Re: [bpf-next PATCH 05/10] bpf: selftests, improve test_sockmap total bytes counter

[Jakub Sitnicki <jakub@cloudflare.com>]

On Tue, 05 May 2020 13:51:14 -0700
John Fastabend <john.fastabend@gmail.com> wrote:

> The recv thread in test_sockmap waits to receive all bytes from sender but
> in the case we use pop data it may wait for more bytes then actually being
> sent. This stalls the test harness for multiple seconds. Because this
> happens in multiple tests it slows time to run the selftest.
> 
> Fix by doing a better job of accounting for total bytes when pop helpers
> are used.
> 
> Signed-off-by: John Fastabend <john.fastabend@gmail.com>
> ---
>  tools/testing/selftests/bpf/test_sockmap.c |    9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/testing/selftests/bpf/test_sockmap.c b/tools/testing/selftests/bpf/test_sockmap.c
> index a81ed5d..36aca86 100644
> --- a/tools/testing/selftests/bpf/test_sockmap.c
> +++ b/tools/testing/selftests/bpf/test_sockmap.c
> @@ -502,9 +502,10 @@ static int msg_loop(int fd, int iov_count, int iov_length, int cnt,
>  		 * paths.
>  		 */
>  		total_bytes = (float)iov_count * (float)iov_length * (float)cnt;
> -		txmsg_pop_total = txmsg_pop;
>  		if (txmsg_apply)
> -			txmsg_pop_total *= (total_bytes / txmsg_apply);
> +			txmsg_pop_total = txmsg_pop * (total_bytes / txmsg_apply);
> +		else
> +			txmsg_pop_total = txmsg_pop * cnt;
>  		total_bytes -= txmsg_pop_total;
>  		err = clock_gettime(CLOCK_MONOTONIC, &s->start);
>  		if (err < 0)
> @@ -638,9 +639,13 @@ static int sendmsg_test(struct sockmap_options *opt)
>  
>  	rxpid = fork();
>  	if (rxpid == 0) {
> +		iov_buf -= (txmsg_pop - txmsg_start_pop + 1);
>  		if (opt->drop_expected)
>  			exit(0);
>  
> +		if (!iov_buf) /* zero bytes sent case */
> +			exit(0);

You probably want to call _exit() from the child to prevent flushing
stdio buffers twice.

> +
>  		if (opt->sendpage)
>  			iov_count = 1;
>  		err = msg_loop(rx_fd, iov_count, iov_buf,
>