s390/bpf: Perform r1 range checking before accessing jit->seen_reg[r1]
authorColin Ian King <colin.king@canonical.com>
Thu, 15 Jul 2021 12:57:12 +0000 (13:57 +0100)
committerDaniel Borkmann <daniel@iogearbox.net>
Thu, 15 Jul 2021 17:47:25 +0000 (19:47 +0200)
Currently array jit->seen_reg[r1] is being accessed before the range
checking of index r1. The range changing on r1 should be performed
first since it will avoid any potential out-of-range accesses on the
array seen_reg[] and also it is more optimal to perform checks on r1
before fetching data from the array. Fix this by swapping the order
of the checks before the array access.

Fixes: 054623105728 ("s390/bpf: Add s390x eBPF JIT compiler backend")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/20210715125712.24690-1-colin.king@canonical.com
arch/s390/net/bpf_jit_comp.c

index 63cae0476bb497e336d0945a8bc1f422895fb054..2ae419f5115a5af1cd6f2af488547191dd9f66f9 100644 (file)
@@ -112,7 +112,7 @@ static inline void reg_set_seen(struct bpf_jit *jit, u32 b1)
 {
        u32 r1 = reg2hex[b1];
 
-       if (!jit->seen_reg[r1] && r1 >= 6 && r1 <= 15)
+       if (r1 >= 6 && r1 <= 15 && !jit->seen_reg[r1])
                jit->seen_reg[r1] = 1;
 }