Skip to contentSkip to conversation

rpc: TCONF when tests aren't compiled + remove kill warning

PATCH New Petr Vorel 4 participants 8 messages May 12, 2020 – May 18, 2020

lore
Share & tools
lore

Petr Vorel

author
#0 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Without this test failure was wrongly reported:
rpc_test 1 TFAIL: tirpc_rpcb_getaddr 10.0.0.2 536875000 failed unexpectedly

Also make sure with variable that server has started before trying to kill it.

Kill server with SIGPIPE, should be enough and SIGKILL
produced warning, which might be confused:
/opt/ltp/testcases/bin/rpc_test.sh: line 61: 3827 Killed $SERVER $PROGNUMNOSVC

---
Hi,

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Also added proper detection if kill is needed (related to the change in the commit) and kill with SIGPIPE (unrelated change, I can add it as separate commit).

Kind regards,
Petr

 testcases/network/rpc/rpc-tirpc/rpc_test.sh | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
testcases/network/rpc/rpc-tirpc/rpc_test.sh +4 -2
index 30cfd2564..dc97213d0 100755
@@ -54,12 +54,13 @@ setup()
 	fi
 
 	[ -n "$CLIENT" ] || tst_brk TBROK "client program not set"
+	tst_check_cmds $CLIENT $SERVER || tst_brk TCONF "LTP compiled without TI-RPC support?"
 }
 
 cleanup()
 {
-	if [ ! -z "$SERVER" ]; then
-		pkill -9 $SERVER > /dev/null 2>&1
+	if [ "$SERVER_STARTED" ]; then
+		pkill -13 -x $SERVER
 		$CLEANER $PROGNUMNOSVC
 	fi
 }
@@ -70,6 +71,7 @@ do_test()
 
 	if [ -n "$SERVER" ]; then
 		$SERVER $PROGNUMNOSVC &
+		SERVER_STARTED=1
 
 		for i in $(seq 1 10); do
 			rpcinfo -p localhost | grep -q $PROGNUMNOSVC && break
Signature · 1 line
2.26.2
author
#1 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Hi,

Quoted 5 lines

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Please any comment for 2x TCONF. It's a bit strange, but just a tiny detail. I'd like to have this fix in the release.

Kind regards,
Petr

#2 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Quoted 9 lines · Petr Vorel

Hi,

Quoted 5 lines

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Please any comment for 2x TCONF. It's a bit strange, but just a tiny detail. I'd like to have this fix in the release.

Hi Petr,

Why don't you use tst_cmd_available? like this:

for i in $CLIENT $SERVER; do

??? tst_cmd_available $i || tst_brk TCONF "$i not found.? LTP compiled
without TI-RPC support?"

done

Best Regards,

Xiao Yang

Quoted 5 lines

Kind regards,
Petr

author
#3 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Quoted 8 lines · Petr Vorel
Quoted 7 lines

Hi,

Quoted 5 lines

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Quoted 4 lines
Quoted 2 lines

Please any comment for 2x TCONF. It's a bit strange, but just a tiny detail. I'd like to have this fix in the release.

Hi Petr,

Hi Yang,

Quoted 6 lines

Why don't you use tst_cmd_available? like this:

for i in $CLIENT $SERVER; do

??? tst_cmd_available $i || tst_brk TCONF "$i not found.? LTP compiled without TI-RPC support?"

Thanks for a review.
Yep, I suggested similar solution in v1
https://patchwork.ozlabs.org/project/ltp/patch/[email protected]/

I like solution from v2 (suggested by Alexey) as more elegant, just 2xTCONF is a bit confusing. But I'm hesitating too much about such a tiny detail :).

Kind regards,
Petr

#4 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

In reply to Petr Vorel

Hi!

Quoted 7 lines
Quoted 5 lines

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Please any comment for 2x TCONF. It's a bit strange, but just a tiny detail. I'd like to have this fix in the release.

I do not really care how many TCONF the test prints as long as the user gets a reasonable message.

With that the patch looks good, acked.

Signature · 2 lines
Cyril Hrubis
[email protected]
author
#5 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Hi Cyril,

Quoted 10 lines
Quoted 7 lines
Quoted 5 lines

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Please any comment for 2x TCONF. It's a bit strange, but just a tiny detail. I'd like to have this fix in the release.

I do not really care how many TCONF the test prints as long as the user gets a reasonable message.

Yep, it's a really tiny detail.

Quoted 1 line

With that the patch looks good, acked.

Thanks for your review, pushed.

Kind regards,
Petr

#6 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

In reply to Petr Vorel
Quoted 2 lines · Petr Vorel

Without this test failure was wrongly reported:
rpc_test 1 TFAIL: tirpc_rpcb_getaddr 10.0.0.2 536875000 failed unexpectedly

Hi Petr,

Do you know which caused this failure?

Best Regards,
Xiao Yang

Quoted 57 lines

Also make sure with variable that server has started before trying to kill it.

Kill server with SIGPIPE, should be enough and SIGKILL
produced warning, which might be confused:
/opt/ltp/testcases/bin/rpc_test.sh: line 61: 3827 Killed $SERVER $PROGNUMNOSVC

---
Hi,

with this implementation (suggested by Alexey), we get TCONF twice. It's a bit strange, but I'll keep it, because I like that required binary is printed:
rpc_test 1 TCONF: 'tirpc_rpcb_getaddr' not found
rpc_test 1 TCONF: LTP compiled without TI-RPC support?

Also added proper detection if kill is needed (related to the change in the commit) and kill with SIGPIPE (unrelated change, I can add it as separate commit).

Kind regards,
Petr

  testcases/network/rpc/rpc-tirpc/rpc_test.sh | 6 ++++--
  1 file changed, 4 insertions(+), 2 deletions(-)
testcases/network/rpc/rpc-tirpc/rpc_test.sh +4 -2
index 30cfd2564..dc97213d0 100755
@@ -54,12 +54,13 @@ setup()
  	fi
 
  	[ -n "$CLIENT" ] || tst_brk TBROK "client program not set"
+	tst_check_cmds $CLIENT $SERVER || tst_brk TCONF "LTP compiled without TI-RPC support?"
  }
 
  cleanup()
  {
-	if [ ! -z "$SERVER" ]; then
-		pkill -9 $SERVER>  /dev/null 2>&1
+	if [ "$SERVER_STARTED" ]; then
+		pkill -13 -x $SERVER
  		$CLEANER $PROGNUMNOSVC
  	fi
  }
@@ -70,6 +71,7 @@ do_test()
 
  	if [ -n "$SERVER" ]; then
  		$SERVER $PROGNUMNOSVC&
+		SERVER_STARTED=1
 
  		for i in $(seq 1 10); do
  			rpcinfo -p localhost | grep -q $PROGNUMNOSVC&&  break
author
#7 lore raw

PATCH v2 1/1 rpc: TCONF when tests aren't compiled + remove kill warning

Hi Xiao,

Quoted 6 lines · Petr Vorel
Quoted 2 lines

Without this test failure was wrongly reported:
rpc_test 1 TFAIL: tirpc_rpcb_getaddr 10.0.0.2 536875000 failed unexpectedly

Hi Petr,

Do you know which caused this failure?

Certain tests compiled with TI-RPC on glibc (not on libtirpc): rpc_pmap_set, rpc_pmap_unset, rpc_svc_register. NOTE glibc implementation is fading out, nothing really important.

Before I just disable these tests, I've asked the experts [1] and posted patchset to show some debug info.

Kind regards,
Petr

[1] https://sourceforge.net/p/libtirpc/mailman/message/37014386/
[2] https://patchwork.ozlabs.org/project/ltp/list/?series=177741&state=*