From 30666d478b44082744c0a1832b68f98ad69a0c9f Mon Sep 17 00:00:00 2001 From: Andre Goddard Rosa Date: Wed, 1 May 2024 11:47:12 -0500 Subject: [PATCH] Add invalid `padding_len` check in `get_pkcs_padding` When trying to decrypt data with an invalid key, we found that `mbedtls` returned `0x6200` (`-25088`), which means "_CIPHER - Input data contains invalid padding and is rejected_" from `mbedtls_cipher_finish`, but it also set the output len as `18446744073709551516`. In case we detect an error with padding, we leave the output len zero'ed and return `MBEDTLS_ERR_CIPHER_INVALID_PADDING`. I believe that the current test cases are sufficient, as they fail if I return the alternative code `MBEDTLS_ERR_CIPHER_BAD_INPUT_DATA`, so they do already expect a padding failure, but now we don't change the output len in the error case. Here's a reference for the way `openssl` checks the padding length: - https://github.com/openssl/openssl/blob/1848c561ec39a9ea91ff1bf740a554be274f98b0/crypto/evp/evp_enc.c#L1023 - https://github.com/openssl/openssl/commit/b554eef43b9ac5b92f590da6a120dbfd9ca0582e Signed-off-by: Andre Goddard Rosa Signed-off-by: Andre Goddard Rosa --- library/cipher.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/library/cipher.c b/library/cipher.c index 0683677eda..f883171921 100644 --- a/library/cipher.c +++ b/library/cipher.c @@ -849,6 +849,9 @@ static int get_pkcs_padding(unsigned char *input, size_t input_len, } padding_len = input[input_len - 1]; + if (padding_len == 0 || padding_len > (int)input_len) { + return MBEDTLS_ERR_CIPHER_INVALID_PADDING; + } *data_len = input_len - padding_len; mbedtls_ct_condition_t bad = mbedtls_ct_uint_gt(padding_len, input_len);