[Tarantool-patches] [PATCH v2] base64: Properly ignore invalid characters

[Alexander V. Tikhonov]

Hi Sergey, thanks for the patch, as I see no new degradation found in
gitlab-ci testing commit criteria pipeline [1], patch LGTM.

[1] - https://gitlab.com/tarantool/tarantool/-/pipelines/230156990

On Tue, Dec 15, 2020 at 05:25:27PM +0300, Sergey Nikiforov via Tarantool-patches wrote:
> Not all invalid characters were ignored by base64 decoder
> causing data corruption and reads beyond decode table
> (faults under ASAN).
> 
> Added corresponding check into base64 unit test.
> 
> Fixes: #5627
> ---
> 
> Branch: https://github.com/tarantool/tarantool/tree/void234/gh-5627-fix-base64-invalid-chars-processing
> Issue: https://github.com/tarantool/tarantool/issues/5627
> 
>  test/unit/base64.c      | 23 ++++++++++++++++++++++-
>  test/unit/base64.result |  5 ++++-
>  third_party/base64.c    |  3 ++-
>  3 files changed, 28 insertions(+), 3 deletions(-)
> 
> diff --git a/test/unit/base64.c b/test/unit/base64.c
> index ada497adf..c0f53a5e1 100644
> --- a/test/unit/base64.c
> +++ b/test/unit/base64.c
> @@ -58,9 +58,28 @@ base64_nowrap_test(const char *str)
>  	base64_test(str, BASE64_NOWRAP, symbols, lengthof(symbols));
>  }
>  
> +static void
> +base64_invalid_chars_test(void)
> +{
> +	/* Upper bit must be cleared */
> +	const char invalid_data[] = { '\x7b', '\x7c', '\x7d', '\x7e' };
> +	char outbuf[8];
> +
> +	plan(1);
> +
> +	/* Invalid chars should be ignored, not decoded into garbage */
> +	is(base64_decode(invalid_data, sizeof(invalid_data),
> +	                 outbuf, sizeof(outbuf)),
> +	   0, "ignoring invalid chars");
> +
> +	check_plan();
> +}
> +
>  int main(int argc, char *argv[])
>  {
> -	plan(28);
> +	plan(28
> +	     + 1 /* invalid chars test */
> +	     );
>  	header();
>  
>  	const char *option_tests[] = {
> @@ -78,6 +97,8 @@ int main(int argc, char *argv[])
>  		base64_nowrap_test(option_tests[i]);
>  	}
>  
> +	base64_invalid_chars_test();
> +
>  	footer();
>  	return check_plan();
>  }
> diff --git a/test/unit/base64.result b/test/unit/base64.result
> index cd1f2b3f6..3bc2c2275 100644
> --- a/test/unit/base64.result
> +++ b/test/unit/base64.result
> @@ -1,4 +1,4 @@
> -1..28
> +1..29
>  	*** main ***
>      1..3
>      ok 1 - length
> @@ -175,4 +175,7 @@ ok 27 - subtests
>      ok 3 - decode length ok
>      ok 4 - encode/decode
>  ok 28 - subtests
> +    1..1
> +    ok 1 - ignoring invalid chars
> +ok 29 - subtests
>  	*** main: done ***
> diff --git a/third_party/base64.c b/third_party/base64.c
> index 8ecab23eb..7c69315ea 100644
> --- a/third_party/base64.c
> +++ b/third_party/base64.c
> @@ -222,7 +222,8 @@ base64_decode_value(int value)
>  		32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
>  		44, 45, 46, 47, 48, 49, 50, 51
>  	};
> -	static const int decoding_size = sizeof(decoding);
> +	static const int decoding_size =
> +		sizeof(decoding) / sizeof(decoding[0]);
>  	int codepos = value;
>  	codepos -= 43;
>  	if (codepos < 0 || codepos >= decoding_size)
> -- 
> 2.25.1
>