Instead of having a mismatch between the requested allocation size and the actual kmalloc bucket size, which is examined later via ksize(), round up proactively so the allocation is explicitly made for the full size, allowing the compiler to correctly reason about the resulting size of the buffer through the existing __alloc_size() hint.
Cc: Pravin B Shelar pshelar@ovn.org Cc: "David S. Miller" davem@davemloft.net Cc: Eric Dumazet edumazet@google.com Cc: Jakub Kicinski kuba@kernel.org Cc: Paolo Abeni pabeni@redhat.com Cc: netdev@vger.kernel.org Cc: dev@openvswitch.org Signed-off-by: Kees Cook keescook@chromium.org --- net/openvswitch/flow_netlink.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink.c index 4c09cf8a0ab2..11b2e2c94c7e 100644 --- a/net/openvswitch/flow_netlink.c +++ b/net/openvswitch/flow_netlink.c @@ -2306,10 +2306,12 @@ int ovs_nla_put_mask(const struct sw_flow *flow, struct sk_buff *skb) static struct sw_flow_actions *nla_alloc_flow_actions(int size) { struct sw_flow_actions *sfa; + int alloc_size;
WARN_ON_ONCE(size > MAX_ACTIONS_BUFSIZE);
- sfa = kmalloc(sizeof(*sfa) + size, GFP_KERNEL); + alloc_size = kmalloc_size_roundup(sizeof(*sfa) + size); + sfa = kmalloc(alloc_size, GFP_KERNEL); if (!sfa) return ERR_PTR(-ENOMEM);