2010-12-07 03:53:01 +08:00
|
|
|
/*
|
|
|
|
* QEMU Block driver for RADOS (Ceph)
|
|
|
|
*
|
2011-05-27 07:07:31 +08:00
|
|
|
* Copyright (C) 2010-2011 Christian Brunner <chb@muc.de>,
|
|
|
|
* Josh Durgin <josh.durgin@dreamhost.com>
|
2010-12-07 03:53:01 +08:00
|
|
|
*
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2. See
|
|
|
|
* the COPYING file in the top-level directory.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
#include <inttypes.h>
|
|
|
|
|
2010-12-07 03:53:01 +08:00
|
|
|
#include "qemu-common.h"
|
|
|
|
#include "qemu-error.h"
|
|
|
|
|
|
|
|
#include "block_int.h"
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
#include <rbd/librbd.h>
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When specifying the image filename use:
|
|
|
|
*
|
2011-05-27 07:07:32 +08:00
|
|
|
* rbd:poolname/devicename[@snapshotname][:option1=value1[:option2=value2...]]
|
2010-12-07 03:53:01 +08:00
|
|
|
*
|
|
|
|
* poolname must be the name of an existing rados pool
|
|
|
|
*
|
|
|
|
* devicename is the basename for all objects used to
|
|
|
|
* emulate the raw device.
|
|
|
|
*
|
2011-05-27 07:07:32 +08:00
|
|
|
* Each option given is used to configure rados, and may be
|
|
|
|
* any Ceph option, or "conf". The "conf" option specifies
|
|
|
|
* a Ceph configuration file to read.
|
|
|
|
*
|
2010-12-07 03:53:01 +08:00
|
|
|
* Metadata information (image size, ...) is stored in an
|
|
|
|
* object with the name "devicename.rbd".
|
|
|
|
*
|
|
|
|
* The raw device is split into 4MB sized objects by default.
|
|
|
|
* The sequencenumber is encoded in a 12 byte long hex-string,
|
|
|
|
* and is attached to the devicename, separated by a dot.
|
|
|
|
* e.g. "devicename.1234567890ab"
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define OBJ_MAX_SIZE (1UL << OBJ_DEFAULT_OBJ_ORDER)
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
#define RBD_MAX_CONF_NAME_SIZE 128
|
|
|
|
#define RBD_MAX_CONF_VAL_SIZE 512
|
|
|
|
#define RBD_MAX_CONF_SIZE 1024
|
|
|
|
#define RBD_MAX_POOL_NAME_SIZE 128
|
|
|
|
#define RBD_MAX_SNAP_NAME_SIZE 128
|
|
|
|
#define RBD_MAX_SNAPS 100
|
|
|
|
|
2010-12-07 03:53:01 +08:00
|
|
|
typedef struct RBDAIOCB {
|
|
|
|
BlockDriverAIOCB common;
|
|
|
|
QEMUBH *bh;
|
|
|
|
int ret;
|
|
|
|
QEMUIOVector *qiov;
|
|
|
|
char *bounce;
|
|
|
|
int write;
|
|
|
|
int64_t sector_num;
|
|
|
|
int error;
|
|
|
|
struct BDRVRBDState *s;
|
|
|
|
int cancelled;
|
|
|
|
} RBDAIOCB;
|
|
|
|
|
|
|
|
typedef struct RADOSCB {
|
|
|
|
int rcbid;
|
|
|
|
RBDAIOCB *acb;
|
|
|
|
struct BDRVRBDState *s;
|
|
|
|
int done;
|
2011-05-27 07:07:31 +08:00
|
|
|
int64_t size;
|
2010-12-07 03:53:01 +08:00
|
|
|
char *buf;
|
|
|
|
int ret;
|
|
|
|
} RADOSCB;
|
|
|
|
|
|
|
|
#define RBD_FD_READ 0
|
|
|
|
#define RBD_FD_WRITE 1
|
|
|
|
|
|
|
|
typedef struct BDRVRBDState {
|
|
|
|
int fds[2];
|
2011-05-27 07:07:31 +08:00
|
|
|
rados_t cluster;
|
|
|
|
rados_ioctx_t io_ctx;
|
|
|
|
rbd_image_t image;
|
|
|
|
char name[RBD_MAX_IMAGE_NAME_SIZE];
|
2010-12-07 03:53:01 +08:00
|
|
|
int qemu_aio_count;
|
2011-05-27 07:07:31 +08:00
|
|
|
char *snap;
|
2010-12-07 03:53:01 +08:00
|
|
|
int event_reader_pos;
|
|
|
|
RADOSCB *event_rcb;
|
|
|
|
} BDRVRBDState;
|
|
|
|
|
|
|
|
static void rbd_aio_bh_cb(void *opaque);
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_next_tok(char *dst, int dst_len,
|
|
|
|
char *src, char delim,
|
|
|
|
const char *name,
|
|
|
|
char **p)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
int l;
|
|
|
|
char *end;
|
|
|
|
|
|
|
|
*p = NULL;
|
|
|
|
|
|
|
|
if (delim != '\0') {
|
|
|
|
end = strchr(src, delim);
|
|
|
|
if (end) {
|
|
|
|
*p = end + 1;
|
|
|
|
*end = '\0';
|
|
|
|
}
|
|
|
|
}
|
|
|
|
l = strlen(src);
|
|
|
|
if (l >= dst_len) {
|
|
|
|
error_report("%s too long", name);
|
|
|
|
return -EINVAL;
|
|
|
|
} else if (l == 0) {
|
|
|
|
error_report("%s too short", name);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
pstrcpy(dst, dst_len, src);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_parsename(const char *filename,
|
|
|
|
char *pool, int pool_len,
|
|
|
|
char *snap, int snap_len,
|
2011-05-27 07:07:32 +08:00
|
|
|
char *name, int name_len,
|
|
|
|
char *conf, int conf_len)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
const char *start;
|
|
|
|
char *p, *buf;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!strstart(filename, "rbd:", &start)) {
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
buf = qemu_strdup(start);
|
|
|
|
p = buf;
|
2011-05-27 07:07:32 +08:00
|
|
|
*snap = '\0';
|
|
|
|
*conf = '\0';
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
ret = qemu_rbd_next_tok(pool, pool_len, p, '/', "pool name", &p);
|
2010-12-07 03:53:01 +08:00
|
|
|
if (ret < 0 || !p) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto done;
|
|
|
|
}
|
2011-05-27 07:07:32 +08:00
|
|
|
|
|
|
|
if (strchr(p, '@')) {
|
|
|
|
ret = qemu_rbd_next_tok(name, name_len, p, '@', "object name", &p);
|
|
|
|
if (ret < 0) {
|
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
ret = qemu_rbd_next_tok(snap, snap_len, p, ':', "snap name", &p);
|
|
|
|
} else {
|
|
|
|
ret = qemu_rbd_next_tok(name, name_len, p, ':', "object name", &p);
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
2011-05-27 07:07:32 +08:00
|
|
|
if (ret < 0 || !p) {
|
2010-12-07 03:53:01 +08:00
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:32 +08:00
|
|
|
ret = qemu_rbd_next_tok(conf, conf_len, p, '\0', "configuration", &p);
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
done:
|
|
|
|
qemu_free(buf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:32 +08:00
|
|
|
static int qemu_rbd_set_conf(rados_t cluster, const char *conf)
|
|
|
|
{
|
|
|
|
char *p, *buf;
|
|
|
|
char name[RBD_MAX_CONF_NAME_SIZE];
|
|
|
|
char value[RBD_MAX_CONF_VAL_SIZE];
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
buf = qemu_strdup(conf);
|
|
|
|
p = buf;
|
|
|
|
|
|
|
|
while (p) {
|
|
|
|
ret = qemu_rbd_next_tok(name, sizeof(name), p,
|
|
|
|
'=', "conf option name", &p);
|
|
|
|
if (ret < 0) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!p) {
|
|
|
|
error_report("conf option %s has no value", name);
|
|
|
|
ret = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = qemu_rbd_next_tok(value, sizeof(value), p,
|
|
|
|
':', "conf option value", &p);
|
|
|
|
if (ret < 0) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (strcmp(name, "conf")) {
|
|
|
|
ret = rados_conf_set(cluster, name, value);
|
|
|
|
if (ret < 0) {
|
|
|
|
error_report("invalid conf option %s", name);
|
|
|
|
ret = -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
ret = rados_conf_read_file(cluster, value);
|
|
|
|
if (ret < 0) {
|
|
|
|
error_report("error reading conf file %s", value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
qemu_free(buf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_create(const char *filename, QEMUOptionParameter *options)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
int64_t bytes = 0;
|
|
|
|
int64_t objsize;
|
2011-05-27 07:07:31 +08:00
|
|
|
int obj_order = 0;
|
|
|
|
char pool[RBD_MAX_POOL_NAME_SIZE];
|
|
|
|
char name[RBD_MAX_IMAGE_NAME_SIZE];
|
|
|
|
char snap_buf[RBD_MAX_SNAP_NAME_SIZE];
|
2011-05-27 07:07:32 +08:00
|
|
|
char conf[RBD_MAX_CONF_SIZE];
|
2010-12-07 03:53:01 +08:00
|
|
|
char *snap = NULL;
|
2011-05-27 07:07:31 +08:00
|
|
|
rados_t cluster;
|
|
|
|
rados_ioctx_t io_ctx;
|
2010-12-07 03:53:01 +08:00
|
|
|
int ret;
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (qemu_rbd_parsename(filename, pool, sizeof(pool),
|
|
|
|
snap_buf, sizeof(snap_buf),
|
2011-05-27 07:07:32 +08:00
|
|
|
name, sizeof(name),
|
|
|
|
conf, sizeof(conf)) < 0) {
|
2010-12-07 03:53:01 +08:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
if (snap_buf[0] != '\0') {
|
|
|
|
snap = snap_buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Read out options */
|
|
|
|
while (options && options->name) {
|
|
|
|
if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
|
|
|
|
bytes = options->value.n;
|
|
|
|
} else if (!strcmp(options->name, BLOCK_OPT_CLUSTER_SIZE)) {
|
|
|
|
if (options->value.n) {
|
|
|
|
objsize = options->value.n;
|
|
|
|
if ((objsize - 1) & objsize) { /* not a power of 2? */
|
|
|
|
error_report("obj size needs to be power of 2");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
if (objsize < 4096) {
|
|
|
|
error_report("obj size too small");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2011-05-27 07:07:31 +08:00
|
|
|
obj_order = ffs(objsize) - 1;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
options++;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (rados_create(&cluster, NULL) < 0) {
|
2010-12-07 03:53:01 +08:00
|
|
|
error_report("error initializing");
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:32 +08:00
|
|
|
if (strstr(conf, "conf=") == NULL) {
|
|
|
|
if (rados_conf_read_file(cluster, NULL) < 0) {
|
|
|
|
error_report("error reading config file");
|
|
|
|
rados_shutdown(cluster);
|
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (conf[0] != '\0' &&
|
|
|
|
qemu_rbd_set_conf(cluster, conf) < 0) {
|
|
|
|
error_report("error setting config options");
|
2011-05-27 07:07:31 +08:00
|
|
|
rados_shutdown(cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (rados_connect(cluster) < 0) {
|
|
|
|
error_report("error connecting");
|
|
|
|
rados_shutdown(cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
return -EIO;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (rados_ioctx_create(cluster, pool, &io_ctx) < 0) {
|
|
|
|
error_report("error opening pool %s", pool);
|
|
|
|
rados_shutdown(cluster);
|
|
|
|
return -EIO;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
ret = rbd_create(io_ctx, name, bytes, &obj_order);
|
|
|
|
rados_ioctx_destroy(io_ctx);
|
|
|
|
rados_shutdown(cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2011-05-27 07:07:31 +08:00
|
|
|
* This aio completion is being called from qemu_rbd_aio_event_reader()
|
|
|
|
* and runs in qemu context. It schedules a bh, but just in case the aio
|
2010-12-07 03:53:01 +08:00
|
|
|
* was not cancelled before.
|
|
|
|
*/
|
2011-05-27 07:07:31 +08:00
|
|
|
static void qemu_rbd_complete_aio(RADOSCB *rcb)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
RBDAIOCB *acb = rcb->acb;
|
|
|
|
int64_t r;
|
|
|
|
|
|
|
|
if (acb->cancelled) {
|
2011-05-27 07:07:31 +08:00
|
|
|
qemu_vfree(acb->bounce);
|
|
|
|
qemu_aio_release(acb);
|
2010-12-07 03:53:01 +08:00
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
|
|
|
|
r = rcb->ret;
|
|
|
|
|
|
|
|
if (acb->write) {
|
|
|
|
if (r < 0) {
|
|
|
|
acb->ret = r;
|
|
|
|
acb->error = 1;
|
|
|
|
} else if (!acb->error) {
|
2011-05-27 07:07:31 +08:00
|
|
|
acb->ret = rcb->size;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
} else {
|
2011-05-27 07:07:31 +08:00
|
|
|
if (r < 0) {
|
|
|
|
memset(rcb->buf, 0, rcb->size);
|
2010-12-07 03:53:01 +08:00
|
|
|
acb->ret = r;
|
|
|
|
acb->error = 1;
|
2011-05-27 07:07:31 +08:00
|
|
|
} else if (r < rcb->size) {
|
|
|
|
memset(rcb->buf + r, 0, rcb->size - r);
|
2010-12-07 03:53:01 +08:00
|
|
|
if (!acb->error) {
|
2011-05-27 07:07:31 +08:00
|
|
|
acb->ret = rcb->size;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
} else if (!acb->error) {
|
2011-05-27 07:07:31 +08:00
|
|
|
acb->ret = r;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
/* Note that acb->bh can be NULL in case where the aio was cancelled */
|
2011-05-27 07:07:31 +08:00
|
|
|
acb->bh = qemu_bh_new(rbd_aio_bh_cb, acb);
|
|
|
|
qemu_bh_schedule(acb->bh);
|
2010-12-07 03:53:01 +08:00
|
|
|
done:
|
|
|
|
qemu_free(rcb);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* aio fd read handler. It runs in the qemu context and calls the
|
|
|
|
* completion handling of completed rados aio operations.
|
|
|
|
*/
|
2011-05-27 07:07:31 +08:00
|
|
|
static void qemu_rbd_aio_event_reader(void *opaque)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = opaque;
|
|
|
|
|
|
|
|
ssize_t ret;
|
|
|
|
|
|
|
|
do {
|
|
|
|
char *p = (char *)&s->event_rcb;
|
|
|
|
|
|
|
|
/* now read the rcb pointer that was sent from a non qemu thread */
|
|
|
|
if ((ret = read(s->fds[RBD_FD_READ], p + s->event_reader_pos,
|
|
|
|
sizeof(s->event_rcb) - s->event_reader_pos)) > 0) {
|
|
|
|
if (ret > 0) {
|
|
|
|
s->event_reader_pos += ret;
|
|
|
|
if (s->event_reader_pos == sizeof(s->event_rcb)) {
|
|
|
|
s->event_reader_pos = 0;
|
2011-05-27 07:07:31 +08:00
|
|
|
qemu_rbd_complete_aio(s->event_rcb);
|
|
|
|
s->qemu_aio_count--;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} while (ret < 0 && errno == EINTR);
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_aio_flush_cb(void *opaque)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = opaque;
|
|
|
|
|
|
|
|
return (s->qemu_aio_count > 0);
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_open(BlockDriverState *bs, const char *filename, int flags)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
2011-05-27 07:07:31 +08:00
|
|
|
char pool[RBD_MAX_POOL_NAME_SIZE];
|
|
|
|
char snap_buf[RBD_MAX_SNAP_NAME_SIZE];
|
2011-05-27 07:07:32 +08:00
|
|
|
char conf[RBD_MAX_CONF_SIZE];
|
2010-12-07 03:53:01 +08:00
|
|
|
int r;
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (qemu_rbd_parsename(filename, pool, sizeof(pool),
|
|
|
|
snap_buf, sizeof(snap_buf),
|
2011-05-27 07:07:32 +08:00
|
|
|
s->name, sizeof(s->name),
|
|
|
|
conf, sizeof(conf)) < 0) {
|
2010-12-07 03:53:01 +08:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
2011-05-27 07:07:31 +08:00
|
|
|
s->snap = NULL;
|
2010-12-07 03:53:01 +08:00
|
|
|
if (snap_buf[0] != '\0') {
|
2011-05-27 07:07:31 +08:00
|
|
|
s->snap = qemu_strdup(snap_buf);
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rados_create(&s->cluster, NULL);
|
|
|
|
if (r < 0) {
|
2010-12-07 03:53:01 +08:00
|
|
|
error_report("error initializing");
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:32 +08:00
|
|
|
if (strstr(conf, "conf=") == NULL) {
|
|
|
|
r = rados_conf_read_file(s->cluster, NULL);
|
|
|
|
if (r < 0) {
|
|
|
|
error_report("error reading config file");
|
|
|
|
rados_shutdown(s->cluster);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (conf[0] != '\0') {
|
|
|
|
r = qemu_rbd_set_conf(s->cluster, conf);
|
|
|
|
if (r < 0) {
|
|
|
|
error_report("error setting config options");
|
|
|
|
rados_shutdown(s->cluster);
|
|
|
|
return r;
|
|
|
|
}
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rados_connect(s->cluster);
|
|
|
|
if (r < 0) {
|
|
|
|
error_report("error connecting");
|
|
|
|
rados_shutdown(s->cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rados_ioctx_create(s->cluster, pool, &s->io_ctx);
|
|
|
|
if (r < 0) {
|
|
|
|
error_report("error opening pool %s", pool);
|
|
|
|
rados_shutdown(s->cluster);
|
|
|
|
return r;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rbd_open(s->io_ctx, s->name, &s->image, s->snap);
|
2010-12-07 03:53:01 +08:00
|
|
|
if (r < 0) {
|
2011-05-27 07:07:31 +08:00
|
|
|
error_report("error reading header from %s", s->name);
|
|
|
|
rados_ioctx_destroy(s->io_ctx);
|
|
|
|
rados_shutdown(s->cluster);
|
|
|
|
return r;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
bs->read_only = (s->snap != NULL);
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
s->event_reader_pos = 0;
|
|
|
|
r = qemu_pipe(s->fds);
|
|
|
|
if (r < 0) {
|
|
|
|
error_report("error opening eventfd");
|
|
|
|
goto failed;
|
|
|
|
}
|
|
|
|
fcntl(s->fds[0], F_SETFL, O_NONBLOCK);
|
|
|
|
fcntl(s->fds[1], F_SETFL, O_NONBLOCK);
|
2011-05-27 07:07:31 +08:00
|
|
|
qemu_aio_set_fd_handler(s->fds[RBD_FD_READ], qemu_rbd_aio_event_reader,
|
|
|
|
NULL, qemu_rbd_aio_flush_cb, NULL, s);
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
failed:
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_close(s->image);
|
|
|
|
rados_ioctx_destroy(s->io_ctx);
|
|
|
|
rados_shutdown(s->cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static void qemu_rbd_close(BlockDriverState *bs)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
|
|
|
|
|
|
|
close(s->fds[0]);
|
|
|
|
close(s->fds[1]);
|
|
|
|
qemu_aio_set_fd_handler(s->fds[RBD_FD_READ], NULL , NULL, NULL, NULL,
|
|
|
|
NULL);
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_close(s->image);
|
|
|
|
rados_ioctx_destroy(s->io_ctx);
|
|
|
|
qemu_free(s->snap);
|
|
|
|
rados_shutdown(s->cluster);
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Cancel aio. Since we don't reference acb in a non qemu threads,
|
|
|
|
* it is safe to access it here.
|
|
|
|
*/
|
2011-05-27 07:07:31 +08:00
|
|
|
static void qemu_rbd_aio_cancel(BlockDriverAIOCB *blockacb)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
RBDAIOCB *acb = (RBDAIOCB *) blockacb;
|
|
|
|
acb->cancelled = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static AIOPool rbd_aio_pool = {
|
|
|
|
.aiocb_size = sizeof(RBDAIOCB),
|
2011-05-27 07:07:31 +08:00
|
|
|
.cancel = qemu_rbd_aio_cancel,
|
2010-12-07 03:53:01 +08:00
|
|
|
};
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_send_pipe(BDRVRBDState *s, RADOSCB *rcb)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
2011-05-27 07:07:31 +08:00
|
|
|
int ret = 0;
|
2010-12-07 03:53:01 +08:00
|
|
|
while (1) {
|
|
|
|
fd_set wfd;
|
2011-05-27 07:07:31 +08:00
|
|
|
int fd = s->fds[RBD_FD_WRITE];
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
/* send the op pointer to the qemu thread that is responsible
|
|
|
|
for the aio/op completion. Must do it in a qemu thread context */
|
2010-12-07 03:53:01 +08:00
|
|
|
ret = write(fd, (void *)&rcb, sizeof(rcb));
|
|
|
|
if (ret >= 0) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (errno == EINTR) {
|
|
|
|
continue;
|
2011-05-27 07:07:31 +08:00
|
|
|
}
|
2010-12-07 03:53:01 +08:00
|
|
|
if (errno != EAGAIN) {
|
|
|
|
break;
|
2011-05-27 07:07:31 +08:00
|
|
|
}
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
FD_ZERO(&wfd);
|
|
|
|
FD_SET(fd, &wfd);
|
|
|
|
do {
|
|
|
|
ret = select(fd + 1, NULL, &wfd, NULL, NULL);
|
|
|
|
} while (ret < 0 && errno == EINTR);
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is the callback function for rbd_aio_read and _write
|
|
|
|
*
|
|
|
|
* Note: this function is being called from a non qemu thread so
|
|
|
|
* we need to be careful about what we do here. Generally we only
|
|
|
|
* write to the block notification pipe, and do the rest of the
|
|
|
|
* io completion handling from qemu_rbd_aio_event_reader() which
|
|
|
|
* runs in a qemu context.
|
|
|
|
*/
|
|
|
|
static void rbd_finish_aiocb(rbd_completion_t c, RADOSCB *rcb)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
rcb->ret = rbd_aio_get_return_value(c);
|
|
|
|
rbd_aio_release(c);
|
|
|
|
ret = qemu_rbd_send_pipe(rcb->s, rcb);
|
2010-12-07 03:53:01 +08:00
|
|
|
if (ret < 0) {
|
2011-05-27 07:07:31 +08:00
|
|
|
error_report("failed writing to acb->s->fds");
|
2010-12-07 03:53:01 +08:00
|
|
|
qemu_free(rcb);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
/* Callback when all queued rbd_aio requests are complete */
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
static void rbd_aio_bh_cb(void *opaque)
|
|
|
|
{
|
|
|
|
RBDAIOCB *acb = opaque;
|
|
|
|
|
|
|
|
if (!acb->write) {
|
|
|
|
qemu_iovec_from_buffer(acb->qiov, acb->bounce, acb->qiov->size);
|
|
|
|
}
|
|
|
|
qemu_vfree(acb->bounce);
|
|
|
|
acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
|
|
|
|
qemu_bh_delete(acb->bh);
|
|
|
|
acb->bh = NULL;
|
|
|
|
|
|
|
|
qemu_aio_release(acb);
|
|
|
|
}
|
|
|
|
|
|
|
|
static BlockDriverAIOCB *rbd_aio_rw_vector(BlockDriverState *bs,
|
|
|
|
int64_t sector_num,
|
|
|
|
QEMUIOVector *qiov,
|
|
|
|
int nb_sectors,
|
|
|
|
BlockDriverCompletionFunc *cb,
|
|
|
|
void *opaque, int write)
|
|
|
|
{
|
|
|
|
RBDAIOCB *acb;
|
|
|
|
RADOSCB *rcb;
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_completion_t c;
|
2010-12-07 03:53:01 +08:00
|
|
|
int64_t off, size;
|
|
|
|
char *buf;
|
2011-05-27 07:07:33 +08:00
|
|
|
int r;
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
|
|
|
|
|
|
|
acb = qemu_aio_get(&rbd_aio_pool, bs, cb, opaque);
|
2011-05-27 07:07:33 +08:00
|
|
|
if (!acb) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2010-12-07 03:53:01 +08:00
|
|
|
acb->write = write;
|
|
|
|
acb->qiov = qiov;
|
|
|
|
acb->bounce = qemu_blockalign(bs, qiov->size);
|
|
|
|
acb->ret = 0;
|
|
|
|
acb->error = 0;
|
|
|
|
acb->s = s;
|
|
|
|
acb->cancelled = 0;
|
|
|
|
acb->bh = NULL;
|
|
|
|
|
|
|
|
if (write) {
|
|
|
|
qemu_iovec_to_buffer(acb->qiov, acb->bounce);
|
|
|
|
}
|
|
|
|
|
|
|
|
buf = acb->bounce;
|
|
|
|
|
|
|
|
off = sector_num * BDRV_SECTOR_SIZE;
|
|
|
|
size = nb_sectors * BDRV_SECTOR_SIZE;
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
s->qemu_aio_count++; /* All the RADOSCB */
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
rcb = qemu_malloc(sizeof(RADOSCB));
|
|
|
|
rcb->done = 0;
|
|
|
|
rcb->acb = acb;
|
|
|
|
rcb->buf = buf;
|
|
|
|
rcb->s = acb->s;
|
|
|
|
rcb->size = size;
|
2011-05-27 07:07:33 +08:00
|
|
|
r = rbd_aio_create_completion(rcb, (rbd_callback_t) rbd_finish_aiocb, &c);
|
|
|
|
if (r < 0) {
|
|
|
|
goto failed;
|
|
|
|
}
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (write) {
|
2011-05-27 07:07:33 +08:00
|
|
|
r = rbd_aio_write(s->image, off, size, buf, c);
|
2011-05-27 07:07:31 +08:00
|
|
|
} else {
|
2011-05-27 07:07:33 +08:00
|
|
|
r = rbd_aio_read(s->image, off, size, buf, c);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (r < 0) {
|
|
|
|
goto failed;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return &acb->common;
|
2011-05-27 07:07:33 +08:00
|
|
|
|
|
|
|
failed:
|
|
|
|
qemu_free(rcb);
|
|
|
|
s->qemu_aio_count--;
|
|
|
|
qemu_aio_release(acb);
|
|
|
|
return NULL;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static BlockDriverAIOCB *qemu_rbd_aio_readv(BlockDriverState *bs,
|
|
|
|
int64_t sector_num,
|
|
|
|
QEMUIOVector *qiov,
|
|
|
|
int nb_sectors,
|
|
|
|
BlockDriverCompletionFunc *cb,
|
|
|
|
void *opaque)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
return rbd_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static BlockDriverAIOCB *qemu_rbd_aio_writev(BlockDriverState *bs,
|
|
|
|
int64_t sector_num,
|
|
|
|
QEMUIOVector *qiov,
|
|
|
|
int nb_sectors,
|
|
|
|
BlockDriverCompletionFunc *cb,
|
|
|
|
void *opaque)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
return rbd_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 1);
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_getinfo(BlockDriverState *bs, BlockDriverInfo *bdi)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_image_info_t info;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
r = rbd_stat(s->image, &info, sizeof(info));
|
|
|
|
if (r < 0) {
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
bdi->cluster_size = info.obj_size;
|
2010-12-07 03:53:01 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int64_t qemu_rbd_getlength(BlockDriverState *bs)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_image_info_t info;
|
|
|
|
int r;
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rbd_stat(s->image, &info, sizeof(info));
|
|
|
|
if (r < 0) {
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
return info.size;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_snap_create(BlockDriverState *bs,
|
|
|
|
QEMUSnapshotInfo *sn_info)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
|
|
|
int r;
|
|
|
|
|
|
|
|
if (sn_info->name[0] == '\0') {
|
|
|
|
return -EINVAL; /* we need a name for rbd snapshots */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* rbd snapshots are using the name as the user controlled unique identifier
|
|
|
|
* we can't use the rbd snapid for that purpose, as it can't be set
|
|
|
|
*/
|
|
|
|
if (sn_info->id_str[0] != '\0' &&
|
|
|
|
strcmp(sn_info->id_str, sn_info->name) != 0) {
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (strlen(sn_info->name) >= sizeof(sn_info->id_str)) {
|
|
|
|
return -ERANGE;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
r = rbd_snap_create(s->image, sn_info->name);
|
2010-12-07 03:53:01 +08:00
|
|
|
if (r < 0) {
|
2011-05-27 07:07:31 +08:00
|
|
|
error_report("failed to create snap: %s", strerror(-r));
|
2010-12-07 03:53:01 +08:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static int qemu_rbd_snap_list(BlockDriverState *bs,
|
|
|
|
QEMUSnapshotInfo **psn_tab)
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
BDRVRBDState *s = bs->opaque;
|
|
|
|
QEMUSnapshotInfo *sn_info, *sn_tab = NULL;
|
2011-05-27 07:07:31 +08:00
|
|
|
int i, snap_count;
|
|
|
|
rbd_snap_info_t *snaps;
|
|
|
|
int max_snaps = RBD_MAX_SNAPS;
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
do {
|
|
|
|
snaps = qemu_malloc(sizeof(*snaps) * max_snaps);
|
|
|
|
snap_count = rbd_snap_list(s->image, snaps, &max_snaps);
|
|
|
|
if (snap_count < 0) {
|
|
|
|
qemu_free(snaps);
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
2011-05-27 07:07:31 +08:00
|
|
|
} while (snap_count == -ERANGE);
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
if (snap_count <= 0) {
|
|
|
|
return snap_count;
|
2010-12-07 03:53:01 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
sn_tab = qemu_mallocz(snap_count * sizeof(QEMUSnapshotInfo));
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
for (i = 0; i < snap_count; i++) {
|
|
|
|
const char *snap_name = snaps[i].name;
|
2010-12-07 03:53:01 +08:00
|
|
|
|
|
|
|
sn_info = sn_tab + i;
|
|
|
|
pstrcpy(sn_info->id_str, sizeof(sn_info->id_str), snap_name);
|
|
|
|
pstrcpy(sn_info->name, sizeof(sn_info->name), snap_name);
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
sn_info->vm_state_size = snaps[i].size;
|
2010-12-07 03:53:01 +08:00
|
|
|
sn_info->date_sec = 0;
|
|
|
|
sn_info->date_nsec = 0;
|
|
|
|
sn_info->vm_clock_nsec = 0;
|
|
|
|
}
|
2011-05-27 07:07:31 +08:00
|
|
|
rbd_snap_list_end(snaps);
|
|
|
|
|
2010-12-07 03:53:01 +08:00
|
|
|
*psn_tab = sn_tab;
|
|
|
|
return snap_count;
|
|
|
|
}
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
static QEMUOptionParameter qemu_rbd_create_options[] = {
|
2010-12-07 03:53:01 +08:00
|
|
|
{
|
|
|
|
.name = BLOCK_OPT_SIZE,
|
|
|
|
.type = OPT_SIZE,
|
|
|
|
.help = "Virtual disk size"
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.name = BLOCK_OPT_CLUSTER_SIZE,
|
|
|
|
.type = OPT_SIZE,
|
|
|
|
.help = "RBD object size"
|
|
|
|
},
|
|
|
|
{NULL}
|
|
|
|
};
|
|
|
|
|
|
|
|
static BlockDriver bdrv_rbd = {
|
|
|
|
.format_name = "rbd",
|
|
|
|
.instance_size = sizeof(BDRVRBDState),
|
2011-05-27 07:07:31 +08:00
|
|
|
.bdrv_file_open = qemu_rbd_open,
|
|
|
|
.bdrv_close = qemu_rbd_close,
|
|
|
|
.bdrv_create = qemu_rbd_create,
|
|
|
|
.bdrv_get_info = qemu_rbd_getinfo,
|
|
|
|
.create_options = qemu_rbd_create_options,
|
|
|
|
.bdrv_getlength = qemu_rbd_getlength,
|
2010-12-07 03:53:01 +08:00
|
|
|
.protocol_name = "rbd",
|
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
.bdrv_aio_readv = qemu_rbd_aio_readv,
|
|
|
|
.bdrv_aio_writev = qemu_rbd_aio_writev,
|
2010-12-07 03:53:01 +08:00
|
|
|
|
2011-05-27 07:07:31 +08:00
|
|
|
.bdrv_snapshot_create = qemu_rbd_snap_create,
|
|
|
|
.bdrv_snapshot_list = qemu_rbd_snap_list,
|
2010-12-07 03:53:01 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static void bdrv_rbd_init(void)
|
|
|
|
{
|
|
|
|
bdrv_register(&bdrv_rbd);
|
|
|
|
}
|
|
|
|
|
|
|
|
block_init(bdrv_rbd_init);
|