crush: pass weight vector size to map function

Pass the size of the weight vector into crush_do_rule() to ensure that we
don't access values past the end.  This can happen if the caller misbehaves
and passes a weight vector that is smaller than max_devices.

Currently the monitor tries to prevent that from happening, but this will
gracefully tolerate previous bad osdmaps that got into this state.  It's
also a bit more defensive.

Reflects ceph.git commit 5922e2c2b8335b5e46c9504349c3a55b7434c01a.

Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>

diff --git a/include/linux/crush/mapper.h b/include/linux/crush/mapper.h
index 5772dee..69310b0 100644 (file)
--- a/include/linux/crush/mapper.h
+++ b/include/linux/crush/mapper.h
@@ -14,6 +14,6 @@ extern int crush_find_rule(const struct crush_map *map, int ruleset, int type, i
 extern int crush_do_rule(const struct crush_map *map,
                         int ruleno,
                         int x, int *result, int result_max,
 extern int crush_do_rule(const struct crush_map *map,
                         int ruleno,
                         int x, int *result, int result_max,

-                        const __u32 *weights);
+                        const __u32 *weights, int weight_max);

 
 #endif
 
 #endif

diff --git a/net/ceph/crush/mapper.c b/net/ceph/crush/mapper.c
index cbd06a9..18d2cf6 100644 (file)
--- a/net/ceph/crush/mapper.c
+++ b/net/ceph/crush/mapper.c
@@ -264,8 +264,12 @@ static int crush_bucket_choose(struct crush_bucket *in, int x, int r)
  * true if device is marked "out" (failed, fully offloaded)
  * of the cluster
  */
  * true if device is marked "out" (failed, fully offloaded)
  * of the cluster
  */

-static int is_out(const struct crush_map *map, const __u32 *weight, int item, int x)
+static int is_out(const struct crush_map *map,
+                 const __u32 *weight, int weight_max,
+                 int item, int x)

 {
 {

+       if (item >= weight_max)
+               return 1;

        if (weight[item] >= 0x10000)
                return 0;
        if (weight[item] == 0)
        if (weight[item] >= 0x10000)
                return 0;
        if (weight[item] == 0)


@@ -292,7 +296,7 @@ static int is_out(const struct crush_map *map, const __u32 *weight, int item, in
  */
 static int crush_choose(const struct crush_map *map,
                        struct crush_bucket *bucket,
  */
 static int crush_choose(const struct crush_map *map,
                        struct crush_bucket *bucket,

-                       const __u32 *weight,
+                       const __u32 *weight, int weight_max,

                        int x, int numrep, int type,
                        int *out, int outpos,
                        int firstn, int recurse_to_leaf,
                        int x, int numrep, int type,
                        int *out, int outpos,
                        int firstn, int recurse_to_leaf,


@@ -396,7 +400,7 @@ static int crush_choose(const struct crush_map *map,
                                        if (item < 0) {
                                                if (crush_choose(map,
                                                         map->buckets[-1-item],
                                        if (item < 0) {
                                                if (crush_choose(map,
                                                         map->buckets[-1-item],

-                                                        weight,
+                                                        weight, weight_max,

                                                         x, outpos+1, 0,
                                                         out2, outpos,
                                                         firstn, 0,
                                                         x, outpos+1, 0,
                                                         out2, outpos,
                                                         firstn, 0,


@@ -414,6 +418,7 @@ static int crush_choose(const struct crush_map *map,
                                        /* out? */
                                        if (itemtype == 0)
                                                reject = is_out(map, weight,
                                        /* out? */
                                        if (itemtype == 0)
                                                reject = is_out(map, weight,

+                                                               weight_max,

                                                                item, x);
                                        else
                                                reject = 0;
                                                                item, x);
                                        else
                                                reject = 0;


@@ -470,10 +475,12 @@ reject:
  * @x: hash input
  * @result: pointer to result vector
  * @result_max: maximum result size
  * @x: hash input
  * @result: pointer to result vector
  * @result_max: maximum result size

+ * @weight: weight vector (for map leaves)
+ * @weight_max: size of weight vector

  */
 int crush_do_rule(const struct crush_map *map,
                  int ruleno, int x, int *result, int result_max,
  */
 int crush_do_rule(const struct crush_map *map,
                  int ruleno, int x, int *result, int result_max,

-                 const __u32 *weight)
+                 const __u32 *weight, int weight_max)

 {
        int result_len;
        int a[CRUSH_MAX_SET];
 {
        int result_len;
        int a[CRUSH_MAX_SET];


@@ -545,7 +552,7 @@ int crush_do_rule(const struct crush_map *map,
                                j = 0;
                                osize += crush_choose(map,
                                                      map->buckets[-1-w[i]],
                                j = 0;
                                osize += crush_choose(map,
                                                      map->buckets[-1-w[i]],

-                                                     weight,
+                                                     weight, weight_max,

                                                      x, numrep,
                                                      curstep->arg2,
                                                      o+osize, j,
                                                      x, numrep,
                                                      curstep->arg2,
                                                      o+osize, j,

diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c
index dbd9a47..6477a68 100644 (file)
--- a/net/ceph/osdmap.c
+++ b/net/ceph/osdmap.c
@@ -1165,7 +1165,7 @@ static int *calc_pg_raw(struct ceph_osdmap *osdmap, struct ceph_pg pgid,
        }
        r = crush_do_rule(osdmap->crush, ruleno, pps, osds,
                          min_t(int, pool->size, *num),
        }
        r = crush_do_rule(osdmap->crush, ruleno, pps, osds,
                          min_t(int, pool->size, *num),

-                         osdmap->osd_weight);
+                         osdmap->osd_weight, osdmap->max_osd);

        if (r < 0) {
                pr_err("error %d from crush rule: pool %lld ruleset %d type %d"
                       " size %d\n", r, pgid.pool, pool->crush_ruleset,
        if (r < 0) {
                pr_err("error %d from crush rule: pool %lld ruleset %d type %d"
                       " size %d\n", r, pgid.pool, pool->crush_ruleset,