From 1604f488ac2dcce33c8218e75a000e8c5fb57e61 Mon Sep 17 00:00:00 2001
From: Jim Schutt <jaschut@sandia.gov>
Date: Fri, 30 Nov 2012 09:15:25 -0700
Subject: libceph: for chooseleaf rules, retry CRUSH map descent from root if
 leaf is failed

Add libceph support for a new CRUSH tunable recently added to Ceph servers.

Consider the CRUSH rule
  step chooseleaf firstn 0 type <node_type>

This rule means that <n> replicas will be chosen in a manner such that
each chosen leaf's branch will contain a unique instance of <node_type>.

When an object is re-replicated after a leaf failure, if the CRUSH map uses
a chooseleaf rule the remapped replica ends up under the <node_type> bucket
that held the failed leaf.  This causes uneven data distribution across the
storage cluster, to the point that when all the leaves but one fail under a
particular <node_type> bucket, that remaining leaf holds all the data from
its failed peers.

This behavior also limits the number of peers that can participate in the
re-replication of the data held by the failed leaf, which increases the
time required to re-replicate after a failure.

For a chooseleaf CRUSH rule, the tree descent has two steps: call them the
inner and outer descents.

If the tree descent down to <node_type> is the outer descent, and the descent
from <node_type> down to a leaf is the inner descent, the issue is that a
down leaf is detected on the inner descent, so only the inner descent is
retried.

In order to disperse re-replicated data as widely as possible across a
storage cluster after a failure, we want to retry the outer descent. So,
fix up crush_choose() to allow the inner descent to return immediately on
choosing a failed leaf.  Wire this up as a new CRUSH tunable.

Note that after this change, for a chooseleaf rule, if the primary OSD
in a placement group has failed, choosing a replacement may result in
one of the other OSDs in the PG colliding with the new primary.  This
requires that OSD's data for that PG to need moving as well.  This
seems unavoidable but should be relatively rare.

This corresponds to ceph.git commit 88f218181a9e6d2292e2697fc93797d0f6d6e5dc.

Signed-off-by: Jim Schutt <jaschut@sandia.gov>
Reviewed-by: Sage Weil <sage@inktank.com>
---
 net/ceph/osdmap.c | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'net/ceph/osdmap.c')

diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index de73214b5d26..ca05871635bc 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -170,6 +170,7 @@ static struct crush_map *crush_decode(void *pbyval, void *end)
         c->choose_local_tries = 2;
         c->choose_local_fallback_tries = 5;
         c->choose_total_tries = 19;
+	c->chooseleaf_descend_once = 0;
 
 	ceph_decode_need(p, end, 4*sizeof(u32), bad);
 	magic = ceph_decode_32(p);
@@ -336,6 +337,11 @@ static struct crush_map *crush_decode(void *pbyval, void *end)
         dout("crush decode tunable choose_total_tries = %d",
              c->choose_total_tries);
 
+	ceph_decode_need(p, end, sizeof(u32), done);
+	c->chooseleaf_descend_once = ceph_decode_32(p);
+	dout("crush decode tunable chooseleaf_descend_once = %d",
+	     c->chooseleaf_descend_once);
+
 done:
 	dout("crush_decode success\n");
 	return c;
-- 
cgit v1.2.3


From e8afad656cbcd06d02a7bacd4b318fa0e2907de0 Mon Sep 17 00:00:00 2001
From: Alex Elder <elder@inktank.com>
Date: Wed, 14 Nov 2012 09:38:19 -0600
Subject: libceph: pass length to ceph_calc_file_object_mapping()

ceph_calc_file_object_mapping() takes (among other things) a "file"
offset and length, and based on the layout, determines the object
number ("bno") backing the affected portion of the file's data and
the offset into that object where the desired range begins.  It also
computes the size that should be used for the request--either the
amount requested or something less if that would exceed the end of
the object.

This patch changes the input length parameter in this function so it
is used only for input.  That is, the argument will be passed by
value rather than by address, so the value provided won't get
updated by the function.

The value would only get updated if the length would surpass the
current object, and in that case the value it got updated to would
be exactly that returned in *oxlen.

Only one of the two callers is affected by this change.  Update
ceph_calc_raw_layout() so it records any updated value.

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
---
 fs/ceph/ioctl.c             | 2 +-
 include/linux/ceph/osdmap.h | 2 +-
 net/ceph/osd_client.c       | 6 ++++--
 net/ceph/osdmap.c           | 9 ++++-----
 4 files changed, 10 insertions(+), 9 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index 36549a46e311..3b22150d3e19 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -194,7 +194,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 		return -EFAULT;
 
 	down_read(&osdc->map_sem);
-	r = ceph_calc_file_object_mapping(&ci->i_layout, dl.file_offset, &len,
+	r = ceph_calc_file_object_mapping(&ci->i_layout, dl.file_offset, len,
 					  &dl.object_no, &dl.object_offset,
 					  &olen);
 	if (r < 0)
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index 5ea57ba69320..1f653e2ff5cc 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -110,7 +110,7 @@ extern void ceph_osdmap_destroy(struct ceph_osdmap *map);
 
 /* calculate mapping of a file extent to an object */
 extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
-					 u64 off, u64 *plen,
+					 u64 off, u64 len,
 					 u64 *bno, u64 *oxoff, u64 *oxlen);
 
 /* calculate mapping of object to a placement group */
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index e29a3ed92958..47e5f5b1f94c 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -53,13 +53,15 @@ int ceph_calc_raw_layout(struct ceph_osd_client *osdc,
 	reqhead->snapid = cpu_to_le64(snapid);
 
 	/* object extent? */
-	r = ceph_calc_file_object_mapping(layout, off, plen, bno,
+	r = ceph_calc_file_object_mapping(layout, off, orig_len, bno,
 					  &objoff, &objlen);
 	if (r < 0)
 		return r;
-	if (*plen < orig_len)
+	if (objlen < orig_len) {
+		*plen = objlen;
 		dout(" skipping last %llu, final file extent %llu~%llu\n",
 		     orig_len - *plen, off, *plen);
+	}
 
 	if (op_has_extent(op->op)) {
 		u32 osize = le32_to_cpu(layout->fl_object_size);
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index ca05871635bc..369f03ba9ee5 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -1016,7 +1016,7 @@ bad:
  * pass a stride back to the caller.
  */
 int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
-				   u64 off, u64 *plen,
+				   u64 off, u64 len,
 				   u64 *ono,
 				   u64 *oxoff, u64 *oxlen)
 {
@@ -1027,7 +1027,7 @@ int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
 	u32 su_per_object;
 	u64 t, su_offset;
 
-	dout("mapping %llu~%llu  osize %u fl_su %u\n", off, *plen,
+	dout("mapping %llu~%llu  osize %u fl_su %u\n", off, len,
 	     osize, su);
 	if (su == 0 || sc == 0)
 		goto invalid;
@@ -1060,11 +1060,10 @@ int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
 
 	/*
 	 * Calculate the length of the extent being written to the selected
-	 * object. This is the minimum of the full length requested (plen) or
+	 * object. This is the minimum of the full length requested (len) or
 	 * the remainder of the current stripe being written to.
 	 */
-	*oxlen = min_t(u64, *plen, su - su_offset);
-	*plen = *oxlen;
+	*oxlen = min_t(u64, len, su - su_offset);
 
 	dout(" obj extent %llu~%llu\n", *oxoff, *oxlen);
 	return 0;
-- 
cgit v1.2.3


From 1ec3911dbd19076bcdfe5540096ff67f91a6ec02 Mon Sep 17 00:00:00 2001
From: Cong Ding <dinggnu@gmail.com>
Date: Fri, 25 Jan 2013 17:48:59 -0600
Subject: libceph: fix undefined behavior when using snprintf()

The variable "str" is used as both the source and destination in
function snprintf(), which is undefined behavior based on C11. The
original description in C11 is:
	"If copying takes place between objects that
	overlap, the behavior is undefined."

And, the function of ceph_osdmap_state_str() is to return the osdmap
state, so it should return "doesn't exist" when all the conditions
are not satisfied. I fix it in this patch.

[elder@inktank.com: shortened the commit message]

Signed-off-by: Cong Ding <dinggnu@gmail.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 net/ceph/osdmap.c | 28 ++++++++++------------------
 1 file changed, 10 insertions(+), 18 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 369f03ba9ee5..3c61e21611d3 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -13,26 +13,18 @@
 
 char *ceph_osdmap_state_str(char *str, int len, int state)
 {
-	int flag = 0;
-
 	if (!len)
-		goto done;
-
-	*str = '\0';
-	if (state) {
-		if (state & CEPH_OSD_EXISTS) {
-			snprintf(str, len, "exists");
-			flag = 1;
-		}
-		if (state & CEPH_OSD_UP) {
-			snprintf(str, len, "%s%s%s", str, (flag ? ", " : ""),
-				 "up");
-			flag = 1;
-		}
-	} else {
+		return str;
+
+	if ((state & CEPH_OSD_EXISTS) && (state & CEPH_OSD_UP))
+		snprintf(str, len, "exists, up");
+	else if (state & CEPH_OSD_EXISTS)
+		snprintf(str, len, "exists");
+	else if (state & CEPH_OSD_UP)
+		snprintf(str, len, "up");
+	else
 		snprintf(str, len, "doesn't exist");
-	}
-done:
+
 	return str;
 }
 
-- 
cgit v1.2.3


From 12979354a1d6ef25d86f381e4d5f9e103f29913a Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Tue, 8 Jan 2013 09:15:10 -0800
Subject: libceph: rename ceph_pg -> ceph_pg_v1

Rename the old version this type to distinguish it from the new version.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 fs/ceph/ioctl.c                 |  2 +-
 include/linux/ceph/osd_client.h |  2 +-
 include/linux/ceph/osdmap.h     |  7 ++++---
 include/linux/ceph/rados.h      |  4 ++--
 net/ceph/osd_client.c           |  2 +-
 net/ceph/osdmap.c               | 18 +++++++++---------
 6 files changed, 18 insertions(+), 17 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index 3b22150d3e19..e831436d6e68 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -186,7 +186,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 	u64 len = 1, olen;
 	u64 tmp;
 	struct ceph_object_layout ol;
-	struct ceph_pg pgid;
+	struct ceph_pg_v1 pgid;
 	int r;
 
 	/* copy and validate */
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index 388158ff0cbc..be2867330e23 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -56,7 +56,7 @@ struct ceph_osd_request {
 	struct list_head r_linger_item;
 	struct list_head r_linger_osd;
 	struct ceph_osd *r_osd;
-	struct ceph_pg   r_pgid;
+	struct ceph_pg_v1 r_pgid;
 	int              r_pg_osds[CEPH_PG_MAX_SIZE];
 	int              r_num_pg_osds;
 
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index c83a838f89f5..eb4989aa48e8 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -28,7 +28,7 @@ struct ceph_pg_pool_info {
 
 struct ceph_pg_mapping {
 	struct rb_node node;
-	struct ceph_pg pgid;
+	struct ceph_pg_v1 pgid;
 	int len;
 	int osds[];
 };
@@ -118,10 +118,11 @@ extern int ceph_calc_object_layout(struct ceph_object_layout *ol,
 				   const char *oid,
 				   struct ceph_file_layout *fl,
 				   struct ceph_osdmap *osdmap);
-extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
+extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
+			       struct ceph_pg_v1 pgid,
 			       int *acting);
 extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
-				struct ceph_pg pgid);
+				struct ceph_pg_v1 pgid);
 
 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
 extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);
diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h
index b65182aba6f7..e7cece69b13f 100644
--- a/include/linux/ceph/rados.h
+++ b/include/linux/ceph/rados.h
@@ -64,7 +64,7 @@ struct ceph_timespec {
  * placement group.
  * we encode this into one __le64.
  */
-struct ceph_pg {
+struct ceph_pg_v1 {
 	__le16 preferred; /* preferred primary osd */
 	__le16 ps;        /* placement seed */
 	__le32 pool;      /* object pool */
@@ -128,7 +128,7 @@ static inline int ceph_stable_mod(int x, int b, int bmask)
  * object layout - how a given object should be stored.
  */
 struct ceph_object_layout {
-	struct ceph_pg ol_pgid;   /* raw pg, with _full_ ps precision. */
+	struct ceph_pg_v1 ol_pgid;   /* raw pg, with _full_ ps precision. */
 	__le32 ol_stripe_unit;    /* for per-object parity, if any */
 } __attribute__ ((packed));
 
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 39629b66f3b1..e3ab8d60d080 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -914,7 +914,7 @@ static int __map_request(struct ceph_osd_client *osdc,
 			 struct ceph_osd_request *req, int force_resend)
 {
 	struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
-	struct ceph_pg pgid;
+	struct ceph_pg_v1 pgid;
 	int acting[CEPH_PG_MAX_SIZE];
 	int o = -1, num = 0;
 	int err;
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 3c61e21611d3..8c89ac25081a 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -350,7 +350,7 @@ bad:
  * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid
  * to a set of osds)
  */
-static int pgid_cmp(struct ceph_pg l, struct ceph_pg r)
+static int pgid_cmp(struct ceph_pg_v1 l, struct ceph_pg_v1 r)
 {
 	u64 a = *(u64 *)&l;
 	u64 b = *(u64 *)&r;
@@ -389,7 +389,7 @@ static int __insert_pg_mapping(struct ceph_pg_mapping *new,
 }
 
 static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root,
-						   struct ceph_pg pgid)
+						   struct ceph_pg_v1 pgid)
 {
 	struct rb_node *n = root->rb_node;
 	struct ceph_pg_mapping *pg;
@@ -411,7 +411,7 @@ static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root,
 	return NULL;
 }
 
-static int __remove_pg_mapping(struct rb_root *root, struct ceph_pg pgid)
+static int __remove_pg_mapping(struct rb_root *root, struct ceph_pg_v1 pgid)
 {
 	struct ceph_pg_mapping *pg = __lookup_pg_mapping(root, pgid);
 
@@ -721,7 +721,7 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 	ceph_decode_32_safe(p, end, len, bad);
 	for (i = 0; i < len; i++) {
 		int n, j;
-		struct ceph_pg pgid;
+		struct ceph_pg_v1 pgid;
 		struct ceph_pg_mapping *pg;
 
 		ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad);
@@ -944,7 +944,7 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	while (len--) {
 		struct ceph_pg_mapping *pg;
 		int j;
-		struct ceph_pg pgid;
+		struct ceph_pg_v1 pgid;
 		u32 pglen;
 		ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
 		ceph_decode_copy(p, &pgid, sizeof(pgid));
@@ -1079,7 +1079,7 @@ int ceph_calc_object_layout(struct ceph_object_layout *ol,
 			    struct ceph_osdmap *osdmap)
 {
 	unsigned int num, num_mask;
-	struct ceph_pg pgid;
+	struct ceph_pg_v1 pgid;
 	int poolid = le32_to_cpu(fl->fl_pg_pool);
 	struct ceph_pg_pool_info *pool;
 	unsigned int ps;
@@ -1108,7 +1108,7 @@ EXPORT_SYMBOL(ceph_calc_object_layout);
  * Calculate raw osd vector for the given pgid.  Return pointer to osd
  * array, or NULL on failure.
  */
-static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
+static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 			int *osds, int *num)
 {
 	struct ceph_pg_mapping *pg;
@@ -1163,7 +1163,7 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 /*
  * Return acting set for given pgid.
  */
-int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
+int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 			int *acting)
 {
 	int rawosds[CEPH_PG_MAX_SIZE], *osds;
@@ -1184,7 +1184,7 @@ int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 /*
  * Return primary osd for given pgid, or -1 if none.
  */
-int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid)
+int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid)
 {
 	int rawosds[CEPH_PG_MAX_SIZE], *osds;
 	int i, num = CEPH_PG_MAX_SIZE;
-- 
cgit v1.2.3


From 5b191d9914eb68257f47de9d5bfe099b77f0687c Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Sat, 23 Feb 2013 10:38:16 -0800
Subject: libceph: decode into cpu-native ceph_pg type

Always decode data into our cpu-native ceph_pg type that has the correct
field widths.  Limit any remaining uses of ceph_pg_v1 to dealing with the
legacy protocol.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 fs/ceph/ioctl.c                 |  5 +--
 include/linux/ceph/osd_client.h |  2 +-
 include/linux/ceph/osdmap.h     | 11 ++++--
 net/ceph/debugfs.c              |  5 ++-
 net/ceph/osd_client.c           |  9 ++---
 net/ceph/osdmap.c               | 78 +++++++++++++++++++++++------------------
 6 files changed, 62 insertions(+), 48 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index e831436d6e68..fb036ed3e129 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -186,7 +186,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 	u64 len = 1, olen;
 	u64 tmp;
 	struct ceph_object_layout ol;
-	struct ceph_pg_v1 pgid;
+	struct ceph_pg pgid;
 	int r;
 
 	/* copy and validate */
@@ -212,7 +212,8 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 	ceph_calc_object_layout(&ol, dl.object_name, &ci->i_layout,
 				osdc->osdmap);
 
-	pgid = ol.ol_pgid;
+	pgid.pool = le32_to_cpu(ol.ol_pgid.pool);
+	pgid.seed = le16_to_cpu(ol.ol_pgid.ps);
 	dl.osd = ceph_calc_pg_primary(osdc->osdmap, pgid);
 	if (dl.osd >= 0) {
 		struct ceph_entity_addr *a =
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index be2867330e23..388158ff0cbc 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -56,7 +56,7 @@ struct ceph_osd_request {
 	struct list_head r_linger_item;
 	struct list_head r_linger_osd;
 	struct ceph_osd *r_osd;
-	struct ceph_pg_v1 r_pgid;
+	struct ceph_pg   r_pgid;
 	int              r_pg_osds[CEPH_PG_MAX_SIZE];
 	int              r_num_pg_osds;
 
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index eb4989aa48e8..8a612df4c248 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -18,6 +18,11 @@
  * The map can be updated either via an incremental map (diff) describing
  * the change between two successive epochs, or as a fully encoded map.
  */
+struct ceph_pg {
+	uint64_t pool;
+	uint32_t seed;
+};
+
 struct ceph_pg_pool_info {
 	struct rb_node node;
 	int id;
@@ -28,7 +33,7 @@ struct ceph_pg_pool_info {
 
 struct ceph_pg_mapping {
 	struct rb_node node;
-	struct ceph_pg_v1 pgid;
+	struct ceph_pg pgid;
 	int len;
 	int osds[];
 };
@@ -119,10 +124,10 @@ extern int ceph_calc_object_layout(struct ceph_object_layout *ol,
 				   struct ceph_file_layout *fl,
 				   struct ceph_osdmap *osdmap);
 extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
-			       struct ceph_pg_v1 pgid,
+			       struct ceph_pg pgid,
 			       int *acting);
 extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
-				struct ceph_pg_v1 pgid);
+				struct ceph_pg pgid);
 
 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
 extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);
diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c
index 38b5dc1823d4..61a9af634f8b 100644
--- a/net/ceph/debugfs.c
+++ b/net/ceph/debugfs.c
@@ -131,10 +131,9 @@ static int osdc_show(struct seq_file *s, void *pp)
 
 		req = rb_entry(p, struct ceph_osd_request, r_node);
 
-		seq_printf(s, "%lld\tosd%d\t%d.%x\t", req->r_tid,
+		seq_printf(s, "%lld\tosd%d\t%lld.%x\t", req->r_tid,
 			   req->r_osd ? req->r_osd->o_osd : -1,
-			   le32_to_cpu(req->r_pgid.pool),
-			   le16_to_cpu(req->r_pgid.ps));
+			   req->r_pgid.pool, req->r_pgid.seed);
 
 		head = req->r_request->front.iov_base;
 		op = (void *)(head + 1);
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index e3ab8d60d080..1990834e518b 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -914,7 +914,7 @@ static int __map_request(struct ceph_osd_client *osdc,
 			 struct ceph_osd_request *req, int force_resend)
 {
 	struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
-	struct ceph_pg_v1 pgid;
+	struct ceph_pg pgid;
 	int acting[CEPH_PG_MAX_SIZE];
 	int o = -1, num = 0;
 	int err;
@@ -926,7 +926,8 @@ static int __map_request(struct ceph_osd_client *osdc,
 		list_move(&req->r_req_lru_item, &osdc->req_notarget);
 		return err;
 	}
-	pgid = reqhead->layout.ol_pgid;
+	pgid.pool = le32_to_cpu(reqhead->layout.ol_pgid.pool);
+	pgid.seed = le16_to_cpu(reqhead->layout.ol_pgid.ps);
 	req->r_pgid = pgid;
 
 	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
@@ -943,8 +944,8 @@ static int __map_request(struct ceph_osd_client *osdc,
 	    (req->r_osd == NULL && o == -1))
 		return 0;  /* no change */
 
-	dout("map_request tid %llu pgid %d.%x osd%d (was osd%d)\n",
-	     req->r_tid, le32_to_cpu(pgid.pool), le16_to_cpu(pgid.ps), o,
+	dout("map_request tid %llu pgid %lld.%x osd%d (was osd%d)\n",
+	     req->r_tid, pgid.pool, pgid.seed, o,
 	     req->r_osd ? req->r_osd->o_osd : -1);
 
 	/* record full pg acting set */
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 8c89ac25081a..81118db5bd11 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -350,14 +350,15 @@ bad:
  * rbtree of pg_mapping for handling pg_temp (explicit mapping of pgid
  * to a set of osds)
  */
-static int pgid_cmp(struct ceph_pg_v1 l, struct ceph_pg_v1 r)
+static int pgid_cmp(struct ceph_pg l, struct ceph_pg r)
 {
-	u64 a = *(u64 *)&l;
-	u64 b = *(u64 *)&r;
-
-	if (a < b)
+	if (l.pool < r.pool)
+		return -1;
+	if (l.pool > r.pool)
+		return 1;
+	if (l.seed < r.seed)
 		return -1;
-	if (a > b)
+	if (l.seed > r.seed)
 		return 1;
 	return 0;
 }
@@ -389,7 +390,7 @@ static int __insert_pg_mapping(struct ceph_pg_mapping *new,
 }
 
 static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root,
-						   struct ceph_pg_v1 pgid)
+						   struct ceph_pg pgid)
 {
 	struct rb_node *n = root->rb_node;
 	struct ceph_pg_mapping *pg;
@@ -403,25 +404,26 @@ static struct ceph_pg_mapping *__lookup_pg_mapping(struct rb_root *root,
 		} else if (c > 0) {
 			n = n->rb_right;
 		} else {
-			dout("__lookup_pg_mapping %llx got %p\n",
-			     *(u64 *)&pgid, pg);
+			dout("__lookup_pg_mapping %lld.%x got %p\n",
+			     pgid.pool, pgid.seed, pg);
 			return pg;
 		}
 	}
 	return NULL;
 }
 
-static int __remove_pg_mapping(struct rb_root *root, struct ceph_pg_v1 pgid)
+static int __remove_pg_mapping(struct rb_root *root, struct ceph_pg pgid)
 {
 	struct ceph_pg_mapping *pg = __lookup_pg_mapping(root, pgid);
 
 	if (pg) {
-		dout("__remove_pg_mapping %llx %p\n", *(u64 *)&pgid, pg);
+		dout("__remove_pg_mapping %lld.%x %p\n", pgid.pool, pgid.seed,
+		     pg);
 		rb_erase(&pg->node, root);
 		kfree(pg);
 		return 0;
 	}
-	dout("__remove_pg_mapping %llx dne\n", *(u64 *)&pgid);
+	dout("__remove_pg_mapping %lld.%x dne\n", pgid.pool, pgid.seed);
 	return -ENOENT;
 }
 
@@ -721,11 +723,14 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 	ceph_decode_32_safe(p, end, len, bad);
 	for (i = 0; i < len; i++) {
 		int n, j;
-		struct ceph_pg_v1 pgid;
+		struct ceph_pg pgid;
+		struct ceph_pg_v1 pgid_v1;
 		struct ceph_pg_mapping *pg;
 
 		ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad);
-		ceph_decode_copy(p, &pgid, sizeof(pgid));
+		ceph_decode_copy(p, &pgid_v1, sizeof(pgid_v1));
+		pgid.pool = le32_to_cpu(pgid_v1.pool);
+		pgid.seed = le16_to_cpu(pgid_v1.ps);
 		n = ceph_decode_32(p);
 		err = -EINVAL;
 		if (n > (UINT_MAX - sizeof(*pg)) / sizeof(u32))
@@ -743,7 +748,8 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 		err = __insert_pg_mapping(pg, &map->pg_temp);
 		if (err)
 			goto bad;
-		dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, len);
+		dout(" added pg_temp %lld.%x len %d\n", pgid.pool, pgid.seed,
+		     len);
 	}
 
 	/* crush */
@@ -944,10 +950,13 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	while (len--) {
 		struct ceph_pg_mapping *pg;
 		int j;
-		struct ceph_pg_v1 pgid;
+		struct ceph_pg_v1 pgid_v1;
+		struct ceph_pg pgid;
 		u32 pglen;
 		ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
-		ceph_decode_copy(p, &pgid, sizeof(pgid));
+		ceph_decode_copy(p, &pgid_v1, sizeof(pgid_v1));
+		pgid.pool = le32_to_cpu(pgid_v1.pool);
+		pgid.seed = le16_to_cpu(pgid_v1.ps);
 		pglen = ceph_decode_32(p);
 
 		if (pglen) {
@@ -973,8 +982,8 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 				kfree(pg);
 				goto bad;
 			}
-			dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid,
-			     pglen);
+			dout(" added pg_temp %lld.%x len %d\n", pgid.pool,
+			     pgid.seed, pglen);
 		} else {
 			/* remove */
 			__remove_pg_mapping(&map->pg_temp, pgid);
@@ -1079,26 +1088,25 @@ int ceph_calc_object_layout(struct ceph_object_layout *ol,
 			    struct ceph_osdmap *osdmap)
 {
 	unsigned int num, num_mask;
-	struct ceph_pg_v1 pgid;
-	int poolid = le32_to_cpu(fl->fl_pg_pool);
+	struct ceph_pg pgid;
 	struct ceph_pg_pool_info *pool;
-	unsigned int ps;
 
 	BUG_ON(!osdmap);
 
-	pool = __lookup_pg_pool(&osdmap->pg_pools, poolid);
+	pgid.pool = le32_to_cpu(fl->fl_pg_pool);
+	pool = __lookup_pg_pool(&osdmap->pg_pools, pgid.pool);
 	if (!pool)
 		return -EIO;
-	ps = ceph_str_hash(pool->v.object_hash, oid, strlen(oid));
+	pgid.seed = ceph_str_hash(pool->v.object_hash, oid, strlen(oid));
 	num = le32_to_cpu(pool->v.pg_num);
 	num_mask = pool->pg_num_mask;
 
-	pgid.ps = cpu_to_le16(ps);
-	pgid.preferred = cpu_to_le16(-1);
-	pgid.pool = fl->fl_pg_pool;
-	dout("calc_object_layout '%s' pgid %d.%x\n", oid, poolid, ps);
+	dout("calc_object_layout '%s' pgid %lld.%x\n", oid, pgid.pool,
+	     pgid.seed);
 
-	ol->ol_pgid = pgid;
+	ol->ol_pgid.ps = cpu_to_le16(pgid.seed);
+	ol->ol_pgid.pool = fl->fl_pg_pool;
+	ol->ol_pgid.preferred = cpu_to_le16(-1);
 	ol->ol_stripe_unit = fl->fl_object_stripe_unit;
 	return 0;
 }
@@ -1108,7 +1116,7 @@ EXPORT_SYMBOL(ceph_calc_object_layout);
  * Calculate raw osd vector for the given pgid.  Return pointer to osd
  * array, or NULL on failure.
  */
-static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
+static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 			int *osds, int *num)
 {
 	struct ceph_pg_mapping *pg;
@@ -1116,8 +1124,8 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 	int ruleno;
 	unsigned int poolid, ps, pps, t, r;
 
-	poolid = le32_to_cpu(pgid.pool);
-	ps = le16_to_cpu(pgid.ps);
+	poolid = pgid.pool;
+	ps = pgid.seed;
 
 	pool = __lookup_pg_pool(&osdmap->pg_pools, poolid);
 	if (!pool)
@@ -1126,7 +1134,7 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 	/* pg_temp? */
 	t = ceph_stable_mod(ps, le32_to_cpu(pool->v.pg_num),
 			    pool->pgp_num_mask);
-	pgid.ps = cpu_to_le16(t);
+	pgid.seed = t;
 	pg = __lookup_pg_mapping(&osdmap->pg_temp, pgid);
 	if (pg) {
 		*num = pg->len;
@@ -1163,7 +1171,7 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 /*
  * Return acting set for given pgid.
  */
-int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
+int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 			int *acting)
 {
 	int rawosds[CEPH_PG_MAX_SIZE], *osds;
@@ -1184,7 +1192,7 @@ int ceph_calc_pg_acting(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid,
 /*
  * Return primary osd for given pgid, or -1 if none.
  */
-int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg_v1 pgid)
+int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid)
 {
 	int rawosds[CEPH_PG_MAX_SIZE], *osds;
 	int i, num = CEPH_PG_MAX_SIZE;
-- 
cgit v1.2.3


From 4f6a7e5ee1393ec4b243b39dac9f36992d161540 Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Sat, 23 Feb 2013 10:41:09 -0800
Subject: ceph: update support for PGID64, PGPOOL3, OSDENC protocol features

Support (and require) the PGID64, PGPOOL3, and OSDENC protocol features.
These have been present in ceph.git since v0.42, Feb 2012.  Require these
features to simplify support; nobody is running older userspace.

Note that the new request and reply encoding is still not in place, so the new
code is not yet functional.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 fs/ceph/mdsmap.c                   |  12 ++-
 include/linux/ceph/ceph_features.h |  14 +++-
 include/linux/ceph/mdsmap.h        |   4 +-
 include/linux/ceph/osdmap.h        |  16 +++-
 include/linux/ceph/rados.h         |  23 ------
 net/ceph/ceph_common.c             |   6 +-
 net/ceph/debugfs.c                 |   6 +-
 net/ceph/osdmap.c                  | 162 ++++++++++++++++++++-----------------
 8 files changed, 124 insertions(+), 119 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
index 73b7d44e8a35..0d3c9240c61b 100644
--- a/fs/ceph/mdsmap.c
+++ b/fs/ceph/mdsmap.c
@@ -59,6 +59,10 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 		return ERR_PTR(-ENOMEM);
 
 	ceph_decode_16_safe(p, end, version, bad);
+	if (version > 3) {
+		pr_warning("got mdsmap version %d > 3, failing", version);
+		goto bad;
+	}
 
 	ceph_decode_need(p, end, 8*sizeof(u32) + sizeof(u64), bad);
 	m->m_epoch = ceph_decode_32(p);
@@ -144,13 +148,13 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 	/* pg_pools */
 	ceph_decode_32_safe(p, end, n, bad);
 	m->m_num_data_pg_pools = n;
-	m->m_data_pg_pools = kcalloc(n, sizeof(u32), GFP_NOFS);
+	m->m_data_pg_pools = kcalloc(n, sizeof(u64), GFP_NOFS);
 	if (!m->m_data_pg_pools)
 		goto badmem;
-	ceph_decode_need(p, end, sizeof(u32)*(n+1), bad);
+	ceph_decode_need(p, end, sizeof(u64)*(n+1), bad);
 	for (i = 0; i < n; i++)
-		m->m_data_pg_pools[i] = ceph_decode_32(p);
-	m->m_cas_pg_pool = ceph_decode_32(p);
+		m->m_data_pg_pools[i] = ceph_decode_64(p);
+	m->m_cas_pg_pool = ceph_decode_64(p);
 
 	/* ok, we don't care about the rest. */
 	dout("mdsmap_decode success epoch %u\n", m->m_epoch);
diff --git a/include/linux/ceph/ceph_features.h b/include/linux/ceph/ceph_features.h
index 9e0f5a8ba247..ab0a54286e0d 100644
--- a/include/linux/ceph/ceph_features.h
+++ b/include/linux/ceph/ceph_features.h
@@ -39,11 +39,17 @@
  * Features supported.
  */
 #define CEPH_FEATURES_SUPPORTED_DEFAULT  \
-	(CEPH_FEATURE_NOSRCADDR |	 \
-	 CEPH_FEATURE_CRUSH_TUNABLES |	  \
-	 CEPH_FEATURE_CRUSH_TUNABLES2 |   \
+	(CEPH_FEATURE_NOSRCADDR |		\
+	 CEPH_FEATURE_PGID64 |			\
+	 CEPH_FEATURE_PGPOOL3 |			\
+	 CEPH_FEATURE_OSDENC |			\
+	 CEPH_FEATURE_CRUSH_TUNABLES |		\
+	 CEPH_FEATURE_CRUSH_TUNABLES2 |		\
 	 CEPH_FEATURE_REPLY_CREATE_INODE)
 
 #define CEPH_FEATURES_REQUIRED_DEFAULT   \
-	(CEPH_FEATURE_NOSRCADDR)
+	(CEPH_FEATURE_NOSRCADDR |	 \
+	 CEPH_FEATURE_PGID64 |		 \
+	 CEPH_FEATURE_PGPOOL3 |		 \
+	 CEPH_FEATURE_OSDENC)
 #endif
diff --git a/include/linux/ceph/mdsmap.h b/include/linux/ceph/mdsmap.h
index cb15b5d867c7..87ed09f54800 100644
--- a/include/linux/ceph/mdsmap.h
+++ b/include/linux/ceph/mdsmap.h
@@ -29,8 +29,8 @@ struct ceph_mdsmap {
 
 	/* which object pools file data can be stored in */
 	int m_num_data_pg_pools;
-	u32 *m_data_pg_pools;
-	u32 m_cas_pg_pool;
+	u64 *m_data_pg_pools;
+	u64 m_cas_pg_pool;
 };
 
 static inline struct ceph_entity_addr *
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index 8a612df4c248..8587746b7f0e 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -25,12 +25,22 @@ struct ceph_pg {
 
 struct ceph_pg_pool_info {
 	struct rb_node node;
-	int id;
-	struct ceph_pg_pool v;
-	int pg_num_mask, pgp_num_mask, lpg_num_mask, lpgp_num_mask;
+	s64 id;
+	u8 type;
+	u8 size;
+	u8 crush_ruleset;
+	u8 object_hash;
+	u32 pg_num, pgp_num;
+	int pg_num_mask, pgp_num_mask;
+	u64 flags;
 	char *name;
 };
 
+struct ceph_object_locator {
+	uint64_t pool;
+	char *key;
+};
+
 struct ceph_pg_mapping {
 	struct rb_node node;
 	struct ceph_pg pgid;
diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h
index e7cece69b13f..d784c8dfb09a 100644
--- a/include/linux/ceph/rados.h
+++ b/include/linux/ceph/rados.h
@@ -8,14 +8,6 @@
 
 #include <linux/ceph/msgr.h>
 
-/*
- * osdmap encoding versions
- */
-#define CEPH_OSDMAP_INC_VERSION     5
-#define CEPH_OSDMAP_INC_VERSION_EXT 6
-#define CEPH_OSDMAP_VERSION         5
-#define CEPH_OSDMAP_VERSION_EXT     6
-
 /*
  * fs id
  */
@@ -91,21 +83,6 @@ struct ceph_pg_v1 {
 
 #define CEPH_PG_TYPE_REP     1
 #define CEPH_PG_TYPE_RAID4   2
-#define CEPH_PG_POOL_VERSION 2
-struct ceph_pg_pool {
-	__u8 type;                /* CEPH_PG_TYPE_* */
-	__u8 size;                /* number of osds in each pg */
-	__u8 crush_ruleset;       /* crush placement rule */
-	__u8 object_hash;         /* hash mapping object name to ps */
-	__le32 pg_num, pgp_num;   /* number of pg's */
-	__le32 lpg_num, lpgp_num; /* number of localized pg's */
-	__le32 last_change;       /* most recent epoch changed */
-	__le64 snap_seq;          /* seq for per-pool snapshot */
-	__le32 snap_epoch;        /* epoch of last snap */
-	__le32 num_snaps;
-	__le32 num_removed_snap_intervals; /* if non-empty, NO per-pool snaps */
-	__le64 auid;               /* who owns the pg */
-} __attribute__ ((packed));
 
 /*
  * stable_mod func is used to control number of placement groups.
diff --git a/net/ceph/ceph_common.c b/net/ceph/ceph_common.c
index c236c235c4a2..c5605ae96714 100644
--- a/net/ceph/ceph_common.c
+++ b/net/ceph/ceph_common.c
@@ -601,10 +601,8 @@ static int __init init_ceph_lib(void)
 	if (ret < 0)
 		goto out_crypto;
 
-	pr_info("loaded (mon/osd proto %d/%d, osdmap %d/%d %d/%d)\n",
-		CEPH_MONC_PROTOCOL, CEPH_OSDC_PROTOCOL,
-		CEPH_OSDMAP_VERSION, CEPH_OSDMAP_VERSION_EXT,
-		CEPH_OSDMAP_INC_VERSION, CEPH_OSDMAP_INC_VERSION_EXT);
+	pr_info("loaded (mon/osd proto %d/%d)\n",
+		CEPH_MONC_PROTOCOL, CEPH_OSDC_PROTOCOL);
 
 	return 0;
 
diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c
index 61a9af634f8b..f4d4b27d6026 100644
--- a/net/ceph/debugfs.c
+++ b/net/ceph/debugfs.c
@@ -66,9 +66,9 @@ static int osdmap_show(struct seq_file *s, void *p)
 	for (n = rb_first(&client->osdc.osdmap->pg_pools); n; n = rb_next(n)) {
 		struct ceph_pg_pool_info *pool =
 			rb_entry(n, struct ceph_pg_pool_info, node);
-		seq_printf(s, "pg_pool %d pg_num %d / %d, lpg_num %d / %d\n",
-			   pool->id, pool->v.pg_num, pool->pg_num_mask,
-			   pool->v.lpg_num, pool->lpg_num_mask);
+		seq_printf(s, "pg_pool %llu pg_num %d / %d\n",
+			   (unsigned long long)pool->id, pool->pg_num,
+			   pool->pg_num_mask);
 	}
 	for (i = 0; i < client->osdc.osdmap->max_osd; i++) {
 		struct ceph_entity_addr *addr =
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 81118db5bd11..911919320d2e 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -45,13 +45,8 @@ static int calc_bits_of(unsigned int t)
  */
 static void calc_pg_masks(struct ceph_pg_pool_info *pi)
 {
-	pi->pg_num_mask = (1 << calc_bits_of(le32_to_cpu(pi->v.pg_num)-1)) - 1;
-	pi->pgp_num_mask =
-		(1 << calc_bits_of(le32_to_cpu(pi->v.pgp_num)-1)) - 1;
-	pi->lpg_num_mask =
-		(1 << calc_bits_of(le32_to_cpu(pi->v.lpg_num)-1)) - 1;
-	pi->lpgp_num_mask =
-		(1 << calc_bits_of(le32_to_cpu(pi->v.lpgp_num)-1)) - 1;
+	pi->pg_num_mask = (1 << calc_bits_of(pi->pg_num-1)) - 1;
+	pi->pgp_num_mask = (1 << calc_bits_of(pi->pgp_num-1)) - 1;
 }
 
 /*
@@ -452,7 +447,7 @@ static int __insert_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *new)
 	return 0;
 }
 
-static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, int id)
+static struct ceph_pg_pool_info *__lookup_pg_pool(struct rb_root *root, u64 id)
 {
 	struct ceph_pg_pool_info *pi;
 	struct rb_node *n = root->rb_node;
@@ -508,24 +503,57 @@ static void __remove_pg_pool(struct rb_root *root, struct ceph_pg_pool_info *pi)
 
 static int __decode_pool(void **p, void *end, struct ceph_pg_pool_info *pi)
 {
-	unsigned int n, m;
+	u8 ev, cv;
+	unsigned len, num;
+	void *pool_end;
+
+	ceph_decode_need(p, end, 2 + 4, bad);
+	ev = ceph_decode_8(p);  /* encoding version */
+	cv = ceph_decode_8(p); /* compat version */
+	if (ev < 5) {
+		pr_warning("got v %d < 5 cv %d of ceph_pg_pool\n", ev, cv);
+		return -EINVAL;
+	}
+	if (cv > 7) {
+		pr_warning("got v %d cv %d > 7 of ceph_pg_pool\n", ev, cv);
+		return -EINVAL;
+	}
+	len = ceph_decode_32(p);
+	ceph_decode_need(p, end, len, bad);
+	pool_end = *p + len;
 
-	ceph_decode_copy(p, &pi->v, sizeof(pi->v));
-	calc_pg_masks(pi);
+	pi->type = ceph_decode_8(p);
+	pi->size = ceph_decode_8(p);
+	pi->crush_ruleset = ceph_decode_8(p);
+	pi->object_hash = ceph_decode_8(p);
 
-	/* num_snaps * snap_info_t */
-	n = le32_to_cpu(pi->v.num_snaps);
-	while (n--) {
-		ceph_decode_need(p, end, sizeof(u64) + 1 + sizeof(u64) +
-				 sizeof(struct ceph_timespec), bad);
-		*p += sizeof(u64) +       /* key */
-			1 + sizeof(u64) + /* u8, snapid */
-			sizeof(struct ceph_timespec);
-		m = ceph_decode_32(p);    /* snap name */
-		*p += m;
+	pi->pg_num = ceph_decode_32(p);
+	pi->pgp_num = ceph_decode_32(p);
+
+	*p += 4 + 4;  /* skip lpg* */
+	*p += 4;      /* skip last_change */
+	*p += 8 + 4;  /* skip snap_seq, snap_epoch */
+
+	/* skip snaps */
+	num = ceph_decode_32(p);
+	while (num--) {
+		*p += 8;  /* snapid key */
+		*p += 1 + 1; /* versions */
+		len = ceph_decode_32(p);
+		*p += len;
 	}
 
-	*p += le32_to_cpu(pi->v.num_removed_snap_intervals) * sizeof(u64) * 2;
+	/* skip removed snaps */
+	num = ceph_decode_32(p);
+	*p += num * (8 + 8);
+
+	*p += 8;  /* skip auid */
+	pi->flags = ceph_decode_64(p);
+
+	/* ignore the rest */
+
+	*p = pool_end;
+	calc_pg_masks(pi);
 	return 0;
 
 bad:
@@ -535,14 +563,15 @@ bad:
 static int __decode_pool_names(void **p, void *end, struct ceph_osdmap *map)
 {
 	struct ceph_pg_pool_info *pi;
-	u32 num, len, pool;
+	u32 num, len;
+	u64 pool;
 
 	ceph_decode_32_safe(p, end, num, bad);
 	dout(" %d pool names\n", num);
 	while (num--) {
-		ceph_decode_32_safe(p, end, pool, bad);
+		ceph_decode_64_safe(p, end, pool, bad);
 		ceph_decode_32_safe(p, end, len, bad);
-		dout("  pool %d len %d\n", pool, len);
+		dout("  pool %llu len %d\n", pool, len);
 		ceph_decode_need(p, end, len, bad);
 		pi = __lookup_pg_pool(&map->pg_pools, pool);
 		if (pi) {
@@ -633,7 +662,6 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 	struct ceph_osdmap *map;
 	u16 version;
 	u32 len, max, i;
-	u8 ev;
 	int err = -EINVAL;
 	void *start = *p;
 	struct ceph_pg_pool_info *pi;
@@ -646,9 +674,12 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 	map->pg_temp = RB_ROOT;
 
 	ceph_decode_16_safe(p, end, version, bad);
-	if (version > CEPH_OSDMAP_VERSION) {
-		pr_warning("got unknown v %d > %d of osdmap\n", version,
-			   CEPH_OSDMAP_VERSION);
+	if (version > 6) {
+		pr_warning("got unknown v %d > 6 of osdmap\n", version);
+		goto bad;
+	}
+	if (version < 6) {
+		pr_warning("got old v %d < 6 of osdmap\n", version);
 		goto bad;
 	}
 
@@ -660,20 +691,12 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 
 	ceph_decode_32_safe(p, end, max, bad);
 	while (max--) {
-		ceph_decode_need(p, end, 4 + 1 + sizeof(pi->v), bad);
+		ceph_decode_need(p, end, 8 + 2, bad);
 		err = -ENOMEM;
 		pi = kzalloc(sizeof(*pi), GFP_NOFS);
 		if (!pi)
 			goto bad;
-		pi->id = ceph_decode_32(p);
-		err = -EINVAL;
-		ev = ceph_decode_8(p); /* encoding version */
-		if (ev > CEPH_PG_POOL_VERSION) {
-			pr_warning("got unknown v %d > %d of ceph_pg_pool\n",
-				   ev, CEPH_PG_POOL_VERSION);
-			kfree(pi);
-			goto bad;
-		}
+		pi->id = ceph_decode_64(p);
 		err = __decode_pool(p, end, pi);
 		if (err < 0) {
 			kfree(pi);
@@ -682,12 +705,10 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 		__insert_pg_pool(&map->pg_pools, pi);
 	}
 
-	if (version >= 5) {
-		err = __decode_pool_names(p, end, map);
-		if (err < 0) {
-			dout("fail to decode pool names");
-			goto bad;
-		}
+	err = __decode_pool_names(p, end, map);
+	if (err < 0) {
+		dout("fail to decode pool names");
+		goto bad;
 	}
 
 	ceph_decode_32_safe(p, end, map->pool_max, bad);
@@ -788,16 +809,17 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	struct ceph_fsid fsid;
 	u32 epoch = 0;
 	struct ceph_timespec modified;
-	u32 len, pool;
-	__s32 new_pool_max, new_flags, max;
+	s32 len;
+	u64 pool;
+	__s64 new_pool_max;
+	__s32 new_flags, max;
 	void *start = *p;
 	int err = -EINVAL;
 	u16 version;
 
 	ceph_decode_16_safe(p, end, version, bad);
-	if (version > CEPH_OSDMAP_INC_VERSION) {
-		pr_warning("got unknown v %d > %d of inc osdmap\n", version,
-			   CEPH_OSDMAP_INC_VERSION);
+	if (version > 6) {
+		pr_warning("got unknown v %d > %d of inc osdmap\n", version, 6);
 		goto bad;
 	}
 
@@ -807,7 +829,7 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	epoch = ceph_decode_32(p);
 	BUG_ON(epoch != map->epoch+1);
 	ceph_decode_copy(p, &modified, sizeof(modified));
-	new_pool_max = ceph_decode_32(p);
+	new_pool_max = ceph_decode_64(p);
 	new_flags = ceph_decode_32(p);
 
 	/* full map? */
@@ -857,18 +879,9 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	/* new_pool */
 	ceph_decode_32_safe(p, end, len, bad);
 	while (len--) {
-		__u8 ev;
 		struct ceph_pg_pool_info *pi;
 
-		ceph_decode_32_safe(p, end, pool, bad);
-		ceph_decode_need(p, end, 1 + sizeof(pi->v), bad);
-		ev = ceph_decode_8(p);  /* encoding version */
-		if (ev > CEPH_PG_POOL_VERSION) {
-			pr_warning("got unknown v %d > %d of ceph_pg_pool\n",
-				   ev, CEPH_PG_POOL_VERSION);
-			err = -EINVAL;
-			goto bad;
-		}
+		ceph_decode_64_safe(p, end, pool, bad);
 		pi = __lookup_pg_pool(&map->pg_pools, pool);
 		if (!pi) {
 			pi = kzalloc(sizeof(*pi), GFP_NOFS);
@@ -894,7 +907,7 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	while (len--) {
 		struct ceph_pg_pool_info *pi;
 
-		ceph_decode_32_safe(p, end, pool, bad);
+		ceph_decode_64_safe(p, end, pool, bad);
 		pi = __lookup_pg_pool(&map->pg_pools, pool);
 		if (pi)
 			__remove_pg_pool(&map->pg_pools, pi);
@@ -1097,8 +1110,8 @@ int ceph_calc_object_layout(struct ceph_object_layout *ol,
 	pool = __lookup_pg_pool(&osdmap->pg_pools, pgid.pool);
 	if (!pool)
 		return -EIO;
-	pgid.seed = ceph_str_hash(pool->v.object_hash, oid, strlen(oid));
-	num = le32_to_cpu(pool->v.pg_num);
+	pgid.seed = ceph_str_hash(pool->object_hash, oid, strlen(oid));
+	num = pool->pg_num;
 	num_mask = pool->pg_num_mask;
 
 	dout("calc_object_layout '%s' pgid %lld.%x\n", oid, pgid.pool,
@@ -1132,8 +1145,7 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 		return NULL;
 
 	/* pg_temp? */
-	t = ceph_stable_mod(ps, le32_to_cpu(pool->v.pg_num),
-			    pool->pgp_num_mask);
+	t = ceph_stable_mod(ps, pool->pg_num, pool->pgp_num_mask);
 	pgid.seed = t;
 	pg = __lookup_pg_mapping(&osdmap->pg_temp, pgid);
 	if (pg) {
@@ -1142,26 +1154,24 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 	}
 
 	/* crush */
-	ruleno = crush_find_rule(osdmap->crush, pool->v.crush_ruleset,
-				 pool->v.type, pool->v.size);
+	ruleno = crush_find_rule(osdmap->crush, pool->crush_ruleset,
+				 pool->type, pool->size);
 	if (ruleno < 0) {
 		pr_err("no crush rule pool %d ruleset %d type %d size %d\n",
-		       poolid, pool->v.crush_ruleset, pool->v.type,
-		       pool->v.size);
+		       poolid, pool->crush_ruleset, pool->type,
+		       pool->size);
 		return NULL;
 	}
 
-	pps = ceph_stable_mod(ps,
-			      le32_to_cpu(pool->v.pgp_num),
-			      pool->pgp_num_mask);
+	pps = ceph_stable_mod(ps, pool->pgp_num, pool->pgp_num_mask);
 	pps += poolid;
 	r = crush_do_rule(osdmap->crush, ruleno, pps, osds,
-			  min_t(int, pool->v.size, *num),
+			  min_t(int, pool->size, *num),
 			  osdmap->osd_weight);
 	if (r < 0) {
 		pr_err("error %d from crush rule: pool %d ruleset %d type %d"
-		       " size %d\n", r, poolid, pool->v.crush_ruleset,
-		       pool->v.type, pool->v.size);
+		       " size %d\n", r, poolid, pool->crush_ruleset,
+		       pool->type, pool->size);
 		return NULL;
 	}
 	*num = r;
-- 
cgit v1.2.3


From 2169aea649c08374bec7d220a3b8f64712275356 Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Mon, 25 Feb 2013 16:13:08 -0800
Subject: libceph: calculate placement based on the internal data types

Instead of using the old ceph_object_layout struct, update our internal
ceph_calc_object_layout method to use the ceph_pg type.  This allows us to
pass the full 32-bit precision of the pgid.seed to the callers.  It also
allows some callers to avoid reaching into the request structures for the
struct ceph_object_layout fields.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 fs/ceph/ioctl.c                 |  5 +----
 include/linux/ceph/osd_client.h |  1 +
 include/linux/ceph/osdmap.h     |  2 +-
 net/ceph/osd_client.c           | 11 +++++++----
 net/ceph/osdmap.c               | 18 +++++-------------
 5 files changed, 15 insertions(+), 22 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index fb036ed3e129..7d85991fd647 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -185,7 +185,6 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 		&ceph_sb_to_client(inode->i_sb)->client->osdc;
 	u64 len = 1, olen;
 	u64 tmp;
-	struct ceph_object_layout ol;
 	struct ceph_pg pgid;
 	int r;
 
@@ -209,11 +208,9 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg)
 
 	snprintf(dl.object_name, sizeof(dl.object_name), "%llx.%08llx",
 		 ceph_ino(inode), dl.object_no);
-	ceph_calc_object_layout(&ol, dl.object_name, &ci->i_layout,
+	ceph_calc_object_layout(&pgid, dl.object_name, &ci->i_layout,
 				osdc->osdmap);
 
-	pgid.pool = le32_to_cpu(ol.ol_pgid.pool);
-	pgid.seed = le16_to_cpu(ol.ol_pgid.ps);
 	dl.osd = ceph_calc_pg_primary(osdc->osdmap, pgid);
 	if (dl.osd >= 0) {
 		struct ceph_entity_addr *a =
diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h
index 388158ff0cbc..ad8899fc3157 100644
--- a/include/linux/ceph/osd_client.h
+++ b/include/linux/ceph/osd_client.h
@@ -82,6 +82,7 @@ struct ceph_osd_request {
 
 	char              r_oid[MAX_OBJ_NAME_SIZE];          /* object name */
 	int               r_oid_len;
+	u64               r_snapid;
 	unsigned long     r_stamp;            /* send OR check time */
 
 	struct ceph_file_layout r_file_layout;
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index 8587746b7f0e..35985125f118 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -129,7 +129,7 @@ extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
 					 u64 *bno, u64 *oxoff, u64 *oxlen);
 
 /* calculate mapping of object to a placement group */
-extern int ceph_calc_object_layout(struct ceph_object_layout *ol,
+extern int ceph_calc_object_layout(struct ceph_pg *pg,
 				   const char *oid,
 				   struct ceph_file_layout *fl,
 				   struct ceph_osdmap *osdmap);
diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
index 1990834e518b..5584f0a08e28 100644
--- a/net/ceph/osd_client.c
+++ b/net/ceph/osd_client.c
@@ -913,21 +913,18 @@ EXPORT_SYMBOL(ceph_osdc_set_request_linger);
 static int __map_request(struct ceph_osd_client *osdc,
 			 struct ceph_osd_request *req, int force_resend)
 {
-	struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
 	struct ceph_pg pgid;
 	int acting[CEPH_PG_MAX_SIZE];
 	int o = -1, num = 0;
 	int err;
 
 	dout("map_request %p tid %lld\n", req, req->r_tid);
-	err = ceph_calc_object_layout(&reqhead->layout, req->r_oid,
+	err = ceph_calc_object_layout(&pgid, req->r_oid,
 				      &req->r_file_layout, osdc->osdmap);
 	if (err) {
 		list_move(&req->r_req_lru_item, &osdc->req_notarget);
 		return err;
 	}
-	pgid.pool = le32_to_cpu(reqhead->layout.ol_pgid.pool);
-	pgid.seed = le16_to_cpu(reqhead->layout.ol_pgid.ps);
 	req->r_pgid = pgid;
 
 	err = ceph_calc_pg_acting(osdc->osdmap, pgid, acting);
@@ -1000,10 +997,16 @@ static void __send_request(struct ceph_osd_client *osdc,
 	     req, req->r_tid, req->r_osd->o_osd, req->r_flags);
 
 	reqhead = req->r_request->front.iov_base;
+	reqhead->snapid = cpu_to_le64(req->r_snapid);
 	reqhead->osdmap_epoch = cpu_to_le32(osdc->osdmap->epoch);
 	reqhead->flags |= cpu_to_le32(req->r_flags);  /* e.g., RETRY */
 	reqhead->reassert_version = req->r_reassert_version;
 
+	reqhead->layout.ol_pgid.ps = cpu_to_le16(req->r_pgid.seed);
+	reqhead->layout.ol_pgid.pool = cpu_to_le32(req->r_pgid.pool);
+	reqhead->layout.ol_pgid.preferred = cpu_to_le16(-1);
+	reqhead->layout.ol_stripe_unit = 0;
+
 	req->r_stamp = jiffies;
 	list_move_tail(&req->r_req_lru_item, &osdc->req_lru);
 
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 911919320d2e..378471644501 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -1095,32 +1095,24 @@ EXPORT_SYMBOL(ceph_calc_file_object_mapping);
  * calculate an object layout (i.e. pgid) from an oid,
  * file_layout, and osdmap
  */
-int ceph_calc_object_layout(struct ceph_object_layout *ol,
+int ceph_calc_object_layout(struct ceph_pg *pg,
 			    const char *oid,
 			    struct ceph_file_layout *fl,
 			    struct ceph_osdmap *osdmap)
 {
 	unsigned int num, num_mask;
-	struct ceph_pg pgid;
 	struct ceph_pg_pool_info *pool;
 
 	BUG_ON(!osdmap);
-
-	pgid.pool = le32_to_cpu(fl->fl_pg_pool);
-	pool = __lookup_pg_pool(&osdmap->pg_pools, pgid.pool);
+	pg->pool = le32_to_cpu(fl->fl_pg_pool);
+	pool = __lookup_pg_pool(&osdmap->pg_pools, pg->pool);
 	if (!pool)
 		return -EIO;
-	pgid.seed = ceph_str_hash(pool->object_hash, oid, strlen(oid));
+	pg->seed = ceph_str_hash(pool->object_hash, oid, strlen(oid));
 	num = pool->pg_num;
 	num_mask = pool->pg_num_mask;
 
-	dout("calc_object_layout '%s' pgid %lld.%x\n", oid, pgid.pool,
-	     pgid.seed);
-
-	ol->ol_pgid.ps = cpu_to_le16(pgid.seed);
-	ol->ol_pgid.pool = fl->fl_pg_pool;
-	ol->ol_pgid.preferred = cpu_to_le16(-1);
-	ol->ol_stripe_unit = fl->fl_object_stripe_unit;
+	dout("calc_object_layout '%s' pgid %lld.%x\n", oid, pg->pool, pg->seed);
 	return 0;
 }
 EXPORT_SYMBOL(ceph_calc_object_layout);
-- 
cgit v1.2.3


From 83ca14fdd35821554058e5fd4fa7b118ee504a33 Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Tue, 26 Feb 2013 10:39:09 -0800
Subject: libceph: add support for HASHPSPOOL pool flag

The legacy behavior adds the pgid seed and pool together as the input for
CRUSH.  That is problematic because each pool's PGs end up mapping to the
same OSDs: 1.5 == 2.4 == 3.3 == ...

Instead, if the HASHPSPOOL flag is set, we has the ps and pool together and
feed that into CRUSH.  This ensures that two adjacent pools will map to
an independent pseudorandom set of OSDs.

Advertise our support for this via a protocol feature flag.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Alex Elder <elder@inktank.com>
---
 include/linux/ceph/ceph_features.h |  4 +++-
 include/linux/ceph/osdmap.h        |  2 ++
 net/ceph/osdmap.c                  | 39 +++++++++++++++++++++++++-------------
 3 files changed, 31 insertions(+), 14 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/include/linux/ceph/ceph_features.h b/include/linux/ceph/ceph_features.h
index ab0a54286e0d..76554cecaab2 100644
--- a/include/linux/ceph/ceph_features.h
+++ b/include/linux/ceph/ceph_features.h
@@ -34,6 +34,7 @@
 #define CEPH_FEATURE_REPLY_CREATE_INODE   (1<<27)
 #define CEPH_FEATURE_OSD_HBMSGS     (1<<28)
 #define CEPH_FEATURE_MDSENC         (1<<29)
+#define CEPH_FEATURE_OSDHASHPSPOOL  (1<<30)
 
 /*
  * Features supported.
@@ -45,7 +46,8 @@
 	 CEPH_FEATURE_OSDENC |			\
 	 CEPH_FEATURE_CRUSH_TUNABLES |		\
 	 CEPH_FEATURE_CRUSH_TUNABLES2 |		\
-	 CEPH_FEATURE_REPLY_CREATE_INODE)
+	 CEPH_FEATURE_REPLY_CREATE_INODE |	\
+	 CEPH_FEATURE_OSDHASHPSPOOL)
 
 #define CEPH_FEATURES_REQUIRED_DEFAULT   \
 	(CEPH_FEATURE_NOSRCADDR |	 \
diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h
index 35985125f118..c819190d1642 100644
--- a/include/linux/ceph/osdmap.h
+++ b/include/linux/ceph/osdmap.h
@@ -23,6 +23,8 @@ struct ceph_pg {
 	uint32_t seed;
 };
 
+#define CEPH_POOL_FLAG_HASHPSPOOL  1
+
 struct ceph_pg_pool_info {
 	struct rb_node node;
 	s64 id;
diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 378471644501..69bc4bf89e3e 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -1127,18 +1127,16 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 	struct ceph_pg_mapping *pg;
 	struct ceph_pg_pool_info *pool;
 	int ruleno;
-	unsigned int poolid, ps, pps, t, r;
+	int r;
+	u32 pps;
 
-	poolid = pgid.pool;
-	ps = pgid.seed;
-
-	pool = __lookup_pg_pool(&osdmap->pg_pools, poolid);
+	pool = __lookup_pg_pool(&osdmap->pg_pools, pgid.pool);
 	if (!pool)
 		return NULL;
 
 	/* pg_temp? */
-	t = ceph_stable_mod(ps, pool->pg_num, pool->pgp_num_mask);
-	pgid.seed = t;
+	pgid.seed = ceph_stable_mod(pgid.seed, pool->pg_num,
+				    pool->pgp_num_mask);
 	pg = __lookup_pg_mapping(&osdmap->pg_temp, pgid);
 	if (pg) {
 		*num = pg->len;
@@ -1149,20 +1147,35 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
 	ruleno = crush_find_rule(osdmap->crush, pool->crush_ruleset,
 				 pool->type, pool->size);
 	if (ruleno < 0) {
-		pr_err("no crush rule pool %d ruleset %d type %d size %d\n",
-		       poolid, pool->crush_ruleset, pool->type,
+		pr_err("no crush rule pool %lld ruleset %d type %d size %d\n",
+		       pgid.pool, pool->crush_ruleset, pool->type,
 		       pool->size);
 		return NULL;
 	}
 
-	pps = ceph_stable_mod(ps, pool->pgp_num, pool->pgp_num_mask);
-	pps += poolid;
+	if (pool->flags & CEPH_POOL_FLAG_HASHPSPOOL) {
+		/* hash pool id and seed sothat pool PGs do not overlap */
+		pps = crush_hash32_2(CRUSH_HASH_RJENKINS1,
+				     ceph_stable_mod(pgid.seed, pool->pgp_num,
+						     pool->pgp_num_mask),
+				     pgid.pool);
+	} else {
+		/*
+		 * legacy ehavior: add ps and pool together.  this is
+		 * not a great approach because the PGs from each pool
+		 * will overlap on top of each other: 0.5 == 1.4 ==
+		 * 2.3 == ...
+		 */
+		pps = ceph_stable_mod(pgid.seed, pool->pgp_num,
+				      pool->pgp_num_mask) +
+			(unsigned)pgid.pool;
+	}
 	r = crush_do_rule(osdmap->crush, ruleno, pps, osds,
 			  min_t(int, pool->size, *num),
 			  osdmap->osd_weight);
 	if (r < 0) {
-		pr_err("error %d from crush rule: pool %d ruleset %d type %d"
-		       " size %d\n", r, poolid, pool->crush_ruleset,
+		pr_err("error %d from crush rule: pool %lld ruleset %d type %d"
+		       " size %d\n", r, pgid.pool, pool->crush_ruleset,
 		       pool->type, pool->size);
 		return NULL;
 	}
-- 
cgit v1.2.3


From d6c0dd6b0c196979fa7b34c1d99432fcb1b7e1df Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Wed, 6 Mar 2013 14:57:03 -0800
Subject: libceph: fix decoding of pgids

In 4f6a7e5ee1393ec4b243b39dac9f36992d161540 we effectively dropped support
for the legacy encoding for the OSDMap and incremental.  However, we didn't
fix the decoding for the pgid.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Yehuda Sadeh <yehuda@inktank.com>
---
 net/ceph/osdmap.c | 42 +++++++++++++++++++++++++++++-------------
 1 file changed, 29 insertions(+), 13 deletions(-)

(limited to 'net/ceph/osdmap.c')

diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index 69bc4bf89e3e..4543b9aba40c 100644
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -654,6 +654,24 @@ static int osdmap_set_max_osd(struct ceph_osdmap *map, int max)
 	return 0;
 }
 
+static int __decode_pgid(void **p, void *end, struct ceph_pg *pg)
+{
+	u8 v;
+
+	ceph_decode_need(p, end, 1+8+4+4, bad);
+	v = ceph_decode_8(p);
+	if (v != 1)
+		goto bad;
+	pg->pool = ceph_decode_64(p);
+	pg->seed = ceph_decode_32(p);
+	*p += 4; /* skip preferred */
+	return 0;
+
+bad:
+	dout("error decoding pgid\n");
+	return -EINVAL;
+}
+
 /*
  * decode a full map.
  */
@@ -745,13 +763,12 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
 	for (i = 0; i < len; i++) {
 		int n, j;
 		struct ceph_pg pgid;
-		struct ceph_pg_v1 pgid_v1;
 		struct ceph_pg_mapping *pg;
 
-		ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad);
-		ceph_decode_copy(p, &pgid_v1, sizeof(pgid_v1));
-		pgid.pool = le32_to_cpu(pgid_v1.pool);
-		pgid.seed = le16_to_cpu(pgid_v1.ps);
+		err = __decode_pgid(p, end, &pgid);
+		if (err)
+			goto bad;
+		ceph_decode_need(p, end, sizeof(u32), bad);
 		n = ceph_decode_32(p);
 		err = -EINVAL;
 		if (n > (UINT_MAX - sizeof(*pg)) / sizeof(u32))
@@ -818,8 +835,8 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	u16 version;
 
 	ceph_decode_16_safe(p, end, version, bad);
-	if (version > 6) {
-		pr_warning("got unknown v %d > %d of inc osdmap\n", version, 6);
+	if (version != 6) {
+		pr_warning("got unknown v %d != 6 of inc osdmap\n", version);
 		goto bad;
 	}
 
@@ -963,15 +980,14 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
 	while (len--) {
 		struct ceph_pg_mapping *pg;
 		int j;
-		struct ceph_pg_v1 pgid_v1;
 		struct ceph_pg pgid;
 		u32 pglen;
-		ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
-		ceph_decode_copy(p, &pgid_v1, sizeof(pgid_v1));
-		pgid.pool = le32_to_cpu(pgid_v1.pool);
-		pgid.seed = le16_to_cpu(pgid_v1.ps);
-		pglen = ceph_decode_32(p);
 
+		err = __decode_pgid(p, end, &pgid);
+		if (err)
+			goto bad;
+		ceph_decode_need(p, end, sizeof(u32), bad);
+		pglen = ceph_decode_32(p);
 		if (pglen) {
 			ceph_decode_need(p, end, pglen*sizeof(u32), bad);
 
-- 
cgit v1.2.3