Re: [PATCH v6] testpmd: cleanup cleanly from signal

[Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>]

On 11/10/22 19:53, Stephen Hemminger wrote:
> Do a clean shutdown of testpmd when a signal is received;
> instead of having testpmd kill itself.
> This fixes the problem where a signal could be received
> in the middle of a PMD and then the signal handler would call
> PMD's close routine leading to locking problems.
> 
> An added benefit is it gets rid of some Windows specific code.
> 
> Fixes: d9a191a00e81 ("app/testpmd: fix quitting in container")
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
> v5 - drop unnecessary print in signal handler.
>       don't cleanup twice
>       don't print message when select() is interrupted.
> 
>   app/test-pmd/testpmd.c | 72 ++++++++++++++++++++----------------------
>   1 file changed, 35 insertions(+), 37 deletions(-)
> 
> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
> index cf5942d0c422..62d87f758ac8 100644
> --- a/app/test-pmd/testpmd.c
> +++ b/app/test-pmd/testpmd.c
> @@ -11,6 +11,7 @@
>   #include <fcntl.h>
>   #ifndef RTE_EXEC_ENV_WINDOWS
>   #include <sys/mman.h>
> +#include <sys/select.h>
>   #endif
>   #include <sys/types.h>
>   #include <errno.h>
> @@ -4224,13 +4225,6 @@ init_port(void)
>   	memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
>   }
>   
> -static void
> -force_quit(void)
> -{
> -	pmd_test_exit();
> -	prompt_exit();
> -}
> -
>   static void
>   print_stats(void)
>   {
> @@ -4249,28 +4243,9 @@ print_stats(void)
>   }
>   
>   static void
> -signal_handler(int signum)
> +signal_handler(int signum __rte_unused)
>   {
> -	if (signum == SIGINT || signum == SIGTERM) {
> -		fprintf(stderr, "\nSignal %d received, preparing to exit...\n",
> -			signum);
> -#ifdef RTE_LIB_PDUMP
> -		/* uninitialize packet capture framework */
> -		rte_pdump_uninit();
> -#endif
> -#ifdef RTE_LIB_LATENCYSTATS
> -		if (latencystats_enabled != 0)
> -			rte_latencystats_uninit();
> -#endif
> -		force_quit();
> -		/* Set flag to indicate the force termination. */
> -		f_quit = 1;
> -		/* exit with the expected status */
> -#ifndef RTE_EXEC_ENV_WINDOWS
> -		signal(signum, SIG_DFL);
> -		kill(getpid(), signum);
> -#endif
> -	}
> +	f_quit = 1;
>   }
>   
>   int
> @@ -4449,9 +4424,6 @@ main(int argc, char** argv)
>   	} else
>   #endif
>   	{
> -		char c;
> -		int rc;
> -
>   		f_quit = 0;
>   
>   		printf("No commandline core given, start packet forwarding\n");
> @@ -4476,15 +4448,41 @@ main(int argc, char** argv)
>   				prev_time = cur_time;
>   				rte_delay_us_sleep(US_PER_S);
>   			}
> -		}
> +		} else {
> +			char c;
> +			fd_set fds;
> +			int rc;
>   
> -		printf("Press enter to exit\n");
> -		rc = read(0, &c, 1);
> -		pmd_test_exit();
> -		if (rc < 0)
> -			return 1;
> +			printf("Press enter to exit\n");
> +
> +			FD_ZERO(&fds);
> +			FD_SET(0, &fds);
> +
> +			rc = select(1, &fds, NULL, NULL, NULL);
> +			if (rc < 0 && errno != EINTR) {
> +				fprintf(stderr, "Select failed: %s\n",
> +					strerror(errno));
> +				return 1;
> +			}
> +			if (rc > 0)
> +				rc = read(0, &c, 1);
> +
> +			pmd_test_exit();

It looks wrong to skip pmd_test_exit() in periodic stats
case (if body above). Earlier it happened in signal handler.
IMHO, pmd_test_exit() should be done just before pdump
uninit below (outside if/else bodies (and removed from
interactive == 1 branch above).

> +			if (rc < 0)
> +				return 1;

It took some time for me to understand what's happening here.
Finally I came to conclusion that it just preserve previous
behaviour to return with failure code immediately if read
fails. Except addition of pmd_test_exit() above.
I think it is a wrong behaviour to skip all cleanup which
is done below, but I agree that it is a separate issue to
fix.

> +			prompt_exit();

prompt_exit() is registered as atexit() callback and if I'm not
mistakes will be called anyway.

> +		}
>   	}
>   
> +#ifdef RTE_LIB_PDUMP
> +	/* uninitialize packet capture framework */
> +	rte_pdump_uninit();
> +#endif
> +#ifdef RTE_LIB_LATENCYSTATS
> +	if (latencystats_enabled != 0)
> +		rte_latencystats_uninit();
> +#endif
> +
>   	ret = rte_eal_cleanup();
>   	if (ret != 0)
>   		rte_exit(EXIT_FAILURE,