[deliverable/linux.git] / arch / powerpc / platforms / pseries / hotplug-memory.c

/*
 * pseries Memory Hotplug infrastructure.
 *
 * Copyright (C) 2008 Badari Pulavarty, IBM Corporation
 *
 *      This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/of.h>
#include <linux/of_address.h>
#include <linux/memblock.h>
#include <linux/vmalloc.h>
#include <linux/memory.h>

#include <asm/firmware.h>
#include <asm/machdep.h>
#include <asm/prom.h>
#include <asm/sparsemem.h>

static unsigned long get_memblock_size(void)
{
	struct device_node *np;
	unsigned int memblock_size = MIN_MEMORY_BLOCK_SIZE;
	struct resource r;

	np = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory");
	if (np) {
		const __be64 *size;

		size = of_get_property(np, "ibm,lmb-size", NULL);
		if (size)
			memblock_size = be64_to_cpup(size);
		of_node_put(np);
	} else  if (machine_is(pseries)) {
		/* This fallback really only applies to pseries */
		unsigned int memzero_size = 0;

		np = of_find_node_by_path("/memory@0");
		if (np) {
			if (!of_address_to_resource(np, 0, &r))
				memzero_size = resource_size(&r);
			of_node_put(np);
		}

		if (memzero_size) {
			/* We now know the size of memory@0, use this to find
			 * the first memoryblock and get its size.
			 */
			char buf[64];

			sprintf(buf, "/memory@%x", memzero_size);
			np = of_find_node_by_path(buf);
			if (np) {
				if (!of_address_to_resource(np, 0, &r))
					memblock_size = resource_size(&r);
				of_node_put(np);
			}
		}
	}
	return memblock_size;
}

/* WARNING: This is going to override the generic definition whenever
 * pseries is built-in regardless of what platform is active at boot
 * time. This is fine for now as this is the only "option" and it
 * should work everywhere. If not, we'll have to turn this into a
 * ppc_md. callback
 */
unsigned long memory_block_size_bytes(void)
{
	return get_memblock_size();
}

#ifdef CONFIG_MEMORY_HOTREMOVE
static int pseries_remove_memblock(unsigned long base, unsigned int memblock_size)
{
	unsigned long start, start_pfn;
	struct zone *zone;
	int ret;
	unsigned long section;
	unsigned long sections_to_remove;

	start_pfn = base >> PAGE_SHIFT;

	if (!pfn_valid(start_pfn)) {
		memblock_remove(base, memblock_size);
		return 0;
	}

	zone = page_zone(pfn_to_page(start_pfn));

	/*
	 * Remove section mappings and sysfs entries for the
	 * section of the memory we are removing.
	 *
	 * NOTE: Ideally, this should be done in generic code like
	 * remove_memory(). But remove_memory() gets called by writing
	 * to sysfs "state" file and we can't remove sysfs entries
	 * while writing to it. So we have to defer it to here.
	 */
	sections_to_remove = (memblock_size >> PAGE_SHIFT) / PAGES_PER_SECTION;
	for (section = 0; section < sections_to_remove; section++) {
		unsigned long pfn = start_pfn + section * PAGES_PER_SECTION;
		ret = __remove_pages(zone, pfn, PAGES_PER_SECTION);
		if (ret)
			return ret;
	}

	/*
	 * Update memory regions for memory remove
	 */
	memblock_remove(base, memblock_size);

	/*
	 * Remove htab bolted mappings for this section of memory
	 */
	start = (unsigned long)__va(base);
	ret = remove_section_mapping(start, start + memblock_size);

	/* Ensure all vmalloc mappings are flushed in case they also
	 * hit that section of memory
	 */
	vm_unmap_aliases();

	return ret;
}

static int pseries_remove_memory(struct device_node *np)
{
	const char *type;
	const unsigned int *regs;
	unsigned long base;
	unsigned int lmb_size;
	int ret = -EINVAL;

	/*
	 * Check to see if we are actually removing memory
	 */
	type = of_get_property(np, "device_type", NULL);
	if (type == NULL || strcmp(type, "memory") != 0)
		return 0;

	/*
	 * Find the bae address and size of the memblock
	 */
	regs = of_get_property(np, "reg", NULL);
	if (!regs)
		return ret;

	base = *(unsigned long *)regs;
	lmb_size = regs[3];

	ret = pseries_remove_memblock(base, lmb_size);
	return ret;
}
#else
static inline int pseries_remove_memblock(unsigned long base,
					  unsigned int memblock_size)
{
	return -EOPNOTSUPP;
}
static inline int pseries_remove_memory(struct device_node *np)
{
	return -EOPNOTSUPP;
}
#endif /* CONFIG_MEMORY_HOTREMOVE */

static int pseries_add_memory(struct device_node *np)
{
	const char *type;
	const unsigned int *regs;
	unsigned long base;
	unsigned int lmb_size;
	int ret = -EINVAL;

	/*
	 * Check to see if we are actually adding memory
	 */
	type = of_get_property(np, "device_type", NULL);
	if (type == NULL || strcmp(type, "memory") != 0)
		return 0;

	/*
	 * Find the base and size of the memblock
	 */
	regs = of_get_property(np, "reg", NULL);
	if (!regs)
		return ret;

	base = *(unsigned long *)regs;
	lmb_size = regs[3];

	/*
	 * Update memory region to represent the memory add
	 */
	ret = memblock_add(base, lmb_size);
	return (ret < 0) ? -EINVAL : 0;
}

static int pseries_update_drconf_memory(struct of_prop_reconfig *pr)
{
	struct of_drconf_cell *new_drmem, *old_drmem;
	unsigned long memblock_size;
	u32 entries;
	u32 *p;
	int i, rc = -EINVAL;

	memblock_size = get_memblock_size();
	if (!memblock_size)
		return -EINVAL;

	p = (u32 *)of_get_property(pr->dn, "ibm,dynamic-memory", NULL);
	if (!p)
		return -EINVAL;

	/* The first int of the property is the number of lmb's described
	 * by the property. This is followed by an array of of_drconf_cell
	 * entries. Get the niumber of entries and skip to the array of
	 * of_drconf_cell's.
	 */
	entries = *p++;
	old_drmem = (struct of_drconf_cell *)p;

	p = (u32 *)pr->prop->value;
	p++;
	new_drmem = (struct of_drconf_cell *)p;

	for (i = 0; i < entries; i++) {
		if ((old_drmem[i].flags & DRCONF_MEM_ASSIGNED) &&
		    (!(new_drmem[i].flags & DRCONF_MEM_ASSIGNED))) {
			rc = pseries_remove_memblock(old_drmem[i].base_addr,
						     memblock_size);
			break;
		} else if ((!(old_drmem[i].flags & DRCONF_MEM_ASSIGNED)) &&
			   (new_drmem[i].flags & DRCONF_MEM_ASSIGNED)) {
			rc = memblock_add(old_drmem[i].base_addr,
					  memblock_size);
			rc = (rc < 0) ? -EINVAL : 0;
			break;
		}
	}

	return rc;
}

static int pseries_memory_notifier(struct notifier_block *nb,
				   unsigned long action, void *node)
{
	struct of_prop_reconfig *pr;
	int err = 0;

	switch (action) {
	case OF_RECONFIG_ATTACH_NODE:
		err = pseries_add_memory(node);
		break;
	case OF_RECONFIG_DETACH_NODE:
		err = pseries_remove_memory(node);
		break;
	case OF_RECONFIG_UPDATE_PROPERTY:
		pr = (struct of_prop_reconfig *)node;
		if (!strcmp(pr->prop->name, "ibm,dynamic-memory"))
			err = pseries_update_drconf_memory(pr);
		break;
	}
	return notifier_from_errno(err);
}

static struct notifier_block pseries_mem_nb = {
	.notifier_call = pseries_memory_notifier,
};

static int __init pseries_memory_hotplug_init(void)
{
	if (firmware_has_feature(FW_FEATURE_LPAR))
		of_reconfig_notifier_register(&pseries_mem_nb);

	return 0;
}
machine_device_initcall(pseries, pseries_memory_hotplug_init);
Commit	Line	Data
57b53926 BP	1	/*
	2	* pseries Memory Hotplug infrastructure.
	3	*
	4	* Copyright (C) 2008 Badari Pulavarty, IBM Corporation
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public License
	8	* as published by the Free Software Foundation; either version
	9	* 2 of the License, or (at your option) any later version.
	10	*/
	11
	12	#include <linux/of.h>
26a2056e	13	#include <linux/of_address.h>
95f72d1e	14	#include <linux/memblock.h>
b4a26be9	15	#include <linux/vmalloc.h>
770e1ac5 BH	16	#include <linux/memory.h>
770e1ac5 BH	17
57b53926 BP	18	#include <asm/firmware.h>
57b53926 BP	19	#include <asm/machdep.h>
26a2056e	20	#include <asm/prom.h>
0b2f8287	21	#include <asm/sparsemem.h>
57b53926	22
c540ada2 NF	23	static unsigned long get_memblock_size(void)
	24	{
	25	struct device_node *np;
770e1ac5 BH	26	unsigned int memblock_size = MIN_MEMORY_BLOCK_SIZE;
770e1ac5 BH	27	struct resource r;
c540ada2 NF	28
	29	np = of_find_node_by_path("/ibm,dynamic-reconfiguration-memory");
	30	if (np) {
770e1ac5	31	const __be64 *size;
c540ada2 NF	32
c540ada2 NF	33	size = of_get_property(np, "ibm,lmb-size", NULL);
770e1ac5 BH	34	if (size)
770e1ac5 BH	35	memblock_size = be64_to_cpup(size);
c540ada2	36	of_node_put(np);
770e1ac5 BH	37	} else if (machine_is(pseries)) {
770e1ac5 BH	38	/* This fallback really only applies to pseries */
c540ada2	39	unsigned int memzero_size = 0;
c540ada2 NF	40
	41	np = of_find_node_by_path("/memory@0");
	42	if (np) {
770e1ac5 BH	43	if (!of_address_to_resource(np, 0, &r))
770e1ac5 BH	44	memzero_size = resource_size(&r);
c540ada2 NF	45	of_node_put(np);
	46	}
	47
	48	if (memzero_size) {
	49	/* We now know the size of memory@0, use this to find
	50	* the first memoryblock and get its size.
	51	*/
	52	char buf[64];
	53
	54	sprintf(buf, "/memory@%x", memzero_size);
	55	np = of_find_node_by_path(buf);
	56	if (np) {
770e1ac5 BH	57	if (!of_address_to_resource(np, 0, &r))
770e1ac5 BH	58	memblock_size = resource_size(&r);
c540ada2 NF	59	of_node_put(np);
	60	}
	61	}
	62	}
c540ada2 NF	63	return memblock_size;
	64	}
	65
770e1ac5 BH	66	/* WARNING: This is going to override the generic definition whenever
	67	* pseries is built-in regardless of what platform is active at boot
	68	* time. This is fine for now as this is the only "option" and it
	69	* should work everywhere. If not, we'll have to turn this into a
	70	* ppc_md. callback
	71	*/
c540ada2 NF	72	unsigned long memory_block_size_bytes(void)
	73	{
	74	return get_memblock_size();
	75	}
	76
4edd7cef	77	#ifdef CONFIG_MEMORY_HOTREMOVE
95f72d1e	78	static int pseries_remove_memblock(unsigned long base, unsigned int memblock_size)
57b53926	79	{
3c3f67ea	80	unsigned long start, start_pfn;
57b53926	81	struct zone *zone;
1633dbba YI	82	int ret;
	83	unsigned long section;
	84	unsigned long sections_to_remove;
92ecd179	85
9fd3f88c	86	start_pfn = base >> PAGE_SHIFT;
04badfd2 NF	87
04badfd2 NF	88	if (!pfn_valid(start_pfn)) {
95f72d1e	89	memblock_remove(base, memblock_size);
04badfd2 NF	90	return 0;
	91	}
	92
57b53926 BP	93	zone = page_zone(pfn_to_page(start_pfn));
	94
	95	/*
	96	* Remove section mappings and sysfs entries for the
	97	* section of the memory we are removing.
	98	*
	99	* NOTE: Ideally, this should be done in generic code like
	100	* remove_memory(). But remove_memory() gets called by writing
	101	* to sysfs "state" file and we can't remove sysfs entries
	102	* while writing to it. So we have to defer it to here.
	103	*/
d760afd4	104	sections_to_remove = (memblock_size >> PAGE_SHIFT) / PAGES_PER_SECTION;
1633dbba YI	105	for (section = 0; section < sections_to_remove; section++) {
1633dbba YI	106	unsigned long pfn = start_pfn + section * PAGES_PER_SECTION;
158544b1	107	ret = __remove_pages(zone, pfn, PAGES_PER_SECTION);
d760afd4 YI	108	if (ret)
	109	return ret;
	110	}
57b53926	111
98d5c21c BP	112	/*
	113	* Update memory regions for memory remove
	114	*/
95f72d1e	115	memblock_remove(base, memblock_size);
98d5c21c	116
57b53926 BP	117	/*
	118	* Remove htab bolted mappings for this section of memory
	119	*/
92ecd179	120	start = (unsigned long)__va(base);
95f72d1e	121	ret = remove_section_mapping(start, start + memblock_size);
b4a26be9 BH	122
	123	/* Ensure all vmalloc mappings are flushed in case they also
	124	* hit that section of memory
	125	*/
	126	vm_unmap_aliases();
	127
57b53926 BP	128	return ret;
	129	}
	130
3c3f67ea NF	131	static int pseries_remove_memory(struct device_node *np)
	132	{
	133	const char *type;
	134	const unsigned int *regs;
	135	unsigned long base;
3fdfd990	136	unsigned int lmb_size;
3c3f67ea NF	137	int ret = -EINVAL;
	138
	139	/*
	140	* Check to see if we are actually removing memory
	141	*/
	142	type = of_get_property(np, "device_type", NULL);
	143	if (type == NULL \|\| strcmp(type, "memory") != 0)
	144	return 0;
	145
	146	/*
95f72d1e	147	* Find the bae address and size of the memblock
3c3f67ea NF	148	*/
	149	regs = of_get_property(np, "reg", NULL);
	150	if (!regs)
	151	return ret;
	152
	153	base = (unsigned long )regs;
3fdfd990	154	lmb_size = regs[3];
3c3f67ea	155
3fdfd990	156	ret = pseries_remove_memblock(base, lmb_size);
3c3f67ea NF	157	return ret;
3c3f67ea NF	158	}
4edd7cef DR	159	#else
	160	static inline int pseries_remove_memblock(unsigned long base,
	161	unsigned int memblock_size)
	162	{
	163	return -EOPNOTSUPP;
	164	}
	165	static inline int pseries_remove_memory(struct device_node *np)
	166	{
	167	return -EOPNOTSUPP;
	168	}
	169	#endif /* CONFIG_MEMORY_HOTREMOVE */
3c3f67ea	170
98d5c21c BP	171	static int pseries_add_memory(struct device_node *np)
	172	{
	173	const char *type;
98d5c21c	174	const unsigned int *regs;
92ecd179	175	unsigned long base;
3fdfd990	176	unsigned int lmb_size;
98d5c21c BP	177	int ret = -EINVAL;
	178
	179	/*
	180	* Check to see if we are actually adding memory
	181	*/
	182	type = of_get_property(np, "device_type", NULL);
	183	if (type == NULL \|\| strcmp(type, "memory") != 0)
	184	return 0;
	185
	186	/*
95f72d1e	187	* Find the base and size of the memblock
98d5c21c	188	*/
98d5c21c BP	189	regs = of_get_property(np, "reg", NULL);
	190	if (!regs)
	191	return ret;
	192
92ecd179	193	base = (unsigned long )regs;
3fdfd990	194	lmb_size = regs[3];
98d5c21c BP	195
	196	/*
	197	* Update memory region to represent the memory add
	198	*/
3fdfd990	199	ret = memblock_add(base, lmb_size);
3c3f67ea NF	200	return (ret < 0) ? -EINVAL : 0;
	201	}
	202
1cf3d8b3	203	static int pseries_update_drconf_memory(struct of_prop_reconfig *pr)
3c3f67ea	204	{
1cf3d8b3	205	struct of_drconf_cell new_drmem, old_drmem;
c540ada2	206	unsigned long memblock_size;
1cf3d8b3 NF	207	u32 entries;
	208	u32 *p;
	209	int i, rc = -EINVAL;
3c3f67ea	210
c540ada2 NF	211	memblock_size = get_memblock_size();
c540ada2 NF	212	if (!memblock_size)
3c3f67ea NF	213	return -EINVAL;
3c3f67ea NF	214
1cf3d8b3 NF	215	p = (u32 *)of_get_property(pr->dn, "ibm,dynamic-memory", NULL);
	216	if (!p)
	217	return -EINVAL;
	218
	219	/* The first int of the property is the number of lmb's described
	220	* by the property. This is followed by an array of of_drconf_cell
	221	* entries. Get the niumber of entries and skip to the array of
	222	* of_drconf_cell's.
	223	*/
	224	entries = *p++;
	225	old_drmem = (struct of_drconf_cell *)p;
	226
	227	p = (u32 *)pr->prop->value;
	228	p++;
	229	new_drmem = (struct of_drconf_cell *)p;
	230
	231	for (i = 0; i < entries; i++) {
	232	if ((old_drmem[i].flags & DRCONF_MEM_ASSIGNED) &&
	233	(!(new_drmem[i].flags & DRCONF_MEM_ASSIGNED))) {
	234	rc = pseries_remove_memblock(old_drmem[i].base_addr,
	235	memblock_size);
	236	break;
	237	} else if ((!(old_drmem[i].flags & DRCONF_MEM_ASSIGNED)) &&
	238	(new_drmem[i].flags & DRCONF_MEM_ASSIGNED)) {
	239	rc = memblock_add(old_drmem[i].base_addr,
	240	memblock_size);
	241	rc = (rc < 0) ? -EINVAL : 0;
	242	break;
	243	}
3c3f67ea NF	244	}
3c3f67ea NF	245
3c3f67ea	246	return rc;
98d5c21c BP	247	}
98d5c21c BP	248
57b53926	249	static int pseries_memory_notifier(struct notifier_block *nb,
1cf3d8b3	250	unsigned long action, void *node)
57b53926	251	{
1cf3d8b3	252	struct of_prop_reconfig *pr;
de2780a3	253	int err = 0;
57b53926 BP	254
57b53926 BP	255	switch (action) {
1cf3d8b3	256	case OF_RECONFIG_ATTACH_NODE:
de2780a3	257	err = pseries_add_memory(node);
57b53926	258	break;
1cf3d8b3	259	case OF_RECONFIG_DETACH_NODE:
de2780a3	260	err = pseries_remove_memory(node);
57b53926	261	break;
1cf3d8b3 NF	262	case OF_RECONFIG_UPDATE_PROPERTY:
	263	pr = (struct of_prop_reconfig *)node;
	264	if (!strcmp(pr->prop->name, "ibm,dynamic-memory"))
	265	err = pseries_update_drconf_memory(pr);
57b53926 BP	266	break;
57b53926 BP	267	}
de2780a3	268	return notifier_from_errno(err);
57b53926 BP	269	}
	270
	271	static struct notifier_block pseries_mem_nb = {
	272	.notifier_call = pseries_memory_notifier,
	273	};
	274
	275	static int __init pseries_memory_hotplug_init(void)
	276	{
	277	if (firmware_has_feature(FW_FEATURE_LPAR))
1cf3d8b3	278	of_reconfig_notifier_register(&pseries_mem_nb);
57b53926 BP	279
	280	return 0;
	281	}
	282	machine_device_initcall(pseries, pseries_memory_hotplug_init);