From mboxrd@z Thu Jan  1 00:00:00 1970
From: Alexey Kodanev <alexey.kodanev@oracle.com>
Date: Mon, 19 Nov 2018 20:19:33 +0300
Subject: [LTP] [PATCH v4 1/2] net: Add tst_net_run helper
In-Reply-To: <20181119154004.GB26645@dell5510>
References: <20181113163008.9093-1-pvorel@suse.cz>
 <09aff1dd-a89f-479b-49da-c0c776b08b91@oracle.com>
 <20181116142411.GA24956@dell5510>
 <37a6f27b-096e-8701-b8ee-4132d10ef889@oracle.com>
 <20181119154004.GB26645@dell5510>
Message-ID: <3a454cb5-64d2-357e-2375-d25df6b1478d@oracle.com>
List-Id: <ltp.lists.linux.it>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: ltp@lists.linux.it

Hi Petr,
On 11/19/2018 06:40 PM, Petr Vorel wrote:
> Hi Alexey,
> 
>>>>> Maybe we could signal where error occurred:
>>>>> * return 0: ok on both lhost and rhost
>>>>> * return 1: failure on lhost
>>>>> * return 2: failure on rhost
>>>>> * return 3: failure on both lhost and rhost
> 
> 
>>>> It is only when there is no "safe" option, right?
>>> yes, as safe exit with TCONF and warning.
> 
>>>> when doing cleanup?
>>> IMHO in any "non-safe" run.
> 
>>>> Could we print TWARN message inside tst_net_run() in such case?
>>> I wanted tst_net_run() to behave similarly as tst_rhost_run(), which doesn't
>>> warn about it. So I propose to add warning to both functions, but also add -q
>>> flag (quiet), which suppress this warning.
> 
>> I think we don't need it for tst_rhost_run() as it is clear where the
>> error has happened. Unlike tst_net_run() without safe option.
> OK, make sense, I'll put it just into tst_net_run().
> 
> 
>>> The reason for quiet mode is, that (I suppose) there are cases when some
>>> features is tested and TWARN isn't wanted. Some examples (looks -s is used often
>>> in cleanup functions):
> 
>> '-s' is used often in cleanup with tst_rhost_run()?
> Sorry, I was wrong. Previously I meant tst_rhost_run *without -s* option.
> 
> tst_rhost_run with -s option in cleanup is only in 2 files:
> testcases/network/mpls/mpls_lib.sh
> testcases/network/tcp_cmds/sendfile/sendfile01
> Maybe that's a mistake (unlike C functions where error in SAFE_* continue with
> warning, in shell it exits during cleanup).
> 

For tst_test.sh we could add something like this to tst_brk()

iff --git a/testcases/lib/tst_test.sh b/testcases/lib/tst_test.sh
index 16081fa..e84cf7f 100644
--- a/testcases/lib/tst_test.sh
+++ b/testcases/lib/tst_test.sh
@@ -41,6 +41,7 @@ trap "tst_brk TBROK 'test interrupted'" INT
 _tst_do_exit()
 {
        local ret=0
+       TST_DO_EXIT=1

        if [ -n "$TST_SETUP_STARTED" -a -n "$TST_CLEANUP" -a \
             -z "$TST_NO_CLEANUP" ]; then
@@ -123,6 +124,11 @@ tst_brk()
        local res=$1
        shift

+       if [ "$TST_DO_EXIT" = 1 ]; then
+               tst_res TWARN "$@"
+               return
+       fi
+
        tst_res "$res" "$@"
        _tst_do_exit
 }


> Without -s we don't get any warning, which is often desirable.
> => I'll put warning and -q switch to suppress it only into tst_net_run().
> 
> 
> Kind regards,
> Petr
>