[#6684] some libstdc++ case fails to be rshed to the target board for timeout after dejagnu updated to 1.5

Document created by Aaronwu Employee on Oct 17, 2013
Version 1Show Document
  • View in full screen mode

[#6684] some libstdc++ case fails to be rshed to the target board for timeout after dejagnu updated to 1.5

Submitted By: Mingquan Pan

Open Date

2011-07-11 04:10:38     Close Date

2012-05-15 01:31:54

Priority:

Medium     Assignee:

Stuart Henderson

Board:

N/A     Silicon Revision:

Resolution:

Assigned (Not Started)     Fixed In Release:

N/A

Processor:

ALL     

Host Operating System:

toolchain rev.:

    kernel rev.:

State:

Closed     Found In Release:

2011R1

Is this bug repeatable?:

N/A     

Summary: some libstdc++ case fails to be rshed to the target board for timeout after dejagnu updated to 1.5

Details:

 

some libstdc++ case fails to be rshed to the target board for timeout after dejagnu updated to 1.5.

 

#########################################

#

# Comparing libstdc++-4.3.sum

#

########################################

Tests that now fail, but worked before:

 

bfin-uclinux: 23_containers/set/modifiers/16728.cc execution test

bfin-uclinux: ext/pb_ds/regression/hash_data_map_rand.cc execution test

bfin-uclinux: ext/pb_ds/regression/tree_data_map_rand.cc execution test

bfin-uclinux: ext/pb_ds/regression/tree_no_data_map_rand.cc execution test

bfin-uclinux: ext/pb_ds/regression/trie_data_map_rand.cc execution test

bfin-uclinux: ext/pb_ds/regression/trie_no_data_map_rand.cc execution test

bfin-uclinux: tr1/5_numerical_facilities/special_functions/20_riemann_zeta/check_value_neg.cc execution test

bfin-uclinux: tr1/5_numerical_facilities/special_functions/20_riemann_zeta/check_value_pos.cc execution test

 

 

detailed log is like:

Executing on host: bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0  -DLOCALEDIR="." -I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util /home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/ext/pb_ds/regression/tree_no_data_map_rand.cc   -DPB_DS_REGRESSION ./libtestc++.a -Wl,-elf2flt=-s80000  -lm  -mcpu=bf527-0.2  -o ./tree_no_data_map_rand.exe    (timeout = 600)

spawn bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0 -DLOCALEDIR="." -I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util /home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/ext/pb_ds/regression/tree_no_data_map_rand.cc -DPB_DS_REGRESSION ./libtestc++.a -Wl,-elf2flt=-s80000 -lm -mcpu=bf527-0.2 -o ./tree_no_data_map_rand.exe^M

PASS: ext/pb_ds/regression/tree_no_data_map_rand.cc (test for excess errors)

Executing on bfin-uclinux: /tmp/tree_no_data_map_rand.exe.27442    (timeout = 300)

spawn [open ...]^M

WARNING: program timed out.

Executing on bfin-uclinux: rm -f  /tmp/tree_no_data_map_rand.exe.27442    (timeout = 300)

spawn [open ...]^M

XYZ0ZYX

Executed ./tree_no_data_map_rand.exe, status 1

rsh to bfin-uclinux failed for /tmp/tree_no_data_map_rand.exe.27442,

FAIL: ext/pb_ds/regression/tree_no_data_map_rand.cc execution test

extra_tool_flags are:

-DPB_DS_REGRESSION

 

 

Follow-ups

 

--- Mingquan Pan                                             2011-09-05 22:43:28

Since these case execables fails for timeout which need to investigate test by

test. I post the left cases found here.

 

Tests that now fail, but worked before:

 

bfin-linux-uclibc: gcc.c-torture/execute/20011008-3.c execution,  -O0

bfin-linux-uclibc: gcc.c-torture/execute/20011008-3.c execution,  -O1

 

 

Executing on host: bfin-linux-uclibc-gcc

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/gcc/testsuite/gcc.c-torture/execute/20011008-3.c

-w  -O0  -DSTACK_SIZE=0x1f000 -DNO_TRAMPOLINES -fno-show-column  -lm

-mcpu=bf561-0.5  -o

/home/test/work/cruise/temp/regtest_build4.3/gcc_build/testsuite/20011008-3.x0 

(timeout = 300)

PASS: gcc.c-torture/execute/20011008-3.c compilation,  -O0

Executing on bfin-linux-uclibc: /tmp/20011008-3.x0.22149    (timeout = 300)

WARNING: program timed out.

Executing on bfin-linux-uclibc: rm -f  /tmp/20011008-3.x0.22149    (timeout =

300)

XYZ0ZYX

Executed

/home/test/work/cruise/temp/regtest_build4.3/gcc_build/testsuite/20011008-3.x0,

status 1

rsh to bfin-linux-uclibc failed for /tmp/20011008-3.x0.22149,

FAIL: gcc.c-torture/execute/20011008-3.c execution,  -O0

Executing on host: bfin-linux-uclibc-gcc

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/gcc/testsuite/gcc.c-torture/execute/20011008-3.c

-w  -O1  -DSTACK_SIZE=0x1f000 -DNO_TRAMPOLINES -fno-show-column  -lm

-mcpu=bf561-0.5  -o

/home/test/work/cruise/temp/regtest_build4.3/gcc_build/testsuite/20011008-3.x1 

(timeout = 300)

PASS: gcc.c-torture/execute/20011008-3.c compilation,  -O1

Executing on bfin-linux-uclibc: /tmp/20011008-3.x1.22149    (timeout = 300)

WARNING: program timed out.

Executing on bfin-linux-uclibc: rm -f  /tmp/20011008-3.x1.22149    (timeout =

300)

XYZ0ZYX

Executed

/home/test/work/cruise/temp/regtest_build4.3/gcc_build/testsuite/20011008-3.x1,

status 1

rsh to bfin-linux-uclibc failed for /tmp/20011008-3.x1.22149,

FAIL: gcc.c-torture/execute/20011008-3.c execution,  -O1

 

--- Mingquan Pan                                             2011-09-05 22:59:41

now sees more libstdc++ case fails for running timeout .

 

#########################################

#

# Comparing libstdc++-4.3.sum

#

########################################

Tests that now fail, but worked before:

 

...

bfin-linux-uclibc: thread/pthread1.cc execution test

bfin-linux-uclibc: thread/pthread2.cc execution test

bfin-linux-uclibc: thread/pthread3.cc execution test

bfin-linux-uclibc: thread/pthread4.cc execution test

bfin-linux-uclibc: thread/pthread6.cc execution test

bfin-linux-uclibc: thread/pthread7-rope.cc execution test

 

Executing on host: bfin-linux-uclibc-g++ -g -O2 -D_GLIBCXX_ASSERT

-fmessage-length=0  -DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/thread/pthread1.cc

  -pthread ./libtestc++.a  -lm  -mcpu=bf561-0.5  -o ./pthread1.exe    (timeout =

600)

PASS: thread/pthread1.cc (test for excess errors)

Executing on bfin-linux-uclibc: /tmp/pthread1.exe.31537    (timeout = 300)

WARNING: program timed out.

Executing on bfin-linux-uclibc: rm -f  /tmp/pthread1.exe.31537    (timeout =

300)

XYZ0ZYX

Executed ./pthread1.exe, status 1

rsh to bfin-linux-uclibc failed for /tmp/pthread1.exe.31537,

FAIL: thread/pthread1.cc execution test

extra_tool_flags are:

-pthread

Executing on host: bfin-linux-uclibc-g++ -g -O2 -D_GLIBCXX_ASSERT

-fmessage-length=0  -DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/thread/pthread2.cc

  -pthread ./libtestc++.a  -lm  -mcpu=bf561-0.5  -o ./pthread2.exe    (timeout =

600)

PASS: thread/pthread2.cc (test for excess errors)

Executing on bfin-linux-uclibc: /tmp/pthread2.exe.31537    (timeout = 300)

WARNING: program timed out.

Executing on bfin-linux-uclibc: rm -f  /tmp/pthread2.exe.31537    (timeout =

300)

XYZ0ZYX

Executed ./pthread2.exe, status 1

rsh to bfin-linux-uclibc failed for /tmp/pthread2.exe.31537,

FAIL: thread/pthread2.cc execution test

extra_tool_flags are:

-pthread

 

--- Stuart Henderson                                         2011-09-16 06:51:38

judging by the various message threads i've found, there's no good way to

increase the remote timeout in dejagnu.  i've asked on the mailing list, but i'm

yet to get a reply.

 

if you want to avoid these timeouts you can make the following change to your

local dejagnu.  if i get a response from the dejagnu mailing with a better

solution, i'll let you know.  however there shouldn't be any harm in making this

change anyway:

 

diff --git a/runtest.exp b/runtest.exp

index 2140020..5ec9f02 100755

--- a/runtest.exp

+++ b/runtest.exp

@@ -1214,7 +1214,7 @@ if { $verbose > 2 } {

     log_user 0

}

 

-set timeout 10

+set timeout 300

 

 

--- Mingquan Pan                                             2011-09-28 22:40:15

Though I have changed the dejagnu runtest.exp as you said above. It looks this

change only works to one case,like :

 

Executing on host: bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0

-DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/23_containers/set/modifiers/16728.cc

    ./libtestc++.a -Wl,-elf2flt=-s80000  -lm  -mcpu=bf561-0.5  -o ./16728.exe  

(timeout = 600)

PASS: 23_containers/set/modifiers/16728.cc (test for excess errors)

Executing on bfin-uclinux: /tmp/16728.exe.20842    (timeout = 300)

XYZ0ZYX

Executing on bfin-uclinux: rm -f  /tmp/16728.exe.20842    (timeout = 300)

XYZ0ZYX

Executed ./16728.exe, status 0

PASS: 23_containers/set/modifiers/16728.cc execution test

extra_tool_flags are:

 

 

The left cases that mentioned above are still run with rsh failure.

 

Executing on host: bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0

-DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/ext/pb_ds/regression/hash_data_map_rand.cc

  -DPB_DS_REGRESSION ./libtestc++.a -Wl,-elf2flt=-s80000  -lm  -mcpu=bf561-0.5

-o ./hash_data_map_rand.exe    (timeout = 600)

PASS: ext/pb_ds/regression/hash_data_map_rand.cc (test for excess errors)

Executing on bfin-uclinux: /tmp/hash_data_map_rand.exe.20842    (timeout =

300)

WARNING: program timed out.

Executing on bfin-uclinux: rm -f  /tmp/hash_data_map_rand.exe.20842    (timeout

= 300)

XYZ0ZYX

Executed ./hash_data_map_rand.exe, status 1

rsh to bfin-uclinux failed for /tmp/hash_data_map_rand.exe.20842,

FAIL: ext/pb_ds/regression/hash_data_map_rand.cc execution test

extra_tool_flags are:

-DPB_DS_REGRESSION

Executing on host: bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0

-DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/ext/pb_ds/regression/hash_no_data_map_rand.cc

  -DPB_DS_REGRESSION ./libtestc++.a -Wl,-elf2flt=-s80000  -lm  -mcpu=bf561-0.5

-o ./hash_no_data_map_rand.exe    (timeout = 600)

PASS: ext/pb_ds/regression/hash_no_data_map_rand.cc (test for excess errors)

Executing on bfin-uclinux: /tmp/hash_no_data_map_rand.exe.20842    (timeout =

300)

XYZ0ZYX

WARNING: program timed out.

Executing on bfin-uclinux: rm -f  /tmp/hash_no_data_map_rand.exe.20842  

(timeout = 300)

XYZ0ZYX

Executed ./hash_no_data_map_rand.exe, status 1

rsh to bfin-uclinux failed for /tmp/hash_no_data_map_rand.exe.20842,

FAIL: ext/pb_ds/regression/hash_no_data_map_rand.cc execution test

extra_tool_flags are:

-DPB_DS_REGRESSION

Executing on host: bfin-uclinux-g++ -g -O2 -D_GLIBCXX_ASSERT -fmessage-length=0

-DLOCALEDIR="."

-I/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/util

/home/test/work/cruise/checkouts/toolchain/gcc-4.3/libstdc++-v3/testsuite/ext/pb_ds/regression/list_update_data_map_rand.cc

  -DPB_DS_REGRESSION ./libtestc++.a -Wl,-elf2flt=-s80000  -lm  -mcpu=bf561-0.5

-o ./list_update_data_map_rand.exe    (timeout = 600)

 

And I wonder the timeout parameter shown in the log previously posted has

already printed as 300s, just the same as the timeout value that in the log

after the change?

 

 

--- Stuart Henderson                                         2011-09-29 06:54:58

yep, i found that the number printed is largely irrelevant in our case.  300s is

5 minutes and i was getting about 10 timeouts (thus 50 minutes), yet the test

run was completing in about 40minutes.  there's something broken there, and i

believe the dejagnu guys are looking at it.

 

as for why the fix is not working for you, i'm afraid i don't know.  either

have a dig about yourself and see if you can find what timeout value affects

change in the results, or wait until i get a response from the dejagnu people.

either way, it's a testing issue and not a problem with the toolchain.

 

--- Sonic Zhang                                              2011-09-30 00:32:13

But, we can't release the toolchain without fixing this testing issue.

 

--- David Gibson                                             2011-09-30 03:25:27

If it's a testing issue rather than a genuine bug, I don't see why we can't

release and document the problem in the release note?

 

--- Mingquan Pan                                             2011-09-30 03:53:24

I have tried to run some of the case manually like following which is passed and

cost only about 1 m.

 

root:/> date

Fri Sep 30 15:56:16 UTC 2011

root:/> ./hash_data_map_rand.exe

root:/> echo $?

0

root:/> date

Fri Sep 30 15:57:17 UTC 2011

root:/>

root:/> version

kernel:    Linux release 3.0.0-ADI-2011R1-pre-svn10094, build #86 Fri Sep 30

11:09:48 GMT 2011

toolchain: bfin-uclinux-gcc release gcc version 4.3.5 (ADI-trunk/svn-5683)

user-dist: release svn-10427, build #12 Fri Sep 30 11:09:07 GMT 2011

root:/>

 

So it looks this issue lies in dejagnu itself.

I can have a try to find where goes wrong but run the whole libstdc++ case is

too long, do you know how to run the specific case alone?

 

--- Mingquan Pan                                             2011-10-11 05:14:01

After trying with the split case independently, the patch you gived works. I

previously apply it to the default path while actually it used another local

path after upgrading.

So the libstdc++ binaries above can rsh and pass now without timing out.

 

As to this bug, it can be left open until real fix arrives.

 

--- Stuart Henderson                                         2012-03-20 10:20:15

i'm still not seeing these failures.  are you still seeing them?

 

--- Mingquan Pan                                             2012-05-15 02:31:55

ok. This workaround fix this issue, these case can be rshed now. So close.

 

 

 

    Files

    Changes

    Commits

    Dependencies

    Duplicates

    Associations

    Tags

 

File Name     File Type     File Size     Posted By

No Files Were Found

Attachments

    Outcomes