Fix yet another locking error in the slab allocator

* kern/slab.c (kmem_cache_free): Relock cache before retrying releasing
an object to the CPU pool layer.

1 files changed, 1 insertions, 0 deletions

diff --git a/kern/slab.c b/kern/slab.c
index d14d07a..56fadbf 100644
--- a/kern/slab.c
+++ b/kern/slab.c
@@ -1229,6 +1229,7 @@ fast_free:
             simple_unlock(&cpu_pool->lock);
             kmem_cache_free(cache->cpu_pool_type->array_cache,
                             (vm_offset_t)array);
+            simple_lock(&cpu_pool->lock);
             goto fast_free;
         }