crypto/scheduler: fix uninitialized capabilities

Capability information is updated as slaves are attached,
but if this information is requested via rte_cryptodev_info_get()
when no slaves have been attached, the structure would not be
initialized, leading to a potential segmentation fault.

Therefore, the structure should be initialized with no
capabilities at device creation.

Fixes: 31439ee72b2c ("crypto/scheduler: add API implementations")
Cc: stable@dpdk.org
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>

diff --git a/drivers/crypto/scheduler/scheduler_pmd.c b/drivers/crypto/scheduler/scheduler_pmd.c
index dfa21c2..0b63c20 100644 (file)
--- a/drivers/crypto/scheduler/scheduler_pmd.c
+++ b/drivers/crypto/scheduler/scheduler_pmd.c
@@ -171,6 +171,20 @@ cryptodev_scheduler_create(const char *name,
                sched_ctx->nb_init_slaves++;
        }
 
+       /*
+        * Initialize capabilities structure as an empty structure,
+        * in case device information is requested when no slaves are attached
+        */
+       sched_ctx->capabilities = rte_zmalloc_socket(NULL,
+                       sizeof(struct rte_cryptodev_capabilities),
+                       0, SOCKET_ID_ANY);
+
+       if (!sched_ctx->capabilities) {
+               RTE_LOG(ERR, PMD, "Not enough memory for capability "
+                               "information\n");
+               return -ENOMEM;
+       }
+
        return 0;
 }