mm/list_lru.c: prefer struct_size over open coded arithmetic
authorLen Baker <len.baker@gmx.com>
Fri, 5 Nov 2021 20:37:40 +0000 (13:37 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Sat, 6 Nov 2021 20:30:35 +0000 (13:30 -0700)
As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing.

This could lead to values wrapping around and a smaller allocation being
made than the caller was expecting.  Using those allocations could lead
to linear overflows of heap memory and other misbehaviors.

So, use the struct_size() helper to do the arithmetic instead of the
argument "size + count * size" in the kvmalloc() functions.

Also, take the opportunity to refactor the memcpy() call to use the
flex_array_size() helper.

This code was detected with the help of Coccinelle and audited and fixed
manually.

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments

Link: https://lkml.kernel.org/r/20211017105929.9284-1-len.baker@gmx.com
Signed-off-by: Len Baker <len.baker@gmx.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/list_lru.c

index cd58790..a6031f1 100644 (file)
@@ -354,8 +354,7 @@ static int memcg_init_list_lru_node(struct list_lru_node *nlru)
        struct list_lru_memcg *memcg_lrus;
        int size = memcg_nr_cache_ids;
 
-       memcg_lrus = kvmalloc(sizeof(*memcg_lrus) +
-                             size * sizeof(void *), GFP_KERNEL);
+       memcg_lrus = kvmalloc(struct_size(memcg_lrus, lru, size), GFP_KERNEL);
        if (!memcg_lrus)
                return -ENOMEM;
 
@@ -389,7 +388,7 @@ static int memcg_update_list_lru_node(struct list_lru_node *nlru,
 
        old = rcu_dereference_protected(nlru->memcg_lrus,
                                        lockdep_is_held(&list_lrus_mutex));
-       new = kvmalloc(sizeof(*new) + new_size * sizeof(void *), GFP_KERNEL);
+       new = kvmalloc(struct_size(new, lru, new_size), GFP_KERNEL);
        if (!new)
                return -ENOMEM;
 
@@ -398,7 +397,7 @@ static int memcg_update_list_lru_node(struct list_lru_node *nlru,
                return -ENOMEM;
        }
 
-       memcpy(&new->lru, &old->lru, old_size * sizeof(void *));
+       memcpy(&new->lru, &old->lru, flex_array_size(new, lru, old_size));
 
        /*
         * The locking below allows readers that hold nlru->lock avoid taking