Re: [LTP] [PATCH] af_alg08: Don't check recv() errno

[Cyril Hrubis <chrubis@suse.cz>]

Hi!
> Recent kernels return EBADMSG from recv() when the crypto operation
> fails. However, older kernels return EINVAL instead. Remove check
> for specific errno and only check that recv() failed.
> 
> Signed-off-by: Martin Doucha <mdoucha@suse.cz>
> ---
>  testcases/kernel/crypto/af_alg08.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/testcases/kernel/crypto/af_alg08.c b/testcases/kernel/crypto/af_alg08.c
> index 5e04b579c..4cc3e5c4e 100644
> --- a/testcases/kernel/crypto/af_alg08.c
> +++ b/testcases/kernel/crypto/af_alg08.c
> @@ -100,7 +100,14 @@ static void try_corrupt(void)
>  	SAFE_SPLICE(pipefd[0], NULL, reqfd, NULL, OVERWRITE_SIZE, 0);
>  
>  	/* Expected to fail (invalid ciphertext); triggers the scratch write */
> -	TST_EXP_FAIL_SILENT(recv(reqfd, recvbuf, sizeof(recvbuf), 0), EBADMSG);
> +	TEST(recv(reqfd, recvbuf, sizeof(recvbuf), 0));
> +
> +	if (TST_RET >= 0) {
> +		tst_res(TFAIL, "Decryption succeeded unexpectedly");
> +	} else if (TST_RET != -1) {
> +		tst_res(TFAIL | TTERRNO, "Invalid recv() return value %ld",
> +			TST_RET);
> +	}

Maybe allowing all errnos is bit to broad.

I wonder why we haven't added TST_EXP_FAIL_ARR_SILENT() since that
should be as simple as:

#define TST_EXP_FAIL_ARR_SILENT(SCALL, EXP_ERRS, EXP_ERRS_CNT, ...) \
	TST_EXP_FAIL_SILENT_(TST_RET == 0, SCALL, #SCALL, EXP_ERRS, EXP_ERRS_CNT, ##__VA_ARGS__)


-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp