summaryrefslogtreecommitdiffstats
path: root/drivers/block
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/block')
-rw-r--r--drivers/block/aoe/aoe.h93
-rw-r--r--drivers/block/aoe/aoeblk.c91
-rw-r--r--drivers/block/aoe/aoechr.c13
-rw-r--r--drivers/block/aoe/aoecmd.c1232
-rw-r--r--drivers/block/aoe/aoedev.c265
-rw-r--r--drivers/block/aoe/aoemain.c10
-rw-r--r--drivers/block/aoe/aoenet.c61
-rw-r--r--drivers/block/drbd/drbd_main.c13
-rw-r--r--drivers/block/nbd.c23
-rw-r--r--drivers/block/osdblk.c3
-rw-r--r--drivers/block/pktcdvd.c52
-rw-r--r--drivers/block/rbd.c1784
-rw-r--r--drivers/block/rbd_types.h27
-rw-r--r--drivers/block/virtio_blk.c306
-rw-r--r--drivers/block/xen-blkback/blkback.c1
15 files changed, 2655 insertions, 1319 deletions
diff --git a/drivers/block/aoe/aoe.h b/drivers/block/aoe/aoe.h
index db195abad69..d2ed7f18d1a 100644
--- a/drivers/block/aoe/aoe.h
+++ b/drivers/block/aoe/aoe.h
@@ -1,5 +1,5 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
-#define VERSION "47"
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
+#define VERSION "50"
#define AOE_MAJOR 152
#define DEVICE_NAME "aoe"
@@ -10,9 +10,6 @@
#define AOE_PARTITIONS (16)
#endif
-#define SYSMINOR(aoemajor, aoeminor) ((aoemajor) * NPERSHELF + (aoeminor))
-#define AOEMAJOR(sysminor) ((sysminor) / NPERSHELF)
-#define AOEMINOR(sysminor) ((sysminor) % NPERSHELF)
#define WHITESPACE " \t\v\f\n"
enum {
@@ -75,72 +72,67 @@ enum {
DEVFL_UP = 1, /* device is installed in system and ready for AoE->ATA commands */
DEVFL_TKILL = (1<<1), /* flag for timer to know when to kill self */
DEVFL_EXT = (1<<2), /* device accepts lba48 commands */
- DEVFL_CLOSEWAIT = (1<<3), /* device is waiting for all closes to revalidate */
- DEVFL_GDALLOC = (1<<4), /* need to alloc gendisk */
- DEVFL_KICKME = (1<<5), /* slow polling network card catch */
- DEVFL_NEWSIZE = (1<<6), /* need to update dev size in block layer */
-
- BUFFL_FAIL = 1,
+ DEVFL_GDALLOC = (1<<3), /* need to alloc gendisk */
+ DEVFL_KICKME = (1<<4), /* slow polling network card catch */
+ DEVFL_NEWSIZE = (1<<5), /* need to update dev size in block layer */
};
enum {
DEFAULTBCNT = 2 * 512, /* 2 sectors */
- NPERSHELF = 16, /* number of slots per shelf address */
- FREETAG = -1,
MIN_BUFS = 16,
NTARGETS = 8,
NAOEIFS = 8,
- NSKBPOOLMAX = 128,
+ NSKBPOOLMAX = 256,
+ NFACTIVE = 61,
TIMERTICK = HZ / 10,
MINTIMER = HZ >> 2,
MAXTIMER = HZ << 1,
- HELPWAIT = 20,
};
struct buf {
- struct list_head bufs;
- ulong stime; /* for disk stats */
- ulong flags;
ulong nframesout;
ulong resid;
ulong bv_resid;
- ulong bv_off;
sector_t sector;
struct bio *bio;
struct bio_vec *bv;
+ struct request *rq;
};
struct frame {
- int tag;
+ struct list_head head;
+ u32 tag;
ulong waited;
+ struct aoetgt *t; /* parent target I belong to */
+ sector_t lba;
+ struct sk_buff *skb; /* command skb freed on module exit */
+ struct sk_buff *r_skb; /* response skb for async processing */
struct buf *buf;
- char *bufaddr;
+ struct bio_vec *bv;
ulong bcnt;
- sector_t lba;
- struct sk_buff *skb;
+ ulong bv_off;
};
struct aoeif {
struct net_device *nd;
- unsigned char lost;
- unsigned char lostjumbo;
- ushort maxbcnt;
+ ulong lost;
+ int bcnt;
};
struct aoetgt {
unsigned char addr[6];
ushort nframes;
- struct frame *frames;
+ struct aoedev *d; /* parent device I belong to */
+ struct list_head ffree; /* list of free frames */
struct aoeif ifs[NAOEIFS];
struct aoeif *ifp; /* current aoeif in use */
ushort nout;
ushort maxout;
- u16 lasttag; /* last tag sent */
- u16 useme;
+ ulong falloc;
ulong lastwadj; /* last window adjustment */
+ int minbcnt;
int wpkts, rpkts;
- int dataref;
};
struct aoedev {
@@ -153,6 +145,9 @@ struct aoedev {
u16 rttavg; /* round trip average of requests/responses */
u16 mintimer;
u16 fw_ver; /* version of blade's firmware */
+ u16 lasttag; /* last tag sent */
+ u16 useme;
+ ulong ref;
struct work_struct work;/* disk create work struct */
struct gendisk *gd;
struct request_queue *blkq;
@@ -160,16 +155,31 @@ struct aoedev {
sector_t ssize;
struct timer_list timer;
spinlock_t lock;
- struct sk_buff_head sendq;
struct sk_buff_head skbpool;
mempool_t *bufpool; /* for deadlock-free Buf allocation */
- struct list_head bufq; /* queue of bios to work on */
- struct buf *inprocess; /* the one we're currently working on */
+ struct { /* pointers to work in progress */
+ struct buf *buf;
+ struct bio *nxbio;
+ struct request *rq;
+ } ip;
+ ulong maxbcnt;
+ struct list_head factive[NFACTIVE]; /* hash of active frames */
struct aoetgt *targets[NTARGETS];
struct aoetgt **tgt; /* target in use when working */
- struct aoetgt **htgt; /* target needing rexmit assistance */
+ struct aoetgt *htgt; /* target needing rexmit assistance */
+ ulong ntargets;
+ ulong kicked;
};
+/* kthread tracking */
+struct ktstate {
+ struct completion rendez;
+ struct task_struct *task;
+ wait_queue_head_t *waitq;
+ int (*fn) (void);
+ char *name;
+ spinlock_t *lock;
+};
int aoeblk_init(void);
void aoeblk_exit(void);
@@ -182,22 +192,29 @@ void aoechr_error(char *);
void aoecmd_work(struct aoedev *d);
void aoecmd_cfg(ushort aoemajor, unsigned char aoeminor);
-void aoecmd_ata_rsp(struct sk_buff *);
+struct sk_buff *aoecmd_ata_rsp(struct sk_buff *);
void aoecmd_cfg_rsp(struct sk_buff *);
void aoecmd_sleepwork(struct work_struct *);
void aoecmd_cleanslate(struct aoedev *);
+void aoecmd_exit(void);
+int aoecmd_init(void);
struct sk_buff *aoecmd_ata_id(struct aoedev *);
+void aoe_freetframe(struct frame *);
+void aoe_flush_iocq(void);
+void aoe_end_request(struct aoedev *, struct request *, int);
+int aoe_ktstart(struct ktstate *k);
+void aoe_ktstop(struct ktstate *k);
int aoedev_init(void);
void aoedev_exit(void);
-struct aoedev *aoedev_by_aoeaddr(int maj, int min);
-struct aoedev *aoedev_by_sysminor_m(ulong sysminor);
+struct aoedev *aoedev_by_aoeaddr(ulong maj, int min, int do_alloc);
void aoedev_downdev(struct aoedev *d);
int aoedev_flush(const char __user *str, size_t size);
+void aoe_failbuf(struct aoedev *, struct buf *);
+void aoedev_put(struct aoedev *);
int aoenet_init(void);
void aoenet_exit(void);
void aoenet_xmit(struct sk_buff_head *);
int is_aoe_netif(struct net_device *ifp);
int set_aoe_iflist(const char __user *str, size_t size);
-
diff --git a/drivers/block/aoe/aoeblk.c b/drivers/block/aoe/aoeblk.c
index 321de7b6c44..00dfc5008ad 100644
--- a/drivers/block/aoe/aoeblk.c
+++ b/drivers/block/aoe/aoeblk.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoeblk.c
* block device routines
@@ -161,68 +161,22 @@ aoeblk_release(struct gendisk *disk, fmode_t mode)
}
static void
-aoeblk_make_request(struct request_queue *q, struct bio *bio)
+aoeblk_request(struct request_queue *q)
{
- struct sk_buff_head queue;
struct aoedev *d;
- struct buf *buf;
- ulong flags;
-
- blk_queue_bounce(q, &bio);
-
- if (bio == NULL) {
- printk(KERN_ERR "aoe: bio is NULL\n");
- BUG();
- return;
- }
- d = bio->bi_bdev->bd_disk->private_data;
- if (d == NULL) {
- printk(KERN_ERR "aoe: bd_disk->private_data is NULL\n");
- BUG();
- bio_endio(bio, -ENXIO);
- return;
- } else if (bio->bi_io_vec == NULL) {
- printk(KERN_ERR "aoe: bi_io_vec is NULL\n");
- BUG();
- bio_endio(bio, -ENXIO);
- return;
- }
- buf = mempool_alloc(d->bufpool, GFP_NOIO);
- if (buf == NULL) {
- printk(KERN_INFO "aoe: buf allocation failure\n");
- bio_endio(bio, -ENOMEM);
- return;
- }
- memset(buf, 0, sizeof(*buf));
- INIT_LIST_HEAD(&buf->bufs);
- buf->stime = jiffies;
- buf->bio = bio;
- buf->resid = bio->bi_size;
- buf->sector = bio->bi_sector;
- buf->bv = &bio->bi_io_vec[bio->bi_idx];
- buf->bv_resid = buf->bv->bv_len;
- WARN_ON(buf->bv_resid == 0);
- buf->bv_off = buf->bv->bv_offset;
-
- spin_lock_irqsave(&d->lock, flags);
+ struct request *rq;
+ d = q->queuedata;
if ((d->flags & DEVFL_UP) == 0) {
pr_info_ratelimited("aoe: device %ld.%d is not up\n",
d->aoemajor, d->aoeminor);
- spin_unlock_irqrestore(&d->lock, flags);
- mempool_free(buf, d->bufpool);
- bio_endio(bio, -ENXIO);
+ while ((rq = blk_peek_request(q))) {
+ blk_start_request(rq);
+ aoe_end_request(d, rq, 1);
+ }
return;
}
-
- list_add_tail(&buf->bufs, &d->bufq);
-
aoecmd_work(d);
- __skb_queue_head_init(&queue);
- skb_queue_splice_init(&d->sendq, &queue);
-
- spin_unlock_irqrestore(&d->lock, flags);
- aoenet_xmit(&queue);
}
static int
@@ -254,41 +208,54 @@ aoeblk_gdalloc(void *vp)
{
struct aoedev *d = vp;
struct gendisk *gd;
+ mempool_t *mp;
+ struct request_queue *q;
+ enum { KB = 1024, MB = KB * KB, READ_AHEAD = 2 * MB, };
ulong flags;
gd = alloc_disk(AOE_PARTITIONS);
if (gd == NULL) {
- printk(KERN_ERR
- "aoe: cannot allocate disk structure for %ld.%d\n",
+ pr_err("aoe: cannot allocate disk structure for %ld.%d\n",
d->aoemajor, d->aoeminor);
goto err;
}
- d->bufpool = mempool_create_slab_pool(MIN_BUFS, buf_pool_cache);
- if (d->bufpool == NULL) {
+ mp = mempool_create(MIN_BUFS, mempool_alloc_slab, mempool_free_slab,
+ buf_pool_cache);
+ if (mp == NULL) {
printk(KERN_ERR "aoe: cannot allocate bufpool for %ld.%d\n",
d->aoemajor, d->aoeminor);
goto err_disk;
}
+ q = blk_init_queue(aoeblk_request, &d->lock);
+ if (q == NULL) {
+ pr_err("aoe: cannot allocate block queue for %ld.%d\n",
+ d->aoemajor, d->aoeminor);
+ mempool_destroy(mp);
+ goto err_disk;
+ }
d->blkq = blk_alloc_queue(GFP_KERNEL);
if (!d->blkq)
goto err_mempool;
- blk_queue_make_request(d->blkq, aoeblk_make_request);
d->blkq->backing_dev_info.name = "aoe";
if (bdi_init(&d->blkq->backing_dev_info))
goto err_blkq;
spin_lock_irqsave(&d->lock, flags);
+ blk_queue_max_hw_sectors(d->blkq, BLK_DEF_MAX_SECTORS);
+ q->backing_dev_info.ra_pages = READ_AHEAD / PAGE_CACHE_SIZE;
+ d->bufpool = mp;
+ d->blkq = gd->queue = q;
+ q->queuedata = d;
+ d->gd = gd;
gd->major = AOE_MAJOR;
- gd->first_minor = d->sysminor * AOE_PARTITIONS;
+ gd->first_minor = d->sysminor;
gd->fops = &aoe_bdops;
gd->private_data = d;
set_capacity(gd, d->ssize);
snprintf(gd->disk_name, sizeof gd->disk_name, "etherd/e%ld.%d",
d->aoemajor, d->aoeminor);
- gd->queue = d->blkq;
- d->gd = gd;
d->flags &= ~DEVFL_GDALLOC;
d->flags |= DEVFL_UP;
diff --git a/drivers/block/aoe/aoechr.c b/drivers/block/aoe/aoechr.c
index e86d2062a16..ed57a890c64 100644
--- a/drivers/block/aoe/aoechr.c
+++ b/drivers/block/aoe/aoechr.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoechr.c
* AoE character device driver
@@ -86,34 +86,34 @@ revalidate(const char __user *str, size_t size)
if (copy_from_user(buf, str, size))
return -EFAULT;
- /* should be e%d.%d format */
n = sscanf(buf, "e%d.%d", &major, &minor);
if (n != 2) {
- printk(KERN_ERR "aoe: invalid device specification\n");
+ pr_err("aoe: invalid device specification %s\n", buf);
return -EINVAL;
}
- d = aoedev_by_aoeaddr(major, minor);
+ d = aoedev_by_aoeaddr(major, minor, 0);
if (!d)
return -EINVAL;
spin_lock_irqsave(&d->lock, flags);
aoecmd_cleanslate(d);
+ aoecmd_cfg(major, minor);
loop:
skb = aoecmd_ata_id(d);
spin_unlock_irqrestore(&d->lock, flags);
/* try again if we are able to sleep a bit,
* otherwise give up this revalidation
*/
- if (!skb && !msleep_interruptible(200)) {
+ if (!skb && !msleep_interruptible(250)) {
spin_lock_irqsave(&d->lock, flags);
goto loop;
}
+ aoedev_put(d);
if (skb) {
struct sk_buff_head queue;
__skb_queue_head_init(&queue);
__skb_queue_tail(&queue, skb);
aoenet_xmit(&queue);
}
- aoecmd_cfg(major, minor);
return 0;
}
@@ -174,6 +174,7 @@ aoechr_write(struct file *filp, const char __user *buf, size_t cnt, loff_t *offp
break;
case MINOR_FLUSH:
ret = aoedev_flush(buf, cnt);
+ break;
}
if (ret == 0)
ret = cnt;
diff --git a/drivers/block/aoe/aoecmd.c b/drivers/block/aoe/aoecmd.c
index 887f68f6d79..3804a0af3ef 100644
--- a/drivers/block/aoe/aoecmd.c
+++ b/drivers/block/aoe/aoecmd.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoecmd.c
* Filesystem request handling methods
@@ -12,10 +12,19 @@
#include <linux/netdevice.h>
#include <linux/genhd.h>
#include <linux/moduleparam.h>
+#include <linux/workqueue.h>
+#include <linux/kthread.h>
#include <net/net_namespace.h>
#include <asm/unaligned.h>
+#include <linux/uio.h>
#include "aoe.h"
+#define MAXIOC (8192) /* default meant to avoid most soft lockups */
+
+static void ktcomplete(struct frame *, struct sk_buff *);
+
+static struct buf *nextbuf(struct aoedev *);
+
static int aoe_deadsecs = 60 * 3;
module_param(aoe_deadsecs, int, 0644);
MODULE_PARM_DESC(aoe_deadsecs, "After aoe_deadsecs seconds, give up and fail dev.");
@@ -25,6 +34,15 @@ module_param(aoe_maxout, int, 0644);
MODULE_PARM_DESC(aoe_maxout,
"Only aoe_maxout outstanding packets for every MAC on eX.Y.");
+static wait_queue_head_t ktiowq;
+static struct ktstate kts;
+
+/* io completion queue */
+static struct {
+ struct list_head head;
+ spinlock_t lock;
+} iocq;
+
static struct sk_buff *
new_skb(ulong len)
{
@@ -41,15 +59,21 @@ new_skb(ulong len)
}
static struct frame *
-getframe(struct aoetgt *t, int tag)
+getframe(struct aoedev *d, u32 tag)
{
- struct frame *f, *e;
+ struct frame *f;
+ struct list_head *head, *pos, *nx;
+ u32 n;
- f = t->frames;
- e = f + t->nframes;
- for (; f<e; f++)
- if (f->tag == tag)
+ n = tag % NFACTIVE;
+ head = &d->factive[n];
+ list_for_each_safe(pos, nx, head) {
+ f = list_entry(pos, struct frame, head);
+ if (f->tag == tag) {
+ list_del(pos);
return f;
+ }
+ }
return NULL;
}
@@ -59,18 +83,18 @@ getframe(struct aoetgt *t, int tag)
* This driver reserves tag -1 to mean "unused frame."
*/
static int
-newtag(struct aoetgt *t)
+newtag(struct aoedev *d)
{
register ulong n;
n = jiffies & 0xffff;
- return n |= (++t->lasttag & 0x7fff) << 16;
+ return n |= (++d->lasttag & 0x7fff) << 16;
}
-static int
+static u32
aoehdr_atainit(struct aoedev *d, struct aoetgt *t, struct aoe_hdr *h)
{
- u32 host_tag = newtag(t);
+ u32 host_tag = newtag(d);
memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src);
memcpy(h->dst, t->addr, sizeof h->dst);
@@ -95,16 +119,18 @@ put_lba(struct aoe_atahdr *ah, sector_t lba)
ah->lba5 = lba >>= 8;
}
-static void
+static struct aoeif *
ifrotate(struct aoetgt *t)
{
- t->ifp++;
- if (t->ifp >= &t->ifs[NAOEIFS] || t->ifp->nd == NULL)
- t->ifp = t->ifs;
- if (t->ifp->nd == NULL) {
- printk(KERN_INFO "aoe: no interface to rotate to\n");
- BUG();
- }
+ struct aoeif *ifp;
+
+ ifp = t->ifp;
+ ifp++;
+ if (ifp >= &t->ifs[NAOEIFS] || ifp->nd == NULL)
+ ifp = t->ifs;
+ if (ifp->nd == NULL)
+ return NULL;
+ return t->ifp = ifp;
}
static void
@@ -129,78 +155,128 @@ skb_pool_get(struct aoedev *d)
return NULL;
}
-/* freeframe is where we do our load balancing so it's a little hairy. */
+void
+aoe_freetframe(struct frame *f)
+{
+ struct aoetgt *t;
+
+ t = f->t;
+ f->buf = NULL;
+ f->bv = NULL;
+ f->r_skb = NULL;
+ list_add(&f->head, &t->ffree);
+}
+
static struct frame *
-freeframe(struct aoedev *d)
+newtframe(struct aoedev *d, struct aoetgt *t)
{
- struct frame *f, *e, *rf;
- struct aoetgt **t;
+ struct frame *f;
struct sk_buff *skb;
+ struct list_head *pos;
+
+ if (list_empty(&t->ffree)) {
+ if (t->falloc >= NSKBPOOLMAX*2)
+ return NULL;
+ f = kcalloc(1, sizeof(*f), GFP_ATOMIC);
+ if (f == NULL)
+ return NULL;
+ t->falloc++;
+ f->t = t;
+ } else {
+ pos = t->ffree.next;
+ list_del(pos);
+ f = list_entry(pos, struct frame, head);
+ }
+
+ skb = f->skb;
+ if (skb == NULL) {
+ f->skb = skb = new_skb(ETH_ZLEN);
+ if (!skb) {
+bail: aoe_freetframe(f);
+ return NULL;
+ }
+ }
+
+ if (atomic_read(&skb_shinfo(skb)->dataref) != 1) {
+ skb = skb_pool_get(d);
+ if (skb == NULL)
+ goto bail;
+ skb_pool_put(d, f->skb);
+ f->skb = skb;
+ }
+
+ skb->truesize -= skb->data_len;
+ skb_shinfo(skb)->nr_frags = skb->data_len = 0;
+ skb_trim(skb, 0);
+ return f;
+}
+
+static struct frame *
+newframe(struct aoedev *d)
+{
+ struct frame *f;
+ struct aoetgt *t, **tt;
+ int totout = 0;
if (d->targets[0] == NULL) { /* shouldn't happen, but I'm paranoid */
printk(KERN_ERR "aoe: NULL TARGETS!\n");
return NULL;
}
- t = d->tgt;
- t++;
- if (t >= &d->targets[NTARGETS] || !*t)
- t = d->targets;
+ tt = d->tgt; /* last used target */
for (;;) {
- if ((*t)->nout < (*t)->maxout
+ tt++;
+ if (tt >= &d->targets[NTARGETS] || !*tt)
+ tt = d->targets;
+ t = *tt;
+ totout += t->nout;
+ if (t->nout < t->maxout
&& t != d->htgt
- && (*t)->ifp->nd) {
- rf = NULL;
- f = (*t)->frames;
- e = f + (*t)->nframes;
- for (; f < e; f++) {
- if (f->tag != FREETAG)
- continue;
- skb = f->skb;
- if (!skb
- && !(f->skb = skb = new_skb(ETH_ZLEN)))
- continue;
- if (atomic_read(&skb_shinfo(skb)->dataref)
- != 1) {
- if (!rf)
- rf = f;
- continue;
- }
-gotone: skb_shinfo(skb)->nr_frags = skb->data_len = 0;
- skb_trim(skb, 0);
- d->tgt = t;
- ifrotate(*t);
+ && t->ifp->nd) {
+ f = newtframe(d, t);
+ if (f) {
+ ifrotate(t);
+ d->tgt = tt;
return f;
}
- /* Work can be done, but the network layer is
- holding our precious packets. Try to grab
- one from the pool. */
- f = rf;
- if (f == NULL) { /* more paranoia */
- printk(KERN_ERR
- "aoe: freeframe: %s.\n",
- "unexpected null rf");
- d->flags |= DEVFL_KICKME;
- return NULL;
- }
- skb = skb_pool_get(d);
- if (skb) {
- skb_pool_put(d, f->skb);
- f->skb = skb;
- goto gotone;
- }
- (*t)->dataref++;
- if ((*t)->nout == 0)
- d->flags |= DEVFL_KICKME;
}
- if (t == d->tgt) /* we've looped and found nada */
+ if (tt == d->tgt) /* we've looped and found nada */
break;
- t++;
- if (t >= &d->targets[NTARGETS] || !*t)
- t = d->targets;
+ }
+ if (totout == 0) {
+ d->kicked++;
+ d->flags |= DEVFL_KICKME;
}
return NULL;
}
+static void
+skb_fillup(struct sk_buff *skb, struct bio_vec *bv, ulong off, ulong cnt)
+{
+ int frag = 0;
+ ulong fcnt;
+loop:
+ fcnt = bv->bv_len - (off - bv->bv_offset);
+ if (fcnt > cnt)
+ fcnt = cnt;
+ skb_fill_page_desc(skb, frag++, bv->bv_page, off, fcnt);
+ cnt -= fcnt;
+ if (cnt <= 0)
+ return;
+ bv++;
+ off = bv->bv_offset;
+ goto loop;
+}
+
+static void
+fhash(struct frame *f)
+{
+ struct aoedev *d = f->t->d;
+ u32 n;
+
+ n = f->tag % NFACTIVE;
+ list_add_tail(&f->head, &d->factive[n]);
+}
+
static int
aoecmd_ata_rw(struct aoedev *d)
{
@@ -208,26 +284,47 @@ aoecmd_ata_rw(struct aoedev *d)
struct aoe_hdr *h;
struct aoe_atahdr *ah;
struct buf *buf;
- struct bio_vec *bv;
struct aoetgt *t;
struct sk_buff *skb;
- ulong bcnt;
+ struct sk_buff_head queue;
+ ulong bcnt, fbcnt;
char writebit, extbit;
writebit = 0x10;
extbit = 0x4;
- f = freeframe(d);
+ buf = nextbuf(d);
+ if (buf == NULL)
+ return 0;
+ f = newframe(d);
if (f == NULL)
return 0;
t = *d->tgt;
- buf = d->inprocess;
- bv = buf->bv;
- bcnt = t->ifp->maxbcnt;
+ bcnt = d->maxbcnt;
if (bcnt == 0)
bcnt = DEFAULTBCNT;
- if (bcnt > buf->bv_resid)
- bcnt = buf->bv_resid;
+ if (bcnt > buf->resid)
+ bcnt = buf->resid;
+ fbcnt = bcnt;
+ f->bv = buf->bv;
+ f->bv_off = f->bv->bv_offset + (f->bv->bv_len - buf->bv_resid);
+ do {
+ if (fbcnt < buf->bv_resid) {
+ buf->bv_resid -= fbcnt;
+ buf->resid -= fbcnt;
+ break;
+ }
+ fbcnt -= buf->bv_resid;
+ buf->resid -= buf->bv_resid;
+ if (buf->resid == 0) {
+ d->ip.buf = NULL;
+ break;
+ }
+ buf->bv++;
+ buf->bv_resid = buf->bv->bv_len;
+ WARN_ON(buf->bv_resid == 0);
+ } while (fbcnt);
+
/* initialize the headers & frame */
skb = f->skb;
h = (struct aoe_hdr *) skb_mac_header(skb);
@@ -235,10 +332,10 @@ aoecmd_ata_rw(struct aoedev *d)
skb_put(skb, sizeof *h + sizeof *ah);
memset(h, 0, skb->len);
f->tag = aoehdr_atainit(d, t, h);
+ fhash(f);
t->nout++;
f->waited = 0;
f->buf = buf;
- f->bufaddr = page_address(bv->bv_page) + buf->bv_off;
f->bcnt = bcnt;
f->lba = buf->sector;
@@ -253,10 +350,11 @@ aoecmd_ata_rw(struct aoedev *d)
ah->lba3 |= 0xe0; /* LBA bit + obsolete 0xa0 */
}
if (bio_data_dir(buf->bio) == WRITE) {
- skb_fill_page_desc(skb, 0, bv->bv_page, buf->bv_off, bcnt);
+ skb_fillup(skb, f->bv, f->bv_off, bcnt);
ah->aflags |= AOEAFL_WRITE;
skb->len += bcnt;
skb->data_len = bcnt;
+ skb->truesize += bcnt;
t->wpkts++;
} else {
t->rpkts++;
@@ -267,23 +365,15 @@ aoecmd_ata_rw(struct aoedev *d)
/* mark all tracking fields and load out */
buf->nframesout += 1;
- buf->bv_off += bcnt;
- buf->bv_resid -= bcnt;
- buf->resid -= bcnt;
buf->sector += bcnt >> 9;
- if (buf->resid == 0) {
- d->inprocess = NULL;
- } else if (buf->bv_resid == 0) {
- buf->bv = ++bv;
- buf->bv_resid = bv->bv_len;
- WARN_ON(buf->bv_resid == 0);
- buf->bv_off = bv->bv_offset;
- }
skb->dev = t->ifp->nd;
skb = skb_clone(skb, GFP_ATOMIC);
- if (skb)
- __skb_queue_tail(&d->sendq, skb);
+ if (skb) {
+ __skb_queue_head_init(&queue);
+ __skb_queue_tail(&queue, skb);
+ aoenet_xmit(&queue);
+ }
return 1;
}
@@ -330,17 +420,25 @@ cont:
}
static void
-resend(struct aoedev *d, struct aoetgt *t, struct frame *f)
+resend(struct aoedev *d, struct frame *f)
{
struct sk_buff *skb;
+ struct sk_buff_head queue;
struct aoe_hdr *h;
struct aoe_atahdr *ah;
+ struct aoetgt *t;
char buf[128];
u32 n;
- ifrotate(t);
- n = newtag(t);
+ t = f->t;
+ n = newtag(d);
skb = f->skb;
+ if (ifrotate(t) == NULL) {
+ /* probably can't happen, but set it up to fail anyway */
+ pr_info("aoe: resend: no interfaces to rotate to.\n");
+ ktcomplete(f, NULL);
+ return;
+ }
h = (struct aoe_hdr *) skb_mac_header(skb);
ah = (struct aoe_atahdr *) (h+1);
@@ -351,39 +449,22 @@ resend(struct aoedev *d, struct aoetgt *t, struct frame *f)
aoechr_error(buf);
f->tag = n;
+ fhash(f);
h->tag = cpu_to_be32(n);
memcpy(h->dst, t->addr, sizeof h->dst);
memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src);
- switch (ah->cmdstat) {
- default:
- break;
- case ATA_CMD_PIO_READ:
- case ATA_CMD_PIO_READ_EXT:
- case ATA_CMD_PIO_WRITE:
- case ATA_CMD_PIO_WRITE_EXT:
- put_lba(ah, f->lba);
-
- n = f->bcnt;
- if (n > DEFAULTBCNT)
- n = DEFAULTBCNT;
- ah->scnt = n >> 9;
- if (ah->aflags & AOEAFL_WRITE) {
- skb_fill_page_desc(skb, 0, virt_to_page(f->bufaddr),
- offset_in_page(f->bufaddr), n);
- skb->len = sizeof *h + sizeof *ah + n;
- skb->data_len = n;
- }
- }
skb->dev = t->ifp->nd;
skb = skb_clone(skb, GFP_ATOMIC);
if (skb == NULL)
return;
- __skb_queue_tail(&d->sendq, skb);
+ __skb_queue_head_init(&queue);
+ __skb_queue_tail(&queue, skb);
+ aoenet_xmit(&queue);
}
static int
-tsince(int tag)
+tsince(u32 tag)
{
int n;
@@ -407,58 +488,65 @@ getif(struct aoetgt *t, struct net_device *nd)
return NULL;
}
-static struct aoeif *
-addif(struct aoetgt *t, struct net_device *nd)
-{
- struct aoeif *p;
-
- p = getif(t, NULL);
- if (!p)
- return NULL;
- p->nd = nd;
- p->maxbcnt = DEFAULTBCNT;
- p->lost = 0;
- p->lostjumbo = 0;
- return p;
-}
-
static void
ejectif(struct aoetgt *t, struct aoeif *ifp)
{
struct aoeif *e;
+ struct net_device *nd;
ulong n;
+ nd = ifp->nd;
e = t->ifs + NAOEIFS - 1;
n = (e - ifp) * sizeof *ifp;
memmove(ifp, ifp+1, n);
e->nd = NULL;
+ dev_put(nd);
}
static int
sthtith(struct aoedev *d)
{
- struct frame *f, *e, *nf;
+ struct frame *f, *nf;
+ struct list_head *nx, *pos, *head;
struct sk_buff *skb;
- struct aoetgt *ht = *d->htgt;
-
- f = ht->frames;
- e = f + ht->nframes;
- for (; f < e; f++) {
- if (f->tag == FREETAG)
- continue;
- nf = freeframe(d);
- if (!nf)
- return 0;
- skb = nf->skb;
- *nf = *f;
- f->skb = skb;
- f->tag = FREETAG;
- nf->waited = 0;
- ht->nout--;
- (*d->tgt)->nout++;
- resend(d, *d->tgt, nf);
+ struct aoetgt *ht = d->htgt;
+ int i;
+
+ for (i = 0; i < NFACTIVE; i++) {
+ head = &d->factive[i];
+ list_for_each_safe(pos, nx, head) {
+ f = list_entry(pos, struct frame, head);
+ if (f->t != ht)
+ continue;
+
+ nf = newframe(d);
+ if (!nf)
+ return 0;
+
+ /* remove frame from active list */
+ list_del(pos);
+
+ /* reassign all pertinent bits to new outbound frame */
+ skb = nf->skb;
+ nf->skb = f->skb;
+ nf->buf = f->buf;
+ nf->bcnt = f->bcnt;
+ nf->lba = f->lba;
+ nf->bv = f->bv;
+ nf->bv_off = f->bv_off;
+ nf->waited = 0;
+ f->skb = skb;
+ aoe_freetframe(f);
+ ht->nout--;
+ nf->t->nout++;
+ resend(d, nf);
+ }
}
- /* he's clean, he's useless. take away his interfaces */
+ /* We've cleaned up the outstanding so take away his
+ * interfaces so he won't be used. We should remove him from
+ * the target array here, but cleaning up a target is
+ * involved. PUNT!
+ */
memset(ht->ifs, 0, sizeof ht->ifs);
d->htgt = NULL;
return 1;
@@ -477,13 +565,15 @@ ata_scnt(unsigned char *packet) {
static void
rexmit_timer(ulong vp)
{
- struct sk_buff_head queue;
struct aoedev *d;
struct aoetgt *t, **tt, **te;
struct aoeif *ifp;
- struct frame *f, *e;
+ struct frame *f;
+ struct list_head *head, *pos, *nx;
+ LIST_HEAD(flist);
register long timeout;
ulong flags, n;
+ int i;
d = (struct aoedev *) vp;
@@ -497,58 +587,22 @@ rexmit_timer(ulong vp)
spin_unlock_irqrestore(&d->lock, flags);
return;
}
- tt = d->targets;
- te = tt + NTARGETS;
- for (; tt < te && *tt; tt++) {
- t = *tt;
- f = t->frames;
- e = f + t->nframes;
- for (; f < e; f++) {
- if (f->tag == FREETAG
- || tsince(f->tag) < timeout)
- continue;
- n = f->waited += timeout;
- n /= HZ;
- if (n > aoe_deadsecs) {
- /* waited too long. device failure. */
- aoedev_downdev(d);
- break;
- }
-
- if (n > HELPWAIT /* see if another target can help */
- && (tt != d->targets || d->targets[1]))
- d->htgt = tt;
-
- if (t->nout == t->maxout) {
- if (t->maxout > 1)
- t->maxout--;
- t->lastwadj = jiffies;
- }
-
- ifp = getif(t, f->skb->dev);
- if (ifp && ++ifp->lost > (t->nframes << 1)
- && (ifp != t->ifs || t->ifs[1].nd)) {
- ejectif(t, ifp);
- ifp = NULL;
- }
- if (ata_scnt(skb_mac_header(f->skb)) > DEFAULTBCNT / 512
- && ifp && ++ifp->lostjumbo > (t->nframes << 1)
- && ifp->maxbcnt != DEFAULTBCNT) {
- printk(KERN_INFO
- "aoe: e%ld.%d: "
- "too many lost jumbo on "
- "%s:%pm - "
- "falling back to %d frames.\n",
- d->aoemajor, d->aoeminor,
- ifp->nd->name, t->addr,
- DEFAULTBCNT);
- ifp->maxbcnt = 0;
- }
- resend(d, t, f);
+ /* collect all frames to rexmit into flist */
+ for (i = 0; i < NFACTIVE; i++) {
+ head = &d->factive[i];
+ list_for_each_safe(pos, nx, head) {
+ f = list_entry(pos, struct frame, head);
+ if (tsince(f->tag) < timeout)
+ break; /* end of expired frames */
+ /* move to flist for later processing */
+ list_move_tail(pos, &flist);
}
-
- /* window check */
+ }
+ /* window check */
+ tt = d->targets;
+ te = tt + d->ntargets;
+ for (; tt < te && (t = *tt); tt++) {
if (t->nout == t->maxout
&& t->maxout < t->nframes
&& (jiffies - t->lastwadj)/HZ > 10) {
@@ -557,45 +611,173 @@ rexmit_timer(ulong vp)
}
}
- if (!skb_queue_empty(&d->sendq)) {
+ if (!list_empty(&flist)) { /* retransmissions necessary */
n = d->rttavg <<= 1;
if (n > MAXTIMER)
d->rttavg = MAXTIMER;
}
- if (d->flags & DEVFL_KICKME || d->htgt) {
- d->flags &= ~DEVFL_KICKME;
- aoecmd_work(d);
+ /* process expired frames */
+ while (!list_empty(&flist)) {
+ pos = flist.next;
+ f = list_entry(pos, struct frame, head);
+ n = f->waited += timeout;
+ n /= HZ;
+ if (n > aoe_deadsecs) {
+ /* Waited too long. Device failure.
+ * Hang all frames on first hash bucket for downdev
+ * to clean up.
+ */
+ list_splice(&flist, &d->factive[0]);
+ aoedev_downdev(d);
+ break;
+ }
+ list_del(pos);
+
+ t = f->t;
+ if (n > aoe_deadsecs/2)
+ d->htgt = t; /* see if another target can help */
+
+ if (t->nout == t->maxout) {
+ if (t->maxout > 1)
+ t->maxout--;
+ t->lastwadj = jiffies;
+ }
+
+ ifp = getif(t, f->skb->dev);
+ if (ifp && ++ifp->lost > (t->nframes << 1)
+ && (ifp != t->ifs || t->ifs[1].nd)) {
+ ejectif(t, ifp);
+ ifp = NULL;
+ }
+ resend(d, f);
}
- __skb_queue_head_init(&queue);
- skb_queue_splice_init(&d->sendq, &queue);
+ if ((d->flags & DEVFL_KICKME || d->htgt) && d->blkq) {
+ d->flags &= ~DEVFL_KICKME;
+ d->blkq->request_fn(d->blkq);
+ }
d->timer.expires = jiffies + TIMERTICK;
add_timer(&d->timer);
spin_unlock_irqrestore(&d->lock, flags);
+}
- aoenet_xmit(&queue);
+static unsigned long
+rqbiocnt(struct request *r)
+{
+ struct bio *bio;
+ unsigned long n = 0;
+
+ __rq_for_each_bio(bio, r)
+ n++;
+ return n;
+}
+
+/* This can be removed if we are certain that no users of the block
+ * layer will ever use zero-count pages in bios. Otherwise we have to
+ * protect against the put_page sometimes done by the network layer.
+ *
+ * See http://oss.sgi.com/archives/xfs/2007-01/msg00594.html for
+ * discussion.
+ *
+ * We cannot use get_page in the workaround, because it insists on a
+ * positive page count as a precondition. So we use _count directly.
+ */
+static void
+bio_pageinc(struct bio *bio)
+{
+ struct bio_vec *bv;
+ struct page *page;
+ int i;
+
+ bio_for_each_segment(bv, bio, i) {
+ page = bv->bv_page;
+ /* Non-zero page count for non-head members of
+ * compound pages is no longer allowed by the kernel,
+ * but this has never been seen here.
+ */
+ if (unlikely(PageCompound(page)))
+ if (compound_trans_head(page) != page) {
+ pr_crit("page tail used for block I/O\n");
+ BUG();
+ }
+ atomic_inc(&page->_count);
+ }
+}
+
+static void
+bio_pagedec(struct bio *bio)
+{
+ struct bio_vec *bv;
+ int i;
+
+ bio_for_each_segment(bv, bio, i)
+ atomic_dec(&bv->bv_page->_count);
+}
+
+static void
+bufinit(struct buf *buf, struct request *rq, struct bio *bio)
+{
+ struct bio_vec *bv;
+
+ memset(buf, 0, sizeof(*buf));
+ buf->rq = rq;
+ buf->bio = bio;
+ buf->resid = bio->bi_size;
+ buf->sector = bio->bi_sector;
+ bio_pageinc(bio);
+ buf->bv = bv = &bio->bi_io_vec[bio->bi_idx];
+ buf->bv_resid = bv->bv_len;
+ WARN_ON(buf->bv_resid == 0);
+}
+
+static struct buf *
+nextbuf(struct aoedev *d)
+{
+ struct request *rq;
+ struct request_queue *q;
+ struct buf *buf;
+ struct bio *bio;
+
+ q = d->blkq;
+ if (q == NULL)
+ return NULL; /* initializing */
+ if (d->ip.buf)
+ return d->ip.buf;
+ rq = d->ip.rq;
+ if (rq == NULL) {
+ rq = blk_peek_request(q);
+ if (rq == NULL)
+ return NULL;
+ blk_start_request(rq);
+ d->ip.rq = rq;
+ d->ip.nxbio = rq->bio;
+ rq->special = (void *) rqbiocnt(rq);
+ }
+ buf = mempool_alloc(d->bufpool, GFP_ATOMIC);
+ if (buf == NULL) {
+ pr_err("aoe: nextbuf: unable to mempool_alloc!\n");
+ return NULL;
+ }
+ bio = d->ip.nxbio;
+ bufinit(buf, rq, bio);
+ bio = bio->bi_next;
+ d->ip.nxbio = bio;
+ if (bio == NULL)
+ d->ip.rq = NULL;
+ return d->ip.buf = buf;
}
/* enters with d->lock held */
void
aoecmd_work(struct aoedev *d)
{
- struct buf *buf;
-loop:
if (d->htgt && !sthtith(d))
return;
- if (d->inprocess == NULL) {
- if (list_empty(&d->bufq))
- return;
- buf = container_of(d->bufq.next, struct buf, bufs);
- list_del(d->bufq.next);
- d->inprocess = buf;
- }
- if (aoecmd_ata_rw(d))
- goto loop;
+ while (aoecmd_ata_rw(d))
+ ;
}
/* this function performs work that has been deferred until sleeping is OK
@@ -604,28 +786,25 @@ void
aoecmd_sleepwork(struct work_struct *work)
{
struct aoedev *d = container_of(work, struct aoedev, work);
+ struct block_device *bd;
+ u64 ssize;
if (d->flags & DEVFL_GDALLOC)
aoeblk_gdalloc(d);
if (d->flags & DEVFL_NEWSIZE) {
- struct block_device *bd;
- unsigned long flags;
- u64 ssize;
-
ssize = get_capacity(d->gd);
bd = bdget_disk(d->gd, 0);
-
if (bd) {
mutex_lock(&bd->bd_inode->i_mutex);
i_size_write(bd->bd_inode, (loff_t)ssize<<9);
mutex_unlock(&bd->bd_inode->i_mutex);
bdput(bd);
}
- spin_lock_irqsave(&d->lock, flags);
+ spin_lock_irq(&d->lock);
d->flags |= DEVFL_UP;
d->flags &= ~DEVFL_NEWSIZE;
- spin_unlock_irqrestore(&d->lock, flags);
+ spin_unlock_irq(&d->lock);
}
}
@@ -718,163 +897,299 @@ gettgt(struct aoedev *d, char *addr)
return NULL;
}
-static inline void
-diskstats(struct gendisk *disk, struct bio *bio, ulong duration, sector_t sector)
+static void
+bvcpy(struct bio_vec *bv, ulong off, struct sk_buff *skb, long cnt)
+{
+ ulong fcnt;
+ char *p;
+ int soff = 0;
+loop:
+ fcnt = bv->bv_len - (off - bv->bv_offset);
+ if (fcnt > cnt)
+ fcnt = cnt;
+ p = page_address(bv->bv_page) + off;
+ skb_copy_bits(skb, soff, p, fcnt);
+ soff += fcnt;
+ cnt -= fcnt;
+ if (cnt <= 0)
+ return;
+ bv++;
+ off = bv->bv_offset;
+ goto loop;
+}
+
+void
+aoe_end_request(struct aoedev *d, struct request *rq, int fastfail)
+{
+ struct bio *bio;
+ int bok;
+ struct request_queue *q;
+
+ q = d->blkq;
+ if (rq == d->ip.rq)
+ d->ip.rq = NULL;
+ do {
+ bio = rq->bio;
+ bok = !fastfail && test_bit(BIO_UPTODATE, &bio->bi_flags);
+ } while (__blk_end_request(rq, bok ? 0 : -EIO, bio->bi_size));
+
+ /* cf. http://lkml.org/lkml/2006/10/31/28 */
+ if (!fastfail)
+ q->request_fn(q);
+}
+
+static void
+aoe_end_buf(struct aoedev *d, struct buf *buf)
+{
+ struct request *rq;
+ unsigned long n;
+
+ if (buf == d->ip.buf)
+ d->ip.buf = NULL;
+ rq = buf->rq;
+ bio_pagedec(buf->bio);
+ mempool_free(buf, d->bufpool);
+ n = (unsigned long) rq->special;
+ rq->special = (void *) --n;
+ if (n == 0)
+ aoe_end_request(d, rq, 0);
+}
+
+static void
+ktiocomplete(struct frame *f)
{
- unsigned long n_sect = bio->bi_size >> 9;
- const int rw = bio_data_dir(bio);
- struct hd_struct *part;
- int cpu;
+ struct aoe_hdr *hin, *hout;
+ struct aoe_atahdr *ahin, *ahout;
+ struct buf *buf;
+ struct sk_buff *skb;
+ struct aoetgt *t;
+ struct aoeif *ifp;
+ struct aoedev *d;
+ long n;
+
+ if (f == NULL)
+ return;
+
+ t = f->t;
+ d = t->d;
+
+ hout = (struct aoe_hdr *) skb_mac_header(f->skb);
+ ahout = (struct aoe_atahdr *) (hout+1);
+ buf = f->buf;
+ skb = f->r_skb;
+ if (skb == NULL)
+ goto noskb; /* just fail the buf. */
+
+ hin = (struct aoe_hdr *) skb->data;
+ skb_pull(skb, sizeof(*hin));
+ ahin = (struct aoe_atahdr *) skb->data;
+ skb_pull(skb, sizeof(*ahin));
+ if (ahin->cmdstat & 0xa9) { /* these bits cleared on success */
+ pr_err("aoe: ata error cmd=%2.2Xh stat=%2.2Xh from e%ld.%d\n",
+ ahout->cmdstat, ahin->cmdstat,
+ d->aoemajor, d->aoeminor);
+noskb: if (buf)
+ clear_bit(BIO_UPTODATE, &buf->bio->bi_flags);
+ goto badrsp;
+ }
- cpu = part_stat_lock();
- part = disk_map_sector_rcu(disk, sector);
+ n = ahout->scnt << 9;
+ switch (ahout->cmdstat) {
+ case ATA_CMD_PIO_READ:
+ case ATA_CMD_PIO_READ_EXT:
+ if (skb->len < n) {
+ pr_err("aoe: runt data size in read. skb->len=%d need=%ld\n",
+ skb->len, n);
+ clear_bit(BIO_UPTODATE, &buf->bio->bi_flags);
+ break;
+ }
+ bvcpy(f->bv, f->bv_off, skb, n);
+ case ATA_CMD_PIO_WRITE:
+ case ATA_CMD_PIO_WRITE_EXT:
+ spin_lock_irq(&d->lock);
+ ifp = getif(t, skb->dev);
+ if (ifp)
+ ifp->lost = 0;
+ if (d->htgt == t) /* I'll help myself, thank you. */
+ d->htgt = NULL;
+ spin_unlock_irq(&d->lock);
+ break;
+ case ATA_CMD_ID_ATA:
+ if (skb->len < 512) {
+ pr_info("aoe: runt data size in ataid. skb->len=%d\n",
+ skb->len);
+ break;
+ }
+ if (skb_linearize(skb))
+ break;
+ spin_lock_irq(&d->lock);
+ ataid_complete(d, t, skb->data);
+ spin_unlock_irq(&d->lock);
+ break;
+ default:
+ pr_info("aoe: unrecognized ata command %2.2Xh for %d.%d\n",
+ ahout->cmdstat,
+ be16_to_cpu(get_unaligned(&hin->major)),
+ hin->minor);
+ }
+badrsp:
+ spin_lock_irq(&d->lock);
+
+ aoe_freetframe(f);
+
+ if (buf && --buf->nframesout == 0 && buf->resid == 0)
+ aoe_end_buf(d, buf);
+
+ aoecmd_work(d);
+
+ spin_unlock_irq(&d->lock);
+ aoedev_put(d);
+ dev_kfree_skb(skb);
+}
+
+/* Enters with iocq.lock held.
+ * Returns true iff responses needing processing remain.
+ */
+static int
+ktio(void)
+{
+ struct frame *f;
+ struct list_head *pos;
+ int i;
- part_stat_inc(cpu, part, ios[rw]);
- part_stat_add(cpu, part, ticks[rw], duration);
- part_stat_add(cpu, part, sectors[rw], n_sect);
- part_stat_add(cpu, part, io_ticks, duration);
+ for (i = 0; ; ++i) {
+ if (i == MAXIOC)
+ return 1;
+ if (list_empty(&iocq.head))
+ return 0;
+ pos = iocq.head.next;
+ list_del(pos);
+ spin_unlock_irq(&iocq.lock);
+ f = list_entry(pos, struct frame, head);
+ ktiocomplete(f);
+ spin_lock_irq(&iocq.lock);
+ }
+}
- part_stat_unlock();
+static int
+kthread(void *vp)
+{
+ struct ktstate *k;
+ DECLARE_WAITQUEUE(wait, current);
+ int more;
+
+ k = vp;
+ current->flags |= PF_NOFREEZE;
+ set_user_nice(current, -10);
+ complete(&k->rendez); /* tell spawner we're running */
+ do {
+ spin_lock_irq(k->lock);
+ more = k->fn();
+ if (!more) {
+ add_wait_queue(k->waitq, &wait);
+ __set_current_state(TASK_INTERRUPTIBLE);
+ }
+ spin_unlock_irq(k->lock);
+ if (!more) {
+ schedule();
+ remove_wait_queue(k->waitq, &wait);
+ } else
+ cond_resched();
+ } while (!kthread_should_stop());
+ complete(&k->rendez); /* tell spawner we're stopping */
+ return 0;
}
void
+aoe_ktstop(struct ktstate *k)
+{
+ kthread_stop(k->task);
+ wait_for_completion(&k->rendez);
+}
+
+int
+aoe_ktstart(struct ktstate *k)
+{
+ struct task_struct *task;
+
+ init_completion(&k->rendez);
+ task = kthread_run(kthread, k, k->name);
+ if (task == NULL || IS_ERR(task))
+ return -ENOMEM;
+ k->task = task;
+ wait_for_completion(&k->rendez); /* allow kthread to start */
+ init_completion(&k->rendez); /* for waiting for exit later */
+ return 0;
+}
+
+/* pass it off to kthreads for processing */
+static void
+ktcomplete(struct frame *f, struct sk_buff *skb)
+{
+ ulong flags;
+
+ f->r_skb = skb;
+ spin_lock_irqsave(&iocq.lock, flags);
+ list_add_tail(&f->head, &iocq.head);
+ spin_unlock_irqrestore(&iocq.lock, flags);
+ wake_up(&ktiowq);
+}
+
+struct sk_buff *
aoecmd_ata_rsp(struct sk_buff *skb)
{
- struct sk_buff_head queue;
struct aoedev *d;
- struct aoe_hdr *hin, *hout;
- struct aoe_atahdr *ahin, *ahout;
+ struct aoe_hdr *h;
struct frame *f;
- struct buf *buf;
struct aoetgt *t;
- struct aoeif *ifp;
- register long n;
+ u32 n;
ulong flags;
char ebuf[128];
u16 aoemajor;
- hin = (struct aoe_hdr *) skb_mac_header(skb);
- aoemajor = get_unaligned_be16(&hin->major);
- d = aoedev_by_aoeaddr(aoemajor, hin->minor);
+ h = (struct aoe_hdr *) skb->data;
+ aoemajor = be16_to_cpu(get_unaligned(&h->major));
+ d = aoedev_by_aoeaddr(aoemajor, h->minor, 0);
if (d == NULL) {
snprintf(ebuf, sizeof ebuf, "aoecmd_ata_rsp: ata response "
"for unknown device %d.%d\n",
- aoemajor, hin->minor);
+ aoemajor, h->minor);
aoechr_error(ebuf);
- return;
+ return skb;
}
spin_lock_irqsave(&d->lock, flags);
- n = get_unaligned_be32(&hin->tag);
- t = gettgt(d, hin->src);
- if (t == NULL) {
- printk(KERN_INFO "aoe: can't find target e%ld.%d:%pm\n",
- d->aoemajor, d->aoeminor, hin->src);
- spin_unlock_irqrestore(&d->lock, flags);
- return;
- }
- f = getframe(t, n);
+ n = be32_to_cpu(get_unaligned(&h->tag));
+ f = getframe(d, n);
if (f == NULL) {
calc_rttavg(d, -tsince(n));
spin_unlock_irqrestore(&d->lock, flags);
+ aoedev_put(d);
snprintf(ebuf, sizeof ebuf,
"%15s e%d.%d tag=%08x@%08lx\n",
"unexpected rsp",
- get_unaligned_be16(&hin->major),
- hin->minor,
- get_unaligned_be32(&hin->tag),
+ get_unaligned_be16(&h->major),
+ h->minor,
+ get_unaligned_be32(&h->tag),
jiffies);
aoechr_error(ebuf);
- return;
+ return skb;
}
-
+ t = f->t;
calc_rttavg(d, tsince(f->tag));
-
- ahin = (struct aoe_atahdr *) (hin+1);
- hout = (struct aoe_hdr *) skb_mac_header(f->skb);
- ahout = (struct aoe_atahdr *) (hout+1);
- buf = f->buf;
-
- if (ahin->cmdstat & 0xa9) { /* these bits cleared on success */
- printk(KERN_ERR
- "aoe: ata error cmd=%2.2Xh stat=%2.2Xh from e%ld.%d\n",
- ahout->cmdstat, ahin->cmdstat,
- d->aoemajor, d->aoeminor);
- if (buf)
- buf->flags |= BUFFL_FAIL;
- } else {
- if (d->htgt && t == *d->htgt) /* I'll help myself, thank you. */
- d->htgt = NULL;
- n = ahout->scnt << 9;
- switch (ahout->cmdstat) {
- case ATA_CMD_PIO_READ:
- case ATA_CMD_PIO_READ_EXT:
- if (skb->len - sizeof *hin - sizeof *ahin < n) {
- printk(KERN_ERR
- "aoe: %s. skb->len=%d need=%ld\n",
- "runt data size in read", skb->len, n);
- /* fail frame f? just returning will rexmit. */
- spin_unlock_irqrestore(&d->lock, flags);
- return;
- }
- memcpy(f->bufaddr, ahin+1, n);
- case ATA_CMD_PIO_WRITE:
- case ATA_CMD_PIO_WRITE_EXT:
- ifp = getif(t, skb->dev);
- if (ifp) {
- ifp->lost = 0;
- if (n > DEFAULTBCNT)
- ifp->lostjumbo = 0;
- }
- if (f->bcnt -= n) {
- f->lba += n >> 9;
- f->bufaddr += n;
- resend(d, t, f);
- goto xmit;
- }
- break;
- case ATA_CMD_ID_ATA:
- if (skb->len - sizeof *hin - sizeof *ahin < 512) {
- printk(KERN_INFO
- "aoe: runt data size in ataid. skb->len=%d\n",
- skb->len);
- spin_unlock_irqrestore(&d->lock, flags);
- return;
- }
- ataid_complete(d, t, (char *) (ahin+1));
- break;
- default:
- printk(KERN_INFO
- "aoe: unrecognized ata command %2.2Xh for %d.%d\n",
- ahout->cmdstat,
- get_unaligned_be16(&hin->major),
- hin->minor);
- }
- }
-
- if (buf && --buf->nframesout == 0 && buf->resid == 0) {
- diskstats(d->gd, buf->bio, jiffies - buf->stime, buf->sector);
- if (buf->flags & BUFFL_FAIL)
- bio_endio(buf->bio, -EIO);
- else {
- bio_flush_dcache_pages(buf->bio);
- bio_endio(buf->bio, 0);
- }
- mempool_free(buf, d->bufpool);
- }
-
- f->buf = NULL;
- f->tag = FREETAG;
t->nout--;
-
aoecmd_work(d);
-xmit:
- __skb_queue_head_init(&queue);
- skb_queue_splice_init(&d->sendq, &queue);
spin_unlock_irqrestore(&d->lock, flags);
- aoenet_xmit(&queue);
+
+ ktcomplete(f, skb);
+
+ /*
+ * Note here that we do not perform an aoedev_put, as we are
+ * leaving this reference for the ktio to release.
+ */
+ return NULL;
}
void
@@ -896,7 +1211,7 @@ aoecmd_ata_id(struct aoedev *d)
struct sk_buff *skb;
struct aoetgt *t;
- f = freeframe(d);
+ f = newframe(d);
if (f == NULL)
return NULL;
@@ -909,6 +1224,7 @@ aoecmd_ata_id(struct aoedev *d)
skb_put(skb, sizeof *h + sizeof *ah);
memset(h, 0, skb->len);
f->tag = aoehdr_atainit(d, t, h);
+ fhash(f);
t->nout++;
f->waited = 0;
@@ -929,7 +1245,6 @@ static struct aoetgt *
addtgt(struct aoedev *d, char *addr, ulong nframes)
{
struct aoetgt *t, **tt, **te;
- struct frame *f, *e;
tt = d->targets;
te = tt + NTARGETS;
@@ -941,26 +1256,73 @@ addtgt(struct aoedev *d, char *addr, ulong nframes)
"aoe: device addtgt failure; too many targets\n");
return NULL;
}
- t = kcalloc(1, sizeof *t, GFP_ATOMIC);
- f = kcalloc(nframes, sizeof *f, GFP_ATOMIC);
- if (!t || !f) {
- kfree(f);
- kfree(t);
+ t = kzalloc(sizeof(*t), GFP_ATOMIC);
+ if (!t) {
printk(KERN_INFO "aoe: cannot allocate memory to add target\n");
return NULL;
}
+ d->ntargets++;
t->nframes = nframes;
- t->frames = f;
- e = f + nframes;
- for (; f < e; f++)
- f->tag = FREETAG;
+ t->d = d;
memcpy(t->addr, addr, sizeof t->addr);
t->ifp = t->ifs;
t->maxout = t->nframes;
+ INIT_LIST_HEAD(&t->ffree);
return *tt = t;
}
+static void
+setdbcnt(struct aoedev *d)
+{
+ struct aoetgt **t, **e;
+ int bcnt = 0;
+
+ t = d->targets;
+ e = t + NTARGETS;
+ for (; t < e && *t; t++)
+ if (bcnt == 0 || bcnt > (*t)->minbcnt)
+ bcnt = (*t)->minbcnt;
+ if (bcnt != d->maxbcnt) {
+ d->maxbcnt = bcnt;
+ pr_info("aoe: e%ld.%d: setting %d byte data frames\n",
+ d->aoemajor, d->aoeminor, bcnt);
+ }
+}
+
+static void
+setifbcnt(struct aoetgt *t, struct net_device *nd, int bcnt)
+{
+ struct aoedev *d;
+ struct aoeif *p, *e;
+ int minbcnt;
+
+ d = t->d;
+ minbcnt = bcnt;
+ p = t->ifs;
+ e = p + NAOEIFS;
+ for (; p < e; p++) {
+ if (p->nd == NULL)
+ break; /* end of the valid interfaces */
+ if (p->nd == nd) {
+ p->bcnt = bcnt; /* we're updating */
+ nd = NULL;
+ } else if (minbcnt > p->bcnt)
+ minbcnt = p->bcnt; /* find the min interface */
+ }
+ if (nd) {
+ if (p == e) {
+ pr_err("aoe: device setifbcnt failure; too many interfaces.\n");
+ return;
+ }
+ dev_hold(nd);
+ p->nd = nd;
+ p->bcnt = bcnt;
+ }
+ t->minbcnt = minbcnt;
+ setdbcnt(d);
+}
+
void
aoecmd_cfg_rsp(struct sk_buff *skb)
{
@@ -968,11 +1330,12 @@ aoecmd_cfg_rsp(struct sk_buff *skb)
struct aoe_hdr *h;
struct aoe_cfghdr *ch;
struct aoetgt *t;
- struct aoeif *ifp;
- ulong flags, sysminor, aoemajor;
+ ulong flags, aoemajor;
struct sk_buff *sl;
+ struct sk_buff_head queue;
u16 n;
+ sl = NULL;
h = (struct aoe_hdr *) skb_mac_header(skb);
ch = (struct aoe_cfghdr *) (h+1);
@@ -986,10 +1349,13 @@ aoecmd_cfg_rsp(struct sk_buff *skb)
"Check shelf dip switches.\n");
return;
}
-
- sysminor = SYSMINOR(aoemajor, h->minor);
- if (sysminor * AOE_PARTITIONS + AOE_PARTITIONS > MINORMASK) {
- printk(KERN_INFO "aoe: e%ld.%d: minor number too large\n",
+ if (aoemajor == 0xffff) {
+ pr_info("aoe: e%ld.%d: broadcast shelf number invalid\n",
+ aoemajor, (int) h->minor);
+ return;
+ }
+ if (h->minor == 0xff) {
+ pr_info("aoe: e%ld.%d: broadcast slot number invalid\n",
aoemajor, (int) h->minor);
return;
}
@@ -998,9 +1364,9 @@ aoecmd_cfg_rsp(struct sk_buff *skb)
if (n > aoe_maxout) /* keep it reasonable */
n = aoe_maxout;
- d = aoedev_by_sysminor_m(sysminor);
+ d = aoedev_by_aoeaddr(aoemajor, h->minor, 1);
if (d == NULL) {
- printk(KERN_INFO "aoe: device sysminor_m failure\n");
+ pr_info("aoe: device allocation failure\n");
return;
}
@@ -1009,52 +1375,26 @@ aoecmd_cfg_rsp(struct sk_buff *skb)
t = gettgt(d, h->src);
if (!t) {
t = addtgt(d, h->src, n);
- if (!t) {
- spin_unlock_irqrestore(&d->lock, flags);
- return;
- }
- }
- ifp = getif(t, skb->dev);
- if (!ifp) {
- ifp = addif(t, skb->dev);
- if (!ifp) {
- printk(KERN_INFO
- "aoe: device addif failure; "
- "too many interfaces?\n");
- spin_unlock_irqrestore(&d->lock, flags);
- return;
- }
- }
- if (ifp->maxbcnt) {
- n = ifp->nd->mtu;
- n -= sizeof (struct aoe_hdr) + sizeof (struct aoe_atahdr);
- n /= 512;
- if (n > ch->scnt)
- n = ch->scnt;
- n = n ? n * 512 : DEFAULTBCNT;
- if (n != ifp->maxbcnt) {
- printk(KERN_INFO
- "aoe: e%ld.%d: setting %d%s%s:%pm\n",
- d->aoemajor, d->aoeminor, n,
- " byte data frames on ", ifp->nd->name,
- t->addr);
- ifp->maxbcnt = n;
- }
+ if (!t)
+ goto bail;
}
+ n = skb->dev->mtu;
+ n -= sizeof(struct aoe_hdr) + sizeof(struct aoe_atahdr);
+ n /= 512;
+ if (n > ch->scnt)
+ n = ch->scnt;
+ n = n ? n * 512 : DEFAULTBCNT;
+ setifbcnt(t, skb->dev, n);
/* don't change users' perspective */
- if (d->nopen) {
- spin_unlock_irqrestore(&d->lock, flags);
- return;
+ if (d->nopen == 0) {
+ d->fw_ver = be16_to_cpu(ch->fwver);
+ sl = aoecmd_ata_id(d);
}
- d->fw_ver = be16_to_cpu(ch->fwver);
-
- sl = aoecmd_ata_id(d);
-
+bail:
spin_unlock_irqrestore(&d->lock, flags);
-
+ aoedev_put(d);
if (sl) {
- struct sk_buff_head queue;
__skb_queue_head_init(&queue);
__skb_queue_tail(&queue, sl);
aoenet_xmit(&queue);
@@ -1065,20 +1405,74 @@ void
aoecmd_cleanslate(struct aoedev *d)
{
struct aoetgt **t, **te;
- struct aoeif *p, *e;
d->mintimer = MINTIMER;
+ d->maxbcnt = 0;
t = d->targets;
te = t + NTARGETS;
- for (; t < te && *t; t++) {
+ for (; t < te && *t; t++)
(*t)->maxout = (*t)->nframes;
- p = (*t)->ifs;
- e = p + NAOEIFS;
- for (; p < e; p++) {
- p->lostjumbo = 0;
- p->lost = 0;
- p->maxbcnt = DEFAULTBCNT;
+}
+
+void
+aoe_failbuf(struct aoedev *d, struct buf *buf)
+{
+ if (buf == NULL)
+ return;
+ buf->resid = 0;
+ clear_bit(BIO_UPTODATE, &buf->bio->bi_flags);
+ if (buf->nframesout == 0)
+ aoe_end_buf(d, buf);
+}
+
+void
+aoe_flush_iocq(void)
+{
+ struct frame *f;
+ struct aoedev *d;
+ LIST_HEAD(flist);
+ struct list_head *pos;
+ struct sk_buff *skb;
+ ulong flags;
+
+ spin_lock_irqsave(&iocq.lock, flags);
+ list_splice_init(&iocq.head, &flist);
+ spin_unlock_irqrestore(&iocq.lock, flags);
+ while (!list_empty(&flist)) {
+ pos = flist.next;
+ list_del(pos);
+ f = list_entry(pos, struct frame, head);
+ d = f->t->d;
+ skb = f->r_skb;
+ spin_lock_irqsave(&d->lock, flags);
+ if (f->buf) {
+ f->buf->nframesout--;
+ aoe_failbuf(d, f->buf);
}
+ aoe_freetframe(f);
+ spin_unlock_irqrestore(&d->lock, flags);
+ dev_kfree_skb(skb);
+ aoedev_put(d);
}
}
+
+int __init
+aoecmd_init(void)
+{
+ INIT_LIST_HEAD(&iocq.head);
+ spin_lock_init(&iocq.lock);
+ init_waitqueue_head(&ktiowq);
+ kts.name = "aoe_ktio";
+ kts.fn = ktio;
+ kts.waitq = &ktiowq;
+ kts.lock = &iocq.lock;
+ return aoe_ktstart(&kts);
+}
+
+void
+aoecmd_exit(void)
+{
+ aoe_ktstop(&kts);
+ aoe_flush_iocq();
+}
diff --git a/drivers/block/aoe/aoedev.c b/drivers/block/aoe/aoedev.c
index 6b5110a4745..90e5b537f94 100644
--- a/drivers/block/aoe/aoedev.c
+++ b/drivers/block/aoe/aoedev.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoedev.c
* AoE device utility functions; maintains device list.
@@ -9,6 +9,9 @@
#include <linux/netdevice.h>
#include <linux/delay.h>
#include <linux/slab.h>
+#include <linux/bitmap.h>
+#include <linux/kdev_t.h>
+#include <linux/moduleparam.h>
#include "aoe.h"
static void dummy_timer(ulong);
@@ -16,23 +19,121 @@ static void aoedev_freedev(struct aoedev *);
static void freetgt(struct aoedev *d, struct aoetgt *t);
static void skbpoolfree(struct aoedev *d);
+static int aoe_dyndevs = 1;
+module_param(aoe_dyndevs, int, 0644);
+MODULE_PARM_DESC(aoe_dyndevs, "Use dynamic minor numbers for devices.");
+
static struct aoedev *devlist;
static DEFINE_SPINLOCK(devlist_lock);
-struct aoedev *
-aoedev_by_aoeaddr(int maj, int min)
+/* Because some systems will have one, many, or no
+ * - partitions,
+ * - slots per shelf,
+ * - or shelves,
+ * we need some flexibility in the way the minor numbers
+ * are allocated. So they are dynamic.
+ */
+#define N_DEVS ((1U<<MINORBITS)/AOE_PARTITIONS)
+
+static DEFINE_SPINLOCK(used_minors_lock);
+static DECLARE_BITMAP(used_minors, N_DEVS);
+
+static int
+minor_get_dyn(ulong *sysminor)
{
- struct aoedev *d;
ulong flags;
+ ulong n;
+ int error = 0;
+
+ spin_lock_irqsave(&used_minors_lock, flags);
+ n = find_first_zero_bit(used_minors, N_DEVS);
+ if (n < N_DEVS)
+ set_bit(n, used_minors);
+ else
+ error = -1;
+ spin_unlock_irqrestore(&used_minors_lock, flags);
+
+ *sysminor = n * AOE_PARTITIONS;
+ return error;
+}
- spin_lock_irqsave(&devlist_lock, flags);
+static int
+minor_get_static(ulong *sysminor, ulong aoemaj, int aoemin)
+{
+ ulong flags;
+ ulong n;
+ int error = 0;
+ enum {
+ /* for backwards compatibility when !aoe_dyndevs,
+ * a static number of supported slots per shelf */
+ NPERSHELF = 16,
+ };
+
+ n = aoemaj * NPERSHELF + aoemin;
+ if (aoemin >= NPERSHELF || n >= N_DEVS) {
+ pr_err("aoe: %s with e%ld.%d\n",
+ "cannot use static minor device numbers",
+ aoemaj, aoemin);
+ error = -1;
+ } else {
+ spin_lock_irqsave(&used_minors_lock, flags);
+ if (test_bit(n, used_minors)) {
+ pr_err("aoe: %s %lu\n",
+ "existing device already has static minor number",
+ n);
+ error = -1;
+ } else
+ set_bit(n, used_minors);
+ spin_unlock_irqrestore(&used_minors_lock, flags);
+ }
- for (d=devlist; d; d=d->next)
- if (d->aoemajor == maj && d->aoeminor == min)
- break;
+ *sysminor = n;
+ return error;
+}
+
+static int
+minor_get(ulong *sysminor, ulong aoemaj, int aoemin)
+{
+ if (aoe_dyndevs)
+ return minor_get_dyn(sysminor);
+ else
+ return minor_get_static(sysminor, aoemaj, aoemin);
+}
+
+static void
+minor_free(ulong minor)
+{
+ ulong flags;
+
+ minor /= AOE_PARTITIONS;
+ BUG_ON(minor >= N_DEVS);
+
+ spin_lock_irqsave(&used_minors_lock, flags);
+ BUG_ON(!test_bit(minor, used_minors));
+ clear_bit(minor, used_minors);
+ spin_unlock_irqrestore(&used_minors_lock, flags);
+}
+
+/*
+ * Users who grab a pointer to the device with aoedev_by_aoeaddr
+ * automatically get a reference count and must be responsible
+ * for performing a aoedev_put. With the addition of async
+ * kthread processing I'm no longer confident that we can
+ * guarantee consistency in the face of device flushes.
+ *
+ * For the time being, we only bother to add extra references for
+ * frames sitting on the iocq. When the kthreads finish processing
+ * these frames, they will aoedev_put the device.
+ */
+
+void
+aoedev_put(struct aoedev *d)
+{
+ ulong flags;
+ spin_lock_irqsave(&devlist_lock, flags);
+ d->ref--;
spin_unlock_irqrestore(&devlist_lock, flags);
- return d;
}
static void
@@ -47,54 +148,74 @@ dummy_timer(ulong vp)
add_timer(&d->timer);
}
+static void
+aoe_failip(struct aoedev *d)
+{
+ struct request *rq;
+ struct bio *bio;
+ unsigned long n;
+
+ aoe_failbuf(d, d->ip.buf);
+
+ rq = d->ip.rq;
+ if (rq == NULL)
+ return;
+ while ((bio = d->ip.nxbio)) {
+ clear_bit(BIO_UPTODATE, &bio->bi_flags);
+ d->ip.nxbio = bio->bi_next;
+ n = (unsigned long) rq->special;
+ rq->special = (void *) --n;
+ }
+ if ((unsigned long) rq->special == 0)
+ aoe_end_request(d, rq, 0);
+}
+
void
aoedev_downdev(struct aoedev *d)
{
- struct aoetgt **t, **te;
- struct frame *f, *e;
- struct buf *buf;
- struct bio *bio;
+ struct aoetgt *t, **tt, **te;
+ struct frame *f;
+ struct list_head *head, *pos, *nx;
+ struct request *rq;
+ int i;
- t = d->targets;
- te = t + NTARGETS;
- for (; t < te && *t; t++) {
- f = (*t)->frames;
- e = f + (*t)->nframes;
- for (; f < e; f->tag = FREETAG, f->buf = NULL, f++) {
- if (f->tag == FREETAG || f->buf == NULL)
- continue;
- buf = f->buf;
- bio = buf->bio;
- if (--buf->nframesout == 0
- && buf != d->inprocess) {
- mempool_free(buf, d->bufpool);
- bio_endio(bio, -EIO);
+ d->flags &= ~DEVFL_UP;
+
+ /* clean out active buffers */
+ for (i = 0; i < NFACTIVE; i++) {
+ head = &d->factive[i];
+ list_for_each_safe(pos, nx, head) {
+ f = list_entry(pos, struct frame, head);
+ list_del(pos);
+ if (f->buf) {
+ f->buf->nframesout--;
+ aoe_failbuf(d, f->buf);
}
+ aoe_freetframe(f);
}
- (*t)->maxout = (*t)->nframes;
- (*t)->nout = 0;
}
- buf = d->inprocess;
- if (buf) {
- bio = buf->bio;
- mempool_free(buf, d->bufpool);
- bio_endio(bio, -EIO);
+ /* reset window dressings */
+ tt = d->targets;
+ te = tt + NTARGETS;
+ for (; tt < te && (t = *tt); tt++) {
+ t->maxout = t->nframes;
+ t->nout = 0;
}
- d->inprocess = NULL;
+
+ /* clean out the in-process request (if any) */
+ aoe_failip(d);
d->htgt = NULL;
- while (!list_empty(&d->bufq)) {
- buf = container_of(d->bufq.next, struct buf, bufs);
- list_del(d->bufq.next);
- bio = buf->bio;
- mempool_free(buf, d->bufpool);
- bio_endio(bio, -EIO);
+ /* fast fail all pending I/O */
+ if (d->blkq) {
+ while ((rq = blk_peek_request(d->blkq))) {
+ blk_start_request(rq);
+ aoe_end_request(d, rq, 1);
+ }
}
if (d->gd)
set_capacity(d->gd, 0);
-
- d->flags &= ~DEVFL_UP;
}
static void
@@ -107,6 +228,7 @@ aoedev_freedev(struct aoedev *d)
aoedisk_rm_sysfs(d);
del_gendisk(d->gd);
put_disk(d->gd);
+ blk_cleanup_queue(d->blkq);
}
t = d->targets;
e = t + NTARGETS;
@@ -115,7 +237,7 @@ aoedev_freedev(struct aoedev *d)
if (d->bufpool)
mempool_destroy(d->bufpool);
skbpoolfree(d);
- blk_cleanup_queue(d->blkq);
+ minor_free(d->sysminor);
kfree(d);
}
@@ -142,7 +264,8 @@ aoedev_flush(const char __user *str, size_t cnt)
spin_lock(&d->lock);
if ((!all && (d->flags & DEVFL_UP))
|| (d->flags & (DEVFL_GDALLOC|DEVFL_NEWSIZE))
- || d->nopen) {
+ || d->nopen
+ || d->ref) {
spin_unlock(&d->lock);
dd = &d->next;
continue;
@@ -163,12 +286,15 @@ aoedev_flush(const char __user *str, size_t cnt)
return 0;
}
-/* I'm not really sure that this is a realistic problem, but if the
-network driver goes gonzo let's just leak memory after complaining. */
+/* This has been confirmed to occur once with Tms=3*1000 due to the
+ * driver changing link and not processing its transmit ring. The
+ * problem is hard enough to solve by returning an error that I'm
+ * still punting on "solving" this.
+ */
static void
skbfree(struct sk_buff *skb)
{
- enum { Sms = 100, Tms = 3*1000};
+ enum { Sms = 250, Tms = 30 * 1000};
int i = Tms / Sms;
if (skb == NULL)
@@ -182,6 +308,7 @@ skbfree(struct sk_buff *skb)
"cannot free skb -- memory leaked.");
return;
}
+ skb->truesize -= skb->data_len;
skb_shinfo(skb)->nr_frags = skb->data_len = 0;
skb_trim(skb, 0);
dev_kfree_skb(skb);
@@ -198,26 +325,29 @@ skbpoolfree(struct aoedev *d)
__skb_queue_head_init(&d->skbpool);
}
-/* find it or malloc it */
+/* find it or allocate it */
struct aoedev *
-aoedev_by_sysminor_m(ulong sysminor)
+aoedev_by_aoeaddr(ulong maj, int min, int do_alloc)
{
struct aoedev *d;
+ int i;
ulong flags;
+ ulong sysminor;
spin_lock_irqsave(&devlist_lock, flags);
for (d=devlist; d; d=d->next)
- if (d->sysminor == sysminor)
+ if (d->aoemajor == maj && d->aoeminor == min) {
+ d->ref++;
break;
- if (d)
+ }
+ if (d || !do_alloc || minor_get(&sysminor, maj, min) < 0)
goto out;
d = kcalloc(1, sizeof *d, GFP_ATOMIC);
if (!d)
goto out;
INIT_WORK(&d->work, aoecmd_sleepwork);
spin_lock_init(&d->lock);
- skb_queue_head_init(&d->sendq);
skb_queue_head_init(&d->skbpool);
init_timer(&d->timer);
d->timer.data = (ulong) d;
@@ -226,10 +356,12 @@ aoedev_by_sysminor_m(ulong sysminor)
add_timer(&d->timer);
d->bufpool = NULL; /* defer to aoeblk_gdalloc */
d->tgt = d->targets;
- INIT_LIST_HEAD(&d->bufq);
+ d->ref = 1;
+ for (i = 0; i < NFACTIVE; i++)
+ INIT_LIST_HEAD(&d->factive[i]);
d->sysminor = sysminor;
- d->aoemajor = AOEMAJOR(sysminor);
- d->aoeminor = AOEMINOR(sysminor);
+ d->aoemajor = maj;
+ d->aoeminor = min;
d->mintimer = MINTIMER;
d->next = devlist;
devlist = d;
@@ -241,13 +373,23 @@ aoedev_by_sysminor_m(ulong sysminor)
static void
freetgt(struct aoedev *d, struct aoetgt *t)
{
- struct frame *f, *e;
+ struct frame *f;
+ struct list_head *pos, *nx, *head;
+ struct aoeif *ifp;
- f = t->frames;
- e = f + t->nframes;
- for (; f < e; f++)
+ for (ifp = t->ifs; ifp < &t->ifs[NAOEIFS]; ++ifp) {
+ if (!ifp->nd)
+ break;
+ dev_put(ifp->nd);
+ }
+
+ head = &t->ffree;
+ list_for_each_safe(pos, nx, head) {
+ list_del(pos);
+ f = list_entry(pos, struct frame, head);
skbfree(f->skb);
- kfree(t->frames);
+ kfree(f);
+ }
kfree(t);
}
@@ -257,6 +399,7 @@ aoedev_exit(void)
struct aoedev *d;
ulong flags;
+ aoe_flush_iocq();
while ((d = devlist)) {
devlist = d->next;
diff --git a/drivers/block/aoe/aoemain.c b/drivers/block/aoe/aoemain.c
index 7f83ad90e76..04793c2c701 100644
--- a/drivers/block/aoe/aoemain.c
+++ b/drivers/block/aoe/aoemain.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoemain.c
* Module initialization routines, discover timer
@@ -61,6 +61,7 @@ aoe_exit(void)
aoenet_exit();
unregister_blkdev(AOE_MAJOR, DEVICE_NAME);
+ aoecmd_exit();
aoechr_exit();
aoedev_exit();
aoeblk_exit(); /* free cache after de-allocating bufs */
@@ -83,17 +84,20 @@ aoe_init(void)
ret = aoenet_init();
if (ret)
goto net_fail;
+ ret = aoecmd_init();
+ if (ret)
+ goto cmd_fail;
ret = register_blkdev(AOE_MAJOR, DEVICE_NAME);
if (ret < 0) {
printk(KERN_ERR "aoe: can't register major\n");
goto blkreg_fail;
}
-
printk(KERN_INFO "aoe: AoE v%s initialised.\n", VERSION);
discover_timer(TINIT);
return 0;
-
blkreg_fail:
+ aoecmd_exit();
+ cmd_fail:
aoenet_exit();
net_fail:
aoeblk_exit();
diff --git a/drivers/block/aoe/aoenet.c b/drivers/block/aoe/aoenet.c
index 4d3bc0d49df..162c6471275 100644
--- a/drivers/block/aoe/aoenet.c
+++ b/drivers/block/aoe/aoenet.c
@@ -1,4 +1,4 @@
-/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
+/* Copyright (c) 2012 Coraid, Inc. See COPYING for GPL terms. */
/*
* aoenet.c
* Ethernet portion of AoE driver
@@ -33,6 +33,9 @@ static char aoe_iflist[IFLISTSZ];
module_param_string(aoe_iflist, aoe_iflist, IFLISTSZ, 0600);
MODULE_PARM_DESC(aoe_iflist, "aoe_iflist=\"dev1 [dev2 ...]\"");
+static wait_queue_head_t txwq;
+static struct ktstate kts;
+
#ifndef MODULE
static int __init aoe_iflist_setup(char *str)
{
@@ -44,6 +47,23 @@ static int __init aoe_iflist_setup(char *str)
__setup("aoe_iflist=", aoe_iflist_setup);
#endif
+static spinlock_t txlock;
+static struct sk_buff_head skbtxq;
+
+/* enters with txlock held */
+static int
+tx(void)
+{
+ struct sk_buff *skb;
+
+ while ((skb = skb_dequeue(&skbtxq))) {
+ spin_unlock_irq(&txlock);
+ dev_queue_xmit(skb);
+ spin_lock_irq(&txlock);
+ }
+ return 0;
+}
+
int
is_aoe_netif(struct net_device *ifp)
{
@@ -88,10 +108,14 @@ void
aoenet_xmit(struct sk_buff_head *queue)
{
struct sk_buff *skb, *tmp;
+ ulong flags;
skb_queue_walk_safe(queue, skb, tmp) {
__skb_unlink(skb, queue);
- dev_queue_xmit(skb);
+ spin_lock_irqsave(&txlock, flags);
+ skb_queue_tail(&skbtxq, skb);
+ spin_unlock_irqrestore(&txlock, flags);
+ wake_up(&txwq);
}
}
@@ -102,7 +126,9 @@ static int
aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt, struct net_device *orig_dev)
{
struct aoe_hdr *h;
+ struct aoe_atahdr *ah;
u32 n;
+ int sn;
if (dev_net(ifp) != &init_net)
goto exit;
@@ -110,13 +136,16 @@ aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt,
skb = skb_share_check(skb, GFP_ATOMIC);
if (skb == NULL)
return 0;
- if (skb_linearize(skb))
- goto exit;
if (!is_aoe_netif(ifp))
goto exit;
skb_push(skb, ETH_HLEN); /* (1) */
-
- h = (struct aoe_hdr *) skb_mac_header(skb);
+ sn = sizeof(*h) + sizeof(*ah);
+ if (skb->len >= sn) {
+ sn -= skb_headlen(skb);
+ if (sn > 0 && !__pskb_pull_tail(skb, sn))
+ goto exit;
+ }
+ h = (struct aoe_hdr *) skb->data;
n = get_unaligned_be32(&h->tag);
if ((h->verfl & AOEFL_RSP) == 0 || (n & 1<<31))
goto exit;
@@ -137,7 +166,8 @@ aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt,
switch (h->cmd) {
case AOECMD_ATA:
- aoecmd_ata_rsp(skb);
+ /* ata_rsp may keep skb for later processing or give it back */
+ skb = aoecmd_ata_rsp(skb);
break;
case AOECMD_CFG:
aoecmd_cfg_rsp(skb);
@@ -145,8 +175,12 @@ aoenet_rcv(struct sk_buff *skb, struct net_device *ifp, struct packet_type *pt,
default:
if (h->cmd >= AOECMD_VEND_MIN)
break; /* don't complain about vendor commands */
- printk(KERN_INFO "aoe: unknown cmd %d\n", h->cmd);
+ pr_info("aoe: unknown AoE command type 0x%02x\n", h->cmd);
+ break;
}
+
+ if (!skb)
+ return 0;
exit:
dev_kfree_skb(skb);
return 0;
@@ -160,6 +194,15 @@ static struct packet_type aoe_pt __read_mostly = {
int __init
aoenet_init(void)
{
+ skb_queue_head_init(&skbtxq);
+ init_waitqueue_head(&txwq);
+ spin_lock_init(&txlock);
+ kts.lock = &txlock;
+ kts.fn = tx;
+ kts.waitq = &txwq;
+ kts.name = "aoe_tx";
+ if (aoe_ktstart(&kts))
+ return -EAGAIN;
dev_add_pack(&aoe_pt);
return 0;
}
@@ -167,6 +210,8 @@ aoenet_init(void)
void
aoenet_exit(void)
{
+ aoe_ktstop(&kts);
+ skb_queue_purge(&skbtxq);
dev_remove_pack(&aoe_pt);
}
diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c
index f93a0320e95..f55683ad4ff 100644
--- a/drivers/block/drbd/drbd_main.c
+++ b/drivers/block/drbd/drbd_main.c
@@ -162,23 +162,12 @@ static const struct block_device_operations drbd_ops = {
.release = drbd_release,
};
-static void bio_destructor_drbd(struct bio *bio)
-{
- bio_free(bio, drbd_md_io_bio_set);
-}
-
struct bio *bio_alloc_drbd(gfp_t gfp_mask)
{
- struct bio *bio;
-
if (!drbd_md_io_bio_set)
return bio_alloc(gfp_mask, 1);
- bio = bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set);
- if (!bio)
- return NULL;
- bio->bi_destructor = bio_destructor_drbd;
- return bio;
+ return bio_alloc_bioset(gfp_mask, 1, drbd_md_io_bio_set);
}
#ifdef __CHECKER__
diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 0c03411c59e..043ddcca4ab 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -78,6 +78,8 @@ static const char *ioctl_cmd_to_ascii(int cmd)
case NBD_SET_SOCK: return "set-sock";
case NBD_SET_BLKSIZE: return "set-blksize";
case NBD_SET_SIZE: return "set-size";
+ case NBD_SET_TIMEOUT: return "set-timeout";
+ case NBD_SET_FLAGS: return "set-flags";
case NBD_DO_IT: return "do-it";
case NBD_CLEAR_SOCK: return "clear-sock";
case NBD_CLEAR_QUE: return "clear-que";
@@ -96,6 +98,7 @@ static const char *nbdcmd_to_ascii(int cmd)
case NBD_CMD_READ: return "read";
case NBD_CMD_WRITE: return "write";
case NBD_CMD_DISC: return "disconnect";
+ case NBD_CMD_TRIM: return "trim/discard";
}
return "invalid";
}
@@ -467,8 +470,12 @@ static void nbd_handle_req(struct nbd_device *nbd, struct request *req)
nbd_cmd(req) = NBD_CMD_READ;
if (rq_data_dir(req) == WRITE) {
- nbd_cmd(req) = NBD_CMD_WRITE;
- if (nbd->flags & NBD_READ_ONLY) {
+ if ((req->cmd_flags & REQ_DISCARD)) {
+ WARN_ON(!(nbd->flags & NBD_FLAG_SEND_TRIM));
+ nbd_cmd(req) = NBD_CMD_TRIM;
+ } else
+ nbd_cmd(req) = NBD_CMD_WRITE;
+ if (nbd->flags & NBD_FLAG_READ_ONLY) {
dev_err(disk_to_dev(nbd->disk),
"Write on read-only\n");
goto error_out;
@@ -651,6 +658,10 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
nbd->xmit_timeout = arg * HZ;
return 0;
+ case NBD_SET_FLAGS:
+ nbd->flags = arg;
+ return 0;
+
case NBD_SET_SIZE_BLOCKS:
nbd->bytesize = ((u64) arg) * nbd->blksize;
bdev->bd_inode->i_size = nbd->bytesize;
@@ -670,6 +681,10 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
mutex_unlock(&nbd->tx_lock);
+ if (nbd->flags & NBD_FLAG_SEND_TRIM)
+ queue_flag_set_unlocked(QUEUE_FLAG_DISCARD,
+ nbd->disk->queue);
+
thread = kthread_create(nbd_thread, nbd, nbd->disk->disk_name);
if (IS_ERR(thread)) {
mutex_lock(&nbd->tx_lock);
@@ -687,6 +702,7 @@ static int __nbd_ioctl(struct block_device *bdev, struct nbd_device *nbd,
nbd->file = NULL;
nbd_clear_que(nbd);
dev_warn(disk_to_dev(nbd->disk), "queue cleared\n");
+ queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, nbd->disk->queue);
if (file)
fput(file);
nbd->bytesize = 0;
@@ -805,6 +821,9 @@ static int __init nbd_init(void)
* Tell the block layer that we are not a rotational device
*/
queue_flag_set_unlocked(QUEUE_FLAG_NONROT, disk->queue);
+ disk->queue->limits.discard_granularity = 512;
+ disk->queue->limits.max_discard_sectors = UINT_MAX;
+ disk->queue->limits.discard_zeroes_data = 0;
}
if (register_blkdev(NBD_MAJOR, "nbd")) {
diff --git a/drivers/block/osdblk.c b/drivers/block/osdblk.c
index 87311ebac0d..1bbc681688e 100644
--- a/drivers/block/osdblk.c
+++ b/drivers/block/osdblk.c
@@ -266,11 +266,10 @@ static struct bio *bio_chain_clone(struct bio *old_chain, gfp_t gfpmask)
struct bio *tmp, *new_chain = NULL, *tail = NULL;
while (old_chain) {
- tmp = bio_kmalloc(gfpmask, old_chain->bi_max_vecs);
+ tmp = bio_clone_kmalloc(old_chain, gfpmask);
if (!tmp)
goto err_out;
- __bio_clone(tmp, old_chain);
tmp->bi_bdev = NULL;
gfpmask &= ~__GFP_WAIT;
tmp->bi_next = NULL;
diff --git a/drivers/block/pktcdvd.c b/drivers/block/pktcdvd.c
index ba66e4445f4..2e7de7a59bf 100644
--- a/drivers/block/pktcdvd.c
+++ b/drivers/block/pktcdvd.c
@@ -522,38 +522,6 @@ static void pkt_bio_finished(struct pktcdvd_device *pd)
}
}
-static void pkt_bio_destructor(struct bio *bio)
-{
- kfree(bio->bi_io_vec);
- kfree(bio);
-}
-
-static struct bio *pkt_bio_alloc(int nr_iovecs)
-{
- struct bio_vec *bvl = NULL;
- struct bio *bio;
-
- bio = kmalloc(sizeof(struct bio), GFP_KERNEL);
- if (!bio)
- goto no_bio;
- bio_init(bio);
-
- bvl = kcalloc(nr_iovecs, sizeof(struct bio_vec), GFP_KERNEL);
- if (!bvl)
- goto no_bvl;
-
- bio->bi_max_vecs = nr_iovecs;
- bio->bi_io_vec = bvl;
- bio->bi_destructor = pkt_bio_destructor;
-
- return bio;
-
- no_bvl:
- kfree(bio);
- no_bio:
- return NULL;
-}
-
/*
* Allocate a packet_data struct
*/
@@ -567,7 +535,7 @@ static struct packet_data *pkt_alloc_packet_data(int frames)
goto no_pkt;
pkt->frames = frames;
- pkt->w_bio = pkt_bio_alloc(frames);
+ pkt->w_bio = bio_kmalloc(GFP_KERNEL, frames);
if (!pkt->w_bio)
goto no_bio;
@@ -581,9 +549,10 @@ static struct packet_data *pkt_alloc_packet_data(int frames)
bio_list_init(&pkt->orig_bios);
for (i = 0; i < frames; i++) {
- struct bio *bio = pkt_bio_alloc(1);
+ struct bio *bio = bio_kmalloc(GFP_KERNEL, 1);
if (!bio)
goto no_rd_bio;
+
pkt->r_bios[i] = bio;
}
@@ -1111,21 +1080,17 @@ static void pkt_gather_data(struct pktcdvd_device *pd, struct packet_data *pkt)
* Schedule reads for missing parts of the packet.
*/
for (f = 0; f < pkt->frames; f++) {
- struct bio_vec *vec;
-
int p, offset;
+
if (written[f])
continue;
+
bio = pkt->r_bios[f];
- vec = bio->bi_io_vec;
- bio_init(bio);
- bio->bi_max_vecs = 1;
+ bio_reset(bio);
bio->bi_sector = pkt->sector + f * (CD_FRAMESIZE >> 9);
bio->bi_bdev = pd->bdev;
bio->bi_end_io = pkt_end_io_read;
bio->bi_private = pkt;
- bio->bi_io_vec = vec;
- bio->bi_destructor = pkt_bio_destructor;
p = (f * CD_FRAMESIZE) / PAGE_SIZE;
offset = (f * CD_FRAMESIZE) % PAGE_SIZE;
@@ -1418,14 +1383,11 @@ static void pkt_start_write(struct pktcdvd_device *pd, struct packet_data *pkt)
}
/* Start the write request */
- bio_init(pkt->w_bio);
- pkt->w_bio->bi_max_vecs = PACKET_MAX_SIZE;
+ bio_reset(pkt->w_bio);
pkt->w_bio->bi_sector = pkt->sector;
pkt->w_bio->bi_bdev = pd->bdev;
pkt->w_bio->bi_end_io = pkt_end_io_packet_write;
pkt->w_bio->bi_private = pkt;
- pkt->w_bio->bi_io_vec = bvec;
- pkt->w_bio->bi_destructor = pkt_bio_destructor;
for (f = 0; f < pkt->frames; f++)
if (!bio_add_page(pkt->w_bio, bvec[f].bv_page, CD_FRAMESIZE, bvec[f].bv_offset))
BUG();
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index 54a55f03115..bb3d9be3b1b 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -41,6 +41,8 @@
#include "rbd_types.h"
+#define RBD_DEBUG /* Activate rbd_assert() calls */
+
/*
* The basic unit of block I/O is a sector. It is interpreted in a
* number of contexts in Linux (blk, bio, genhd), but the default is
@@ -50,16 +52,24 @@
#define SECTOR_SHIFT 9
#define SECTOR_SIZE (1ULL << SECTOR_SHIFT)
+/* It might be useful to have this defined elsewhere too */
+
+#define U64_MAX ((u64) (~0ULL))
+
#define RBD_DRV_NAME "rbd"
#define RBD_DRV_NAME_LONG "rbd (rados block device)"
#define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */
#define RBD_MAX_SNAP_NAME_LEN 32
+#define RBD_MAX_SNAP_COUNT 510 /* allows max snapc to fit in 4KB */
#define RBD_MAX_OPT_LEN 1024
#define RBD_SNAP_HEAD_NAME "-"
+#define RBD_IMAGE_ID_LEN_MAX 64
+#define RBD_OBJ_PREFIX_LEN_MAX 64
+
/*
* An RBD device name will be "rbd#", where the "rbd" comes from
* RBD_DRV_NAME above, and # is a unique integer identifier.
@@ -69,21 +79,22 @@
#define DEV_NAME_LEN 32
#define MAX_INT_FORMAT_WIDTH ((5 * sizeof (int)) / 2 + 1)
-#define RBD_NOTIFY_TIMEOUT_DEFAULT 10
+#define RBD_READ_ONLY_DEFAULT false
/*
* block device image metadata (in-memory version)
*/
struct rbd_image_header {
- u64 image_size;
+ /* These four fields never change for a given rbd image */
char *object_prefix;
+ u64 features;
__u8 obj_order;
__u8 crypt_type;
__u8 comp_type;
- struct ceph_snap_context *snapc;
- size_t snap_names_len;
- u32 total_snaps;
+ /* The remaining fields need to be updated occasionally */
+ u64 image_size;
+ struct ceph_snap_context *snapc;
char *snap_names;
u64 *snap_sizes;
@@ -91,7 +102,7 @@ struct rbd_image_header {
};
struct rbd_options {
- int notify_timeout;
+ bool read_only;
};
/*
@@ -99,7 +110,6 @@ struct rbd_options {
*/
struct rbd_client {
struct ceph_client *client;
- struct rbd_options *rbd_opts;
struct kref kref;
struct list_head node;
};
@@ -141,6 +151,16 @@ struct rbd_snap {
u64 size;
struct list_head node;
u64 id;
+ u64 features;
+};
+
+struct rbd_mapping {
+ char *snap_name;
+ u64 snap_id;
+ u64 size;
+ u64 features;
+ bool snap_exists;
+ bool read_only;
};
/*
@@ -151,8 +171,9 @@ struct rbd_device {
int major; /* blkdev assigned major */
struct gendisk *disk; /* blkdev's gendisk and rq */
- struct request_queue *q;
+ u32 image_format; /* Either 1 or 2 */
+ struct rbd_options rbd_opts;
struct rbd_client *rbd_client;
char name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */
@@ -160,6 +181,8 @@ struct rbd_device {
spinlock_t lock; /* queue lock */
struct rbd_image_header header;
+ char *image_id;
+ size_t image_id_len;
char *image_name;
size_t image_name_len;
char *header_name;
@@ -171,13 +194,8 @@ struct rbd_device {
/* protects updating the header */
struct rw_semaphore header_rwsem;
- /* name of the snapshot this device reads from */
- char *snap_name;
- /* id of the snapshot this device reads from */
- u64 snap_id; /* current snapshot id */
- /* whether the snap_id this device reads from still exists */
- bool snap_exists;
- int read_only;
+
+ struct rbd_mapping mapping;
struct list_head node;
@@ -196,12 +214,10 @@ static DEFINE_SPINLOCK(rbd_dev_list_lock);
static LIST_HEAD(rbd_client_list); /* clients */
static DEFINE_SPINLOCK(rbd_client_list_lock);
-static int __rbd_init_snaps_header(struct rbd_device *rbd_dev);
+static int rbd_dev_snaps_update(struct rbd_device *rbd_dev);
+static int rbd_dev_snaps_register(struct rbd_device *rbd_dev);
+
static void rbd_dev_release(struct device *dev);
-static ssize_t rbd_snap_add(struct device *dev,
- struct device_attribute *attr,
- const char *buf,
- size_t count);
static void __rbd_remove_snap_dev(struct rbd_snap *snap);
static ssize_t rbd_add(struct bus_type *bus, const char *buf,
@@ -229,6 +245,18 @@ static struct device rbd_root_dev = {
.release = rbd_root_dev_release,
};
+#ifdef RBD_DEBUG
+#define rbd_assert(expr) \
+ if (unlikely(!(expr))) { \
+ printk(KERN_ERR "\nAssertion failure in %s() " \
+ "at line %d:\n\n" \
+ "\trbd_assert(%s);\n\n", \
+ __func__, __LINE__, #expr); \
+ BUG(); \
+ }
+#else /* !RBD_DEBUG */
+# define rbd_assert(expr) ((void) 0)
+#endif /* !RBD_DEBUG */
static struct device *rbd_get_dev(struct rbd_device *rbd_dev)
{
@@ -246,11 +274,11 @@ static int rbd_open(struct block_device *bdev, fmode_t mode)
{
struct rbd_device *rbd_dev = bdev->bd_disk->private_data;
- if ((mode & FMODE_WRITE) && rbd_dev->read_only)
+ if ((mode & FMODE_WRITE) && rbd_dev->mapping.read_only)
return -EROFS;
rbd_get_dev(rbd_dev);
- set_device_ro(bdev, rbd_dev->read_only);
+ set_device_ro(bdev, rbd_dev->mapping.read_only);
return 0;
}
@@ -274,8 +302,7 @@ static const struct block_device_operations rbd_bd_ops = {
* Initialize an rbd client instance.
* We own *ceph_opts.
*/
-static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts,
- struct rbd_options *rbd_opts)
+static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
{
struct rbd_client *rbdc;
int ret = -ENOMEM;
@@ -299,8 +326,6 @@ static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts,
if (ret < 0)
goto out_err;
- rbdc->rbd_opts = rbd_opts;
-
spin_lock(&rbd_client_list_lock);
list_add_tail(&rbdc->node, &rbd_client_list);
spin_unlock(&rbd_client_list_lock);
@@ -322,36 +347,52 @@ out_opt:
}
/*
- * Find a ceph client with specific addr and configuration.
+ * Find a ceph client with specific addr and configuration. If
+ * found, bump its reference count.
*/
-static struct rbd_client *__rbd_client_find(struct ceph_options *ceph_opts)
+static struct rbd_client *rbd_client_find(struct ceph_options *ceph_opts)
{
struct rbd_client *client_node;
+ bool found = false;
if (ceph_opts->flags & CEPH_OPT_NOSHARE)
return NULL;
- list_for_each_entry(client_node, &rbd_client_list, node)
- if (!ceph_compare_options(ceph_opts, client_node->client))
- return client_node;
- return NULL;
+ spin_lock(&rbd_client_list_lock);
+ list_for_each_entry(client_node, &rbd_client_list, node) {
+ if (!ceph_compare_options(ceph_opts, client_node->client)) {
+ kref_get(&client_node->kref);
+ found = true;
+ break;
+ }
+ }
+ spin_unlock(&rbd_client_list_lock);
+
+ return found ? client_node : NULL;
}
/*
* mount options
*/
enum {
- Opt_notify_timeout,
Opt_last_int,
/* int args above */
Opt_last_string,
/* string args above */
+ Opt_read_only,
+ Opt_read_write,
+ /* Boolean args above */
+ Opt_last_bool,
};
static match_table_t rbd_opts_tokens = {
- {Opt_notify_timeout, "notify_timeout=%d"},
/* int args above */
/* string args above */
+ {Opt_read_only, "mapping.read_only"},
+ {Opt_read_only, "ro"}, /* Alternate spelling */
+ {Opt_read_write, "read_write"},
+ {Opt_read_write, "rw"}, /* Alternate spelling */
+ /* Boolean args above */
{-1, NULL}
};
@@ -376,16 +417,22 @@ static int parse_rbd_opts_token(char *c, void *private)
} else if (token > Opt_last_int && token < Opt_last_string) {
dout("got string token %d val %s\n", token,
argstr[0].from);
+ } else if (token > Opt_last_string && token < Opt_last_bool) {
+ dout("got Boolean token %d\n", token);
} else {
dout("got token %d\n", token);
}
switch (token) {
- case Opt_notify_timeout:
- rbd_opts->notify_timeout = intval;
+ case Opt_read_only:
+ rbd_opts->read_only = true;
+ break;
+ case Opt_read_write:
+ rbd_opts->read_only = false;
break;
default:
- BUG_ON(token);
+ rbd_assert(false);
+ break;
}
return 0;
}
@@ -394,48 +441,33 @@ static int parse_rbd_opts_token(char *c, void *private)
* Get a ceph client with specific addr and configuration, if one does
* not exist create it.
*/
-static struct rbd_client *rbd_get_client(const char *mon_addr,
- size_t mon_addr_len,
- char *options)
+static int rbd_get_client(struct rbd_device *rbd_dev, const char *mon_addr,
+ size_t mon_addr_len, char *options)
{
- struct rbd_client *rbdc;
+ struct rbd_options *rbd_opts = &rbd_dev->rbd_opts;
struct ceph_options *ceph_opts;
- struct rbd_options *rbd_opts;
-
- rbd_opts = kzalloc(sizeof(*rbd_opts), GFP_KERNEL);
- if (!rbd_opts)
- return ERR_PTR(-ENOMEM);
+ struct rbd_client *rbdc;
- rbd_opts->notify_timeout = RBD_NOTIFY_TIMEOUT_DEFAULT;
+ rbd_opts->read_only = RBD_READ_ONLY_DEFAULT;
ceph_opts = ceph_parse_options(options, mon_addr,
mon_addr + mon_addr_len,
parse_rbd_opts_token, rbd_opts);
- if (IS_ERR(ceph_opts)) {
- kfree(rbd_opts);
- return ERR_CAST(ceph_opts);
- }
+ if (IS_ERR(ceph_opts))
+ return PTR_ERR(ceph_opts);
- spin_lock(&rbd_client_list_lock);
- rbdc = __rbd_client_find(ceph_opts);
+ rbdc = rbd_client_find(ceph_opts);
if (rbdc) {
/* using an existing client */
- kref_get(&rbdc->kref);
- spin_unlock(&rbd_client_list_lock);
-
ceph_destroy_options(ceph_opts);
- kfree(rbd_opts);
-
- return rbdc;
+ } else {
+ rbdc = rbd_client_create(ceph_opts);
+ if (IS_ERR(rbdc))
+ return PTR_ERR(rbdc);
}
- spin_unlock(&rbd_client_list_lock);
-
- rbdc = rbd_client_create(ceph_opts, rbd_opts);
+ rbd_dev->rbd_client = rbdc;
- if (IS_ERR(rbdc))
- kfree(rbd_opts);
-
- return rbdc;
+ return 0;
}
/*
@@ -453,7 +485,6 @@ static void rbd_client_release(struct kref *kref)
spin_unlock(&rbd_client_list_lock);
ceph_destroy_client(rbdc->client);
- kfree(rbdc->rbd_opts);
kfree(rbdc);
}
@@ -479,10 +510,38 @@ static void rbd_coll_release(struct kref *kref)
kfree(coll);
}
+static bool rbd_image_format_valid(u32 image_format)
+{
+ return image_format == 1 || image_format == 2;
+}
+
static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
{
- return !memcmp(&ondisk->text,
- RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT));
+ size_t size;
+ u32 snap_count;
+
+ /* The header has to start with the magic rbd header text */
+ if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT)))
+ return false;
+
+ /*
+ * The size of a snapshot header has to fit in a size_t, and
+ * that limits the number of snapshots.
+ */
+ snap_count = le32_to_cpu(ondisk->snap_count);
+ size = SIZE_MAX - sizeof (struct ceph_snap_context);
+ if (snap_count > size / sizeof (__le64))
+ return false;
+
+ /*
+ * Not only that, but the size of the entire the snapshot
+ * header must also be representable in a size_t.
+ */
+ size -= snap_count * sizeof (__le64);
+ if ((u64) size < le64_to_cpu(ondisk->snap_names_len))
+ return false;
+
+ return true;
}
/*
@@ -490,179 +549,203 @@ static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
* header.
*/
static int rbd_header_from_disk(struct rbd_image_header *header,
- struct rbd_image_header_ondisk *ondisk,
- u32 allocated_snaps)
+ struct rbd_image_header_ondisk *ondisk)
{
u32 snap_count;
+ size_t len;
+ size_t size;
+ u32 i;
- if (!rbd_dev_ondisk_valid(ondisk))
- return -ENXIO;
+ memset(header, 0, sizeof (*header));
snap_count = le32_to_cpu(ondisk->snap_count);
- if (snap_count > (SIZE_MAX - sizeof(struct ceph_snap_context))
- / sizeof (u64))
- return -EINVAL;
- header->snapc = kmalloc(sizeof(struct ceph_snap_context) +
- snap_count * sizeof(u64),
- GFP_KERNEL);
- if (!header->snapc)
+
+ len = strnlen(ondisk->object_prefix, sizeof (ondisk->object_prefix));
+ header->object_prefix = kmalloc(len + 1, GFP_KERNEL);
+ if (!header->object_prefix)
return -ENOMEM;
+ memcpy(header->object_prefix, ondisk->object_prefix, len);
+ header->object_prefix[len] = '\0';
if (snap_count) {
- header->snap_names_len = le64_to_cpu(ondisk->snap_names_len);
- header->snap_names = kmalloc(header->snap_names_len,
- GFP_KERNEL);
+ u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len);
+
+ /* Save a copy of the snapshot names */
+
+ if (snap_names_len > (u64) SIZE_MAX)
+ return -EIO;
+ header->snap_names = kmalloc(snap_names_len, GFP_KERNEL);
if (!header->snap_names)
- goto err_snapc;
- header->snap_sizes = kmalloc(snap_count * sizeof(u64),
- GFP_KERNEL);
+ goto out_err;
+ /*
+ * Note that rbd_dev_v1_header_read() guarantees
+ * the ondisk buffer we're working with has
+ * snap_names_len bytes beyond the end of the
+ * snapshot id array, this memcpy() is safe.
+ */
+ memcpy(header->snap_names, &ondisk->snaps[snap_count],
+ snap_names_len);
+
+ /* Record each snapshot's size */
+
+ size = snap_count * sizeof (*header->snap_sizes);
+ header->snap_sizes = kmalloc(size, GFP_KERNEL);
if (!header->snap_sizes)
- goto err_names;
+ goto out_err;
+ for (i = 0; i < snap_count; i++)
+ header->snap_sizes[i] =
+ le64_to_cpu(ondisk->snaps[i].image_size);
} else {
WARN_ON(ondisk->snap_names_len);
- header->snap_names_len = 0;
header->snap_names = NULL;
header->snap_sizes = NULL;
}
- header->object_prefix = kmalloc(sizeof (ondisk->block_name) + 1,
- GFP_KERNEL);
- if (!header->object_prefix)
- goto err_sizes;
-
- memcpy(header->object_prefix, ondisk->block_name,
- sizeof(ondisk->block_name));
- header->object_prefix[sizeof (ondisk->block_name)] = '\0';
-
- header->image_size = le64_to_cpu(ondisk->image_size);
+ header->features = 0; /* No features support in v1 images */
header->obj_order = ondisk->options.order;
header->crypt_type = ondisk->options.crypt_type;
header->comp_type = ondisk->options.comp_type;
+ /* Allocate and fill in the snapshot context */
+
+ header->image_size = le64_to_cpu(ondisk->image_size);
+ size = sizeof (struct ceph_snap_context);
+ size += snap_count * sizeof (header->snapc->snaps[0]);
+ header->snapc = kzalloc(size, GFP_KERNEL);
+ if (!header->snapc)
+ goto out_err;
+
atomic_set(&header->snapc->nref, 1);
header->snapc->seq = le64_to_cpu(ondisk->snap_seq);
header->snapc->num_snaps = snap_count;
- header->total_snaps = snap_count;
-
- if (snap_count && allocated_snaps == snap_count) {
- int i;
-
- for (i = 0; i < snap_count; i++) {
- header->snapc->snaps[i] =
- le64_to_cpu(ondisk->snaps[i].id);
- header->snap_sizes[i] =
- le64_to_cpu(ondisk->snaps[i].image_size);
- }
-
- /* copy snapshot names */
- memcpy(header->snap_names, &ondisk->snaps[snap_count],
- header->snap_names_len);
- }
+ for (i = 0; i < snap_count; i++)
+ header->snapc->snaps[i] =
+ le64_to_cpu(ondisk->snaps[i].id);
return 0;
-err_sizes:
+out_err:
kfree(header->snap_sizes);
header->snap_sizes = NULL;
-err_names:
kfree(header->snap_names);
header->snap_names = NULL;
-err_snapc:
- kfree(header->snapc);
- header->snapc = NULL;
+ kfree(header->object_prefix);
+ header->object_prefix = NULL;
return -ENOMEM;
}
-static int snap_by_name(struct rbd_image_header *header, const char *snap_name,
- u64 *seq, u64 *size)
+static int snap_by_name(struct rbd_device *rbd_dev, const char *snap_name)
{
- int i;
- char *p = header->snap_names;
- for (i = 0; i < header->total_snaps; i++) {
- if (!strcmp(snap_name, p)) {
+ struct rbd_snap *snap;
- /* Found it. Pass back its id and/or size */
+ list_for_each_entry(snap, &rbd_dev->snaps, node) {
+ if (!strcmp(snap_name, snap->name)) {
+ rbd_dev->mapping.snap_id = snap->id;
+ rbd_dev->mapping.size = snap->size;
+ rbd_dev->mapping.features = snap->features;
- if (seq)
- *seq = header->snapc->snaps[i];
- if (size)
- *size = header->snap_sizes[i];
- return i;
+ return 0;
}
- p += strlen(p) + 1; /* Skip ahead to the next name */
}
+
return -ENOENT;
}
-static int rbd_header_set_snap(struct rbd_device *rbd_dev, u64 *size)
+static int rbd_dev_set_mapping(struct rbd_device *rbd_dev, char *snap_name)
{
int ret;
- down_write(&rbd_dev->header_rwsem);
-
- if (!memcmp(rbd_dev->snap_name, RBD_SNAP_HEAD_NAME,
+ if (!memcmp(snap_name, RBD_SNAP_HEAD_NAME,
sizeof (RBD_SNAP_HEAD_NAME))) {
- rbd_dev->snap_id = CEPH_NOSNAP;
- rbd_dev->snap_exists = false;
- rbd_dev->read_only = 0;
- if (size)
- *size = rbd_dev->header.image_size;
+ rbd_dev->mapping.snap_id = CEPH_NOSNAP;
+ rbd_dev->mapping.size = rbd_dev->header.image_size;
+ rbd_dev->mapping.features = rbd_dev->header.features;
+ rbd_dev->mapping.snap_exists = false;
+ rbd_dev->mapping.read_only = rbd_dev->rbd_opts.read_only;
+ ret = 0;
} else {
- u64 snap_id = 0;
-
- ret = snap_by_name(&rbd_dev->header, rbd_dev->snap_name,
- &snap_id, size);
+ ret = snap_by_name(rbd_dev, snap_name);
if (ret < 0)
goto done;
- rbd_dev->snap_id = snap_id;
- rbd_dev->snap_exists = true;
- rbd_dev->read_only = 1;
+ rbd_dev->mapping.snap_exists = true;
+ rbd_dev->mapping.read_only = true;
}
-
- ret = 0;
+ rbd_dev->mapping.snap_name = snap_name;
done:
- up_write(&rbd_dev->header_rwsem);
return ret;
}
static void rbd_header_free(struct rbd_image_header *header)
{
kfree(header->object_prefix);
+ header->object_prefix = NULL;
kfree(header->snap_sizes);
+ header->snap_sizes = NULL;
kfree(header->snap_names);
+ header->snap_names = NULL;
ceph_put_snap_context(header->snapc);
+ header->snapc = NULL;
}
-/*
- * get the actual striped segment name, offset and length
- */
-static u64 rbd_get_segment(struct rbd_image_header *header,
- const char *object_prefix,
- u64 ofs, u64 len,
- char *seg_name, u64 *segofs)
+static char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset)
+{
+ char *name;
+ u64 segment;
+ int ret;
+
+ name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO);
+ if (!name)
+ return NULL;
+ segment = offset >> rbd_dev->header.obj_order;
+ ret = snprintf(name, RBD_MAX_SEG_NAME_LEN, "%s.%012llx",
+ rbd_dev->header.object_prefix, segment);
+ if (ret < 0 || ret >= RBD_MAX_SEG_NAME_LEN) {
+ pr_err("error formatting segment name for #%llu (%d)\n",
+ segment, ret);
+ kfree(name);
+ name = NULL;
+ }
+
+ return name;
+}
+
+static u64 rbd_segment_offset(struct rbd_device *rbd_dev, u64 offset)
{
- u64 seg = ofs >> header->obj_order;
+ u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
- if (seg_name)
- snprintf(seg_name, RBD_MAX_SEG_NAME_LEN,
- "%s.%012llx", object_prefix, seg);
+ return offset & (segment_size - 1);
+}
+
+static u64 rbd_segment_length(struct rbd_device *rbd_dev,
+ u64 offset, u64 length)
+{
+ u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
- ofs = ofs & ((1 << header->obj_order) - 1);
- len = min_t(u64, len, (1 << header->obj_order) - ofs);
+ offset &= segment_size - 1;
- if (segofs)
- *segofs = ofs;
+ rbd_assert(length <= U64_MAX - offset);
+ if (offset + length > segment_size)
+ length = segment_size - offset;
- return len;
+ return length;
}
static int rbd_get_num_segments(struct rbd_image_header *header,
u64 ofs, u64 len)
{
- u64 start_seg = ofs >> header->obj_order;
- u64 end_seg = (ofs + len - 1) >> header->obj_order;
+ u64 start_seg;
+ u64 end_seg;
+
+ if (!len)
+ return 0;
+ if (len - 1 > U64_MAX - ofs)
+ return -ERANGE;
+
+ start_seg = ofs >> header->obj_order;
+ end_seg = (ofs + len - 1) >> header->obj_order;
+
return end_seg - start_seg + 1;
}
@@ -724,7 +807,9 @@ static struct bio *bio_chain_clone(struct bio **old, struct bio **next,
struct bio_pair **bp,
int len, gfp_t gfpmask)
{
- struct bio *tmp, *old_chain = *old, *new_chain = NULL, *tail = NULL;
+ struct bio *old_chain = *old;
+ struct bio *new_chain = NULL;
+ struct bio *tail;
int total = 0;
if (*bp) {
@@ -733,9 +818,12 @@ static struct bio *bio_chain_clone(struct bio **old, struct bio **next,
}
while (old_chain && (total < len)) {
+ struct bio *tmp;
+
tmp = bio_kmalloc(gfpmask, old_chain->bi_max_vecs);
if (!tmp)
goto err_out;
+ gfpmask &= ~__GFP_WAIT; /* can't wait after the first */
if (total + old_chain->bi_size > len) {
struct bio_pair *bp;
@@ -763,24 +851,18 @@ static struct bio *bio_chain_clone(struct bio **old, struct bio **next,
}
tmp->bi_bdev = NULL;
- gfpmask &= ~__GFP_WAIT;
tmp->bi_next = NULL;
-
- if (!new_chain) {
- new_chain = tail = tmp;
- } else {
+ if (new_chain)
tail->bi_next = tmp;
- tail = tmp;
- }
+ else
+ new_chain = tmp;
+ tail = tmp;
old_chain = old_chain->bi_next;
total += tmp->bi_size;
}
- BUG_ON(total < len);
-
- if (tail)
- tail->bi_next = NULL;
+ rbd_assert(total == len);
*old = old_chain;
@@ -938,8 +1020,9 @@ static int rbd_do_request(struct request *rq,
layout->fl_stripe_count = cpu_to_le32(1);
layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
layout->fl_pg_pool = cpu_to_le32(rbd_dev->pool_id);
- ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
- req, ops);
+ ret = ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
+ req, ops);
+ rbd_assert(ret == 0);
ceph_osdc_build_request(req, ofs, &len,
ops,
@@ -1030,8 +1113,8 @@ static int rbd_req_sync_op(struct rbd_device *rbd_dev,
int flags,
struct ceph_osd_req_op *ops,
const char *object_name,
- u64 ofs, u64 len,
- char *buf,
+ u64 ofs, u64 inbound_size,
+ char *inbound,
struct ceph_osd_request **linger_req,
u64 *ver)
{
@@ -1039,15 +1122,15 @@ static int rbd_req_sync_op(struct rbd_device *rbd_dev,
struct page **pages;
int num_pages;
- BUG_ON(ops == NULL);
+ rbd_assert(ops != NULL);
- num_pages = calc_pages_for(ofs , len);
+ num_pages = calc_pages_for(ofs, inbound_size);
pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
if (IS_ERR(pages))
return PTR_ERR(pages);
ret = rbd_do_request(NULL, rbd_dev, snapc, snapid,
- object_name, ofs, len, NULL,
+ object_name, ofs, inbound_size, NULL,
pages, num_pages,
flags,
ops,
@@ -1057,8 +1140,8 @@ static int rbd_req_sync_op(struct rbd_device *rbd_dev,
if (ret < 0)
goto done;
- if ((flags & CEPH_OSD_FLAG_READ) && buf)
- ret = ceph_copy_from_page_vector(pages, buf, ofs, ret);
+ if ((flags & CEPH_OSD_FLAG_READ) && inbound)
+ ret = ceph_copy_from_page_vector(pages, inbound, ofs, ret);
done:
ceph_release_page_vector(pages, num_pages);
@@ -1085,14 +1168,11 @@ static int rbd_do_op(struct request *rq,
struct ceph_osd_req_op *ops;
u32 payload_len;
- seg_name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO);
+ seg_name = rbd_segment_name(rbd_dev, ofs);
if (!seg_name)
return -ENOMEM;
-
- seg_len = rbd_get_segment(&rbd_dev->header,
- rbd_dev->header.object_prefix,
- ofs, len,
- seg_name, &seg_ofs);
+ seg_len = rbd_segment_length(rbd_dev, ofs, len);
+ seg_ofs = rbd_segment_offset(rbd_dev, ofs);
payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0);
@@ -1104,7 +1184,7 @@ static int rbd_do_op(struct request *rq,
/* we've taken care of segment sizes earlier when we
cloned the bios. We should never have a segment
truncated at this point */
- BUG_ON(seg_len < len);
+ rbd_assert(seg_len == len);
ret = rbd_do_request(rq, rbd_dev, snapc, snapid,
seg_name, seg_ofs, seg_len,
@@ -1306,89 +1386,36 @@ static int rbd_req_sync_unwatch(struct rbd_device *rbd_dev)
return ret;
}
-struct rbd_notify_info {
- struct rbd_device *rbd_dev;
-};
-
-static void rbd_notify_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
-{
- struct rbd_device *rbd_dev = (struct rbd_device *)data;
- if (!rbd_dev)
- return;
-
- dout("rbd_notify_cb %s notify_id=%llu opcode=%u\n",
- rbd_dev->header_name, (unsigned long long) notify_id,
- (unsigned int) opcode);
-}
-
/*
- * Request sync osd notify
- */
-static int rbd_req_sync_notify(struct rbd_device *rbd_dev)
-{
- struct ceph_osd_req_op *ops;
- struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
- struct ceph_osd_event *event;
- struct rbd_notify_info info;
- int payload_len = sizeof(u32) + sizeof(u32);
- int ret;
-
- ops = rbd_create_rw_ops(1, CEPH_OSD_OP_NOTIFY, payload_len);
- if (!ops)
- return -ENOMEM;
-
- info.rbd_dev = rbd_dev;
-
- ret = ceph_osdc_create_event(osdc, rbd_notify_cb, 1,
- (void *)&info, &event);
- if (ret < 0)
- goto fail;
-
- ops[0].watch.ver = 1;
- ops[0].watch.flag = 1;
- ops[0].watch.cookie = event->cookie;
- ops[0].watch.prot_ver = RADOS_NOTIFY_VER;
- ops[0].watch.timeout = 12;
-
- ret = rbd_req_sync_op(rbd_dev, NULL,
- CEPH_NOSNAP,
- CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
- ops,
- rbd_dev->header_name,
- 0, 0, NULL, NULL, NULL);
- if (ret < 0)
- goto fail_event;
-
- ret = ceph_osdc_wait_event(event, CEPH_OSD_TIMEOUT_DEFAULT);
- dout("ceph_osdc_wait_event returned %d\n", ret);
- rbd_destroy_ops(ops);
- return 0;
-
-fail_event:
- ceph_osdc_cancel_event(event);
-fail:
- rbd_destroy_ops(ops);
- return ret;
-}
-
-/*
- * Request sync osd read
+ * Synchronous osd object method call
*/
static int rbd_req_sync_exec(struct rbd_device *rbd_dev,
const char *object_name,
const char *class_name,
const char *method_name,
- const char *data,
- int len,
+ const char *outbound,
+ size_t outbound_size,
+ char *inbound,
+ size_t inbound_size,
+ int flags,
u64 *ver)
{
struct ceph_osd_req_op *ops;
int class_name_len = strlen(class_name);
int method_name_len = strlen(method_name);
+ int payload_size;
int ret;
- ops = rbd_create_rw_ops(1, CEPH_OSD_OP_CALL,
- class_name_len + method_name_len + len);
+ /*
+ * Any input parameters required by the method we're calling
+ * will be sent along with the class and method names as
+ * part of the message payload. That data and its size are
+ * supplied via the indata and indata_len fields (named from
+ * the perspective of the server side) in the OSD request
+ * operation.
+ */
+ payload_size = class_name_len + method_name_len + outbound_size;
+ ops = rbd_create_rw_ops(1, CEPH_OSD_OP_CALL, payload_size);
if (!ops)
return -ENOMEM;
@@ -1397,14 +1424,14 @@ static int rbd_req_sync_exec(struct rbd_device *rbd_dev,
ops[0].cls.method_name = method_name;
ops[0].cls.method_len = (__u8) method_name_len;
ops[0].cls.argc = 0;
- ops[0].cls.indata = data;
- ops[0].cls.indata_len = len;
+ ops[0].cls.indata = outbound;
+ ops[0].cls.indata_len = outbound_size;
ret = rbd_req_sync_op(rbd_dev, NULL,
CEPH_NOSNAP,
- CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
- ops,
- object_name, 0, 0, NULL, NULL, ver);
+ flags, ops,
+ object_name, 0, inbound_size, inbound,
+ NULL, ver);
rbd_destroy_ops(ops);
@@ -1446,10 +1473,6 @@ static void rbd_rq_fn(struct request_queue *q)
struct rbd_req_coll *coll;
struct ceph_snap_context *snapc;
- /* peek at request from block layer */
- if (!rq)
- break;
-
dout("fetched request\n");
/* filter out block requests we don't understand */
@@ -1464,7 +1487,7 @@ static void rbd_rq_fn(struct request_queue *q)
size = blk_rq_bytes(rq);
ofs = blk_rq_pos(rq) * SECTOR_SIZE;
rq_bio = rq->bio;
- if (do_write && rbd_dev->read_only) {
+ if (do_write && rbd_dev->mapping.read_only) {
__blk_end_request_all(rq, -EROFS);
continue;
}
@@ -1473,7 +1496,8 @@ static void rbd_rq_fn(struct request_queue *q)
down_read(&rbd_dev->header_rwsem);
- if (rbd_dev->snap_id != CEPH_NOSNAP && !rbd_dev->snap_exists) {
+ if (rbd_dev->mapping.snap_id != CEPH_NOSNAP &&
+ !rbd_dev->mapping.snap_exists) {
up_read(&rbd_dev->header_rwsem);
dout("request for non-existent snapshot");
spin_lock_irq(q->queue_lock);
@@ -1490,6 +1514,12 @@ static void rbd_rq_fn(struct request_queue *q)
size, (unsigned long long) blk_rq_pos(rq) * SECTOR_SIZE);
num_segs = rbd_get_num_segments(&rbd_dev->header, ofs, size);
+ if (num_segs <= 0) {
+ spin_lock_irq(q->queue_lock);
+ __blk_end_request_all(rq, num_segs);
+ ceph_put_snap_context(snapc);
+ continue;
+ }
coll = rbd_alloc_coll(num_segs);
if (!coll) {
spin_lock_irq(q->queue_lock);
@@ -1501,10 +1531,7 @@ static void rbd_rq_fn(struct request_queue *q)
do {
/* a bio clone to be passed down to OSD req */
dout("rq->bio->bi_vcnt=%hu\n", rq->bio->bi_vcnt);
- op_size = rbd_get_segment(&rbd_dev->header,
- rbd_dev->header.object_prefix,
- ofs, size,
- NULL, NULL);
+ op_size = rbd_segment_length(rbd_dev, ofs, size);
kref_get(&coll->kref);
bio = bio_chain_clone(&rq_bio, &next_bio, &bp,
op_size, GFP_ATOMIC);
@@ -1524,7 +1551,7 @@ static void rbd_rq_fn(struct request_queue *q)
coll, cur_seg);
else
rbd_req_read(rq, rbd_dev,
- rbd_dev->snap_id,
+ rbd_dev->mapping.snap_id,
ofs,
op_size, bio,
coll, cur_seg);
@@ -1580,8 +1607,6 @@ static void rbd_free_disk(struct rbd_device *rbd_dev)
if (!disk)
return;
- rbd_header_free(&rbd_dev->header);
-
if (disk->flags & GENHD_FL_UP)
del_gendisk(disk);
if (disk->queue)
@@ -1590,105 +1615,96 @@ static void rbd_free_disk(struct rbd_device *rbd_dev)
}
/*
- * reload the ondisk the header
+ * Read the complete header for the given rbd device.
+ *
+ * Returns a pointer to a dynamically-allocated buffer containing
+ * the complete and validated header. Caller can pass the address
+ * of a variable that will be filled in with the version of the
+ * header object at the time it was read.
+ *
+ * Returns a pointer-coded errno if a failure occurs.
*/
-static int rbd_read_header(struct rbd_device *rbd_dev,
- struct rbd_image_header *header)
+static struct rbd_image_header_ondisk *
+rbd_dev_v1_header_read(struct rbd_device *rbd_dev, u64 *version)
{
- ssize_t rc;
- struct rbd_image_header_ondisk *dh;
+ struct rbd_image_header_ondisk *ondisk = NULL;
u32 snap_count = 0;
- u64 ver;
- size_t len;
+ u64 names_size = 0;
+ u32 want_count;
+ int ret;
/*
- * First reads the fixed-size header to determine the number
- * of snapshots, then re-reads it, along with all snapshot
- * records as well as their stored names.
+ * The complete header will include an array of its 64-bit
+ * snapshot ids, followed by the names of those snapshots as
+ * a contiguous block of NUL-terminated strings. Note that
+ * the number of snapshots could change by the time we read
+ * it in, in which case we re-read it.
*/
- len = sizeof (*dh);
- while (1) {
- dh = kmalloc(len, GFP_KERNEL);
- if (!dh)
- return -ENOMEM;
-
- rc = rbd_req_sync_read(rbd_dev,
- CEPH_NOSNAP,
+ do {
+ size_t size;
+
+ kfree(ondisk);
+
+ size = sizeof (*ondisk);
+ size += snap_count * sizeof (struct rbd_image_snap_ondisk);
+ size += names_size;
+ ondisk = kmalloc(size, GFP_KERNEL);
+ if (!ondisk)
+ return ERR_PTR(-ENOMEM);
+
+ ret = rbd_req_sync_read(rbd_dev, CEPH_NOSNAP,
rbd_dev->header_name,
- 0, len,
- (char *)dh, &ver);
- if (rc < 0)
- goto out_dh;
-
- rc = rbd_header_from_disk(header, dh, snap_count);
- if (rc < 0) {
- if (rc == -ENXIO)
- pr_warning("unrecognized header format"
- " for image %s\n",
- rbd_dev->image_name);
- goto out_dh;
+ 0, size,
+ (char *) ondisk, version);
+
+ if (ret < 0)
+ goto out_err;
+ if (WARN_ON((size_t) ret < size)) {
+ ret = -ENXIO;
+ pr_warning("short header read for image %s"
+ " (want %zd got %d)\n",
+ rbd_dev->image_name, size, ret);
+ goto out_err;
+ }
+ if (!rbd_dev_ondisk_valid(ondisk)) {
+ ret = -ENXIO;
+ pr_warning("invalid header for image %s\n",
+ rbd_dev->image_name);
+ goto out_err;
}
- if (snap_count == header->total_snaps)
- break;
+ names_size = le64_to_cpu(ondisk->snap_names_len);
+ want_count = snap_count;
+ snap_count = le32_to_cpu(ondisk->snap_count);
+ } while (snap_count != want_count);
- snap_count = header->total_snaps;
- len = sizeof (*dh) +
- snap_count * sizeof(struct rbd_image_snap_ondisk) +
- header->snap_names_len;
+ return ondisk;
- rbd_header_free(header);
- kfree(dh);
- }
- header->obj_version = ver;
+out_err:
+ kfree(ondisk);
-out_dh:
- kfree(dh);
- return rc;
+ return ERR_PTR(ret);
}
/*
- * create a snapshot
+ * reload the ondisk the header
*/
-static int rbd_header_add_snap(struct rbd_device *rbd_dev,
- const char *snap_name,
- gfp_t gfp_flags)
+static int rbd_read_header(struct rbd_device *rbd_dev,
+ struct rbd_image_header *header)
{
- int name_len = strlen(snap_name);
- u64 new_snapid;
+ struct rbd_image_header_ondisk *ondisk;
+ u64 ver = 0;
int ret;
- void *data, *p, *e;
- struct ceph_mon_client *monc;
-
- /* we should create a snapshot only if we're pointing at the head */
- if (rbd_dev->snap_id != CEPH_NOSNAP)
- return -EINVAL;
- monc = &rbd_dev->rbd_client->client->monc;
- ret = ceph_monc_create_snapid(monc, rbd_dev->pool_id, &new_snapid);
- dout("created snapid=%llu\n", (unsigned long long) new_snapid);
- if (ret < 0)
- return ret;
-
- data = kmalloc(name_len + 16, gfp_flags);
- if (!data)
- return -ENOMEM;
+ ondisk = rbd_dev_v1_header_read(rbd_dev, &ver);
+ if (IS_ERR(ondisk))
+ return PTR_ERR(ondisk);
+ ret = rbd_header_from_disk(header, ondisk);
+ if (ret >= 0)
+ header->obj_version = ver;
+ kfree(ondisk);
- p = data;
- e = data + name_len + 16;
-
- ceph_encode_string_safe(&p, e, snap_name, name_len, bad);
- ceph_encode_64_safe(&p, e, new_snapid, bad);
-
- ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
- "rbd", "snap_add",
- data, p - data, NULL);
-
- kfree(data);
-
- return ret < 0 ? ret : 0;
-bad:
- return -ERANGE;
+ return ret;
}
static void __rbd_remove_all_snaps(struct rbd_device *rbd_dev)
@@ -1715,11 +1731,15 @@ static int __rbd_refresh_header(struct rbd_device *rbd_dev, u64 *hver)
down_write(&rbd_dev->header_rwsem);
/* resized? */
- if (rbd_dev->snap_id == CEPH_NOSNAP) {
+ if (rbd_dev->mapping.snap_id == CEPH_NOSNAP) {
sector_t size = (sector_t) h.image_size / SECTOR_SIZE;
- dout("setting size to %llu sectors", (unsigned long long) size);
- set_capacity(rbd_dev->disk, size);
+ if (size != (sector_t) rbd_dev->mapping.size) {
+ dout("setting size to %llu sectors",
+ (unsigned long long) size);
+ rbd_dev->mapping.size = (u64) size;
+ set_capacity(rbd_dev->disk, size);
+ }
}
/* rbd_dev->header.object_prefix shouldn't change */
@@ -1732,16 +1752,16 @@ static int __rbd_refresh_header(struct rbd_device *rbd_dev, u64 *hver)
*hver = h.obj_version;
rbd_dev->header.obj_version = h.obj_version;
rbd_dev->header.image_size = h.image_size;
- rbd_dev->header.total_snaps = h.total_snaps;
rbd_dev->header.snapc = h.snapc;
rbd_dev->header.snap_names = h.snap_names;
- rbd_dev->header.snap_names_len = h.snap_names_len;
rbd_dev->header.snap_sizes = h.snap_sizes;
/* Free the extra copy of the object prefix */
WARN_ON(strcmp(rbd_dev->header.object_prefix, h.object_prefix));
kfree(h.object_prefix);
- ret = __rbd_init_snaps_header(rbd_dev);
+ ret = rbd_dev_snaps_update(rbd_dev);
+ if (!ret)
+ ret = rbd_dev_snaps_register(rbd_dev);
up_write(&rbd_dev->header_rwsem);
@@ -1763,29 +1783,12 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
{
struct gendisk *disk;
struct request_queue *q;
- int rc;
u64 segment_size;
- u64 total_size = 0;
-
- /* contact OSD, request size info about the object being mapped */
- rc = rbd_read_header(rbd_dev, &rbd_dev->header);
- if (rc)
- return rc;
-
- /* no need to lock here, as rbd_dev is not registered yet */
- rc = __rbd_init_snaps_header(rbd_dev);
- if (rc)
- return rc;
-
- rc = rbd_header_set_snap(rbd_dev, &total_size);
- if (rc)
- return rc;
/* create gendisk info */
- rc = -ENOMEM;
disk = alloc_disk(RBD_MINORS_PER_MAJOR);
if (!disk)
- goto out;
+ return -ENOMEM;
snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
rbd_dev->dev_id);
@@ -1795,7 +1798,6 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
disk->private_data = rbd_dev;
/* init rq */
- rc = -ENOMEM;
q = blk_init_queue(rbd_rq_fn, &rbd_dev->lock);
if (!q)
goto out_disk;
@@ -1816,20 +1818,14 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
q->queuedata = rbd_dev;
rbd_dev->disk = disk;
- rbd_dev->q = q;
- /* finally, announce the disk to the world */
- set_capacity(disk, total_size / SECTOR_SIZE);
- add_disk(disk);
+ set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
- pr_info("%s: added with size 0x%llx\n",
- disk->disk_name, (unsigned long long)total_size);
return 0;
-
out_disk:
put_disk(disk);
-out:
- return rc;
+
+ return -ENOMEM;
}
/*
@@ -1854,6 +1850,19 @@ static ssize_t rbd_size_show(struct device *dev,
return sprintf(buf, "%llu\n", (unsigned long long) size * SECTOR_SIZE);
}
+/*
+ * Note this shows the features for whatever's mapped, which is not
+ * necessarily the base image.
+ */
+static ssize_t rbd_features_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
+
+ return sprintf(buf, "0x%016llx\n",
+ (unsigned long long) rbd_dev->mapping.features);
+}
+
static ssize_t rbd_major_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
@@ -1895,13 +1904,25 @@ static ssize_t rbd_name_show(struct device *dev,
return sprintf(buf, "%s\n", rbd_dev->image_name);
}
+static ssize_t rbd_image_id_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
+
+ return sprintf(buf, "%s\n", rbd_dev->image_id);
+}
+
+/*
+ * Shows the name of the currently-mapped snapshot (or
+ * RBD_SNAP_HEAD_NAME for the base image).
+ */
static ssize_t rbd_snap_show(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- return sprintf(buf, "%s\n", rbd_dev->snap_name);
+ return sprintf(buf, "%s\n", rbd_dev->mapping.snap_name);
}
static ssize_t rbd_image_refresh(struct device *dev,
@@ -1918,25 +1939,27 @@ static ssize_t rbd_image_refresh(struct device *dev,
}
static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
+static DEVICE_ATTR(features, S_IRUGO, rbd_features_show, NULL);
static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
static DEVICE_ATTR(pool_id, S_IRUGO, rbd_pool_id_show, NULL);
static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
+static DEVICE_ATTR(image_id, S_IRUGO, rbd_image_id_show, NULL);
static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
-static DEVICE_ATTR(create_snap, S_IWUSR, NULL, rbd_snap_add);
static struct attribute *rbd_attrs[] = {
&dev_attr_size.attr,
+ &dev_attr_features.attr,
&dev_attr_major.attr,
&dev_attr_client_id.attr,
&dev_attr_pool.attr,
&dev_attr_pool_id.attr,
&dev_attr_name.attr,
+ &dev_attr_image_id.attr,
&dev_attr_current_snap.attr,
&dev_attr_refresh.attr,
- &dev_attr_create_snap.attr,
NULL
};
@@ -1982,12 +2005,24 @@ static ssize_t rbd_snap_id_show(struct device *dev,
return sprintf(buf, "%llu\n", (unsigned long long)snap->id);
}
+static ssize_t rbd_snap_features_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
+
+ return sprintf(buf, "0x%016llx\n",
+ (unsigned long long) snap->features);
+}
+
static DEVICE_ATTR(snap_size, S_IRUGO, rbd_snap_size_show, NULL);
static DEVICE_ATTR(snap_id, S_IRUGO, rbd_snap_id_show, NULL);
+static DEVICE_ATTR(snap_features, S_IRUGO, rbd_snap_features_show, NULL);
static struct attribute *rbd_snap_attrs[] = {
&dev_attr_snap_size.attr,
&dev_attr_snap_id.attr,
+ &dev_attr_snap_features.attr,
NULL,
};
@@ -2012,10 +2047,21 @@ static struct device_type rbd_snap_device_type = {
.release = rbd_snap_dev_release,
};
+static bool rbd_snap_registered(struct rbd_snap *snap)
+{
+ bool ret = snap->dev.type == &rbd_snap_device_type;
+ bool reg = device_is_registered(&snap->dev);
+
+ rbd_assert(!ret ^ reg);
+
+ return ret;
+}
+
static void __rbd_remove_snap_dev(struct rbd_snap *snap)
{
list_del(&snap->node);
- device_unregister(&snap->dev);
+ if (device_is_registered(&snap->dev))
+ device_unregister(&snap->dev);
}
static int rbd_register_snap_dev(struct rbd_snap *snap,
@@ -2028,13 +2074,17 @@ static int rbd_register_snap_dev(struct rbd_snap *snap,
dev->parent = parent;
dev->release = rbd_snap_dev_release;
dev_set_name(dev, "snap_%s", snap->name);
+ dout("%s: registering device for snapshot %s\n", __func__, snap->name);
+
ret = device_register(dev);
return ret;
}
static struct rbd_snap *__rbd_add_snap_dev(struct rbd_device *rbd_dev,
- int i, const char *name)
+ const char *snap_name,
+ u64 snap_id, u64 snap_size,
+ u64 snap_features)
{
struct rbd_snap *snap;
int ret;
@@ -2044,17 +2094,13 @@ static struct rbd_snap *__rbd_add_snap_dev(struct rbd_device *rbd_dev,
return ERR_PTR(-ENOMEM);
ret = -ENOMEM;
- snap->name = kstrdup(name, GFP_KERNEL);
+ snap->name = kstrdup(snap_name, GFP_KERNEL);
if (!snap->name)
goto err;
- snap->size = rbd_dev->header.snap_sizes[i];
- snap->id = rbd_dev->header.snapc->snaps[i];
- if (device_is_registered(&rbd_dev->dev)) {
- ret = rbd_register_snap_dev(snap, &rbd_dev->dev);
- if (ret < 0)
- goto err;
- }
+ snap->id = snap_id;
+ snap->size = snap_size;
+ snap->features = snap_features;
return snap;
@@ -2065,128 +2111,439 @@ err:
return ERR_PTR(ret);
}
+static char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ char *snap_name;
+
+ rbd_assert(which < rbd_dev->header.snapc->num_snaps);
+
+ *snap_size = rbd_dev->header.snap_sizes[which];
+ *snap_features = 0; /* No features for v1 */
+
+ /* Skip over names until we find the one we are looking for */
+
+ snap_name = rbd_dev->header.snap_names;
+ while (which--)
+ snap_name += strlen(snap_name) + 1;
+
+ return snap_name;
+}
+
/*
- * search for the previous snap in a null delimited string list
+ * Get the size and object order for an image snapshot, or if
+ * snap_id is CEPH_NOSNAP, gets this information for the base
+ * image.
*/
-const char *rbd_prev_snap_name(const char *name, const char *start)
+static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
+ u8 *order, u64 *snap_size)
{
- if (name < start + 2)
- return NULL;
+ __le64 snapid = cpu_to_le64(snap_id);
+ int ret;
+ struct {
+ u8 order;
+ __le64 size;
+ } __attribute__ ((packed)) size_buf = { 0 };
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_size",
+ (char *) &snapid, sizeof (snapid),
+ (char *) &size_buf, sizeof (size_buf),
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ return ret;
+
+ *order = size_buf.order;
+ *snap_size = le64_to_cpu(size_buf.size);
+
+ dout(" snap_id 0x%016llx order = %u, snap_size = %llu\n",
+ (unsigned long long) snap_id, (unsigned int) *order,
+ (unsigned long long) *snap_size);
+
+ return 0;
+}
+
+static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
+{
+ return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
+ &rbd_dev->header.obj_order,
+ &rbd_dev->header.image_size);
+}
+
+static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
+{
+ void *reply_buf;
+ int ret;
+ void *p;
+
+ reply_buf = kzalloc(RBD_OBJ_PREFIX_LEN_MAX, GFP_KERNEL);
+ if (!reply_buf)
+ return -ENOMEM;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_object_prefix",
+ NULL, 0,
+ reply_buf, RBD_OBJ_PREFIX_LEN_MAX,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ p = reply_buf;
+ rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
+ p + RBD_OBJ_PREFIX_LEN_MAX,
+ NULL, GFP_NOIO);
+
+ if (IS_ERR(rbd_dev->header.object_prefix)) {
+ ret = PTR_ERR(rbd_dev->header.object_prefix);
+ rbd_dev->header.object_prefix = NULL;
+ } else {
+ dout(" object_prefix = %s\n", rbd_dev->header.object_prefix);
+ }
- name -= 2;
- while (*name) {
- if (name == start)
- return start;
- name--;
+out:
+ kfree(reply_buf);
+
+ return ret;
+}
+
+static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
+ u64 *snap_features)
+{
+ __le64 snapid = cpu_to_le64(snap_id);
+ struct {
+ __le64 features;
+ __le64 incompat;
+ } features_buf = { 0 };
+ int ret;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_features",
+ (char *) &snapid, sizeof (snapid),
+ (char *) &features_buf, sizeof (features_buf),
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ return ret;
+ *snap_features = le64_to_cpu(features_buf.features);
+
+ dout(" snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
+ (unsigned long long) snap_id,
+ (unsigned long long) *snap_features,
+ (unsigned long long) le64_to_cpu(features_buf.incompat));
+
+ return 0;
+}
+
+static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
+{
+ return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
+ &rbd_dev->header.features);
+}
+
+static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev, u64 *ver)
+{
+ size_t size;
+ int ret;
+ void *reply_buf;
+ void *p;
+ void *end;
+ u64 seq;
+ u32 snap_count;
+ struct ceph_snap_context *snapc;
+ u32 i;
+
+ /*
+ * We'll need room for the seq value (maximum snapshot id),
+ * snapshot count, and array of that many snapshot ids.
+ * For now we have a fixed upper limit on the number we're
+ * prepared to receive.
+ */
+ size = sizeof (__le64) + sizeof (__le32) +
+ RBD_MAX_SNAP_COUNT * sizeof (__le64);
+ reply_buf = kzalloc(size, GFP_KERNEL);
+ if (!reply_buf)
+ return -ENOMEM;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_snapcontext",
+ NULL, 0,
+ reply_buf, size,
+ CEPH_OSD_FLAG_READ, ver);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ ret = -ERANGE;
+ p = reply_buf;
+ end = (char *) reply_buf + size;
+ ceph_decode_64_safe(&p, end, seq, out);
+ ceph_decode_32_safe(&p, end, snap_count, out);
+
+ /*
+ * Make sure the reported number of snapshot ids wouldn't go
+ * beyond the end of our buffer. But before checking that,
+ * make sure the computed size of the snapshot context we
+ * allocate is representable in a size_t.
+ */
+ if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
+ / sizeof (u64)) {
+ ret = -EINVAL;
+ goto out;
}
- return name + 1;
+ if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
+ goto out;
+
+ size = sizeof (struct ceph_snap_context) +
+ snap_count * sizeof (snapc->snaps[0]);
+ snapc = kmalloc(size, GFP_KERNEL);
+ if (!snapc) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ atomic_set(&snapc->nref, 1);
+ snapc->seq = seq;
+ snapc->num_snaps = snap_count;
+ for (i = 0; i < snap_count; i++)
+ snapc->snaps[i] = ceph_decode_64(&p);
+
+ rbd_dev->header.snapc = snapc;
+
+ dout(" snap context seq = %llu, snap_count = %u\n",
+ (unsigned long long) seq, (unsigned int) snap_count);
+
+out:
+ kfree(reply_buf);
+
+ return 0;
}
-/*
- * compare the old list of snapshots that we have to what's in the header
- * and update it accordingly. Note that the header holds the snapshots
- * in a reverse order (from newest to oldest) and we need to go from
- * older to new so that we don't get a duplicate snap name when
- * doing the process (e.g., removed snapshot and recreated a new
- * one with the same name.
- */
-static int __rbd_init_snaps_header(struct rbd_device *rbd_dev)
+static char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
{
- const char *name, *first_name;
- int i = rbd_dev->header.total_snaps;
- struct rbd_snap *snap, *old_snap = NULL;
- struct list_head *p, *n;
+ size_t size;
+ void *reply_buf;
+ __le64 snap_id;
+ int ret;
+ void *p;
+ void *end;
+ size_t snap_name_len;
+ char *snap_name;
+
+ size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
+ reply_buf = kmalloc(size, GFP_KERNEL);
+ if (!reply_buf)
+ return ERR_PTR(-ENOMEM);
- first_name = rbd_dev->header.snap_names;
- name = first_name + rbd_dev->header.snap_names_len;
+ snap_id = cpu_to_le64(rbd_dev->header.snapc->snaps[which]);
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_snapshot_name",
+ (char *) &snap_id, sizeof (snap_id),
+ reply_buf, size,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
- list_for_each_prev_safe(p, n, &rbd_dev->snaps) {
- u64 cur_id;
+ p = reply_buf;
+ end = (char *) reply_buf + size;
+ snap_name_len = 0;
+ snap_name = ceph_extract_encoded_string(&p, end, &snap_name_len,
+ GFP_KERNEL);
+ if (IS_ERR(snap_name)) {
+ ret = PTR_ERR(snap_name);
+ goto out;
+ } else {
+ dout(" snap_id 0x%016llx snap_name = %s\n",
+ (unsigned long long) le64_to_cpu(snap_id), snap_name);
+ }
+ kfree(reply_buf);
- old_snap = list_entry(p, struct rbd_snap, node);
+ return snap_name;
+out:
+ kfree(reply_buf);
- if (i)
- cur_id = rbd_dev->header.snapc->snaps[i - 1];
+ return ERR_PTR(ret);
+}
- if (!i || old_snap->id < cur_id) {
- /*
- * old_snap->id was skipped, thus was
- * removed. If this rbd_dev is mapped to
- * the removed snapshot, record that it no
- * longer exists, to prevent further I/O.
- */
- if (rbd_dev->snap_id == old_snap->id)
- rbd_dev->snap_exists = false;
- __rbd_remove_snap_dev(old_snap);
- continue;
- }
- if (old_snap->id == cur_id) {
- /* we have this snapshot already */
- i--;
- name = rbd_prev_snap_name(name, first_name);
+static char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ __le64 snap_id;
+ u8 order;
+ int ret;
+
+ snap_id = rbd_dev->header.snapc->snaps[which];
+ ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, &order, snap_size);
+ if (ret)
+ return ERR_PTR(ret);
+ ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, snap_features);
+ if (ret)
+ return ERR_PTR(ret);
+
+ return rbd_dev_v2_snap_name(rbd_dev, which);
+}
+
+static char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ if (rbd_dev->image_format == 1)
+ return rbd_dev_v1_snap_info(rbd_dev, which,
+ snap_size, snap_features);
+ if (rbd_dev->image_format == 2)
+ return rbd_dev_v2_snap_info(rbd_dev, which,
+ snap_size, snap_features);
+ return ERR_PTR(-EINVAL);
+}
+
+/*
+ * Scan the rbd device's current snapshot list and compare it to the
+ * newly-received snapshot context. Remove any existing snapshots
+ * not present in the new snapshot context. Add a new snapshot for
+ * any snaphots in the snapshot context not in the current list.
+ * And verify there are no changes to snapshots we already know
+ * about.
+ *
+ * Assumes the snapshots in the snapshot context are sorted by
+ * snapshot id, highest id first. (Snapshots in the rbd_dev's list
+ * are also maintained in that order.)
+ */
+static int rbd_dev_snaps_update(struct rbd_device *rbd_dev)
+{
+ struct ceph_snap_context *snapc = rbd_dev->header.snapc;
+ const u32 snap_count = snapc->num_snaps;
+ struct list_head *head = &rbd_dev->snaps;
+ struct list_head *links = head->next;
+ u32 index = 0;
+
+ dout("%s: snap count is %u\n", __func__, (unsigned int) snap_count);
+ while (index < snap_count || links != head) {
+ u64 snap_id;
+ struct rbd_snap *snap;
+ char *snap_name;
+ u64 snap_size = 0;
+ u64 snap_features = 0;
+
+ snap_id = index < snap_count ? snapc->snaps[index]
+ : CEPH_NOSNAP;
+ snap = links != head ? list_entry(links, struct rbd_snap, node)
+ : NULL;
+ rbd_assert(!snap || snap->id != CEPH_NOSNAP);
+
+ if (snap_id == CEPH_NOSNAP || (snap && snap->id > snap_id)) {
+ struct list_head *next = links->next;
+
+ /* Existing snapshot not in the new snap context */
+
+ if (rbd_dev->mapping.snap_id == snap->id)
+ rbd_dev->mapping.snap_exists = false;
+ __rbd_remove_snap_dev(snap);
+ dout("%ssnap id %llu has been removed\n",
+ rbd_dev->mapping.snap_id == snap->id ?
+ "mapped " : "",
+ (unsigned long long) snap->id);
+
+ /* Done with this list entry; advance */
+
+ links = next;
continue;
}
- for (; i > 0;
- i--, name = rbd_prev_snap_name(name, first_name)) {
- if (!name) {
- WARN_ON(1);
- return -EINVAL;
+
+ snap_name = rbd_dev_snap_info(rbd_dev, index,
+ &snap_size, &snap_features);
+ if (IS_ERR(snap_name))
+ return PTR_ERR(snap_name);
+
+ dout("entry %u: snap_id = %llu\n", (unsigned int) snap_count,
+ (unsigned long long) snap_id);
+ if (!snap || (snap_id != CEPH_NOSNAP && snap->id < snap_id)) {
+ struct rbd_snap *new_snap;
+
+ /* We haven't seen this snapshot before */
+
+ new_snap = __rbd_add_snap_dev(rbd_dev, snap_name,
+ snap_id, snap_size, snap_features);
+ if (IS_ERR(new_snap)) {
+ int err = PTR_ERR(new_snap);
+
+ dout(" failed to add dev, error %d\n", err);
+
+ return err;
}
- cur_id = rbd_dev->header.snapc->snaps[i];
- /* snapshot removal? handle it above */
- if (cur_id >= old_snap->id)
- break;
- /* a new snapshot */
- snap = __rbd_add_snap_dev(rbd_dev, i - 1, name);
- if (IS_ERR(snap))
- return PTR_ERR(snap);
-
- /* note that we add it backward so using n and not p */
- list_add(&snap->node, n);
- p = &snap->node;
+
+ /* New goes before existing, or at end of list */
+
+ dout(" added dev%s\n", snap ? "" : " at end\n");
+ if (snap)
+ list_add_tail(&new_snap->node, &snap->node);
+ else
+ list_add_tail(&new_snap->node, head);
+ } else {
+ /* Already have this one */
+
+ dout(" already present\n");
+
+ rbd_assert(snap->size == snap_size);
+ rbd_assert(!strcmp(snap->name, snap_name));
+ rbd_assert(snap->features == snap_features);
+
+ /* Done with this list entry; advance */
+
+ links = links->next;
}
+
+ /* Advance to the next entry in the snapshot context */
+
+ index++;
}
- /* we're done going over the old snap list, just add what's left */
- for (; i > 0; i--) {
- name = rbd_prev_snap_name(name, first_name);
- if (!name) {
- WARN_ON(1);
- return -EINVAL;
+ dout("%s: done\n", __func__);
+
+ return 0;
+}
+
+/*
+ * Scan the list of snapshots and register the devices for any that
+ * have not already been registered.
+ */
+static int rbd_dev_snaps_register(struct rbd_device *rbd_dev)
+{
+ struct rbd_snap *snap;
+ int ret = 0;
+
+ dout("%s called\n", __func__);
+ if (WARN_ON(!device_is_registered(&rbd_dev->dev)))
+ return -EIO;
+
+ list_for_each_entry(snap, &rbd_dev->snaps, node) {
+ if (!rbd_snap_registered(snap)) {
+ ret = rbd_register_snap_dev(snap, &rbd_dev->dev);
+ if (ret < 0)
+ break;
}
- snap = __rbd_add_snap_dev(rbd_dev, i - 1, name);
- if (IS_ERR(snap))
- return PTR_ERR(snap);
- list_add(&snap->node, &rbd_dev->snaps);
}
+ dout("%s: returning %d\n", __func__, ret);
- return 0;
+ return ret;
}
static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
{
- int ret;
struct device *dev;
- struct rbd_snap *snap;
+ int ret;
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
- dev = &rbd_dev->dev;
+ dev = &rbd_dev->dev;
dev->bus = &rbd_bus_type;
dev->type = &rbd_device_type;
dev->parent = &rbd_root_dev;
dev->release = rbd_dev_release;
dev_set_name(dev, "%d", rbd_dev->dev_id);
ret = device_register(dev);
- if (ret < 0)
- goto out;
- list_for_each_entry(snap, &rbd_dev->snaps, node) {
- ret = rbd_register_snap_dev(snap, &rbd_dev->dev);
- if (ret < 0)
- break;
- }
-out:
mutex_unlock(&ctl_mutex);
+
return ret;
}
@@ -2211,33 +2568,37 @@ static int rbd_init_watch_dev(struct rbd_device *rbd_dev)
return ret;
}
-static atomic64_t rbd_id_max = ATOMIC64_INIT(0);
+static atomic64_t rbd_dev_id_max = ATOMIC64_INIT(0);
/*
* Get a unique rbd identifier for the given new rbd_dev, and add
* the rbd_dev to the global list. The minimum rbd id is 1.
*/
-static void rbd_id_get(struct rbd_device *rbd_dev)
+static void rbd_dev_id_get(struct rbd_device *rbd_dev)
{
- rbd_dev->dev_id = atomic64_inc_return(&rbd_id_max);
+ rbd_dev->dev_id = atomic64_inc_return(&rbd_dev_id_max);
spin_lock(&rbd_dev_list_lock);
list_add_tail(&rbd_dev->node, &rbd_dev_list);
spin_unlock(&rbd_dev_list_lock);
+ dout("rbd_dev %p given dev id %llu\n", rbd_dev,
+ (unsigned long long) rbd_dev->dev_id);
}
/*
* Remove an rbd_dev from the global list, and record that its
* identifier is no longer in use.
*/
-static void rbd_id_put(struct rbd_device *rbd_dev)
+static void rbd_dev_id_put(struct rbd_device *rbd_dev)
{
struct list_head *tmp;
int rbd_id = rbd_dev->dev_id;
int max_id;
- BUG_ON(rbd_id < 1);
+ rbd_assert(rbd_id > 0);
+ dout("rbd_dev %p released dev id %llu\n", rbd_dev,
+ (unsigned long long) rbd_dev->dev_id);
spin_lock(&rbd_dev_list_lock);
list_del_init(&rbd_dev->node);
@@ -2245,7 +2606,7 @@ static void rbd_id_put(struct rbd_device *rbd_dev)
* If the id being "put" is not the current maximum, there
* is nothing special we need to do.
*/
- if (rbd_id != atomic64_read(&rbd_id_max)) {
+ if (rbd_id != atomic64_read(&rbd_dev_id_max)) {
spin_unlock(&rbd_dev_list_lock);
return;
}
@@ -2266,12 +2627,13 @@ static void rbd_id_put(struct rbd_device *rbd_dev)
spin_unlock(&rbd_dev_list_lock);
/*
- * The max id could have been updated by rbd_id_get(), in
+ * The max id could have been updated by rbd_dev_id_get(), in
* which case it now accurately reflects the new maximum.
* Be careful not to overwrite the maximum value in that
* case.
*/
- atomic64_cmpxchg(&rbd_id_max, rbd_id, max_id);
+ atomic64_cmpxchg(&rbd_dev_id_max, rbd_id, max_id);
+ dout(" max dev id has been reset\n");
}
/*
@@ -2360,28 +2722,31 @@ static inline char *dup_token(const char **buf, size_t *lenp)
}
/*
- * This fills in the pool_name, image_name, image_name_len, snap_name,
- * rbd_dev, rbd_md_name, and name fields of the given rbd_dev, based
- * on the list of monitor addresses and other options provided via
- * /sys/bus/rbd/add.
+ * This fills in the pool_name, image_name, image_name_len, rbd_dev,
+ * rbd_md_name, and name fields of the given rbd_dev, based on the
+ * list of monitor addresses and other options provided via
+ * /sys/bus/rbd/add. Returns a pointer to a dynamically-allocated
+ * copy of the snapshot name to map if successful, or a
+ * pointer-coded error otherwise.
*
* Note: rbd_dev is assumed to have been initially zero-filled.
*/
-static int rbd_add_parse_args(struct rbd_device *rbd_dev,
- const char *buf,
- const char **mon_addrs,
- size_t *mon_addrs_size,
- char *options,
- size_t options_size)
+static char *rbd_add_parse_args(struct rbd_device *rbd_dev,
+ const char *buf,
+ const char **mon_addrs,
+ size_t *mon_addrs_size,
+ char *options,
+ size_t options_size)
{
size_t len;
- int ret;
+ char *err_ptr = ERR_PTR(-EINVAL);
+ char *snap_name;
/* The first four tokens are required */
len = next_token(&buf);
if (!len)
- return -EINVAL;
+ return err_ptr;
*mon_addrs_size = len + 1;
*mon_addrs = buf;
@@ -2389,9 +2754,9 @@ static int rbd_add_parse_args(struct rbd_device *rbd_dev,
len = copy_token(&buf, options, options_size);
if (!len || len >= options_size)
- return -EINVAL;
+ return err_ptr;
- ret = -ENOMEM;
+ err_ptr = ERR_PTR(-ENOMEM);
rbd_dev->pool_name = dup_token(&buf, NULL);
if (!rbd_dev->pool_name)
goto out_err;
@@ -2400,41 +2765,227 @@ static int rbd_add_parse_args(struct rbd_device *rbd_dev,
if (!rbd_dev->image_name)
goto out_err;
- /* Create the name of the header object */
+ /* Snapshot name is optional */
+ len = next_token(&buf);
+ if (!len) {
+ buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */
+ len = sizeof (RBD_SNAP_HEAD_NAME) - 1;
+ }
+ snap_name = kmalloc(len + 1, GFP_KERNEL);
+ if (!snap_name)
+ goto out_err;
+ memcpy(snap_name, buf, len);
+ *(snap_name + len) = '\0';
- rbd_dev->header_name = kmalloc(rbd_dev->image_name_len
- + sizeof (RBD_SUFFIX),
- GFP_KERNEL);
- if (!rbd_dev->header_name)
+dout(" SNAP_NAME is <%s>, len is %zd\n", snap_name, len);
+
+ return snap_name;
+
+out_err:
+ kfree(rbd_dev->image_name);
+ rbd_dev->image_name = NULL;
+ rbd_dev->image_name_len = 0;
+ kfree(rbd_dev->pool_name);
+ rbd_dev->pool_name = NULL;
+
+ return err_ptr;
+}
+
+/*
+ * An rbd format 2 image has a unique identifier, distinct from the
+ * name given to it by the user. Internally, that identifier is
+ * what's used to specify the names of objects related to the image.
+ *
+ * A special "rbd id" object is used to map an rbd image name to its
+ * id. If that object doesn't exist, then there is no v2 rbd image
+ * with the supplied name.
+ *
+ * This function will record the given rbd_dev's image_id field if
+ * it can be determined, and in that case will return 0. If any
+ * errors occur a negative errno will be returned and the rbd_dev's
+ * image_id field will be unchanged (and should be NULL).
+ */
+static int rbd_dev_image_id(struct rbd_device *rbd_dev)
+{
+ int ret;
+ size_t size;
+ char *object_name;
+ void *response;
+ void *p;
+
+ /*
+ * First, see if the format 2 image id file exists, and if
+ * so, get the image's persistent id from it.
+ */
+ size = sizeof (RBD_ID_PREFIX) + rbd_dev->image_name_len;
+ object_name = kmalloc(size, GFP_NOIO);
+ if (!object_name)
+ return -ENOMEM;
+ sprintf(object_name, "%s%s", RBD_ID_PREFIX, rbd_dev->image_name);
+ dout("rbd id object name is %s\n", object_name);
+
+ /* Response will be an encoded string, which includes a length */
+
+ size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
+ response = kzalloc(size, GFP_NOIO);
+ if (!response) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ ret = rbd_req_sync_exec(rbd_dev, object_name,
+ "rbd", "get_id",
+ NULL, 0,
+ response, RBD_IMAGE_ID_LEN_MAX,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ p = response;
+ rbd_dev->image_id = ceph_extract_encoded_string(&p,
+ p + RBD_IMAGE_ID_LEN_MAX,
+ &rbd_dev->image_id_len,
+ GFP_NOIO);
+ if (IS_ERR(rbd_dev->image_id)) {
+ ret = PTR_ERR(rbd_dev->image_id);
+ rbd_dev->image_id = NULL;
+ } else {
+ dout("image_id is %s\n", rbd_dev->image_id);
+ }
+out:
+ kfree(response);
+ kfree(object_name);
+
+ return ret;
+}
+
+static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
+{
+ int ret;
+ size_t size;
+
+ /* Version 1 images have no id; empty string is used */
+
+ rbd_dev->image_id = kstrdup("", GFP_KERNEL);
+ if (!rbd_dev->image_id)
+ return -ENOMEM;
+ rbd_dev->image_id_len = 0;
+
+ /* Record the header object name for this rbd image. */
+
+ size = rbd_dev->image_name_len + sizeof (RBD_SUFFIX);
+ rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
+ if (!rbd_dev->header_name) {
+ ret = -ENOMEM;
goto out_err;
+ }
sprintf(rbd_dev->header_name, "%s%s", rbd_dev->image_name, RBD_SUFFIX);
+ /* Populate rbd image metadata */
+
+ ret = rbd_read_header(rbd_dev, &rbd_dev->header);
+ if (ret < 0)
+ goto out_err;
+ rbd_dev->image_format = 1;
+
+ dout("discovered version 1 image, header name is %s\n",
+ rbd_dev->header_name);
+
+ return 0;
+
+out_err:
+ kfree(rbd_dev->header_name);
+ rbd_dev->header_name = NULL;
+ kfree(rbd_dev->image_id);
+ rbd_dev->image_id = NULL;
+
+ return ret;
+}
+
+static int rbd_dev_v2_probe(struct rbd_device *rbd_dev)
+{
+ size_t size;
+ int ret;
+ u64 ver = 0;
+
/*
- * The snapshot name is optional. If none is is supplied,
- * we use the default value.
+ * Image id was filled in by the caller. Record the header
+ * object name for this rbd image.
*/
- rbd_dev->snap_name = dup_token(&buf, &len);
- if (!rbd_dev->snap_name)
+ size = sizeof (RBD_HEADER_PREFIX) + rbd_dev->image_id_len;
+ rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
+ if (!rbd_dev->header_name)
+ return -ENOMEM;
+ sprintf(rbd_dev->header_name, "%s%s",
+ RBD_HEADER_PREFIX, rbd_dev->image_id);
+
+ /* Get the size and object order for the image */
+
+ ret = rbd_dev_v2_image_size(rbd_dev);
+ if (ret < 0)
goto out_err;
- if (!len) {
- /* Replace the empty name with the default */
- kfree(rbd_dev->snap_name);
- rbd_dev->snap_name
- = kmalloc(sizeof (RBD_SNAP_HEAD_NAME), GFP_KERNEL);
- if (!rbd_dev->snap_name)
- goto out_err;
- memcpy(rbd_dev->snap_name, RBD_SNAP_HEAD_NAME,
- sizeof (RBD_SNAP_HEAD_NAME));
- }
+ /* Get the object prefix (a.k.a. block_name) for the image */
- return 0;
+ ret = rbd_dev_v2_object_prefix(rbd_dev);
+ if (ret < 0)
+ goto out_err;
+
+ /* Get the features for the image */
+ ret = rbd_dev_v2_features(rbd_dev);
+ if (ret < 0)
+ goto out_err;
+
+ /* crypto and compression type aren't (yet) supported for v2 images */
+
+ rbd_dev->header.crypt_type = 0;
+ rbd_dev->header.comp_type = 0;
+
+ /* Get the snapshot context, plus the header version */
+
+ ret = rbd_dev_v2_snap_context(rbd_dev, &ver);
+ if (ret)
+ goto out_err;
+ rbd_dev->header.obj_version = ver;
+
+ rbd_dev->image_format = 2;
+
+ dout("discovered version 2 image, header name is %s\n",
+ rbd_dev->header_name);
+
+ return -ENOTSUPP;
out_err:
kfree(rbd_dev->header_name);
- kfree(rbd_dev->image_name);
- kfree(rbd_dev->pool_name);
- rbd_dev->pool_name = NULL;
+ rbd_dev->header_name = NULL;
+ kfree(rbd_dev->header.object_prefix);
+ rbd_dev->header.object_prefix = NULL;
+
+ return ret;
+}
+
+/*
+ * Probe for the existence of the header object for the given rbd
+ * device. For format 2 images this includes determining the image
+ * id.
+ */
+static int rbd_dev_probe(struct rbd_device *rbd_dev)
+{
+ int ret;
+
+ /*
+ * Get the id from the image id object. If it's not a
+ * format 2 image, we'll get ENOENT back, and we'll assume
+ * it's a format 1 image.
+ */
+ ret = rbd_dev_image_id(rbd_dev);
+ if (ret)
+ ret = rbd_dev_v1_probe(rbd_dev);
+ else
+ ret = rbd_dev_v2_probe(rbd_dev);
+ if (ret)
+ dout("probe failed, returning %d\n", ret);
return ret;
}
@@ -2449,16 +3000,17 @@ static ssize_t rbd_add(struct bus_type *bus,
size_t mon_addrs_size = 0;
struct ceph_osd_client *osdc;
int rc = -ENOMEM;
+ char *snap_name;
if (!try_module_get(THIS_MODULE))
return -ENODEV;
options = kmalloc(count, GFP_KERNEL);
if (!options)
- goto err_nomem;
+ goto err_out_mem;
rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL);
if (!rbd_dev)
- goto err_nomem;
+ goto err_out_mem;
/* static rbd_device initialization */
spin_lock_init(&rbd_dev->lock);
@@ -2466,27 +3018,18 @@ static ssize_t rbd_add(struct bus_type *bus,
INIT_LIST_HEAD(&rbd_dev->snaps);
init_rwsem(&rbd_dev->header_rwsem);
- /* generate unique id: find highest unique id, add one */
- rbd_id_get(rbd_dev);
-
- /* Fill in the device name, now that we have its id. */
- BUILD_BUG_ON(DEV_NAME_LEN
- < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
- sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->dev_id);
-
/* parse add command */
- rc = rbd_add_parse_args(rbd_dev, buf, &mon_addrs, &mon_addrs_size,
- options, count);
- if (rc)
- goto err_put_id;
-
- rbd_dev->rbd_client = rbd_get_client(mon_addrs, mon_addrs_size - 1,
- options);
- if (IS_ERR(rbd_dev->rbd_client)) {
- rc = PTR_ERR(rbd_dev->rbd_client);
- goto err_put_id;
+ snap_name = rbd_add_parse_args(rbd_dev, buf,
+ &mon_addrs, &mon_addrs_size, options, count);
+ if (IS_ERR(snap_name)) {
+ rc = PTR_ERR(snap_name);
+ goto err_out_mem;
}
+ rc = rbd_get_client(rbd_dev, mon_addrs, mon_addrs_size - 1, options);
+ if (rc < 0)
+ goto err_out_args;
+
/* pick the pool */
osdc = &rbd_dev->rbd_client->client->osdc;
rc = ceph_pg_poolid_by_name(osdc->osdmap, rbd_dev->pool_name);
@@ -2494,23 +3037,53 @@ static ssize_t rbd_add(struct bus_type *bus,
goto err_out_client;
rbd_dev->pool_id = rc;
- /* register our block device */
- rc = register_blkdev(0, rbd_dev->name);
+ rc = rbd_dev_probe(rbd_dev);
if (rc < 0)
goto err_out_client;
+ rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+
+ /* no need to lock here, as rbd_dev is not registered yet */
+ rc = rbd_dev_snaps_update(rbd_dev);
+ if (rc)
+ goto err_out_header;
+
+ rc = rbd_dev_set_mapping(rbd_dev, snap_name);
+ if (rc)
+ goto err_out_header;
+
+ /* generate unique id: find highest unique id, add one */
+ rbd_dev_id_get(rbd_dev);
+
+ /* Fill in the device name, now that we have its id. */
+ BUILD_BUG_ON(DEV_NAME_LEN
+ < sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
+ sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->dev_id);
+
+ /* Get our block major device number. */
+
+ rc = register_blkdev(0, rbd_dev->name);
+ if (rc < 0)
+ goto err_out_id;
rbd_dev->major = rc;
- rc = rbd_bus_add_dev(rbd_dev);
+ /* Set up the blkdev mapping. */
+
+ rc = rbd_init_disk(rbd_dev);
if (rc)
goto err_out_blkdev;
+ rc = rbd_bus_add_dev(rbd_dev);
+ if (rc)
+ goto err_out_disk;
+
/*
* At this point cleanup in the event of an error is the job
* of the sysfs code (initiated by rbd_bus_del_dev()).
- *
- * Set up and announce blkdev mapping.
*/
- rc = rbd_init_disk(rbd_dev);
+
+ down_write(&rbd_dev->header_rwsem);
+ rc = rbd_dev_snaps_register(rbd_dev);
+ up_write(&rbd_dev->header_rwsem);
if (rc)
goto err_out_bus;
@@ -2518,6 +3091,13 @@ static ssize_t rbd_add(struct bus_type *bus,
if (rc)
goto err_out_bus;
+ /* Everything's ready. Announce the disk to the world. */
+
+ add_disk(rbd_dev->disk);
+
+ pr_info("%s: added with size 0x%llx\n", rbd_dev->disk->disk_name,
+ (unsigned long long) rbd_dev->mapping.size);
+
return count;
err_out_bus:
@@ -2527,19 +3107,23 @@ err_out_bus:
kfree(options);
return rc;
+err_out_disk:
+ rbd_free_disk(rbd_dev);
err_out_blkdev:
unregister_blkdev(rbd_dev->major, rbd_dev->name);
+err_out_id:
+ rbd_dev_id_put(rbd_dev);
+err_out_header:
+ rbd_header_free(&rbd_dev->header);
err_out_client:
+ kfree(rbd_dev->header_name);
rbd_put_client(rbd_dev);
-err_put_id:
- if (rbd_dev->pool_name) {
- kfree(rbd_dev->snap_name);
- kfree(rbd_dev->header_name);
- kfree(rbd_dev->image_name);
- kfree(rbd_dev->pool_name);
- }
- rbd_id_put(rbd_dev);
-err_nomem:
+ kfree(rbd_dev->image_id);
+err_out_args:
+ kfree(rbd_dev->mapping.snap_name);
+ kfree(rbd_dev->image_name);
+ kfree(rbd_dev->pool_name);
+err_out_mem:
kfree(rbd_dev);
kfree(options);
@@ -2585,12 +3169,16 @@ static void rbd_dev_release(struct device *dev)
rbd_free_disk(rbd_dev);
unregister_blkdev(rbd_dev->major, rbd_dev->name);
+ /* release allocated disk header fields */
+ rbd_header_free(&rbd_dev->header);
+
/* done with the id, and with the rbd_dev */
- kfree(rbd_dev->snap_name);
+ kfree(rbd_dev->mapping.snap_name);
+ kfree(rbd_dev->image_id);
kfree(rbd_dev->header_name);
kfree(rbd_dev->pool_name);
kfree(rbd_dev->image_name);
- rbd_id_put(rbd_dev);
+ rbd_dev_id_put(rbd_dev);
kfree(rbd_dev);
/* release module ref */
@@ -2628,47 +3216,7 @@ static ssize_t rbd_remove(struct bus_type *bus,
done:
mutex_unlock(&ctl_mutex);
- return ret;
-}
-static ssize_t rbd_snap_add(struct device *dev,
- struct device_attribute *attr,
- const char *buf,
- size_t count)
-{
- struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- int ret;
- char *name = kmalloc(count + 1, GFP_KERNEL);
- if (!name)
- return -ENOMEM;
-
- snprintf(name, count, "%s", buf);
-
- mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
-
- ret = rbd_header_add_snap(rbd_dev,
- name, GFP_KERNEL);
- if (ret < 0)
- goto err_unlock;
-
- ret = __rbd_refresh_header(rbd_dev, NULL);
- if (ret < 0)
- goto err_unlock;
-
- /* shouldn't hold ctl_mutex when notifying.. notify might
- trigger a watch callback that would need to get that mutex */
- mutex_unlock(&ctl_mutex);
-
- /* make a best effort, don't error if failed */
- rbd_req_sync_notify(rbd_dev);
-
- ret = count;
- kfree(name);
- return ret;
-
-err_unlock:
- mutex_unlock(&ctl_mutex);
- kfree(name);
return ret;
}
diff --git a/drivers/block/rbd_types.h b/drivers/block/rbd_types.h
index 0924e9e41a6..cbe77fa105b 100644
--- a/drivers/block/rbd_types.h
+++ b/drivers/block/rbd_types.h
@@ -15,15 +15,30 @@
#include <linux/types.h>
+/* For format version 2, rbd image 'foo' consists of objects
+ * rbd_id.foo - id of image
+ * rbd_header.<id> - image metadata
+ * rbd_data.<id>.0000000000000000
+ * rbd_data.<id>.0000000000000001
+ * ... - data
+ * Clients do not access header data directly in rbd format 2.
+ */
+
+#define RBD_HEADER_PREFIX "rbd_header."
+#define RBD_DATA_PREFIX "rbd_data."
+#define RBD_ID_PREFIX "rbd_id."
+
/*
- * rbd image 'foo' consists of objects
- * foo.rbd - image metadata
- * foo.00000000
- * foo.00000001
- * ... - data
+ * For format version 1, rbd image 'foo' consists of objects
+ * foo.rbd - image metadata
+ * rb.<idhi>.<idlo>.00000000
+ * rb.<idhi>.<idlo>.00000001
+ * ... - data
+ * There is no notion of a persistent image id in rbd format 1.
*/
#define RBD_SUFFIX ".rbd"
+
#define RBD_DIRECTORY "rbd_directory"
#define RBD_INFO "rbd_info"
@@ -47,7 +62,7 @@ struct rbd_image_snap_ondisk {
struct rbd_image_header_ondisk {
char text[40];
- char block_name[24];
+ char object_prefix[24];
char signature[4];
char version[8];
struct {
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index c0bbeb47075..0bdde8fba39 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -14,6 +14,9 @@
#define PART_BITS 4
+static bool use_bio;
+module_param(use_bio, bool, S_IRUGO);
+
static int major;
static DEFINE_IDA(vd_index_ida);
@@ -23,6 +26,7 @@ struct virtio_blk
{
struct virtio_device *vdev;
struct virtqueue *vq;
+ wait_queue_head_t queue_wait;
/* The disk structure for the kernel. */
struct gendisk *disk;
@@ -51,53 +55,244 @@ struct virtio_blk
struct virtblk_req
{
struct request *req;
+ struct bio *bio;
struct virtio_blk_outhdr out_hdr;
struct virtio_scsi_inhdr in_hdr;
+ struct work_struct work;
+ struct virtio_blk *vblk;
+ int flags;
u8 status;
+ struct scatterlist sg[];
+};
+
+enum {
+ VBLK_IS_FLUSH = 1,
+ VBLK_REQ_FLUSH = 2,
+ VBLK_REQ_DATA = 4,
+ VBLK_REQ_FUA = 8,
};
-static void blk_done(struct virtqueue *vq)
+static inline int virtblk_result(struct virtblk_req *vbr)
+{
+ switch (vbr->status) {
+ case VIRTIO_BLK_S_OK:
+ return 0;
+ case VIRTIO_BLK_S_UNSUPP:
+ return -ENOTTY;
+ default:
+ return -EIO;
+ }
+}
+
+static inline struct virtblk_req *virtblk_alloc_req(struct virtio_blk *vblk,
+ gfp_t gfp_mask)
{
- struct virtio_blk *vblk = vq->vdev->priv;
struct virtblk_req *vbr;
- unsigned int len;
- unsigned long flags;
- spin_lock_irqsave(vblk->disk->queue->queue_lock, flags);
- while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
- int error;
+ vbr = mempool_alloc(vblk->pool, gfp_mask);
+ if (!vbr)
+ return NULL;
- switch (vbr->status) {
- case VIRTIO_BLK_S_OK:
- error = 0;
- break;
- case VIRTIO_BLK_S_UNSUPP:
- error = -ENOTTY;
- break;
- default:
- error = -EIO;
+ vbr->vblk = vblk;
+ if (use_bio)
+ sg_init_table(vbr->sg, vblk->sg_elems);
+
+ return vbr;
+}
+
+static void virtblk_add_buf_wait(struct virtio_blk *vblk,
+ struct virtblk_req *vbr,
+ unsigned long out,
+ unsigned long in)
+{
+ DEFINE_WAIT(wait);
+
+ for (;;) {
+ prepare_to_wait_exclusive(&vblk->queue_wait, &wait,
+ TASK_UNINTERRUPTIBLE);
+
+ spin_lock_irq(vblk->disk->queue->queue_lock);
+ if (virtqueue_add_buf(vblk->vq, vbr->sg, out, in, vbr,
+ GFP_ATOMIC) < 0) {
+ spin_unlock_irq(vblk->disk->queue->queue_lock);
+ io_schedule();
+ } else {
+ virtqueue_kick(vblk->vq);
+ spin_unlock_irq(vblk->disk->queue->queue_lock);
break;
}
- switch (vbr->req->cmd_type) {
- case REQ_TYPE_BLOCK_PC:
- vbr->req->resid_len = vbr->in_hdr.residual;
- vbr->req->sense_len = vbr->in_hdr.sense_len;
- vbr->req->errors = vbr->in_hdr.errors;
- break;
- case REQ_TYPE_SPECIAL:
- vbr->req->errors = (error != 0);
- break;
- default:
- break;
+ }
+
+ finish_wait(&vblk->queue_wait, &wait);
+}
+
+static inline void virtblk_add_req(struct virtblk_req *vbr,
+ unsigned int out, unsigned int in)
+{
+ struct virtio_blk *vblk = vbr->vblk;
+
+ spin_lock_irq(vblk->disk->queue->queue_lock);
+ if (unlikely(virtqueue_add_buf(vblk->vq, vbr->sg, out, in, vbr,
+ GFP_ATOMIC) < 0)) {
+ spin_unlock_irq(vblk->disk->queue->queue_lock);
+ virtblk_add_buf_wait(vblk, vbr, out, in);
+ return;
+ }
+ virtqueue_kick(vblk->vq);
+ spin_unlock_irq(vblk->disk->queue->queue_lock);
+}
+
+static int virtblk_bio_send_flush(struct virtblk_req *vbr)
+{
+ unsigned int out = 0, in = 0;
+
+ vbr->flags |= VBLK_IS_FLUSH;
+ vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
+ vbr->out_hdr.sector = 0;
+ vbr->out_hdr.ioprio = 0;
+ sg_set_buf(&vbr->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
+ sg_set_buf(&vbr->sg[out + in++], &vbr->status, sizeof(vbr->status));
+
+ virtblk_add_req(vbr, out, in);
+
+ return 0;
+}
+
+static int virtblk_bio_send_data(struct virtblk_req *vbr)
+{
+ struct virtio_blk *vblk = vbr->vblk;
+ unsigned int num, out = 0, in = 0;
+ struct bio *bio = vbr->bio;
+
+ vbr->flags &= ~VBLK_IS_FLUSH;
+ vbr->out_hdr.type = 0;
+ vbr->out_hdr.sector = bio->bi_sector;
+ vbr->out_hdr.ioprio = bio_prio(bio);
+
+ sg_set_buf(&vbr->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
+
+ num = blk_bio_map_sg(vblk->disk->queue, bio, vbr->sg + out);
+
+ sg_set_buf(&vbr->sg[num + out + in++], &vbr->status,
+ sizeof(vbr->status));
+
+ if (num) {
+ if (bio->bi_rw & REQ_WRITE) {
+ vbr->out_hdr.type |= VIRTIO_BLK_T_OUT;
+ out += num;
+ } else {
+ vbr->out_hdr.type |= VIRTIO_BLK_T_IN;
+ in += num;
}
+ }
+
+ virtblk_add_req(vbr, out, in);
+
+ return 0;
+}
+
+static void virtblk_bio_send_data_work(struct work_struct *work)
+{
+ struct virtblk_req *vbr;
+
+ vbr = container_of(work, struct virtblk_req, work);
+
+ virtblk_bio_send_data(vbr);
+}
+
+static void virtblk_bio_send_flush_work(struct work_struct *work)
+{
+ struct virtblk_req *vbr;
+
+ vbr = container_of(work, struct virtblk_req, work);
+
+ virtblk_bio_send_flush(vbr);
+}
+
+static inline void virtblk_request_done(struct virtblk_req *vbr)
+{
+ struct virtio_blk *vblk = vbr->vblk;
+ struct request *req = vbr->req;
+ int error = virtblk_result(vbr);
+
+ if (req->cmd_type == REQ_TYPE_BLOCK_PC) {
+ req->resid_len = vbr->in_hdr.residual;
+ req->sense_len = vbr->in_hdr.sense_len;
+ req->errors = vbr->in_hdr.errors;
+ } else if (req->cmd_type == REQ_TYPE_SPECIAL) {
+ req->errors = (error != 0);
+ }
+
+ __blk_end_request_all(req, error);
+ mempool_free(vbr, vblk->pool);
+}
+
+static inline void virtblk_bio_flush_done(struct virtblk_req *vbr)
+{
+ struct virtio_blk *vblk = vbr->vblk;
+
+ if (vbr->flags & VBLK_REQ_DATA) {
+ /* Send out the actual write data */
+ INIT_WORK(&vbr->work, virtblk_bio_send_data_work);
+ queue_work(virtblk_wq, &vbr->work);
+ } else {
+ bio_endio(vbr->bio, virtblk_result(vbr));
+ mempool_free(vbr, vblk->pool);
+ }
+}
+
+static inline void virtblk_bio_data_done(struct virtblk_req *vbr)
+{
+ struct virtio_blk *vblk = vbr->vblk;
- __blk_end_request_all(vbr->req, error);
+ if (unlikely(vbr->flags & VBLK_REQ_FUA)) {
+ /* Send out a flush before end the bio */
+ vbr->flags &= ~VBLK_REQ_DATA;
+ INIT_WORK(&vbr->work, virtblk_bio_send_flush_work);
+ queue_work(virtblk_wq, &vbr->work);
+ } else {
+ bio_endio(vbr->bio, virtblk_result(vbr));
mempool_free(vbr, vblk->pool);
}
+}
+
+static inline void virtblk_bio_done(struct virtblk_req *vbr)
+{
+ if (unlikely(vbr->flags & VBLK_IS_FLUSH))
+ virtblk_bio_flush_done(vbr);
+ else
+ virtblk_bio_data_done(vbr);
+}
+
+static void virtblk_done(struct virtqueue *vq)
+{
+ struct virtio_blk *vblk = vq->vdev->priv;
+ bool bio_done = false, req_done = false;
+ struct virtblk_req *vbr;
+ unsigned long flags;
+ unsigned int len;
+
+ spin_lock_irqsave(vblk->disk->queue->queue_lock, flags);
+ do {
+ virtqueue_disable_cb(vq);
+ while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
+ if (vbr->bio) {
+ virtblk_bio_done(vbr);
+ bio_done = true;
+ } else {
+ virtblk_request_done(vbr);
+ req_done = true;
+ }
+ }
+ } while (!virtqueue_enable_cb(vq));
/* In case queue is stopped waiting for more buffers. */
- blk_start_queue(vblk->disk->queue);
+ if (req_done)
+ blk_start_queue(vblk->disk->queue);
spin_unlock_irqrestore(vblk->disk->queue->queue_lock, flags);
+
+ if (bio_done)
+ wake_up(&vblk->queue_wait);
}
static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
@@ -106,13 +301,13 @@ static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
unsigned long num, out = 0, in = 0;
struct virtblk_req *vbr;
- vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
+ vbr = virtblk_alloc_req(vblk, GFP_ATOMIC);
if (!vbr)
/* When another request finishes we'll try again. */
return false;
vbr->req = req;
-
+ vbr->bio = NULL;
if (req->cmd_flags & REQ_FLUSH) {
vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
vbr->out_hdr.sector = 0;
@@ -172,7 +367,8 @@ static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
}
}
- if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr, GFP_ATOMIC)<0) {
+ if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr,
+ GFP_ATOMIC) < 0) {
mempool_free(vbr, vblk->pool);
return false;
}
@@ -180,7 +376,7 @@ static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
return true;
}
-static void do_virtblk_request(struct request_queue *q)
+static void virtblk_request(struct request_queue *q)
{
struct virtio_blk *vblk = q->queuedata;
struct request *req;
@@ -203,6 +399,34 @@ static void do_virtblk_request(struct request_queue *q)
virtqueue_kick(vblk->vq);
}
+static void virtblk_make_request(struct request_queue *q, struct bio *bio)
+{
+ struct virtio_blk *vblk = q->queuedata;
+ struct virtblk_req *vbr;
+
+ BUG_ON(bio->bi_phys_segments + 2 > vblk->sg_elems);
+
+ vbr = virtblk_alloc_req(vblk, GFP_NOIO);
+ if (!vbr) {
+ bio_endio(bio, -ENOMEM);
+ return;
+ }
+
+ vbr->bio = bio;
+ vbr->flags = 0;
+ if (bio->bi_rw & REQ_FLUSH)
+ vbr->flags |= VBLK_REQ_FLUSH;
+ if (bio->bi_rw & REQ_FUA)
+ vbr->flags |= VBLK_REQ_FUA;
+ if (bio->bi_size)
+ vbr->flags |= VBLK_REQ_DATA;
+
+ if (unlikely(vbr->flags & VBLK_REQ_FLUSH))
+ virtblk_bio_send_flush(vbr);
+ else
+ virtblk_bio_send_data(vbr);
+}
+
/* return id (s/n) string for *disk to *id_str
*/
static int virtblk_get_id(struct gendisk *disk, char *id_str)
@@ -360,7 +584,7 @@ static int init_vq(struct virtio_blk *vblk)
int err = 0;
/* We expect one virtqueue, for output. */
- vblk->vq = virtio_find_single_vq(vblk->vdev, blk_done, "requests");
+ vblk->vq = virtio_find_single_vq(vblk->vdev, virtblk_done, "requests");
if (IS_ERR(vblk->vq))
err = PTR_ERR(vblk->vq);
@@ -477,6 +701,8 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
struct virtio_blk *vblk;
struct request_queue *q;
int err, index;
+ int pool_size;
+
u64 cap;
u32 v, blk_size, sg_elems, opt_io_size;
u16 min_io_size;
@@ -506,10 +732,12 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
goto out_free_index;
}
+ init_waitqueue_head(&vblk->queue_wait);
vblk->vdev = vdev;
vblk->sg_elems = sg_elems;
sg_init_table(vblk->sg, vblk->sg_elems);
mutex_init(&vblk->config_lock);
+
INIT_WORK(&vblk->config_work, virtblk_config_changed_work);
vblk->config_enable = true;
@@ -517,7 +745,10 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
if (err)
goto out_free_vblk;
- vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
+ pool_size = sizeof(struct virtblk_req);
+ if (use_bio)
+ pool_size += sizeof(struct scatterlist) * sg_elems;
+ vblk->pool = mempool_create_kmalloc_pool(1, pool_size);
if (!vblk->pool) {
err = -ENOMEM;
goto out_free_vq;
@@ -530,12 +761,14 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
goto out_mempool;
}
- q = vblk->disk->queue = blk_init_queue(do_virtblk_request, NULL);
+ q = vblk->disk->queue = blk_init_queue(virtblk_request, NULL);
if (!q) {
err = -ENOMEM;
goto out_put_disk;
}
+ if (use_bio)
+ blk_queue_make_request(q, virtblk_make_request);
q->queuedata = vblk;
virtblk_name_format("vd", index, vblk->disk->disk_name, DISK_NAME_LEN);
@@ -620,7 +853,6 @@ static int __devinit virtblk_probe(struct virtio_device *vdev)
if (!err && opt_io_size)
blk_queue_io_opt(q, blk_size * opt_io_size);
-
add_disk(vblk->disk);
err = device_create_file(disk_to_dev(vblk->disk), &dev_attr_serial);
if (err)
diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c
index c6decb901e5..280a13846e6 100644
--- a/drivers/block/xen-blkback/blkback.c
+++ b/drivers/block/xen-blkback/blkback.c
@@ -42,6 +42,7 @@
#include <xen/events.h>
#include <xen/page.h>
+#include <xen/xen.h>
#include <asm/xen/hypervisor.h>
#include <asm/xen/hypercall.h>
#include "common.h"