}
+static inline void zone_span_writelock(struct zone *zone) {}
+static inline void zone_span_writeunlock(struct zone *zone) {}
+static inline void zone_seqlock_init(struct zone *zone) {}
+#endif /* ! CONFIG_MEMORY_HOTPLUG */
 #endif /* __LINUX_MEMORY_HOTPLUG_H */
diff -puN include/linux/mmzone.h~C6-zone-span_seqlock include/linux/mmzone.h
--- memhotplug/include/linux/mmzone.h~C6-zone-span_seqlock	2005-09-02 12:42:11.000000000 -0700
+++ memhotplug-dave/include/linux/mmzone.h	2005-09-02 12:42:11.000000000 -0700
@@ -12,6 +12,7 @@
 #include <linux/threads.h>
 #include <linux/numa.h>
 #include <linux/init.h>
+#include <linux/seqlock.h>
 #include <asm/atomic.h>
 
 /* Free memory management - zoned buddy allocator.  */
@@ -137,6 +138,10 @@ struct zone {
 	 * free areas of different sizes
 	 */
 	spinlock_t		lock;
+#ifdef CONFIG_MEMORY_HOTPLUG
+	/* see spanned/present_pages for more description */
+	seqlock_t		span_seqlock;
+#endif
 	struct free_area	free_area[MAX_ORDER];
 
 
@@ -220,6 +225,16 @@ struct zone {
 	/* zone_start_pfn == zone_start_paddr >> PAGE_SHIFT */
 	unsigned long		zone_start_pfn;
 
+	/*
+	 * zone_start_pfn, spanned_pages and present_pages are all
+	 * protected by span_seqlock.  It is a seqlock because it has
+	 * to be read outside of zone->lock, and it is done in the main
+	 * allocator path.  But, it is written quite infrequently.
+	 *
+	 * The lock is declared along with zone->lock because it is
+	 * frequently read in proximity to zone->lock.  It's good to
+	 * give them a chance of being in the same cacheline.
+	 */
 	unsigned long		spanned_pages;	/* total size, including holes */
 	unsigned long		present_pages;	/* amount of memory (excluding holes) */
 
diff -puN mm/page_alloc.c~C6-zone-span_seqlock mm/page_alloc.c
--- memhotplug/mm/page_alloc.c~C6-zone-span_seqlock	2005-09-02 12:42:11.000000000 -0700
+++ memhotplug-dave/mm/page_alloc.c	2005-09-02 13:43:10.000000000 -0700
@@ -32,6 +32,7 @@
 #include <linux/sysctl.h>
 #include <linux/cpu.h>
 #include <linux/cpuset.h>
+#include <linux/memory_hotplug.h>
 #include <linux/nodemask.h>
 #include <linux/vmalloc.h>
 
@@ -79,12 +80,19 @@ unsigned long __initdata nr_all_pages;
 
 static int page_outside_zone_boundaries(struct zone *zone, struct page *page)
 {
-	if (page_to_pfn(page) >= zone->zone_start_pfn + zone->spanned_pages)
-		return 1;
-	if (page_to_pfn(page) < zone->zone_start_pfn)
-		return 1;
+	int ret = 0;
+	unsigned seq;
+	unsigned long pfn = page_to_pfn(page);
 
-	return 0;
+	do {
+		seq = zone_span_seqbegin(zone);
+		if (pfn >= zone->zone_start_pfn + zone->spanned_pages)
+			ret = 1;
+		else if (pfn < zone->zone_start_pfn)
+			ret = 1;
+	} while (zone_span_seqretry(zone, seq));
+
+	return ret;
 }
 
 static int page_is_consistent(struct zone *zone, struct page *page)
@@ -1970,6 +1978,7 @@ static void __init free_area_init_core(s
 		zone->name = zone_names[j];
 		spin_lock_init(&zone->lock);
 		spin_lock_init(&zone->lru_lock);
+		zone_seqlock_init(zone);
 		zone->zone_pgdat = pgdat;
 		zone->free_pages = 0;
 
_

[PATCH 07/11] memory hotplug: sysfs and add/remove functions

[Dave Hansen]

This adds generic memory add/remove and supporting functions 
for memory hotplug into a new file as well as a memory hotplug
kernel config option.

Individual architecture patches will follow.

For now, disable memory hotplug when swsusp is enabled.  There's
a lot of churn there right now.  We'll fix it up properly once
it calms down.

Signed-off-by: Matt Tolentino <matthew.e.tolentino@intel.com>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/drivers/base/Makefile          |    1 
 memhotplug-dave/drivers/base/init.c            |    2 
 memhotplug-dave/drivers/base/memory.c          |  455 +++++++++++++++++++++++++
 memhotplug-dave/include/linux/memory.h         |   94 +++++
 memhotplug-dave/include/linux/memory_hotplug.h |   35 +
 memhotplug-dave/include/linux/mm.h             |    1 
 memhotplug-dave/mm/Kconfig                     |    8 
 memhotplug-dave/mm/Makefile                    |    2 
 memhotplug-dave/mm/memory_hotplug.c            |  178 +++++++++
 memhotplug-dave/mm/page_alloc.c                |    4 
 10 files changed, 777 insertions(+), 3 deletions(-)

diff -puN drivers/base/Makefile~D0-sysfs-memory-class drivers/base/Makefile
--- memhotplug/drivers/base/Makefile~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/drivers/base/Makefile	2005-08-18 14:59:48.000000000 -0700
@@ -7,6 +7,7 @@ obj-y			:= core.o sys.o bus.o dd.o \
 obj-y			+= power/
 obj-$(CONFIG_FW_LOADER)	+= firmware_class.o
 obj-$(CONFIG_NUMA)	+= node.o
+obj-$(CONFIG_MEMORY_HOTPLUG) += memory.o
 
 ifeq ($(CONFIG_DEBUG_DRIVER),y)
 EXTRA_CFLAGS += -DDEBUG
diff -puN drivers/base/init.c~D0-sysfs-memory-class drivers/base/init.c
--- memhotplug/drivers/base/init.c~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/drivers/base/init.c	2005-08-18 14:59:48.000000000 -0700
@@ -9,6 +9,7 @@
 
 #include <linux/device.h>
 #include <linux/init.h>
+#include <linux/memory.h>
 
 extern int devices_init(void);
 extern int buses_init(void);
@@ -39,5 +40,6 @@ void __init driver_init(void)
 	platform_bus_init();
 	system_bus_init();
 	cpu_dev_init();
+	memory_dev_init();
 	attribute_container_init();
 }
diff -puN /dev/null drivers/base/memory.c
--- /dev/null	2005-03-30 22:36:15.000000000 -0800
+++ memhotplug-dave/drivers/base/memory.c	2005-08-18 14:59:48.000000000 -0700
@@ -0,0 +1,455 @@
+/*
+ * drivers/base/memory.c - basic Memory class support
+ *
+ * Written by Matt Tolentino <matthew.e.tolentino@intel.com>
+ *            Dave Hansen <haveblue@us.ibm.com>
+ *
+ * This file provides the necessary infrastructure to represent
+ * a SPARSEMEM-memory-model system's physical memory in /sysfs.
+ * All arch-independent code that assumes MEMORY_HOTPLUG requires
+ * SPARSEMEM should be contained here, or in mm/memory_hotplug.c.
+ */
+
+#include <linux/sysdev.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/sched.h>	/* capable() */
+#include <linux/topology.h>
+#include <linux/device.h>
+#include <linux/memory.h>
+#include <linux/kobject.h>
+#include <linux/memory_hotplug.h>
+#include <linux/mm.h>
+#include <asm/atomic.h>
+#include <asm/uaccess.h>
+
+#define MEMORY_CLASS_NAME	"memory"
+
+struct sysdev_class memory_sysdev_class = {
+	set_kset_name(MEMORY_CLASS_NAME),
+};
+EXPORT_SYMBOL(memory_sysdev_class);
+
+static char *memory_hotplug_name(struct kset *kset, struct kobject *kobj)
+{
+	return MEMORY_CLASS_NAME;
+}
+
+static int memory_hotplug(struct kset *kset, struct kobject *kobj, char **envp,
+			int num_envp, char *buffer, int buffer_size)
+{
+	int retval = 0;
+
+	return retval;
+}
+
+static struct kset_hotplug_ops memory_hotplug_ops = {
+	.name		= memory_hotplug_name,
+	.hotplug	= memory_hotplug,
+};
+
+static struct notifier_block *memory_chain;
+
+int register_memory_notifier(struct notifier_block *nb)
+{
+        return notifier_chain_register(&memory_chain, nb);
+}
+
+void unregister_memory_notifier(struct notifier_block *nb)
+{
+        notifier_chain_unregister(&memory_chain, nb);
+}
+
+/*
+ * register_memory - Setup a sysfs device for a memory block
+ */
+int
+register_memory(struct memory_block *memory, struct mem_section *section,
+		struct node *root)
+{
+	int error;
+
+	memory->sysdev.cls = &memory_sysdev_class;
+	memory->sysdev.id = __section_nr(section);
+
+	error = sysdev_register(&memory->sysdev);
+
+	if (root && !error)
+		error = sysfs_create_link(&root->sysdev.kobj,
+					  &memory->sysdev.kobj,
+					  kobject_name(&memory->sysdev.kobj));
+
+	return error;
+}
+
+void
+unregister_memory(struct memory_block *memory, struct mem_section *section,
+		struct node *root)
+{
+	BUG_ON(memory->sysdev.cls != &memory_sysdev_class);
+	BUG_ON(memory->sysdev.id != __section_nr(section));
+
+	sysdev_unregister(&memory->sysdev);
+	if (root)
+		sysfs_remove_link(&root->sysdev.kobj,
+				  kobject_name(&memory->sysdev.kobj));
+}
+
+/*
+ * use this as the physical section index that this memsection
+ * uses.
+ */
+
+static ssize_t show_mem_phys_index(struct sys_device *dev, char *buf)
+{
+	struct memory_block *mem =
+		container_of(dev, struct memory_block, sysdev);
+	return sprintf(buf, "%08lx\n", mem->phys_index);
+}
+
+/*
+ * online, offline, going offline, etc.
+ */
+static ssize_t show_mem_state(struct sys_device *dev, char *buf)
+{
+	struct memory_block *mem =
+		container_of(dev, struct memory_block, sysdev);
+	ssize_t len = 0;
+
+	/*
+	 * We can probably put these states in a nice little array
+	 * so that they're not open-coded
+	 */
+	switch (mem->state) {
+		case MEM_ONLINE:
+			len = sprintf(buf, "online\n");
+			break;
+		case MEM_OFFLINE:
+			len = sprintf(buf, "offline\n");
+			break;
+		case MEM_GOING_OFFLINE:
+			len = sprintf(buf, "going-offline\n");
+			break;
+		default:
+			len = sprintf(buf, "ERROR-UNKNOWN-%ld\n",
+					mem->state);
+			WARN_ON(1);
+			break;
+	}
+
+	return len;
+}
+
+static inline int memory_notify(unsigned long val, void *v)
+{
+	return notifier_call_chain(&memory_chain, val, v);
+}
+
+/*
+ * MEMORY_HOTPLUG depends on SPARSEMEM in mm/Kconfig, so it is
+ * OK to have direct references to sparsemem variables in here.
+ */
+static int
+memory_block_action(struct memory_block *mem, unsigned long action)
+{
+	int i;
+	unsigned long psection;
+	unsigned long start_pfn, start_paddr;
+	struct page *first_page;
+	int ret;
+	int old_state = mem->state;
+
+	psection = mem->phys_index;
+	first_page = pfn_to_page(psection << PFN_SECTION_SHIFT);
+
+	/*
+	 * The probe routines leave the pages reserved, just
+	 * as the bootmem code does.  Make sure they're still
+	 * that way.
+	 */
+	if (action == MEM_ONLINE) {
+		for (i = 0; i < PAGES_PER_SECTION; i++) {
+			if (PageReserved(first_page+i))
+				continue;
+
+			printk(KERN_WARNING "section number %ld page number %d "
+				"not reserved, was it already online? \n",
+				psection, i);
+			return -EBUSY;
+		}
+	}
+
+	switch (action) {
+		case MEM_ONLINE:
+			start_pfn = page_to_pfn(first_page);
+			ret = online_pages(start_pfn, PAGES_PER_SECTION);
+			break;
+		case MEM_OFFLINE:
+			mem->state = MEM_GOING_OFFLINE;
+			memory_notify(MEM_GOING_OFFLINE, NULL);
+			start_paddr = page_to_pfn(first_page) << PAGE_SHIFT;
+			ret = remove_memory(start_paddr,
+					    PAGES_PER_SECTION << PAGE_SHIFT);
+			if (ret) {
+				mem->state = old_state;
+				break;
+			}
+			memory_notify(MEM_MAPPING_INVALID, NULL);
+			break;
+		default:
+			printk(KERN_WARNING "%s(%p, %ld) unknown action: %ld\n",
+					__FUNCTION__, mem, action, action);
+			WARN_ON(1);
+			ret = -EINVAL;
+	}
+	/*
+	 * For now, only notify on successful memory operations
+	 */
+	if (!ret)
+		memory_notify(action, NULL);
+
+	return ret;
+}
+
+static int memory_block_change_state(struct memory_block *mem,
+		unsigned long to_state, unsigned long from_state_req)
+{
+	int ret = 0;
+	down(&mem->state_sem);
+
+	if (mem->state != from_state_req) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	ret = memory_block_action(mem, to_state);
+	if (!ret)
+		mem->state = to_state;
+
+out:
+	up(&mem->state_sem);
+	return ret;
+}
+
+static ssize_t
+store_mem_state(struct sys_device *dev, const char *buf, size_t count)
+{
+	struct memory_block *mem;
+	unsigned int phys_section_nr;
+	int ret = -EINVAL;
+
+	mem = container_of(dev, struct memory_block, sysdev);
+	phys_section_nr = mem->phys_index;
+
+	if (!valid_section_nr(phys_section_nr))
+		goto out;
+
+	if (!strncmp(buf, "online", min((int)count, 6)))
+		ret = memory_block_change_state(mem, MEM_ONLINE, MEM_OFFLINE);
+	else if(!strncmp(buf, "offline", min((int)count, 7)))
+		ret = memory_block_change_state(mem, MEM_OFFLINE, MEM_ONLINE);
+out:
+	if (ret)
+		return ret;
+	return count;
+}
+
+/*
+ * phys_device is a bad name for this.  What I really want
+ * is a way to differentiate between memory ranges that
+ * are part of physical devices that constitute
+ * a complete removable unit or fru.
+ * i.e. do these ranges belong to the same physical device,
+ * s.t. if I offline all of these sections I can then
+ * remove the physical device?
+ */
+static ssize_t show_phys_device(struct sys_device *dev, char *buf)
+{
+	struct memory_block *mem =
+		container_of(dev, struct memory_block, sysdev);
+	return sprintf(buf, "%d\n", mem->phys_device);
+}
+
+SYSDEV_ATTR(phys_index, 0444, show_mem_phys_index, NULL);
+SYSDEV_ATTR(state, 0644, show_mem_state, store_mem_state);
+SYSDEV_ATTR(phys_device, 0444, show_phys_device, NULL);
+
+#define mem_create_simple_file(mem, attr_name)	\
+	sysdev_create_file(&mem->sysdev, &attr_##attr_name)
+#define mem_remove_simple_file(mem, attr_name)	\
+	sysdev_remove_file(&mem->sysdev, &attr_##attr_name)
+
+/*
+ * Block size attribute stuff
+ */
+static ssize_t
+print_block_size(struct class *class, char *buf)
+{
+	return sprintf(buf, "%lx\n", (unsigned long)PAGES_PER_SECTION * PAGE_SIZE);
+}
+
+static CLASS_ATTR(block_size_bytes, 0444, print_block_size, NULL);
+
+static int block_size_init(void)
+{
+	sysfs_create_file(&memory_sysdev_class.kset.kobj,
+		&class_attr_block_size_bytes.attr);
+	return 0;
+}
+
+/*
+ * Some architectures will have custom drivers to do this, and
+ * will not need to do it from userspace.  The fake hot-add code
+ * as well as ppc64 will do all of their discovery in userspace
+ * and will require this interface.
+ */
+#ifdef CONFIG_ARCH_MEMORY_PROBE
+static ssize_t
+memory_probe_store(struct class *class, const char __user *buf, size_t count)
+{
+	u64 phys_addr;
+	int ret;
+
+	phys_addr = simple_strtoull(buf, NULL, 0);
+
+	ret = add_memory(phys_addr, PAGES_PER_SECTION << PAGE_SHIFT);
+
+	if (ret)
+		count = ret;
+
+	return count;
+}
+static CLASS_ATTR(probe, 0700, NULL, memory_probe_store);
+
+static int memory_probe_init(void)
+{
+	sysfs_create_file(&memory_sysdev_class.kset.kobj,
+		&class_attr_probe.attr);
+	return 0;
+}
+#else
+#define memory_probe_init(...)	do {} while (0)
+#endif
+
+/*
+ * Note that phys_device is optional.  It is here to allow for
+ * differentiation between which *physical* devices each
+ * section belongs to...
+ */
+
+int add_memory_block(unsigned long node_id, struct mem_section *section,
+		     unsigned long state, int phys_device)
+{
+	size_t size = sizeof(struct memory_block);
+	struct memory_block *mem = kmalloc(size, GFP_KERNEL);
+	int ret = 0;
+
+	if (!mem)
+		return -ENOMEM;
+
+	memset(mem, 0, size);
+
+	mem->phys_index = __section_nr(section);
+	mem->state = state;
+	init_MUTEX(&mem->state_sem);
+	mem->phys_device = phys_device;
+
+	ret = register_memory(mem, section, NULL);
+	if (!ret)
+		ret = mem_create_simple_file(mem, phys_index);
+	if (!ret)
+		ret = mem_create_simple_file(mem, state);
+	if (!ret)
+		ret = mem_create_simple_file(mem, phys_device);
+
+	return ret;
+}
+
+/*
+ * For now, we have a linear search to go find the appropriate
+ * memory_block corresponding to a particular phys_index. If
+ * this gets to be a real problem, we can always use a radix
+ * tree or something here.
+ *
+ * This could be made generic for all sysdev classes.
+ */
+struct memory_block *find_memory_block(struct mem_section *section)
+{
+	struct kobject *kobj;
+	struct sys_device *sysdev;
+	struct memory_block *mem;
+	char name[sizeof(MEMORY_CLASS_NAME) + 9 + 1];
+
+	/*
+	 * This only works because we know that section == sysdev->id
+	 * slightly redundant with sysdev_register()
+	 */
+	sprintf(&name[0], "%s%d", MEMORY_CLASS_NAME, __section_nr(section));
+
+	kobj = kset_find_obj(&memory_sysdev_class.kset, name);
+	if (!kobj)
+		return NULL;
+
+	sysdev = container_of(kobj, struct sys_device, kobj);
+	mem = container_of(sysdev, struct memory_block, sysdev);
+
+	return mem;
+}
+
+int remove_memory_block(unsigned long node_id, struct mem_section *section,
+		int phys_device)
+{
+	struct memory_block *mem;
+
+	mem = find_memory_block(section);
+	mem_remove_simple_file(mem, phys_index);
+	mem_remove_simple_file(mem, state);
+	mem_remove_simple_file(mem, phys_device);
+	unregister_memory(mem, section, NULL);
+
+	return 0;
+}
+
+/*
+ * need an interface for the VM to add new memory regions,
+ * but without onlining it.
+ */
+int register_new_memory(struct mem_section *section)
+{
+	return add_memory_block(0, section, MEM_OFFLINE, 0);
+}
+
+int unregister_memory_section(struct mem_section *section)
+{
+	if (!valid_section(section))
+		return -EINVAL;
+
+	return remove_memory_block(0, section, 0);
+}
+
+/*
+ * Initialize the sysfs support for memory devices...
+ */
+int __init memory_dev_init(void)
+{
+	unsigned int i;
+	int ret;
+
+	memory_sysdev_class.kset.hotplug_ops = &memory_hotplug_ops;
+	ret = sysdev_class_register(&memory_sysdev_class);
+
+	/*
+	 * Create entries for memory sections that were found
+	 * during boot and have been initialized
+	 */
+	for (i = 0; i < NR_MEM_SECTIONS; i++) {
+		if (!valid_section_nr(i))
+			continue;
+		add_memory_block(0, __nr_to_section(i), MEM_ONLINE, 0);
+	}
+
+	memory_probe_init();
+	block_size_init();
+
+	return ret;
+}
diff -puN /dev/null include/linux/memory.h
--- /dev/null	2005-03-30 22:36:15.000000000 -0800
+++ memhotplug-dave/include/linux/memory.h	2005-08-18 14:59:48.000000000 -0700
@@ -0,0 +1,94 @@
+/*
+ * include/linux/memory.h - generic memory definition
+ *
+ * This is mainly for topological representation. We define the
+ * basic "struct memory_block" here, which can be embedded in per-arch
+ * definitions or NUMA information.
+ *
+ * Basic handling of the devices is done in drivers/base/memory.c
+ * and system devices are handled in drivers/base/sys.c.
+ *
+ * Memory block are exported via sysfs in the class/memory/devices/
+ * directory.
+ *
+ */
+#ifndef _LINUX_MEMORY_H_
+#define _LINUX_MEMORY_H_
+
+#include <linux/sysdev.h>
+#include <linux/node.h>
+#include <linux/compiler.h>
+
+#include <asm/semaphore.h>
+
+struct memory_block {
+	unsigned long phys_index;
+	unsigned long state;
+	/*
+	 * This serializes all state change requests.  It isn't
+	 * held during creation because the control files are
+	 * created long after the critical areas during
+	 * initialization.
+	 */
+	struct semaphore state_sem;
+	int phys_device;		/* to which fru does this belong? */
+	void *hw;			/* optional pointer to fw/hw data */
+	int (*phys_callback)(struct memory_block *);
+	struct sys_device sysdev;
+};
+
+/* These states are exposed to userspace as text strings in sysfs */
+#define	MEM_ONLINE		(1<<0) /* exposed to userspace */
+#define	MEM_GOING_OFFLINE	(1<<1) /* exposed to userspace */
+#define	MEM_OFFLINE		(1<<2) /* exposed to userspace */
+
+/*
+ * All of these states are currently kernel-internal for notifying
+ * kernel components and architectures.
+ *
+ * For MEM_MAPPING_INVALID, all notifier chains with priority >0
+ * are called before pfn_to_page() becomes invalid.  The priority=0
+ * entry is reserved for the function that actually makes
+ * pfn_to_page() stop working.  Any notifiers that want to be called
+ * after that should have priority <0.
+ */
+#define	MEM_MAPPING_INVALID	(1<<3)
+
+#ifndef CONFIG_MEMORY_HOTPLUG
+static inline int memory_dev_init(void)
+{
+	return 0;
+}
+static inline int register_memory_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
+static inline void unregister_memory_notifier(struct notifier_block *nb)
+{
+}
+#else
+extern int register_memory(struct memory_block *, struct mem_section *section, struct node *);
+extern int register_new_memory(struct mem_section *);
+extern int unregister_memory_section(struct mem_section *);
+extern int memory_dev_init(void);
+extern int register_memory_notifier(struct notifier_block *nb);
+extern void unregister_memory_notifier(struct notifier_block *nb);
+
+#define CONFIG_MEM_BLOCK_SIZE	(PAGES_PER_SECTION<<PAGE_SHIFT)
+
+extern int invalidate_phys_mapping(unsigned long, unsigned long);
+struct notifier_block;
+
+extern int register_memory_notifier(struct notifier_block *nb);
+extern void unregister_memory_notifier(struct notifier_block *nb);
+
+extern struct sysdev_class memory_sysdev_class;
+#endif /* CONFIG_MEMORY_HOTPLUG */
+
+#define hotplug_memory_notifier(fn, pri) {			\
+	static struct notifier_block fn##_mem_nb =		\
+		{ .notifier_call = fn, .priority = pri };	\
+	register_memory_notifier(&fn##_mem_nb);			\
+}
+
+#endif /* _LINUX_MEMORY_H_ */
diff -puN include/linux/memory_hotplug.h~D0-sysfs-memory-class include/linux/memory_hotplug.h
--- memhotplug/include/linux/memory_hotplug.h~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/include/linux/memory_hotplug.h	2005-08-18 14:59:48.000000000 -0700
@@ -3,6 +3,8 @@
 
 #include <linux/mmzone.h>
 #include <linux/spinlock.h>
+#include <linux/mmzone.h>
+#include <linux/notifier.h>
 
 #ifdef CONFIG_MEMORY_HOTPLUG
 /*
@@ -46,6 +48,19 @@ static inline void zone_seqlock_init(str
 {
 	seqlock_init(&zone->span_seqlock);
 }
+extern int zone_grow_free_lists(struct zone *zone, unsigned long new_nr_pages);
+extern int zone_grow_waitqueues(struct zone *zone, unsigned long nr_pages);
+extern int add_one_highpage(struct page *page, int pfn, int bad_ppro);
+/* need some defines for these for archs that don't support it */
+extern void online_page(struct page *page);
+/* VM interface that may be used by firmware interface */
+extern int add_memory(u64 start, u64 size);
+extern int remove_memory(u64 start, u64 size);
+extern int online_pages(unsigned long, unsigned long);
+
+/* reasonably generic interface to expand the physical pages in a zone  */
+extern int __add_pages(struct zone *zone, unsigned long start_pfn,
+	unsigned long nr_pages);
 #else /* ! CONFIG_MEMORY_HOTPLUG */
 /*
  * Stub functions for when hotplug is off
@@ -65,5 +80,25 @@ static inline int zone_span_seqretry(str
 static inline void zone_span_writelock(struct zone *zone) {}
 static inline void zone_span_writeunlock(struct zone *zone) {}
 static inline void zone_seqlock_init(struct zone *zone) {}
+
+static inline int mhp_notimplemented(const char *func)
+{
+	printk(KERN_WARNING "%s() called, with CONFIG_MEMORY_HOTPLUG disabled\n", func);
+	dump_stack();
+	return -ENOSYS;
+}
+
+static inline int __add_pages(struct zone *zone, unsigned long start_pfn,
+	unsigned long nr_pages)
+{
+	return mhp_notimplemented(__FUNCTION__);
+}
 #endif /* ! CONFIG_MEMORY_HOTPLUG */
+static inline int __remove_pages(struct zone *zone, unsigned long start_pfn,
+	unsigned long nr_pages)
+{
+	printk(KERN_WARNING "%s() called, not yet supported\n", __FUNCTION__);
+	dump_stack();
+	return -ENOSYS;
+}
 #endif /* __LINUX_MEMORY_HOTPLUG_H */
diff -puN include/linux/mm.h~D0-sysfs-memory-class include/linux/mm.h
--- memhotplug/include/linux/mm.h~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/include/linux/mm.h	2005-08-18 14:59:48.000000000 -0700
@@ -791,6 +791,7 @@ extern void free_area_init_node(int nid,
 	unsigned long * zones_size, unsigned long zone_start_pfn, 
 	unsigned long *zholes_size);
 extern void memmap_init_zone(unsigned long, int, unsigned long, unsigned long);
+extern void setup_per_zone_pages_min(void);
 extern void mem_init(void);
 extern void show_mem(void);
 extern void si_meminfo(struct sysinfo * val);
diff -puN mm/Kconfig~D0-sysfs-memory-class mm/Kconfig
--- memhotplug/mm/Kconfig~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/mm/Kconfig	2005-08-18 14:59:48.000000000 -0700
@@ -111,3 +111,11 @@ config SPARSEMEM_STATIC
 config SPARSEMEM_EXTREME
 	def_bool y
 	depends on SPARSEMEM && !SPARSEMEM_STATIC
+
+# eventually, we can have this option just 'select SPARSEMEM'
+config MEMORY_HOTPLUG
+	bool "Allow for memory hot-add"
+	depends on SPARSEMEM && HOTPLUG && !SOFTWARE_SUSPEND
+
+comment "Memory hotplug is currently incompatible with Software Suspend"
+	depends on SPARSEMEM && HOTPLUG && SOFTWARE_SUSPEND
diff -puN mm/Makefile~D0-sysfs-memory-class mm/Makefile
--- memhotplug/mm/Makefile~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/mm/Makefile	2005-08-18 14:59:48.000000000 -0700
@@ -18,5 +18,5 @@ obj-$(CONFIG_NUMA) 	+= mempolicy.o
 obj-$(CONFIG_SPARSEMEM)	+= sparse.o
 obj-$(CONFIG_SHMEM) += shmem.o
 obj-$(CONFIG_TINY_SHMEM) += tiny-shmem.o
-
+obj-$(CONFIG_MEMORY_HOTPLUG) += memory_hotplug.o
 obj-$(CONFIG_FS_XIP) += filemap_xip.o
diff -puN /dev/null mm/memory_hotplug.c
--- /dev/null	2005-03-30 22:36:15.000000000 -0800
+++ memhotplug-dave/mm/memory_hotplug.c	2005-08-18 14:59:48.000000000 -0700
@@ -0,0 +1,178 @@
+/*
+ *  linux/mm/memory_hotplug.c
+ *
+ *  Copyright (C)
+ */
+
+#include <linux/config.h>
+#include <linux/stddef.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/interrupt.h>
+#include <linux/pagemap.h>
+#include <linux/bootmem.h>
+#include <linux/compiler.h>
+#include <linux/module.h>
+#include <linux/pagevec.h>
+#include <linux/slab.h>
+#include <linux/sysctl.h>
+#include <linux/cpu.h>
+#include <linux/memory.h>
+#include <linux/memory_hotplug.h>
+#include <linux/highmem.h>
+#include <linux/vmalloc.h>
+
+#include <asm/tlbflush.h>
+
+static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
+{
+	struct page *page, *ret;
+	unsigned long memmap_size = sizeof(struct page) * nr_pages;
+
+	page = alloc_pages(GFP_KERNEL, get_order(memmap_size));
+	if (page)
+		goto got_map_page;
+
+	ret = vmalloc(memmap_size);
+	if (ret)
+		goto got_map_ptr;
+
+	return NULL;
+got_map_page:
+	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
+got_map_ptr:
+	memset(ret, 0, memmap_size);
+
+	return ret;
+}
+
+extern void zonetable_add(struct zone *zone, int nid, int zid, unsigned long pfn,
+			  unsigned long size);
+static void __add_zone(struct zone *zone, unsigned long phys_start_pfn)
+{
+	struct pglist_data *pgdat = zone->zone_pgdat;
+	int nr_pages = PAGES_PER_SECTION;
+	int nid = pgdat->node_id;
+	int zone_type;
+
+	zone_type = zone - pgdat->node_zones;
+	memmap_init_zone(nr_pages, nid, zone_type, phys_start_pfn);
+	zonetable_add(zone, nid, zone_type, phys_start_pfn, nr_pages);
+}
+
+extern int sparse_add_one_section(struct zone *, unsigned long,
+				  struct page *mem_map);
+int __add_section(struct zone *zone, unsigned long phys_start_pfn)
+{
+	struct pglist_data *pgdat = zone->zone_pgdat;
+	int nr_pages = PAGES_PER_SECTION;
+	struct page *memmap;
+	int ret;
+
+	/*
+	 * This can potentially allocate memory, and does its own
+	 * internal locking.
+	 */
+	sparse_index_init(pfn_to_section_nr(phys_start_pfn), pgdat->node_id);
+
+	pgdat_resize_lock(pgdat, &flags);
+	memmap = __kmalloc_section_memmap(nr_pages);
+	ret = sparse_add_one_section(zone, phys_start_pfn, memmap);
+	pgdat_resize_unlock(pgdat, &flags);
+
+	if (ret <= 0) {
+		/* the mem_map didn't get used */
+		if (memmap >= (struct page *)VMALLOC_START &&
+		    memmap < (struct page *)VMALLOC_END)
+			vfree(memmap);
+		else
+			free_pages((unsigned long)memmap,
+				   get_order(sizeof(struct page) * nr_pages));
+	}
+
+	if (ret < 0)
+		return ret;
+
+	__add_zone(zone, phys_start_pfn);
+	return register_new_memory(__pfn_to_section(phys_start_pfn));
+}
+
+/*
+ * Reasonably generic function for adding memory.  It is
+ * expected that archs that support memory hotplug will
+ * call this function after deciding the zone to which to
+ * add the new pages.
+ */
+int __add_pages(struct zone *zone, unsigned long phys_start_pfn,
+		 unsigned long nr_pages)
+{
+	unsigned long i;
+	int err = 0;
+
+	for (i = 0; i < nr_pages; i += PAGES_PER_SECTION) {
+		err = __add_section(zone, phys_start_pfn + i);
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static void grow_zone_span(struct zone *zone,
+		unsigned long start_pfn, unsigned long end_pfn)
+{
+	unsigned long old_zone_end_pfn;
+
+	zone_span_writelock(zone);
+
+	old_zone_end_pfn = zone->zone_start_pfn + zone->spanned_pages;
+	if (start_pfn < zone->zone_start_pfn)
+		zone->zone_start_pfn = start_pfn;
+
+	if (end_pfn > old_zone_end_pfn)
+		zone->spanned_pages = end_pfn - zone->zone_start_pfn;
+
+	zone_span_writeunlock(zone);
+}
+
+static void grow_pgdat_span(struct pglist_data *pgdat,
+		unsigned long start_pfn, unsigned long end_pfn)
+{
+	unsigned long old_pgdat_end_pfn =
+		pgdat->node_start_pfn + pgdat->node_spanned_pages;
+
+	if (start_pfn < pgdat->node_start_pfn)
+		pgdat->node_start_pfn = start_pfn;
+
+	if (end_pfn > old_pgdat_end_pfn)
+		pgdat->node_spanned_pages = end_pfn - pgdat->node_spanned_pages;
+}
+
+int online_pages(unsigned long pfn, unsigned long nr_pages)
+{
+	unsigned long i;
+	unsigned long flags;
+	unsigned long onlined_pages = 0;
+	struct zone *zone;
+
+	/*
+	 * This doesn't need a lock to do pfn_to_page().
+	 * The section can't be removed here because of the
+	 * memory_block->state_sem.
+	 */
+	zone = page_zone(pfn_to_page(pfn));
+	pgdat_resize_lock(zone->zone_pgdat, &flags);
+	grow_zone_span(zone, pfn, pfn + nr_pages);
+	grow_pgdat_span(zone->zone_pgdat, pfn, pfn + nr_pages);
+	pgdat_resize_unlock(zone->zone_pgdat, &flags);
+
+	for (i = 0; i < nr_pages; i++) {
+		struct page *page = pfn_to_page(pfn + i);
+		online_page(page);
+		onlined_pages++;
+	}
+	zone->present_pages += onlined_pages;
+
+	return 0;
+}
diff -puN mm/page_alloc.c~D0-sysfs-memory-class mm/page_alloc.c
--- memhotplug/mm/page_alloc.c~D0-sysfs-memory-class	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/mm/page_alloc.c	2005-08-18 14:59:48.000000000 -0700
@@ -1663,7 +1663,7 @@ static void __init calculate_zone_totalp
  * up by free_all_bootmem() once the early boot process is
  * done. Non-atomic initialization, single-pass.
  */
-void __init memmap_init_zone(unsigned long size, int nid, unsigned long zone,
+void __devinit memmap_init_zone(unsigned long size, int nid, unsigned long zone,
 		unsigned long start_pfn)
 {
 	struct page *page;
@@ -2385,7 +2385,7 @@ static void setup_per_zone_lowmem_reserv
  *	that the pages_{min,low,high} values for each zone are set correctly 
  *	with respect to min_free_kbytes.
  */
-static void setup_per_zone_pages_min(void)
+void setup_per_zone_pages_min(void)
 {
 	unsigned long pages_min = min_free_kbytes >> (PAGE_SHIFT - 10);
 	unsigned long lowmem_pages = 0;
_

[PATCH 08/11] memory hotplug: move section_mem_map alloc to sparse.c

[Dave Hansen]

This basically keeps up from having to extern __kmalloc_section_memmap().

The vaddr_in_vmalloc_area() helper could go in a vmalloc header, but
that header gets hard to work with, because it needs some arch-specific
macros.  Just stick it in here for now, instead of creating another header.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/mm/memory_hotplug.c |   43 --------------------
 memhotplug-dave/mm/sparse.c         |   74 +++++++++++++++++++++++++++++++++---
 2 files changed, 70 insertions(+), 47 deletions(-)

diff -puN mm/memory_hotplug.c~D0.6-move_memmap_kmalloc_to_sparse.c mm/memory_hotplug.c
--- memhotplug/mm/memory_hotplug.c~D0.6-move_memmap_kmalloc_to_sparse.c	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/mm/memory_hotplug.c	2005-08-18 14:59:48.000000000 -0700
@@ -24,28 +24,6 @@
 
 #include <asm/tlbflush.h>
 
-static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
-{
-	struct page *page, *ret;
-	unsigned long memmap_size = sizeof(struct page) * nr_pages;
-
-	page = alloc_pages(GFP_KERNEL, get_order(memmap_size));
-	if (page)
-		goto got_map_page;
-
-	ret = vmalloc(memmap_size);
-	if (ret)
-		goto got_map_ptr;
-
-	return NULL;
-got_map_page:
-	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
-got_map_ptr:
-	memset(ret, 0, memmap_size);
-
-	return ret;
-}
-
 extern void zonetable_add(struct zone *zone, int nid, int zid, unsigned long pfn,
 			  unsigned long size);
 static void __add_zone(struct zone *zone, unsigned long phys_start_pfn)
@@ -60,7 +38,7 @@ static void __add_zone(struct zone *zone
 	zonetable_add(zone, nid, zone_type, phys_start_pfn, nr_pages);
 }
 
-extern int sparse_add_one_section(struct zone *, unsigned long,
+extern int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
 				  struct page *mem_map);
 int __add_section(struct zone *zone, unsigned long phys_start_pfn)
 {
@@ -69,26 +47,7 @@ int __add_section(struct zone *zone, uns
 	struct page *memmap;
 	int ret;
 
-	/*
-	 * This can potentially allocate memory, and does its own
-	 * internal locking.
-	 */
-	sparse_index_init(pfn_to_section_nr(phys_start_pfn), pgdat->node_id);
-
-	pgdat_resize_lock(pgdat, &flags);
-	memmap = __kmalloc_section_memmap(nr_pages);
 	ret = sparse_add_one_section(zone, phys_start_pfn, memmap);
-	pgdat_resize_unlock(pgdat, &flags);
-
-	if (ret <= 0) {
-		/* the mem_map didn't get used */
-		if (memmap >= (struct page *)VMALLOC_START &&
-		    memmap < (struct page *)VMALLOC_END)
-			vfree(memmap);
-		else
-			free_pages((unsigned long)memmap,
-				   get_order(sizeof(struct page) * nr_pages));
-	}
 
 	if (ret < 0)
 		return ret;
diff -puN mm/sparse.c~D0.6-move_memmap_kmalloc_to_sparse.c mm/sparse.c
--- memhotplug/mm/sparse.c~D0.6-move_memmap_kmalloc_to_sparse.c	2005-08-18 14:59:48.000000000 -0700
+++ memhotplug-dave/mm/sparse.c	2005-08-18 14:59:48.000000000 -0700
@@ -5,8 +5,10 @@
 #include <linux/mm.h>
 #include <linux/mmzone.h>
 #include <linux/bootmem.h>
+#include <linux/highmem.h>
 #include <linux/module.h>
 #include <linux/spinlock.h>
+#include <linux/vmalloc.h>
 #include <asm/dma.h>
 
 /*
@@ -164,6 +166,45 @@ static struct page *sparse_early_mem_map
 	return NULL;
 }
 
+static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
+{
+	struct page *page, *ret;
+	unsigned long memmap_size = sizeof(struct page) * nr_pages;
+
+	page = alloc_pages(GFP_KERNEL, get_order(memmap_size));
+	if (page)
+		goto got_map_page;
+
+	ret = vmalloc(memmap_size);
+	if (ret)
+		goto got_map_ptr;
+
+	return NULL;
+got_map_page:
+	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
+got_map_ptr:
+	memset(ret, 0, memmap_size);
+
+	return ret;
+}
+
+static int vaddr_in_vmalloc_area(void *addr)
+{
+	if (addr >= (void *)VMALLOC_START &&
+	    addr < (void *)VMALLOC_END)
+		return 1;
+	return 0;
+}
+
+static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
+{
+	if (vaddr_in_vmalloc_area(memmap))
+		vfree(memmap);
+	else
+		free_pages((unsigned long)memmap,
+			   get_order(sizeof(struct page) * nr_pages));
+}
+
 /*
  * Allocate the accumulated non-linear sections, allocate a mem_map
  * for each and record the physical to section mapping.
@@ -189,14 +230,37 @@ void sparse_init(void)
  * set.  If this is <=0, then that means that the passed-in
  * map was not consumed and must be freed.
  */
-int sparse_add_one_section(unsigned long start_pfn, int nr_pages, struct page *map)
+int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
+			   int nr_pages)
 {
-	struct mem_section *ms = __pfn_to_section(start_pfn);
+	unsigned long section_nr = pfn_to_section_nr(start_pfn);
+	struct pglist_data *pgdat = zone->zone_pgdat;
+	struct mem_section *ms;
+	struct page *memmap;
+	unsigned long flags;
+	int ret;
+
+	/*
+	 * no locking for this, because it does its own
+	 * plus, it does a kmalloc
+	 */
+	sparse_index_init(section_nr, pgdat->node_id);
+	memmap = __kmalloc_section_memmap(nr_pages);
 
-	if (ms->section_mem_map & SECTION_MARKED_PRESENT)
-		return -EEXIST;
+	pgdat_resize_lock(pgdat, &flags);
 
+	ms = __pfn_to_section(start_pfn);
+	if (ms->section_mem_map & SECTION_MARKED_PRESENT) {
+		ret = -EEXIST;
+		goto out;
+	}
 	ms->section_mem_map |= SECTION_MARKED_PRESENT;
 
-	return sparse_init_one_section(ms, pfn_to_section_nr(start_pfn), map);
+	ret = sparse_init_one_section(ms, section_nr, memmap);
+
+	if (ret <= 0)
+		__kfree_section_memmap(memmap, nr_pages);
+out:
+	pgdat_resize_unlock(pgdat, &flags);
+	return ret;
 }
_

[PATCH 09/11] memory hotplug: call setup_per_zone_pages_min after hotplug

[Dave Hansen]

From: IWAMOTO Toshihiro <iwamoto@valinux.co.jp>
> I found the tests does not work well with Dave's patchset.
> I've found the followings:
> 
> 	- setup_per_zone_pages_min() calls should be added in
> 	   capture_page_range() and online_pages()
> 	- lru_add_drain() should be called before try_to_migrate_pages()

The following patch deals with the first item.

Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/mm/memory_hotplug.c |    2 ++
 1 files changed, 2 insertions(+)

diff -puN mm/memory_hotplug.c~D0.7-call_setup_per_zone_pages_min_after_memory_size_change mm/memory_hotplug.c
--- memhotplug/mm/memory_hotplug.c~D0.7-call_setup_per_zone_pages_min_after_memory_size_change	2005-08-18 14:59:49.000000000 -0700
+++ memhotplug-dave/mm/memory_hotplug.c	2005-08-18 14:59:49.000000000 -0700
@@ -133,5 +133,7 @@ int online_pages(unsigned long pfn, unsi
 	}
 	zone->present_pages += onlined_pages;
 
+	setup_per_zone_pages_min();
+
 	return 0;
 }
_

[PATCH 10/11] memory hotplug: i386 addition functions

[Dave Hansen]

Adds the necessary for non-NUMA hot-add of highmem
to an existing zone on i386.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
---

 memhotplug-dave/arch/i386/mm/discontig.c |    4 +-
 memhotplug-dave/arch/i386/mm/init.c      |   61 ++++++++++++++++++++++++++++---
 2 files changed, 58 insertions(+), 7 deletions(-)

diff -puN arch/i386/mm/discontig.c~D1-i386-hotplug-functions arch/i386/mm/discontig.c
--- memhotplug/arch/i386/mm/discontig.c~D1-i386-hotplug-functions	2005-08-18 14:59:50.000000000 -0700
+++ memhotplug-dave/arch/i386/mm/discontig.c	2005-08-18 14:59:50.000000000 -0700
@@ -98,7 +98,7 @@ unsigned long node_memmap_size_bytes(int
 
 extern unsigned long find_max_low_pfn(void);
 extern void find_max_pfn(void);
-extern void one_highpage_init(struct page *, int, int);
+extern void add_one_highpage_init(struct page *, int, int);
 
 extern struct e820map e820;
 extern unsigned long init_pg_tables_end;
@@ -427,7 +427,7 @@ void __init set_highmem_pages_init(int b
 			if (!pfn_valid(node_pfn))
 				continue;
 			page = pfn_to_page(node_pfn);
-			one_highpage_init(page, node_pfn, bad_ppro);
+			add_one_highpage_init(page, node_pfn, bad_ppro);
 		}
 	}
 	totalram_pages += totalhigh_pages;
diff -puN arch/i386/mm/init.c~D1-i386-hotplug-functions arch/i386/mm/init.c
--- memhotplug/arch/i386/mm/init.c~D1-i386-hotplug-functions	2005-08-18 14:59:50.000000000 -0700
+++ memhotplug-dave/arch/i386/mm/init.c	2005-08-18 14:59:50.000000000 -0700
@@ -27,6 +27,7 @@
 #include <linux/slab.h>
 #include <linux/proc_fs.h>
 #include <linux/efi.h>
+#include <linux/memory_hotplug.h>
 
 #include <asm/processor.h>
 #include <asm/system.h>
@@ -265,17 +266,45 @@ static void __init permanent_kmaps_init(
 	pkmap_page_table = pte;	
 }
 
-void __init one_highpage_init(struct page *page, int pfn, int bad_ppro)
+void __devinit free_new_highpage(struct page *page)
+{
+	set_page_count(page, 1);
+	__free_page(page);
+	totalhigh_pages++;
+}
+
+void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro)
 {
 	if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
 		ClearPageReserved(page);
-		set_page_count(page, 1);
-		__free_page(page);
-		totalhigh_pages++;
 	} else
 		SetPageReserved(page);
 }
 
+int add_one_highpage_hotplug(struct page *page, int pfn)
+{
+	free_new_highpage(page);
+	totalram_pages++;
+#ifdef CONFIG_FLATMEM
+	max_mapnr = max(pfn, max_mapnr);
+#endif
+	num_physpages++;
+	return 0;
+}
+
+/*
+ * Not currently handling the NUMA case.
+ * Assuming single node and all memory that
+ * has been added dynamically that would be
+ * onlined here is in HIGHMEM
+ */
+void online_page(struct page *page)
+{
+	ClearPageReserved(page);
+	add_one_highpage_hotplug(page, page_to_pfn(page));
+}
+
+
 #ifdef CONFIG_NUMA
 extern void set_highmem_pages_init(int);
 #else
@@ -283,7 +312,7 @@ static void __init set_highmem_pages_ini
 {
 	int pfn;
 	for (pfn = highstart_pfn; pfn < highend_pfn; pfn++)
-		one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
+		add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
 	totalram_pages += totalhigh_pages;
 }
 #endif /* CONFIG_FLATMEM */
@@ -614,6 +643,28 @@ void __init mem_init(void)
 #endif
 }
 
+/*
+ * this is for the non-NUMA, single node SMP system case.
+ * Specifically, in the case of x86, we will always add
+ * memory to the highmem for now.
+ */
+#ifndef CONFIG_NEED_MULTIPLE_NODES
+int add_memory(u64 start, u64 size)
+{
+	struct pglist_data *pgdata = &contig_page_data;
+	struct zone *zone = pgdata->node_zones + MAX_NR_ZONES-1;
+	unsigned long start_pfn = start >> PAGE_SHIFT;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+
+	return __add_pages(zone, start_pfn, nr_pages);
+}
+
+int remove_memory(u64 start, u64 size)
+{
+	return -EINVAL;
+}
+#endif
+
 kmem_cache_t *pgd_cache;
 kmem_cache_t *pmd_cache;
 
_

[PATCH 11/11] memory hotplug: ppc64 specific hot-add functions

[Dave Hansen]

Here is a set of ppc64 specific patches that at least allow
compilation/booting with the following configurations:

FLATMEM
SPARSEMEN
SPARSEMEM + MEMORY_HOTPLUG

Signed-off-by: Mike Kravetz <kravetz@us.ibm.com>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>

---

 memhotplug-dave/arch/ppc64/mm/init.c         |   77 +++++++++++++++++++++++++++
 memhotplug-dave/include/asm-ppc64/abs_addr.h |    2 
 2 files changed, 79 insertions(+)

diff -puN arch/ppc64/mm/init.c~D2-ppc64-hotplug-functions arch/ppc64/mm/init.c
--- memhotplug/arch/ppc64/mm/init.c~D2-ppc64-hotplug-functions	2005-08-18 14:59:50.000000000 -0700
+++ memhotplug-dave/arch/ppc64/mm/init.c	2005-08-18 14:59:50.000000000 -0700
@@ -870,3 +870,80 @@ pgprot_t phys_mem_access_prot(struct fil
 	return vma_prot;
 }
 EXPORT_SYMBOL(phys_mem_access_prot);
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+
+void online_page(struct page *page)
+{
+	ClearPageReserved(page);
+	free_cold_page(page);
+	totalram_pages++;
+	num_physpages++;
+}
+
+/*
+ * This works only for the non-NUMA case.  Later, we'll need a lookup
+ * to convert from real physical addresses to nid, that doesn't use
+ * pfn_to_nid().
+ */
+int __devinit add_memory(u64 start, u64 size)
+{
+	struct pglist_data *pgdata = NODE_DATA(0);
+	struct zone *zone;
+	unsigned long start_pfn = start >> PAGE_SHIFT;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+
+	/* this should work for most non-highmem platforms */
+	zone = pgdata->node_zones;
+
+	return __add_pages(zone, start_pfn, nr_pages);
+
+	return 0;
+}
+
+/*
+ * First pass at this code will check to determine if the remove
+ * request is within the RMO.  Do not allow removal within the RMO.
+ */
+int __devinit remove_memory(u64 start, u64 size)
+{
+	struct zone *zone;
+	unsigned long start_pfn, end_pfn, nr_pages;
+
+	start_pfn = start >> PAGE_SHIFT;
+	nr_pages = size >> PAGE_SHIFT;
+	end_pfn = start_pfn + nr_pages;
+
+	printk("%s(): Attempting to remove memoy in range "
+			"%lx to %lx\n", __func__, start, start+size);
+	/*
+	 * check for range within RMO
+	 */
+	zone = page_zone(pfn_to_page(start_pfn));
+
+	printk("%s(): memory will be removed from "
+			"the %s zone\n", __func__, zone->name);
+
+	/*
+	 * not handling removing memory ranges that
+	 * overlap multiple zones yet
+	 */
+	if (end_pfn > (zone->zone_start_pfn + zone->spanned_pages))
+		goto overlap;
+
+	/* make sure it is NOT in RMO */
+	if ((start < lmb.rmo_size) || ((start+size) < lmb.rmo_size)) {
+		printk("%s(): range to be removed must NOT be in RMO!\n",
+			__func__);
+		goto in_rmo;
+	}
+
+	return __remove_pages(zone, start_pfn, nr_pages);
+
+overlap:
+	printk("%s(): memory range to be removed overlaps "
+		"multiple zones!!!\n", __func__);
+in_rmo:
+	return -1;
+}
+#endif /* CONFIG_MEMORY_HOTPLUG */

Re: [PATCH 07/11] memory hotplug: sysfs and add/remove functions

[Andrew Morton]

Dave Hansen <haveblue@us.ibm.com> wrote:
>
> +		for (i = 0; i < PAGES_PER_SECTION; i++) {
> +			if (PageReserved(first_page+i))
> +				continue;

How intimate do these patches get with PageReserved()?  Bear in mind that
we're slowly working toward making PageReserved go away.

Re: [PATCH 07/11] memory hotplug: sysfs and add/remove functions

[Dave Hansen]

On Fri, 2005-09-02 at 15:13 -0700, Andrew Morton wrote:
> Dave Hansen <haveblue@us.ibm.com> wrote:
> >
> > +		for (i = 0; i < PAGES_PER_SECTION; i++) {
> > +			if (PageReserved(first_page+i))
> > +				continue;
> 
> How intimate do these patches get with PageReserved()?  Bear in mind that
> we're slowly working toward making PageReserved go away.

It's basically the same way that the init code uses it.  When
initialized, a struct page has it set.  In theory, an architecture could
decide to keep the bit set when it is doing online_pages().  However, I
don't think any do that today.  Nobody would really notice if we killed
that.  That check could probably instead be something like
page_is_ram().

-- Dave

Re: [PATCH 03/11] memory hotplug prep: __section_nr helper

[Denis Vlasenko]

On Friday 02 September 2005 23:56, Dave Hansen wrote:
> 
> A little helper that we use in the hotplug code.
> 
> Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
> ---
> 
>  memhotplug-dave/include/linux/mmzone.h |   25 +++++++++++++++++++++++++
>  1 files changed, 25 insertions(+)
> 
> diff -puN include/linux/mmzone.h~C3-__section_nr include/linux/mmzone.h
> --- memhotplug/include/linux/mmzone.h~C3-__section_nr	2005-08-18 14:59:45.000000000 -0700
> +++ memhotplug-dave/include/linux/mmzone.h	2005-08-18 14:59:45.000000000 -0700
> @@ -511,6 +511,31 @@ static inline struct mem_section *__nr_t
>  }
>  
>  /*
> + * Although written for the SPARSEMEM_EXTREME case, this happens
> + * to also work for the flat array case becase
> + * NR_SECTION_ROOTS==NR_MEM_SECTIONS.
> + */
> +static inline int __section_nr(struct mem_section* ms)
> +{
> +	unsigned long root_nr;
> +	struct mem_section* root;
> +
> +	for (root_nr = 0;
> +	     root_nr < NR_MEM_SECTIONS;
> +	     root_nr += SECTIONS_PER_ROOT) {
> +		root = __nr_to_section(root_nr);
> +
> +		if (!root)
> +			continue;
> +
> +		if ((ms >= root) && (ms < (root + SECTIONS_PER_ROOT)))
> +		     break;
> +	}
> +
> +	return (root_nr * SECTIONS_PER_ROOT) + (ms - root);
> +}
> +
> +/*

isn't it too much for the inlining?
--
vda

Re: [PATCH 01/11] memory hotplug prep: kill local_mapnr

[Andrew Morton]

Dave Hansen <haveblue@us.ibm.com> wrote:
>
>  --- memhotplug/include/asm-x86_64/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
>  +++ memhotplug-dave/include/asm-x86_64/mmzone.h	2005-08-18 14:59:43.000000000 -0700
>  @@ -38,8 +38,6 @@ static inline __attribute__((pure)) int 
>   
>   #ifdef CONFIG_DISCONTIGMEM
>   
>  -#define pfn_to_nid(pfn) phys_to_nid((unsigned long)(pfn) << PAGE_SHIFT)
>  -#define kvaddr_to_nid(kaddr)	phys_to_nid(__pa(kaddr))
>   
>   /* AK: this currently doesn't deal with invalid addresses. We'll see 
>      if the 2.5 kernel doesn't pass them
>  _

What's this bit doing here?   It breaks the x86_64 build all over the place.

I'll drop that chunk and see how we go...

Re: [PATCH 01/11] memory hotplug prep: kill local_mapnr

[Dave Hansen]

On Wed, 2005-09-07 at 02:37 -0700, Andrew Morton wrote:
> Dave Hansen <haveblue@us.ibm.com> wrote:
> >
> >  --- memhotplug/include/asm-x86_64/mmzone.h~C0-kill-local_mapnr	2005-08-18 14:59:43.000000000 -0700
> >  +++ memhotplug-dave/include/asm-x86_64/mmzone.h	2005-08-18 14:59:43.000000000 -0700
> >  @@ -38,8 +38,6 @@ static inline __attribute__((pure)) int 
> >   
> >   #ifdef CONFIG_DISCONTIGMEM
> >   
> >  -#define pfn_to_nid(pfn) phys_to_nid((unsigned long)(pfn) << PAGE_SHIFT)
> >  -#define kvaddr_to_nid(kaddr)	phys_to_nid(__pa(kaddr))
> >   
> >   /* AK: this currently doesn't deal with invalid addresses. We'll see 
> >      if the 2.5 kernel doesn't pass them
> >  _
> 
> What's this bit doing here?   It breaks the x86_64 build all over the place.
> 
> I'll drop that chunk and see how we go...

That could have easily been some merge borkage on my part.  I don't
think that hunk is valid, so dropping it is the right move.

-- Dave