[mirror_qemu.git] / block / qed.h

/*
 * QEMU Enhanced Disk Format
 *
 * Copyright IBM, Corp. 2010
 *
 * Authors:
 *  Stefan Hajnoczi   <stefanha@linux.vnet.ibm.com>
 *  Anthony Liguori   <aliguori@us.ibm.com>
 *
 * This work is licensed under the terms of the GNU LGPL, version 2 or later.
 * See the COPYING.LIB file in the top-level directory.
 *
 */

#ifndef BLOCK_QED_H
#define BLOCK_QED_H

#include "block/block_int.h"
#include "qemu/cutils.h"

/* The layout of a QED file is as follows:
 *
 * +--------+----------+----------+----------+-----+
 * | header | L1 table | cluster0 | cluster1 | ... |
 * +--------+----------+----------+----------+-----+
 *
 * There is a 2-level pagetable for cluster allocation:
 *
 *                     +----------+
 *                     | L1 table |
 *                     +----------+
 *                ,------'  |  '------.
 *           +----------+   |    +----------+
 *           | L2 table |  ...   | L2 table |
 *           +----------+        +----------+
 *       ,------'  |  '------.
 *  +----------+   |    +----------+
 *  |   Data   |  ...   |   Data   |
 *  +----------+        +----------+
 *
 * The L1 table is fixed size and always present.  L2 tables are allocated on
 * demand.  The L1 table size determines the maximum possible image size; it
 * can be influenced using the cluster_size and table_size values.
 *
 * All fields are little-endian on disk.
 */
#define  QED_DEFAULT_CLUSTER_SIZE  65536
enum {
    QED_MAGIC = 'Q' | 'E' << 8 | 'D' << 16 | '\0' << 24,

    /* The image supports a backing file */
    QED_F_BACKING_FILE = 0x01,

    /* The image needs a consistency check before use */
    QED_F_NEED_CHECK = 0x02,

    /* The backing file format must not be probed, treat as raw image */
    QED_F_BACKING_FORMAT_NO_PROBE = 0x04,

    /* Feature bits must be used when the on-disk format changes */
    QED_FEATURE_MASK = QED_F_BACKING_FILE | /* supported feature bits */
                       QED_F_NEED_CHECK |
                       QED_F_BACKING_FORMAT_NO_PROBE,
    QED_COMPAT_FEATURE_MASK = 0,            /* supported compat feature bits */
    QED_AUTOCLEAR_FEATURE_MASK = 0,         /* supported autoclear feature bits */

    /* Data is stored in groups of sectors called clusters.  Cluster size must
     * be large to avoid keeping too much metadata.  I/O requests that have
     * sub-cluster size will require read-modify-write.
     */
    QED_MIN_CLUSTER_SIZE = 4 * 1024, /* in bytes */
    QED_MAX_CLUSTER_SIZE = 64 * 1024 * 1024,

    /* Allocated clusters are tracked using a 2-level pagetable.  Table size is
     * a multiple of clusters so large maximum image sizes can be supported
     * without jacking up the cluster size too much.
     */
    QED_MIN_TABLE_SIZE = 1,        /* in clusters */
    QED_MAX_TABLE_SIZE = 16,
    QED_DEFAULT_TABLE_SIZE = 4,

    /* Delay to flush and clean image after last allocating write completes */
    QED_NEED_CHECK_TIMEOUT = 5,    /* in seconds */
};

typedef struct {
    uint32_t magic;                 /* QED\0 */

    uint32_t cluster_size;          /* in bytes */
    uint32_t table_size;            /* for L1 and L2 tables, in clusters */
    uint32_t header_size;           /* in clusters */

    uint64_t features;              /* format feature bits */
    uint64_t compat_features;       /* compatible feature bits */
    uint64_t autoclear_features;    /* self-resetting feature bits */

    uint64_t l1_table_offset;       /* in bytes */
    uint64_t image_size;            /* total logical image size, in bytes */

    /* if (features & QED_F_BACKING_FILE) */
    uint32_t backing_filename_offset; /* in bytes from start of header */
    uint32_t backing_filename_size;   /* in bytes */
} QEMU_PACKED QEDHeader;

typedef struct {
    uint64_t offsets[0];            /* in bytes */
} QEDTable;

/* The L2 cache is a simple write-through cache for L2 structures */
typedef struct CachedL2Table {
    QEDTable *table;
    uint64_t offset;    /* offset=0 indicates an invalidate entry */
    QTAILQ_ENTRY(CachedL2Table) node;
    int ref;
} CachedL2Table;

typedef struct {
    QTAILQ_HEAD(, CachedL2Table) entries;
    unsigned int n_entries;
} L2TableCache;

typedef struct QEDRequest {
    CachedL2Table *l2_table;
} QEDRequest;

enum {
    QED_AIOCB_WRITE = 0x0001,       /* read or write? */
    QED_AIOCB_ZERO  = 0x0002,       /* zero write, used with QED_AIOCB_WRITE */
};

typedef struct QEDAIOCB {
    BlockAIOCB common;
    int bh_ret;                     /* final return status for completion bh */
    QSIMPLEQ_ENTRY(QEDAIOCB) next;  /* next request */
    int flags;                      /* QED_AIOCB_* bits ORed together */
    uint64_t end_pos;               /* request end on block device, in bytes */

    /* User scatter-gather list */
    QEMUIOVector *qiov;
    size_t qiov_offset;             /* byte count already processed */

    /* Current cluster scatter-gather list */
    QEMUIOVector cur_qiov;
    QEMUIOVector *backing_qiov;
    uint64_t cur_pos;               /* position on block device, in bytes */
    uint64_t cur_cluster;           /* cluster offset in image file */
    unsigned int cur_nclusters;     /* number of clusters being accessed */
    int find_cluster_ret;           /* used for L1/L2 update */

    QEDRequest request;
} QEDAIOCB;

typedef struct {
    BlockDriverState *bs;           /* device */
    uint64_t file_size;             /* length of image file, in bytes */

    QEDHeader header;               /* always cpu-endian */
    QEDTable *l1_table;
    L2TableCache l2_cache;          /* l2 table cache */
    uint32_t table_nelems;
    uint32_t l1_shift;
    uint32_t l2_shift;
    uint32_t l2_mask;

    /* Allocating write request queue */
    QSIMPLEQ_HEAD(, QEDAIOCB) allocating_write_reqs;
    bool allocating_write_reqs_plugged;

    /* Periodic flush and clear need check flag */
    QEMUTimer *need_check_timer;
} BDRVQEDState;

enum {
    QED_CLUSTER_FOUND,         /* cluster found */
    QED_CLUSTER_ZERO,          /* zero cluster found */
    QED_CLUSTER_L2,            /* cluster missing in L2 */
    QED_CLUSTER_L1,            /* cluster missing in L1 */
};

void qed_acquire(BDRVQEDState *s);
void qed_release(BDRVQEDState *s);

/**
 * Header functions
 */
int qed_write_header_sync(BDRVQEDState *s);

/**
 * L2 cache functions
 */
void qed_init_l2_cache(L2TableCache *l2_cache);
void qed_free_l2_cache(L2TableCache *l2_cache);
CachedL2Table *qed_alloc_l2_cache_entry(L2TableCache *l2_cache);
void qed_unref_l2_cache_entry(CachedL2Table *entry);
CachedL2Table *qed_find_l2_cache_entry(L2TableCache *l2_cache, uint64_t offset);
void qed_commit_l2_cache_entry(L2TableCache *l2_cache, CachedL2Table *l2_table);

/**
 * Table I/O functions
 */
int qed_read_l1_table_sync(BDRVQEDState *s);
int qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n);
int qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index,
                            unsigned int n);
int qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request,
                           uint64_t offset);
int qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, uint64_t offset);
int qed_write_l2_table(BDRVQEDState *s, QEDRequest *request,
                       unsigned int index, unsigned int n, bool flush);
int qed_write_l2_table_sync(BDRVQEDState *s, QEDRequest *request,
                            unsigned int index, unsigned int n, bool flush);

/**
 * Cluster functions
 */
int qed_find_cluster(BDRVQEDState *s, QEDRequest *request, uint64_t pos,
                     size_t *len, uint64_t *img_offset);

/**
 * Consistency check
 */
int qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix);

QEDTable *qed_alloc_table(BDRVQEDState *s);

/**
 * Round down to the start of a cluster
 */
static inline uint64_t qed_start_of_cluster(BDRVQEDState *s, uint64_t offset)
{
    return offset & ~(uint64_t)(s->header.cluster_size - 1);
}

static inline uint64_t qed_offset_into_cluster(BDRVQEDState *s, uint64_t offset)
{
    return offset & (s->header.cluster_size - 1);
}

static inline uint64_t qed_bytes_to_clusters(BDRVQEDState *s, uint64_t bytes)
{
    return qed_start_of_cluster(s, bytes + (s->header.cluster_size - 1)) /
           (s->header.cluster_size - 1);
}

static inline unsigned int qed_l1_index(BDRVQEDState *s, uint64_t pos)
{
    return pos >> s->l1_shift;
}

static inline unsigned int qed_l2_index(BDRVQEDState *s, uint64_t pos)
{
    return (pos >> s->l2_shift) & s->l2_mask;
}

/**
 * Test if a cluster offset is valid
 */
static inline bool qed_check_cluster_offset(BDRVQEDState *s, uint64_t offset)
{
    uint64_t header_size = (uint64_t)s->header.header_size *
                           s->header.cluster_size;

    if (offset & (s->header.cluster_size - 1)) {
        return false;
    }
    return offset >= header_size && offset < s->file_size;
}

/**
 * Test if a table offset is valid
 */
static inline bool qed_check_table_offset(BDRVQEDState *s, uint64_t offset)
{
    uint64_t end_offset = offset + (s->header.table_size - 1) *
                          s->header.cluster_size;

    /* Overflow check */
    if (end_offset <= offset) {
        return false;
    }

    return qed_check_cluster_offset(s, offset) &&
           qed_check_cluster_offset(s, end_offset);
}

static inline bool qed_offset_is_cluster_aligned(BDRVQEDState *s,
                                                 uint64_t offset)
{
    if (qed_offset_into_cluster(s, offset)) {
        return false;
    }
    return true;
}

static inline bool qed_offset_is_unalloc_cluster(uint64_t offset)
{
    if (offset == 0) {
        return true;
    }
    return false;
}

static inline bool qed_offset_is_zero_cluster(uint64_t offset)
{
    if (offset == 1) {
        return true;
    }
    return false;
}

#endif /* BLOCK_QED_H */
Commit	Line	Data
75411d23 SH	1	/*
	2	* QEMU Enhanced Disk Format
	3	*
	4	* Copyright IBM, Corp. 2010
	5	*
	6	* Authors:
	7	* Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
	8	* Anthony Liguori <aliguori@us.ibm.com>
	9	*
	10	* This work is licensed under the terms of the GNU LGPL, version 2 or later.
	11	* See the COPYING.LIB file in the top-level directory.
	12	*
	13	*/
	14
	15	#ifndef BLOCK_QED_H
	16	#define BLOCK_QED_H
	17
737e150e	18	#include "block/block_int.h"
f348b6d1	19	#include "qemu/cutils.h"
75411d23 SH	20
	21	/* The layout of a QED file is as follows:
	22	*
	23	* +--------+----------+----------+----------+-----+
	24	* \| header \| L1 table \| cluster0 \| cluster1 \| ... \|
	25	* +--------+----------+----------+----------+-----+
	26	*
	27	* There is a 2-level pagetable for cluster allocation:
	28	*
	29	* +----------+
	30	* \| L1 table \|
	31	* +----------+
	32	* ,------' \| '------.
	33	* +----------+ \| +----------+
	34	* \| L2 table \| ... \| L2 table \|
	35	* +----------+ +----------+
	36	* ,------' \| '------.
	37	* +----------+ \| +----------+
	38	* \| Data \| ... \| Data \|
	39	* +----------+ +----------+
	40	*
	41	* The L1 table is fixed size and always present. L2 tables are allocated on
	42	* demand. The L1 table size determines the maximum possible image size; it
	43	* can be influenced using the cluster_size and table_size values.
	44	*
	45	* All fields are little-endian on disk.
	46	*/
7ab74849	47	#define QED_DEFAULT_CLUSTER_SIZE 65536
75411d23 SH	48	enum {
	49	QED_MAGIC = 'Q' \| 'E' << 8 \| 'D' << 16 \| '\0' << 24,
	50
	51	/* The image supports a backing file */
	52	QED_F_BACKING_FILE = 0x01,
	53
01979a98 SH	54	/* The image needs a consistency check before use */
	55	QED_F_NEED_CHECK = 0x02,
	56
75411d23 SH	57	/* The backing file format must not be probed, treat as raw image */
	58	QED_F_BACKING_FORMAT_NO_PROBE = 0x04,
	59
	60	/* Feature bits must be used when the on-disk format changes */
	61	QED_FEATURE_MASK = QED_F_BACKING_FILE \| /* supported feature bits */
01979a98	62	QED_F_NEED_CHECK \|
75411d23 SH	63	QED_F_BACKING_FORMAT_NO_PROBE,
	64	QED_COMPAT_FEATURE_MASK = 0, /* supported compat feature bits */
	65	QED_AUTOCLEAR_FEATURE_MASK = 0, /* supported autoclear feature bits */
	66
	67	/* Data is stored in groups of sectors called clusters. Cluster size must
	68	* be large to avoid keeping too much metadata. I/O requests that have
	69	* sub-cluster size will require read-modify-write.
	70	*/
	71	QED_MIN_CLUSTER_SIZE = 4 * 1024, /* in bytes */
	72	QED_MAX_CLUSTER_SIZE = 64 * 1024 * 1024,
75411d23 SH	73
	74	/* Allocated clusters are tracked using a 2-level pagetable. Table size is
	75	* a multiple of clusters so large maximum image sizes can be supported
	76	* without jacking up the cluster size too much.
	77	*/
	78	QED_MIN_TABLE_SIZE = 1, /* in clusters */
	79	QED_MAX_TABLE_SIZE = 16,
	80	QED_DEFAULT_TABLE_SIZE = 4,
6f321e93 SH	81
	82	/* Delay to flush and clean image after last allocating write completes */
	83	QED_NEED_CHECK_TIMEOUT = 5, /* in seconds */
75411d23 SH	84	};
	85
	86	typedef struct {
	87	uint32_t magic; /* QED\0 */
	88
	89	uint32_t cluster_size; /* in bytes */
	90	uint32_t table_size; /* for L1 and L2 tables, in clusters */
	91	uint32_t header_size; /* in clusters */
	92
	93	uint64_t features; /* format feature bits */
	94	uint64_t compat_features; /* compatible feature bits */
	95	uint64_t autoclear_features; /* self-resetting feature bits */
	96
	97	uint64_t l1_table_offset; /* in bytes */
	98	uint64_t image_size; /* total logical image size, in bytes */
	99
	100	/* if (features & QED_F_BACKING_FILE) */
	101	uint32_t backing_filename_offset; /* in bytes from start of header */
	102	uint32_t backing_filename_size; /* in bytes */
687fb893	103	} QEMU_PACKED QEDHeader;
75411d23	104
298800ca SH	105	typedef struct {
	106	uint64_t offsets[0]; /* in bytes */
	107	} QEDTable;
	108
	109	/* The L2 cache is a simple write-through cache for L2 structures */
	110	typedef struct CachedL2Table {
	111	QEDTable *table;
	112	uint64_t offset; /* offset=0 indicates an invalidate entry */
	113	QTAILQ_ENTRY(CachedL2Table) node;
	114	int ref;
	115	} CachedL2Table;
	116
	117	typedef struct {
	118	QTAILQ_HEAD(, CachedL2Table) entries;
	119	unsigned int n_entries;
	120	} L2TableCache;
	121
	122	typedef struct QEDRequest {
	123	CachedL2Table *l2_table;
	124	} QEDRequest;
	125
6e4f59bd SH	126	enum {
6e4f59bd SH	127	QED_AIOCB_WRITE = 0x0001, /* read or write? */
0e71be19	128	QED_AIOCB_ZERO = 0x0002, /* zero write, used with QED_AIOCB_WRITE */
6e4f59bd SH	129	};
6e4f59bd SH	130
eabba580	131	typedef struct QEDAIOCB {
7c84b1b8	132	BlockAIOCB common;
eabba580 SH	133	int bh_ret; /* final return status for completion bh */
eabba580 SH	134	QSIMPLEQ_ENTRY(QEDAIOCB) next; /* next request */
6e4f59bd	135	int flags; /* QED_AIOCB_* bits ORed together */
eabba580 SH	136	uint64_t end_pos; /* request end on block device, in bytes */
	137
	138	/* User scatter-gather list */
	139	QEMUIOVector *qiov;
	140	size_t qiov_offset; /* byte count already processed */
	141
	142	/* Current cluster scatter-gather list */
	143	QEMUIOVector cur_qiov;
f06ee3d4	144	QEMUIOVector *backing_qiov;
eabba580 SH	145	uint64_t cur_pos; /* position on block device, in bytes */
	146	uint64_t cur_cluster; /* cluster offset in image file */
	147	unsigned int cur_nclusters; /* number of clusters being accessed */
	148	int find_cluster_ret; /* used for L1/L2 update */
	149
	150	QEDRequest request;
	151	} QEDAIOCB;
	152
75411d23 SH	153	typedef struct {
	154	BlockDriverState bs; / device */
	155	uint64_t file_size; /* length of image file, in bytes */
	156
	157	QEDHeader header; /* always cpu-endian */
298800ca SH	158	QEDTable *l1_table;
298800ca SH	159	L2TableCache l2_cache; /* l2 table cache */
75411d23 SH	160	uint32_t table_nelems;
	161	uint32_t l1_shift;
	162	uint32_t l2_shift;
	163	uint32_t l2_mask;
eabba580 SH	164
	165	/* Allocating write request queue */
	166	QSIMPLEQ_HEAD(, QEDAIOCB) allocating_write_reqs;
6f321e93 SH	167	bool allocating_write_reqs_plugged;
	168
	169	/* Periodic flush and clear need check flag */
	170	QEMUTimer *need_check_timer;
75411d23 SH	171	} BDRVQEDState;
75411d23 SH	172
298800ca SH	173	enum {
298800ca SH	174	QED_CLUSTER_FOUND, /* cluster found */
21df65b6	175	QED_CLUSTER_ZERO, /* zero cluster found */
298800ca SH	176	QED_CLUSTER_L2, /* cluster missing in L2 */
	177	QED_CLUSTER_L1, /* cluster missing in L1 */
	178	};
	179
2f47da5f PB	180	void qed_acquire(BDRVQEDState *s);
	181	void qed_release(BDRVQEDState *s);
	182
b10170ac SH	183	/**
	184	* Header functions
	185	*/
	186	int qed_write_header_sync(BDRVQEDState *s);
	187
298800ca SH	188	/**
	189	* L2 cache functions
	190	*/
	191	void qed_init_l2_cache(L2TableCache *l2_cache);
	192	void qed_free_l2_cache(L2TableCache *l2_cache);
	193	CachedL2Table qed_alloc_l2_cache_entry(L2TableCache l2_cache);
	194	void qed_unref_l2_cache_entry(CachedL2Table *entry);
	195	CachedL2Table qed_find_l2_cache_entry(L2TableCache l2_cache, uint64_t offset);
	196	void qed_commit_l2_cache_entry(L2TableCache l2_cache, CachedL2Table l2_table);
	197
	198	/**
	199	* Table I/O functions
	200	*/
	201	int qed_read_l1_table_sync(BDRVQEDState *s);
453e53e2	202	int qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n);
298800ca SH	203	int qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index,
	204	unsigned int n);
	205	int qed_read_l2_table_sync(BDRVQEDState s, QEDRequest request,
	206	uint64_t offset);
a8165d2d	207	int qed_read_l2_table(BDRVQEDState s, QEDRequest request, uint64_t offset);
453e53e2 KW	208	int qed_write_l2_table(BDRVQEDState s, QEDRequest request,
453e53e2 KW	209	unsigned int index, unsigned int n, bool flush);
298800ca SH	210	int qed_write_l2_table_sync(BDRVQEDState s, QEDRequest request,
	211	unsigned int index, unsigned int n, bool flush);
	212
	213	/**
	214	* Cluster functions
	215	*/
0f21b7a1 KW	216	int qed_find_cluster(BDRVQEDState s, QEDRequest request, uint64_t pos,
0f21b7a1 KW	217	size_t len, uint64_t img_offset);
298800ca SH	218
	219	/**
	220	* Consistency check
	221	*/
	222	int qed_check(BDRVQEDState s, BdrvCheckResult result, bool fix);
	223
	224	QEDTable qed_alloc_table(BDRVQEDState s);
	225
75411d23 SH	226	/**
	227	* Round down to the start of a cluster
	228	*/
	229	static inline uint64_t qed_start_of_cluster(BDRVQEDState *s, uint64_t offset)
	230	{
	231	return offset & ~(uint64_t)(s->header.cluster_size - 1);
	232	}
	233
298800ca SH	234	static inline uint64_t qed_offset_into_cluster(BDRVQEDState *s, uint64_t offset)
	235	{
	236	return offset & (s->header.cluster_size - 1);
	237	}
	238
19dfc44a	239	static inline uint64_t qed_bytes_to_clusters(BDRVQEDState *s, uint64_t bytes)
298800ca SH	240	{
	241	return qed_start_of_cluster(s, bytes + (s->header.cluster_size - 1)) /
	242	(s->header.cluster_size - 1);
	243	}
	244
	245	static inline unsigned int qed_l1_index(BDRVQEDState *s, uint64_t pos)
	246	{
	247	return pos >> s->l1_shift;
	248	}
	249
	250	static inline unsigned int qed_l2_index(BDRVQEDState *s, uint64_t pos)
	251	{
	252	return (pos >> s->l2_shift) & s->l2_mask;
	253	}
	254
75411d23 SH	255	/**
	256	* Test if a cluster offset is valid
	257	*/
	258	static inline bool qed_check_cluster_offset(BDRVQEDState *s, uint64_t offset)
	259	{
	260	uint64_t header_size = (uint64_t)s->header.header_size *
	261	s->header.cluster_size;
	262
	263	if (offset & (s->header.cluster_size - 1)) {
	264	return false;
	265	}
	266	return offset >= header_size && offset < s->file_size;
	267	}
	268
	269	/**
	270	* Test if a table offset is valid
	271	*/
	272	static inline bool qed_check_table_offset(BDRVQEDState *s, uint64_t offset)
	273	{
	274	uint64_t end_offset = offset + (s->header.table_size - 1) *
	275	s->header.cluster_size;
	276
	277	/* Overflow check */
	278	if (end_offset <= offset) {
	279	return false;
	280	}
	281
	282	return qed_check_cluster_offset(s, offset) &&
	283	qed_check_cluster_offset(s, end_offset);
	284	}
	285
21df65b6 AL	286	static inline bool qed_offset_is_cluster_aligned(BDRVQEDState *s,
	287	uint64_t offset)
	288	{
	289	if (qed_offset_into_cluster(s, offset)) {
	290	return false;
	291	}
	292	return true;
	293	}
	294
	295	static inline bool qed_offset_is_unalloc_cluster(uint64_t offset)
	296	{
	297	if (offset == 0) {
	298	return true;
	299	}
	300	return false;
	301	}
	302
	303	static inline bool qed_offset_is_zero_cluster(uint64_t offset)
	304	{
	305	if (offset == 1) {
	306	return true;
	307	}
	308	return false;
	309	}
	310
75411d23	311	#endif /* BLOCK_QED_H */