[LTP] [PATCH v1] kernel/pty06: Skip test when TTY not available

[blezhepekov]

On 2022-03-11 17:33, Cyril Hrubis wrote:
> Hi!
>> Since TTYs under test are hardcoded, they
>> might not be available in some systems.
> 
> Is the system you are running the tests on completely without ttys?
> 
>> Signed-off-by: Bogdan Lezhepekov <blezhepekov@suse.de>
>> ---
>>  testcases/kernel/pty/pty06.c | 9 ++++++---
>>  1 file changed, 6 insertions(+), 3 deletions(-)
>> 
>> diff --git a/testcases/kernel/pty/pty06.c 
>> b/testcases/kernel/pty/pty06.c
>> index b621f584f..250c38f6b 100644
>> --- a/testcases/kernel/pty/pty06.c
>> +++ b/testcases/kernel/pty/pty06.c
>> @@ -45,8 +45,6 @@ static struct tst_fzsync_pair fzp;
>> 
>>  static void *open_close(void *unused)
>>  {
>> -	sprintf(tty_path_b, "/dev/tty%d", test_tty_port);
>> -
>>  	while (tst_fzsync_run_b(&fzp)) {
>>  		tst_fzsync_start_race_b(&fzp);
>>  		int fd = SAFE_OPEN(tty_path_b, O_RDWR);
>> @@ -60,7 +58,6 @@ static void *open_close(void *unused)
>> 
>>  static void do_test(void)
>>  {
>> -	sprintf(tty_path_a, "/dev/tty%d", test_tty_port + 1);
>>  	int fd = SAFE_OPEN(tty_path_a, O_RDWR);
>> 
>>  	tst_fzsync_pair_reset(&fzp, open_close);
>> @@ -80,6 +77,12 @@ static void do_test(void)
>> 
>>  static void setup(void)
>>  {
>> +	sprintf(tty_path_a, "/dev/tty%d", test_tty_port + 1);
>> +	sprintf(tty_path_b, "/dev/tty%d", test_tty_port);
>> +
>> +	if (access(tty_path_a, F_OK) != 0 || access(tty_path_b, F_OK) != 0)
>> +		tst_brk(TCONF, "TTY(s) under test not available in system");
> 
> There is no need for the != 0 could be simplified just to:
> 
> 	if (access(...) || access(...))
> 		tst_brk(TCONF, ...);
> 
>>  	tst_fzsync_pair_init(&fzp);
>>  }
>> 
>> --
>> 2.35.1
>> 
>> 
>> --
>> Mailing list info: https://lists.linux.it/listinfo/ltp

Hi Cyril,

Sorry for the delayed reply. I see the test requires an emulated 
terminal (ttyX), whilst our system supports serial console only (ttySX).

What would be your recommendation, does it make sense to switch to 
ttySX? I wasn't sure about it, so disabling seemed more logical.

Thanks,
Bogdan