[linux.git] / drivers / xen / tmem.c

/*
 * Xen implementation for transcendent memory (tmem)
 *
 * Copyright (C) 2009-2011 Oracle Corp.  All rights reserved.
 * Author: Dan Magenheimer
 */

#define pr_fmt(fmt) "xen:" KBUILD_MODNAME ": " fmt

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/types.h>
#include <linux/init.h>
#include <linux/pagemap.h>
#include <linux/cleancache.h>
#include <linux/frontswap.h>

#include <xen/xen.h>
#include <xen/interface/xen.h>
#include <asm/xen/hypercall.h>
#include <asm/xen/page.h>
#include <asm/xen/hypervisor.h>
#include <xen/tmem.h>

#ifndef CONFIG_XEN_TMEM_MODULE
bool __read_mostly tmem_enabled = false;

static int __init enable_tmem(char *s)
{
	tmem_enabled = true;
	return 1;
}
__setup("tmem", enable_tmem);
#endif

#ifdef CONFIG_CLEANCACHE
static bool cleancache __read_mostly = true;
module_param(cleancache, bool, S_IRUGO);
static bool selfballooning __read_mostly = true;
module_param(selfballooning, bool, S_IRUGO);
#endif /* CONFIG_CLEANCACHE */

#ifdef CONFIG_FRONTSWAP
static bool frontswap __read_mostly = true;
module_param(frontswap, bool, S_IRUGO);
#else /* CONFIG_FRONTSWAP */
#define frontswap (0)
#endif /* CONFIG_FRONTSWAP */

#ifdef CONFIG_XEN_SELFBALLOONING
static bool selfshrinking __read_mostly = true;
module_param(selfshrinking, bool, S_IRUGO);
#endif /* CONFIG_XEN_SELFBALLOONING */

#define TMEM_CONTROL               0
#define TMEM_NEW_POOL              1
#define TMEM_DESTROY_POOL          2
#define TMEM_NEW_PAGE              3
#define TMEM_PUT_PAGE              4
#define TMEM_GET_PAGE              5
#define TMEM_FLUSH_PAGE            6
#define TMEM_FLUSH_OBJECT          7
#define TMEM_READ                  8
#define TMEM_WRITE                 9
#define TMEM_XCHG                 10

/* Bits for HYPERVISOR_tmem_op(TMEM_NEW_POOL) */
#define TMEM_POOL_PERSIST          1
#define TMEM_POOL_SHARED           2
#define TMEM_POOL_PAGESIZE_SHIFT   4
#define TMEM_VERSION_SHIFT        24


struct tmem_pool_uuid {
	u64 uuid_lo;
	u64 uuid_hi;
};

struct tmem_oid {
	u64 oid[3];
};

#define TMEM_POOL_PRIVATE_UUID	{ 0, 0 }

/* flags for tmem_ops.new_pool */
#define TMEM_POOL_PERSIST          1
#define TMEM_POOL_SHARED           2

/* xen tmem foundation ops/hypercalls */

static inline int xen_tmem_op(u32 tmem_cmd, u32 tmem_pool, struct tmem_oid oid,
	u32 index, unsigned long gmfn, u32 tmem_offset, u32 pfn_offset, u32 len)
{
	struct tmem_op op;
	int rc = 0;

	op.cmd = tmem_cmd;
	op.pool_id = tmem_pool;
	op.u.gen.oid[0] = oid.oid[0];
	op.u.gen.oid[1] = oid.oid[1];
	op.u.gen.oid[2] = oid.oid[2];
	op.u.gen.index = index;
	op.u.gen.tmem_offset = tmem_offset;
	op.u.gen.pfn_offset = pfn_offset;
	op.u.gen.len = len;
	set_xen_guest_handle(op.u.gen.gmfn, (void *)gmfn);
	rc = HYPERVISOR_tmem_op(&op);
	return rc;
}

static int xen_tmem_new_pool(struct tmem_pool_uuid uuid,
				u32 flags, unsigned long pagesize)
{
	struct tmem_op op;
	int rc = 0, pageshift;

	for (pageshift = 0; pagesize != 1; pageshift++)
		pagesize >>= 1;
	flags |= (pageshift - 12) << TMEM_POOL_PAGESIZE_SHIFT;
	flags |= TMEM_SPEC_VERSION << TMEM_VERSION_SHIFT;
	op.cmd = TMEM_NEW_POOL;
	op.u.new.uuid[0] = uuid.uuid_lo;
	op.u.new.uuid[1] = uuid.uuid_hi;
	op.u.new.flags = flags;
	rc = HYPERVISOR_tmem_op(&op);
	return rc;
}

/* xen generic tmem ops */

static int xen_tmem_put_page(u32 pool_id, struct tmem_oid oid,
			     u32 index, unsigned long pfn)
{
	unsigned long gmfn = xen_pv_domain() ? pfn_to_mfn(pfn) : pfn;

	return xen_tmem_op(TMEM_PUT_PAGE, pool_id, oid, index,
		gmfn, 0, 0, 0);
}

static int xen_tmem_get_page(u32 pool_id, struct tmem_oid oid,
			     u32 index, unsigned long pfn)
{
	unsigned long gmfn = xen_pv_domain() ? pfn_to_mfn(pfn) : pfn;

	return xen_tmem_op(TMEM_GET_PAGE, pool_id, oid, index,
		gmfn, 0, 0, 0);
}

static int xen_tmem_flush_page(u32 pool_id, struct tmem_oid oid, u32 index)
{
	return xen_tmem_op(TMEM_FLUSH_PAGE, pool_id, oid, index,
		0, 0, 0, 0);
}

static int xen_tmem_flush_object(u32 pool_id, struct tmem_oid oid)
{
	return xen_tmem_op(TMEM_FLUSH_OBJECT, pool_id, oid, 0, 0, 0, 0, 0);
}


#ifdef CONFIG_CLEANCACHE
static int xen_tmem_destroy_pool(u32 pool_id)
{
	struct tmem_oid oid = { { 0 } };

	return xen_tmem_op(TMEM_DESTROY_POOL, pool_id, oid, 0, 0, 0, 0, 0);
}

/* cleancache ops */

static void tmem_cleancache_put_page(int pool, struct cleancache_filekey key,
				     pgoff_t index, struct page *page)
{
	u32 ind = (u32) index;
	struct tmem_oid oid = *(struct tmem_oid *)&key;
	unsigned long pfn = page_to_pfn(page);

	if (pool < 0)
		return;
	if (ind != index)
		return;
	mb(); /* ensure page is quiescent; tmem may address it with an alias */
	(void)xen_tmem_put_page((u32)pool, oid, ind, pfn);
}

static int tmem_cleancache_get_page(int pool, struct cleancache_filekey key,
				    pgoff_t index, struct page *page)
{
	u32 ind = (u32) index;
	struct tmem_oid oid = *(struct tmem_oid *)&key;
	unsigned long pfn = page_to_pfn(page);
	int ret;

	/* translate return values to linux semantics */
	if (pool < 0)
		return -1;
	if (ind != index)
		return -1;
	ret = xen_tmem_get_page((u32)pool, oid, ind, pfn);
	if (ret == 1)
		return 0;
	else
		return -1;
}

static void tmem_cleancache_flush_page(int pool, struct cleancache_filekey key,
				       pgoff_t index)
{
	u32 ind = (u32) index;
	struct tmem_oid oid = *(struct tmem_oid *)&key;

	if (pool < 0)
		return;
	if (ind != index)
		return;
	(void)xen_tmem_flush_page((u32)pool, oid, ind);
}

static void tmem_cleancache_flush_inode(int pool, struct cleancache_filekey key)
{
	struct tmem_oid oid = *(struct tmem_oid *)&key;

	if (pool < 0)
		return;
	(void)xen_tmem_flush_object((u32)pool, oid);
}

static void tmem_cleancache_flush_fs(int pool)
{
	if (pool < 0)
		return;
	(void)xen_tmem_destroy_pool((u32)pool);
}

static int tmem_cleancache_init_fs(size_t pagesize)
{
	struct tmem_pool_uuid uuid_private = TMEM_POOL_PRIVATE_UUID;

	return xen_tmem_new_pool(uuid_private, 0, pagesize);
}

static int tmem_cleancache_init_shared_fs(char *uuid, size_t pagesize)
{
	struct tmem_pool_uuid shared_uuid;

	shared_uuid.uuid_lo = *(u64 *)uuid;
	shared_uuid.uuid_hi = *(u64 *)(&uuid[8]);
	return xen_tmem_new_pool(shared_uuid, TMEM_POOL_SHARED, pagesize);
}

static struct cleancache_ops tmem_cleancache_ops = {
	.put_page = tmem_cleancache_put_page,
	.get_page = tmem_cleancache_get_page,
	.invalidate_page = tmem_cleancache_flush_page,
	.invalidate_inode = tmem_cleancache_flush_inode,
	.invalidate_fs = tmem_cleancache_flush_fs,
	.init_shared_fs = tmem_cleancache_init_shared_fs,
	.init_fs = tmem_cleancache_init_fs
};
#endif

#ifdef CONFIG_FRONTSWAP
/* frontswap tmem operations */

/* a single tmem poolid is used for all frontswap "types" (swapfiles) */
static int tmem_frontswap_poolid;

/*
 * Swizzling increases objects per swaptype, increasing tmem concurrency
 * for heavy swaploads.  Later, larger nr_cpus -> larger SWIZ_BITS
 */
#define SWIZ_BITS		4
#define SWIZ_MASK		((1 << SWIZ_BITS) - 1)
#define _oswiz(_type, _ind)	((_type << SWIZ_BITS) | (_ind & SWIZ_MASK))
#define iswiz(_ind)		(_ind >> SWIZ_BITS)

static inline struct tmem_oid oswiz(unsigned type, u32 ind)
{
	struct tmem_oid oid = { .oid = { 0 } };
	oid.oid[0] = _oswiz(type, ind);
	return oid;
}

/* returns 0 if the page was successfully put into frontswap, -1 if not */
static int tmem_frontswap_store(unsigned type, pgoff_t offset,
				   struct page *page)
{
	u64 ind64 = (u64)offset;
	u32 ind = (u32)offset;
	unsigned long pfn = page_to_pfn(page);
	int pool = tmem_frontswap_poolid;
	int ret;

	if (pool < 0)
		return -1;
	if (ind64 != ind)
		return -1;
	mb(); /* ensure page is quiescent; tmem may address it with an alias */
	ret = xen_tmem_put_page(pool, oswiz(type, ind), iswiz(ind), pfn);
	/* translate Xen tmem return values to linux semantics */
	if (ret == 1)
		return 0;
	else
		return -1;
}

/*
 * returns 0 if the page was successfully gotten from frontswap, -1 if
 * was not present (should never happen!)
 */
static int tmem_frontswap_load(unsigned type, pgoff_t offset,
				   struct page *page)
{
	u64 ind64 = (u64)offset;
	u32 ind = (u32)offset;
	unsigned long pfn = page_to_pfn(page);
	int pool = tmem_frontswap_poolid;
	int ret;

	if (pool < 0)
		return -1;
	if (ind64 != ind)
		return -1;
	ret = xen_tmem_get_page(pool, oswiz(type, ind), iswiz(ind), pfn);
	/* translate Xen tmem return values to linux semantics */
	if (ret == 1)
		return 0;
	else
		return -1;
}

/* flush a single page from frontswap */
static void tmem_frontswap_flush_page(unsigned type, pgoff_t offset)
{
	u64 ind64 = (u64)offset;
	u32 ind = (u32)offset;
	int pool = tmem_frontswap_poolid;

	if (pool < 0)
		return;
	if (ind64 != ind)
		return;
	(void) xen_tmem_flush_page(pool, oswiz(type, ind), iswiz(ind));
}

/* flush all pages from the passed swaptype */
static void tmem_frontswap_flush_area(unsigned type)
{
	int pool = tmem_frontswap_poolid;
	int ind;

	if (pool < 0)
		return;
	for (ind = SWIZ_MASK; ind >= 0; ind--)
		(void)xen_tmem_flush_object(pool, oswiz(type, ind));
}

static void tmem_frontswap_init(unsigned ignored)
{
	struct tmem_pool_uuid private = TMEM_POOL_PRIVATE_UUID;

	/* a single tmem poolid is used for all frontswap "types" (swapfiles) */
	if (tmem_frontswap_poolid < 0)
		tmem_frontswap_poolid =
		    xen_tmem_new_pool(private, TMEM_POOL_PERSIST, PAGE_SIZE);
}

static struct frontswap_ops tmem_frontswap_ops = {
	.store = tmem_frontswap_store,
	.load = tmem_frontswap_load,
	.invalidate_page = tmem_frontswap_flush_page,
	.invalidate_area = tmem_frontswap_flush_area,
	.init = tmem_frontswap_init
};
#endif

static int __init xen_tmem_init(void)
{
	if (!xen_domain())
		return 0;
#ifdef CONFIG_FRONTSWAP
	if (tmem_enabled && frontswap) {
		char *s = "";
		struct frontswap_ops *old_ops;

		tmem_frontswap_poolid = -1;
		old_ops = frontswap_register_ops(&tmem_frontswap_ops);
		if (IS_ERR(old_ops) || old_ops) {
			if (IS_ERR(old_ops))
				return PTR_ERR(old_ops);
			s = " (WARNING: frontswap_ops overridden)";
		}
		pr_info("frontswap enabled, RAM provided by Xen Transcendent Memory%s\n",
			s);
	}
#endif
#ifdef CONFIG_CLEANCACHE
	BUILD_BUG_ON(sizeof(struct cleancache_filekey) != sizeof(struct tmem_oid));
	if (tmem_enabled && cleancache) {
		int err;

		err = cleancache_register_ops(&tmem_cleancache_ops);
		if (err)
			pr_warn("xen-tmem: failed to enable cleancache: %d\n",
				err);
		else
			pr_info("cleancache enabled, RAM provided by "
				"Xen Transcendent Memory\n");
	}
#endif
#ifdef CONFIG_XEN_SELFBALLOONING
	/*
	 * There is no point of driving pages to the swap system if they
	 * aren't going anywhere in tmem universe.
	 */
	if (!frontswap) {
		selfshrinking = false;
		selfballooning = false;
	}
	xen_selfballoon_init(selfballooning, selfshrinking);
#endif
	return 0;
}

module_init(xen_tmem_init)
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Dan Magenheimer <[email protected]>");
MODULE_DESCRIPTION("Shim to Xen transcendent memory");
Commit	Line	Data
5bc20fc5 DM	1	/*
	2	* Xen implementation for transcendent memory (tmem)
	3	*
afec6e04	4	* Copyright (C) 2009-2011 Oracle Corp. All rights reserved.
5bc20fc5 DM	5	* Author: Dan Magenheimer
	6	*/
	7
283c0972 JP	8	#define pr_fmt(fmt) "xen:" KBUILD_MODNAME ": " fmt
283c0972 JP	9
10a7a077	10	#include <linux/module.h>
5bc20fc5 DM	11	#include <linux/kernel.h>
	12	#include <linux/types.h>
	13	#include <linux/init.h>
	14	#include <linux/pagemap.h>
	15	#include <linux/cleancache.h>
afec6e04	16	#include <linux/frontswap.h>
afec6e04	17
5bc20fc5 DM	18	#include <xen/xen.h>
	19	#include <xen/interface/xen.h>
	20	#include <asm/xen/hypercall.h>
	21	#include <asm/xen/page.h>
	22	#include <asm/xen/hypervisor.h>
b8b0f559	23	#include <xen/tmem.h>
5bc20fc5	24
0cb401d4 KRW	25	#ifndef CONFIG_XEN_TMEM_MODULE
	26	bool __read_mostly tmem_enabled = false;
	27
	28	static int __init enable_tmem(char *s)
	29	{
	30	tmem_enabled = true;
	31	return 1;
	32	}
	33	__setup("tmem", enable_tmem);
	34	#endif
	35
	36	#ifdef CONFIG_CLEANCACHE
9fd19653	37	static bool cleancache __read_mostly = true;
9fd19653	38	module_param(cleancache, bool, S_IRUGO);
2ca62b04	39	static bool selfballooning __read_mostly = true;
9fd19653	40	module_param(selfballooning, bool, S_IRUGO);
0cb401d4 KRW	41	#endif /* CONFIG_CLEANCACHE */
	42
	43	#ifdef CONFIG_FRONTSWAP
9fd19653	44	static bool frontswap __read_mostly = true;
9fd19653	45	module_param(frontswap, bool, S_IRUGO);
1d7004f0 FC	46	#else /* CONFIG_FRONTSWAP */
1d7004f0 FC	47	#define frontswap (0)
e8f9cb03 KRW	48	#endif /* CONFIG_FRONTSWAP */
e8f9cb03 KRW	49
23972c63	50	#ifdef CONFIG_XEN_SELFBALLOONING
2ca62b04 KRW	51	static bool selfshrinking __read_mostly = true;
2ca62b04 KRW	52	module_param(selfshrinking, bool, S_IRUGO);
23972c63	53	#endif /* CONFIG_XEN_SELFBALLOONING */
0cb401d4	54
5bc20fc5 DM	55	#define TMEM_CONTROL 0
	56	#define TMEM_NEW_POOL 1
	57	#define TMEM_DESTROY_POOL 2
	58	#define TMEM_NEW_PAGE 3
	59	#define TMEM_PUT_PAGE 4
	60	#define TMEM_GET_PAGE 5
	61	#define TMEM_FLUSH_PAGE 6
	62	#define TMEM_FLUSH_OBJECT 7
	63	#define TMEM_READ 8
	64	#define TMEM_WRITE 9
	65	#define TMEM_XCHG 10
	66
	67	/* Bits for HYPERVISOR_tmem_op(TMEM_NEW_POOL) */
	68	#define TMEM_POOL_PERSIST 1
	69	#define TMEM_POOL_SHARED 2
	70	#define TMEM_POOL_PAGESIZE_SHIFT 4
	71	#define TMEM_VERSION_SHIFT 24
	72
	73
	74	struct tmem_pool_uuid {
	75	u64 uuid_lo;
	76	u64 uuid_hi;
	77	};
	78
	79	struct tmem_oid {
	80	u64 oid[3];
	81	};
	82
	83	#define TMEM_POOL_PRIVATE_UUID { 0, 0 }
	84
	85	/* flags for tmem_ops.new_pool */
	86	#define TMEM_POOL_PERSIST 1
	87	#define TMEM_POOL_SHARED 2
	88
	89	/* xen tmem foundation ops/hypercalls */
	90
	91	static inline int xen_tmem_op(u32 tmem_cmd, u32 tmem_pool, struct tmem_oid oid,
	92	u32 index, unsigned long gmfn, u32 tmem_offset, u32 pfn_offset, u32 len)
	93	{
	94	struct tmem_op op;
	95	int rc = 0;
	96
	97	op.cmd = tmem_cmd;
	98	op.pool_id = tmem_pool;
	99	op.u.gen.oid[0] = oid.oid[0];
	100	op.u.gen.oid[1] = oid.oid[1];
	101	op.u.gen.oid[2] = oid.oid[2];
	102	op.u.gen.index = index;
	103	op.u.gen.tmem_offset = tmem_offset;
	104	op.u.gen.pfn_offset = pfn_offset;
	105	op.u.gen.len = len;
	106	set_xen_guest_handle(op.u.gen.gmfn, (void *)gmfn);
	107	rc = HYPERVISOR_tmem_op(&op);
	108	return rc;
	109	}
	110
	111	static int xen_tmem_new_pool(struct tmem_pool_uuid uuid,
	112	u32 flags, unsigned long pagesize)
	113	{
	114	struct tmem_op op;
	115	int rc = 0, pageshift;
	116
	117	for (pageshift = 0; pagesize != 1; pageshift++)
	118	pagesize >>= 1;
119	flags \|= (pageshift - 12) << TMEM_POOL_PAGESIZE_SHIFT;
120	flags \|= TMEM_SPEC_VERSION << TMEM_VERSION_SHIFT;
121	op.cmd = TMEM_NEW_POOL;
122	op.u.new.uuid[0] = uuid.uuid_lo;
123	op.u.new.uuid[1] = uuid.uuid_hi;
124	op.u.new.flags = flags;
125	rc = HYPERVISOR_tmem_op(&op);
126	return rc;
127	}
128
129	/* xen generic tmem ops */
130
131	static int xen_tmem_put_page(u32 pool_id, struct tmem_oid oid,
132	u32 index, unsigned long pfn)
133	{
134	unsigned long gmfn = xen_pv_domain() ? pfn_to_mfn(pfn) : pfn;
135
136	return xen_tmem_op(TMEM_PUT_PAGE, pool_id, oid, index,
137	gmfn, 0, 0, 0);
138	}
139
140	static int xen_tmem_get_page(u32 pool_id, struct tmem_oid oid,
141	u32 index, unsigned long pfn)
142	{
143	unsigned long gmfn = xen_pv_domain() ? pfn_to_mfn(pfn) : pfn;
144
145	return xen_tmem_op(TMEM_GET_PAGE, pool_id, oid, index,
146	gmfn, 0, 0, 0);
147	}
148
149	static int xen_tmem_flush_page(u32 pool_id, struct tmem_oid oid, u32 index)
150	{
151	return xen_tmem_op(TMEM_FLUSH_PAGE, pool_id, oid, index,
152	0, 0, 0, 0);
153	}
154
155	static int xen_tmem_flush_object(u32 pool_id, struct tmem_oid oid)
156	{
157	return xen_tmem_op(TMEM_FLUSH_OBJECT, pool_id, oid, 0, 0, 0, 0, 0);
158	}
159
5bc20fc5	160
afec6e04 DM	161	#ifdef CONFIG_CLEANCACHE
	162	static int xen_tmem_destroy_pool(u32 pool_id)
	163	{
	164	struct tmem_oid oid = { { 0 } };
	165
	166	return xen_tmem_op(TMEM_DESTROY_POOL, pool_id, oid, 0, 0, 0, 0, 0);
	167	}
	168
5bc20fc5 DM	169	/* cleancache ops */
	170
	171	static void tmem_cleancache_put_page(int pool, struct cleancache_filekey key,
	172	pgoff_t index, struct page *page)
	173	{
	174	u32 ind = (u32) index;
	175	struct tmem_oid oid = (struct tmem_oid )&key;
	176	unsigned long pfn = page_to_pfn(page);
	177
	178	if (pool < 0)
	179	return;
	180	if (ind != index)
	181	return;
	182	mb(); /* ensure page is quiescent; tmem may address it with an alias */
	183	(void)xen_tmem_put_page((u32)pool, oid, ind, pfn);
	184	}
	185
	186	static int tmem_cleancache_get_page(int pool, struct cleancache_filekey key,
	187	pgoff_t index, struct page *page)
	188	{
	189	u32 ind = (u32) index;
	190	struct tmem_oid oid = (struct tmem_oid )&key;
	191	unsigned long pfn = page_to_pfn(page);
	192	int ret;
	193
	194	/* translate return values to linux semantics */
	195	if (pool < 0)
	196	return -1;
	197	if (ind != index)
	198	return -1;
	199	ret = xen_tmem_get_page((u32)pool, oid, ind, pfn);
	200	if (ret == 1)
	201	return 0;
	202	else
	203	return -1;
	204	}
	205
	206	static void tmem_cleancache_flush_page(int pool, struct cleancache_filekey key,
	207	pgoff_t index)
	208	{
	209	u32 ind = (u32) index;
	210	struct tmem_oid oid = (struct tmem_oid )&key;
	211
	212	if (pool < 0)
	213	return;
	214	if (ind != index)
	215	return;
	216	(void)xen_tmem_flush_page((u32)pool, oid, ind);
	217	}
	218
	219	static void tmem_cleancache_flush_inode(int pool, struct cleancache_filekey key)
	220	{
	221	struct tmem_oid oid = (struct tmem_oid )&key;
	222
	223	if (pool < 0)
	224	return;
	225	(void)xen_tmem_flush_object((u32)pool, oid);
	226	}
	227
	228	static void tmem_cleancache_flush_fs(int pool)
	229	{
	230	if (pool < 0)
	231	return;
	232	(void)xen_tmem_destroy_pool((u32)pool);
233	}
234
235	static int tmem_cleancache_init_fs(size_t pagesize)
236	{
237	struct tmem_pool_uuid uuid_private = TMEM_POOL_PRIVATE_UUID;
238
239	return xen_tmem_new_pool(uuid_private, 0, pagesize);
240	}
241
242	static int tmem_cleancache_init_shared_fs(char *uuid, size_t pagesize)
243	{
244	struct tmem_pool_uuid shared_uuid;
245
246	shared_uuid.uuid_lo = (u64 )uuid;
247	shared_uuid.uuid_hi = (u64 )(&uuid[8]);
248	return xen_tmem_new_pool(shared_uuid, TMEM_POOL_SHARED, pagesize);
249	}
250
833f8662	251	static struct cleancache_ops tmem_cleancache_ops = {
5bc20fc5 DM	252	.put_page = tmem_cleancache_put_page,
5bc20fc5 DM	253	.get_page = tmem_cleancache_get_page,
91c6cc9b DM	254	.invalidate_page = tmem_cleancache_flush_page,
	255	.invalidate_inode = tmem_cleancache_flush_inode,
	256	.invalidate_fs = tmem_cleancache_flush_fs,
5bc20fc5 DM	257	.init_shared_fs = tmem_cleancache_init_shared_fs,
	258	.init_fs = tmem_cleancache_init_fs
	259	};
afec6e04	260	#endif
5bc20fc5	261
afec6e04 DM	262	#ifdef CONFIG_FRONTSWAP
	263	/* frontswap tmem operations */
	264
	265	/* a single tmem poolid is used for all frontswap "types" (swapfiles) */
	266	static int tmem_frontswap_poolid;
	267
	268	/*
	269	* Swizzling increases objects per swaptype, increasing tmem concurrency
	270	* for heavy swaploads. Later, larger nr_cpus -> larger SWIZ_BITS
	271	*/
	272	#define SWIZ_BITS 4
	273	#define SWIZ_MASK ((1 << SWIZ_BITS) - 1)
	274	#define _oswiz(_type, _ind) ((_type << SWIZ_BITS) \| (_ind & SWIZ_MASK))
	275	#define iswiz(_ind) (_ind >> SWIZ_BITS)
	276
	277	static inline struct tmem_oid oswiz(unsigned type, u32 ind)
5bc20fc5	278	{
afec6e04 DM	279	struct tmem_oid oid = { .oid = { 0 } };
	280	oid.oid[0] = _oswiz(type, ind);
	281	return oid;
	282	}
5bc20fc5	283
afec6e04	284	/* returns 0 if the page was successfully put into frontswap, -1 if not */
165c8aed	285	static int tmem_frontswap_store(unsigned type, pgoff_t offset,
afec6e04 DM	286	struct page *page)
	287	{
	288	u64 ind64 = (u64)offset;
	289	u32 ind = (u32)offset;
	290	unsigned long pfn = page_to_pfn(page);
	291	int pool = tmem_frontswap_poolid;
	292	int ret;
	293
	294	if (pool < 0)
	295	return -1;
	296	if (ind64 != ind)
	297	return -1;
	298	mb(); /* ensure page is quiescent; tmem may address it with an alias */
	299	ret = xen_tmem_put_page(pool, oswiz(type, ind), iswiz(ind), pfn);
	300	/* translate Xen tmem return values to linux semantics */
	301	if (ret == 1)
	302	return 0;
	303	else
	304	return -1;
	305	}
	306
	307	/*
	308	* returns 0 if the page was successfully gotten from frontswap, -1 if
	309	* was not present (should never happen!)
	310	*/
165c8aed	311	static int tmem_frontswap_load(unsigned type, pgoff_t offset,
afec6e04 DM	312	struct page *page)
	313	{
	314	u64 ind64 = (u64)offset;
	315	u32 ind = (u32)offset;
	316	unsigned long pfn = page_to_pfn(page);
	317	int pool = tmem_frontswap_poolid;
	318	int ret;
	319
	320	if (pool < 0)
	321	return -1;
	322	if (ind64 != ind)
	323	return -1;
	324	ret = xen_tmem_get_page(pool, oswiz(type, ind), iswiz(ind), pfn);
	325	/* translate Xen tmem return values to linux semantics */
	326	if (ret == 1)
	327	return 0;
	328	else
	329	return -1;
	330	}
	331
	332	/* flush a single page from frontswap */
	333	static void tmem_frontswap_flush_page(unsigned type, pgoff_t offset)
	334	{
	335	u64 ind64 = (u64)offset;
	336	u32 ind = (u32)offset;
	337	int pool = tmem_frontswap_poolid;
	338
	339	if (pool < 0)
	340	return;
	341	if (ind64 != ind)
	342	return;
	343	(void) xen_tmem_flush_page(pool, oswiz(type, ind), iswiz(ind));
	344	}
	345
	346	/* flush all pages from the passed swaptype */
	347	static void tmem_frontswap_flush_area(unsigned type)
	348	{
	349	int pool = tmem_frontswap_poolid;
	350	int ind;
	351
	352	if (pool < 0)
	353	return;
	354	for (ind = SWIZ_MASK; ind >= 0; ind--)
	355	(void)xen_tmem_flush_object(pool, oswiz(type, ind));
	356	}
	357
	358	static void tmem_frontswap_init(unsigned ignored)
	359	{
	360	struct tmem_pool_uuid private = TMEM_POOL_PRIVATE_UUID;
	361
	362	/* a single tmem poolid is used for all frontswap "types" (swapfiles) */
	363	if (tmem_frontswap_poolid < 0)
	364	tmem_frontswap_poolid =
	365	xen_tmem_new_pool(private, TMEM_POOL_PERSIST, PAGE_SIZE);
	366	}
	367
1e01c968	368	static struct frontswap_ops tmem_frontswap_ops = {
165c8aed KRW	369	.store = tmem_frontswap_store,
165c8aed KRW	370	.load = tmem_frontswap_load,
91c6cc9b DM	371	.invalidate_page = tmem_frontswap_flush_page,
91c6cc9b DM	372	.invalidate_area = tmem_frontswap_flush_area,
afec6e04 DM	373	.init = tmem_frontswap_init
	374	};
	375	#endif
	376
144c61ad	377	static int __init xen_tmem_init(void)
afec6e04	378	{
5bc20fc5 DM	379	if (!xen_domain())
5bc20fc5 DM	380	return 0;
afec6e04	381	#ifdef CONFIG_FRONTSWAP
9fd19653	382	if (tmem_enabled && frontswap) {
afec6e04	383	char *s = "";
b2c75c44	384	struct frontswap_ops *old_ops;
afec6e04 DM	385
afec6e04 DM	386	tmem_frontswap_poolid = -1;
b2c75c44	387	old_ops = frontswap_register_ops(&tmem_frontswap_ops);
f42158fe KRW	388	if (IS_ERR(old_ops) \|\| old_ops) {
	389	if (IS_ERR(old_ops))
	390	return PTR_ERR(old_ops);
afec6e04	391	s = " (WARNING: frontswap_ops overridden)";
f42158fe	392	}
283c0972 JP	393	pr_info("frontswap enabled, RAM provided by Xen Transcendent Memory%s\n",
283c0972 JP	394	s);
afec6e04 DM	395	}
afec6e04 DM	396	#endif
5bc20fc5	397	#ifdef CONFIG_CLEANCACHE
01b720f3	398	BUILD_BUG_ON(sizeof(struct cleancache_filekey) != sizeof(struct tmem_oid));
9fd19653	399	if (tmem_enabled && cleancache) {
53d85c98 VD	400	int err;
	401
	402	err = cleancache_register_ops(&tmem_cleancache_ops);
	403	if (err)
	404	pr_warn("xen-tmem: failed to enable cleancache: %d\n",
	405	err);
	406	else
	407	pr_info("cleancache enabled, RAM provided by "
	408	"Xen Transcendent Memory\n");
5bc20fc5	409	}
10a7a077 DM	410	#endif
10a7a077 DM	411	#ifdef CONFIG_XEN_SELFBALLOONING
37d46e15 KRW	412	/*
	413	* There is no point of driving pages to the swap system if they
	414	* aren't going anywhere in tmem universe.
	415	*/
	416	if (!frontswap) {
	417	selfshrinking = false;
	418	selfballooning = false;
	419	}
2ca62b04	420	xen_selfballoon_init(selfballooning, selfshrinking);
5bc20fc5 DM	421	#endif
	422	return 0;
	423	}
	424
	425	module_init(xen_tmem_init)
10a7a077 DM	426	MODULE_LICENSE("GPL");
	427	MODULE_AUTHOR("Dan Magenheimer <[email protected]>");
	428	MODULE_DESCRIPTION("Shim to Xen transcendent memory");