ceph: fix endian conversions for ceph_pg
The endian conversions don't quite work with the old union ceph_pg. Just
make it a regular struct, and make each field __le. This is simpler and it
has the added bonus of actually working.
Signed-off-by: Sage Weil <sage@newdream.net>
diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index e4f99ef..4c33e19 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -99,7 +99,7 @@
u64 len = 1, olen;
u64 tmp;
struct ceph_object_layout ol;
- union ceph_pg pgid;
+ struct ceph_pg pgid;
/* copy and validate */
if (copy_from_user(&dl, arg, sizeof(dl)))
@@ -121,7 +121,7 @@
ceph_calc_object_layout(&ol, dl.object_name, &ci->i_layout,
osdc->osdmap);
- pgid.pg64 = le64_to_cpu(ol.ol_pgid);
+ pgid = ol.ol_pgid;
dl.osd = ceph_calc_pg_primary(osdc->osdmap, pgid);
if (dl.osd >= 0) {
struct ceph_entity_addr *a =
diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c
index 7dc0f62..7db14ba 100644
--- a/fs/ceph/osd_client.c
+++ b/fs/ceph/osd_client.c
@@ -520,7 +520,7 @@
struct ceph_osd_request *req)
{
struct ceph_osd_request_head *reqhead = req->r_request->front.iov_base;
- union ceph_pg pgid;
+ struct ceph_pg pgid;
int o = -1;
int err;
struct ceph_osd *newosd = NULL;
@@ -530,7 +530,7 @@
&req->r_file_layout, osdc->osdmap);
if (err)
return err;
- pgid.pg64 = le64_to_cpu(reqhead->layout.ol_pgid);
+ pgid = reqhead->layout.ol_pgid;
o = ceph_calc_pg_primary(osdc->osdmap, pgid);
if ((req->r_osd && req->r_osd->o_osd == o &&
@@ -538,8 +538,8 @@
(req->r_osd == NULL && o == -1))
return 0; /* no change */
- dout("map_osds tid %llu pgid %llx pool %d osd%d (was osd%d)\n",
- req->r_tid, pgid.pg64, pgid.pg.pool, o,
+ dout("map_osds 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,
req->r_osd ? req->r_osd->o_osd : -1);
if (req->r_osd) {
diff --git a/fs/ceph/osdmap.c b/fs/ceph/osdmap.c
index cd7bb26..8b0cd11 100644
--- a/fs/ceph/osdmap.c
+++ b/fs/ceph/osdmap.c
@@ -366,19 +366,33 @@
/*
* Insert a new pg_temp mapping
*/
+static int pgid_cmp(struct ceph_pg l, struct ceph_pg r)
+{
+ u64 a = *(u64 *)&l;
+ u64 b = *(u64 *)&r;
+
+ if (a < b)
+ return -1;
+ if (a > b)
+ return 1;
+ return 0;
+}
+
static int __insert_pg_mapping(struct ceph_pg_mapping *new,
struct rb_root *root)
{
struct rb_node **p = &root->rb_node;
struct rb_node *parent = NULL;
struct ceph_pg_mapping *pg = NULL;
+ int c;
while (*p) {
parent = *p;
pg = rb_entry(parent, struct ceph_pg_mapping, node);
- if (new->pgid < pg->pgid)
+ c = pgid_cmp(new->pgid, pg->pgid);
+ if (c < 0)
p = &(*p)->rb_left;
- else if (new->pgid > pg->pgid)
+ else if (c > 0)
p = &(*p)->rb_right;
else
return -EEXIST;
@@ -467,11 +481,11 @@
ceph_decode_32_safe(p, end, len, bad);
for (i = 0; i < len; i++) {
int n, j;
- u64 pgid;
+ struct ceph_pg pgid;
struct ceph_pg_mapping *pg;
ceph_decode_need(p, end, sizeof(u32) + sizeof(u64), bad);
- pgid = ceph_decode_64(p);
+ ceph_decode_copy(p, &pgid, sizeof(pgid));
n = ceph_decode_32(p);
ceph_decode_need(p, end, n * sizeof(u32), bad);
pg = kmalloc(sizeof(*pg) + n*sizeof(u32), GFP_NOFS);
@@ -487,7 +501,7 @@
err = __insert_pg_mapping(pg, &map->pg_temp);
if (err)
goto bad;
- dout(" added pg_temp %llx len %d\n", pgid, len);
+ dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid, len);
}
/* crush */
@@ -659,19 +673,20 @@
while (len--) {
struct ceph_pg_mapping *pg;
int j;
- u64 pgid;
+ struct ceph_pg pgid;
u32 pglen;
ceph_decode_need(p, end, sizeof(u64) + sizeof(u32), bad);
- pgid = ceph_decode_64(p);
+ ceph_decode_copy(p, &pgid, sizeof(pgid));
pglen = ceph_decode_32(p);
/* remove any? */
- while (rbp && rb_entry(rbp, struct ceph_pg_mapping,
- node)->pgid <= pgid) {
+ while (rbp && pgid_cmp(rb_entry(rbp, struct ceph_pg_mapping,
+ node)->pgid, pgid) <= 0) {
struct rb_node *cur = rbp;
rbp = rb_next(rbp);
dout(" removed pg_temp %llx\n",
- rb_entry(cur, struct ceph_pg_mapping, node)->pgid);
+ *(u64 *)&rb_entry(cur, struct ceph_pg_mapping,
+ node)->pgid);
rb_erase(cur, &map->pg_temp);
}
@@ -690,14 +705,16 @@
err = __insert_pg_mapping(pg, &map->pg_temp);
if (err)
goto bad;
- dout(" added pg_temp %llx len %d\n", pgid, pglen);
+ dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid,
+ pglen);
}
}
while (rbp) {
struct rb_node *cur = rbp;
rbp = rb_next(rbp);
dout(" removed pg_temp %llx\n",
- rb_entry(cur, struct ceph_pg_mapping, node)->pgid);
+ *(u64 *)&rb_entry(cur, struct ceph_pg_mapping,
+ node)->pgid);
rb_erase(cur, &map->pg_temp);
}
@@ -782,16 +799,19 @@
struct ceph_osdmap *osdmap)
{
unsigned num, num_mask;
- union ceph_pg pgid;
+ struct ceph_pg pgid;
s32 preferred = (s32)le32_to_cpu(fl->fl_pg_preferred);
int poolid = le32_to_cpu(fl->fl_pg_pool);
struct ceph_pg_pool_info *pool;
+ unsigned ps;
if (poolid >= osdmap->num_pools)
return -EIO;
- pool = &osdmap->pg_pool[poolid];
+ pool = &osdmap->pg_pool[poolid];
+ ps = ceph_full_name_hash(oid, strlen(oid));
if (preferred >= 0) {
+ ps += preferred;
num = le32_to_cpu(pool->v.lpg_num);
num_mask = pool->lpg_num_mask;
} else {
@@ -799,22 +819,17 @@
num_mask = pool->pg_num_mask;
}
- pgid.pg64 = 0; /* start with it zeroed out */
- pgid.pg.ps = ceph_full_name_hash(oid, strlen(oid));
- pgid.pg.preferred = preferred;
+ pgid.ps = cpu_to_le16(ps);
+ pgid.preferred = cpu_to_le16(preferred);
+ pgid.pool = fl->fl_pg_pool;
if (preferred >= 0)
- pgid.pg.ps += preferred;
- pgid.pg.pool = le32_to_cpu(fl->fl_pg_pool);
- if (preferred >= 0)
- dout("calc_object_layout '%s' pgid %d.%xp%d (%llx)\n", oid,
- pgid.pg.pool, pgid.pg.ps, (int)preferred, pgid.pg64);
+ dout("calc_object_layout '%s' pgid %d.%xp%d\n", oid, poolid, ps,
+ (int)preferred);
else
- dout("calc_object_layout '%s' pgid %d.%x (%llx)\n", oid,
- pgid.pg.pool, pgid.pg.ps, pgid.pg64);
+ dout("calc_object_layout '%s' pgid %d.%x\n", oid, poolid, ps);
- ol->ol_pgid = cpu_to_le64(pgid.pg64);
+ ol->ol_pgid = pgid;
ol->ol_stripe_unit = fl->fl_object_stripe_unit;
-
return 0;
}
@@ -822,21 +837,24 @@
* 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, union ceph_pg pgid,
+static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
int *osds, int *num)
{
struct rb_node *n = osdmap->pg_temp.rb_node;
struct ceph_pg_mapping *pg;
struct ceph_pg_pool_info *pool;
int ruleno;
- unsigned pps; /* placement ps */
+ unsigned poolid, ps, pps;
+ int preferred;
+ int c;
/* pg_temp? */
while (n) {
pg = rb_entry(n, struct ceph_pg_mapping, node);
- if (pgid.pg64 < pg->pgid)
+ c = pgid_cmp(pgid, pg->pgid);
+ if (c < 0)
n = n->rb_left;
- else if (pgid.pg64 > pg->pgid)
+ else if (c > 0)
n = n->rb_right;
else {
*num = pg->len;
@@ -845,36 +863,40 @@
}
/* crush */
- if (pgid.pg.pool >= osdmap->num_pools)
+ poolid = le32_to_cpu(pgid.pool);
+ ps = le16_to_cpu(pgid.ps);
+ preferred = (s16)le16_to_cpu(pgid.preferred);
+
+ if (poolid >= osdmap->num_pools)
return NULL;
- pool = &osdmap->pg_pool[pgid.pg.pool];
+ pool = &osdmap->pg_pool[poolid];
ruleno = crush_find_rule(osdmap->crush, pool->v.crush_ruleset,
pool->v.type, pool->v.size);
if (ruleno < 0) {
pr_err("no crush rule pool %d type %d size %d\n",
- pgid.pg.pool, pool->v.type, pool->v.size);
+ poolid, pool->v.type, pool->v.size);
return NULL;
}
- if (pgid.pg.preferred >= 0)
- pps = ceph_stable_mod(pgid.pg.ps,
+ if (preferred >= 0)
+ pps = ceph_stable_mod(ps,
le32_to_cpu(pool->v.lpgp_num),
pool->lpgp_num_mask);
else
- pps = ceph_stable_mod(pgid.pg.ps,
+ pps = ceph_stable_mod(ps,
le32_to_cpu(pool->v.pgp_num),
pool->pgp_num_mask);
- pps += pgid.pg.pool;
+ pps += poolid;
*num = crush_do_rule(osdmap->crush, ruleno, pps, osds,
min_t(int, pool->v.size, *num),
- pgid.pg.preferred, osdmap->osd_weight);
+ preferred, osdmap->osd_weight);
return osds;
}
/*
* Return primary osd for given pgid, or -1 if none.
*/
-int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, union ceph_pg pgid)
+int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, struct ceph_pg pgid)
{
int rawosds[10], *osds;
int i, num = ARRAY_SIZE(rawosds);
diff --git a/fs/ceph/osdmap.h b/fs/ceph/osdmap.h
index 07127c6..c4af841 100644
--- a/fs/ceph/osdmap.h
+++ b/fs/ceph/osdmap.h
@@ -25,7 +25,7 @@
struct ceph_pg_mapping {
struct rb_node node;
- u64 pgid;
+ struct ceph_pg pgid;
int len;
int osds[];
};
@@ -118,6 +118,7 @@
const char *oid,
struct ceph_file_layout *fl,
struct ceph_osdmap *osdmap);
-extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap, union ceph_pg pgid);
+extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
+ struct ceph_pg pgid);
#endif
diff --git a/fs/ceph/rados.h b/fs/ceph/rados.h
index a48cf4a..85bdef7 100644
--- a/fs/ceph/rados.h
+++ b/fs/ceph/rados.h
@@ -55,13 +55,10 @@
* placement group.
* we encode this into one __le64.
*/
-union ceph_pg {
- __u64 pg64;
- struct {
- __s16 preferred; /* preferred primary osd */
- __u16 ps; /* placement seed */
- __u32 pool; /* object pool */
- } __attribute__ ((packed)) pg;
+struct ceph_pg {
+ __le16 preferred; /* preferred primary osd */
+ __le16 ps; /* placement seed */
+ __le32 pool; /* object pool */
} __attribute__ ((packed));
/*
@@ -117,7 +114,7 @@
* object layout - how a given object should be stored.
*/
struct ceph_object_layout {
- __le64 ol_pgid; /* raw pg, with _full_ ps precision. */
+ struct ceph_pg ol_pgid; /* raw pg, with _full_ ps precision. */
__le32 ol_stripe_unit; /* for per-object parity, if any */
} __attribute__ ((packed));