Hi Ken

I confirm that with your patch, the soak test now pass on a RHEL 5.3
with valgrind 3.2.1.  As you suspected, there were a couple of
supression reports.  I attach them all below.

Thanks again, Keith.

/usr/local/python2.7/bin/python ./tests/python/proton-test
proton_tests.soak.* 2>&1 > valgrind_output_with-patch-from-ken
==11569==
==11569== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_realloc
   fun:CRYPTO_realloc
   fun:lh_insert
   obj:/lib64/libcrypto.so.0.9.8e
   fun:ERR_load_strings
   fun:ERR_load_X509V3_strings
   fun:ERR_load_crypto_strings
   fun:SSL_load_error_strings
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
}
==11569==
==11569== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_malloc
   fun:CRYPTO_malloc
   fun:lh_new
   fun:OBJ_NAME_init
   fun:OBJ_NAME_add
   fun:EVP_add_cipher
   fun:SSL_library_init
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
   fun:main
}
==11626==
==11626== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_realloc
   fun:CRYPTO_realloc
   fun:lh_insert
   obj:/lib64/libcrypto.so.0.9.8e
   fun:ERR_load_strings
   fun:ERR_load_X509V3_strings
   fun:ERR_load_crypto_strings
   fun:SSL_load_error_strings
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
}
==11626==
==11626== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_malloc
   fun:CRYPTO_malloc
   fun:lh_new
   fun:OBJ_NAME_init
   fun:OBJ_NAME_add
   fun:EVP_add_cipher
   fun:SSL_library_init
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
   fun:main
}
==11659==
==11659== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_realloc
   fun:CRYPTO_realloc
   fun:lh_insert
   obj:/lib64/libcrypto.so.0.9.8e
   fun:ERR_load_strings
   fun:ERR_load_X509V3_strings
   fun:ERR_load_crypto_strings
   fun:SSL_load_error_strings
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
}
==11659==
==11659== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_malloc
   fun:CRYPTO_malloc
   fun:lh_new
   fun:OBJ_NAME_init
   fun:OBJ_NAME_add
   fun:EVP_add_cipher
   fun:SSL_library_init
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
   fun:main
}
==11767==
==11767== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_realloc
   fun:CRYPTO_realloc
   fun:lh_insert
   obj:/lib64/libcrypto.so.0.9.8e
   fun:ERR_load_strings
   fun:ERR_load_X509V3_strings
   fun:ERR_load_crypto_strings
   fun:SSL_load_error_strings
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
}
==11767==
==11767== ---- Print suppression ? --- [Return/N/n/Y/y/C/c] ---- y
{
   <insert a suppression name here>
   Memcheck:Leak
   fun:_vgrZU_libcZdsoZa_malloc
   fun:CRYPTO_malloc
   fun:lh_new
   fun:OBJ_NAME_init
   fun:OBJ_NAME_add
   fun:EVP_add_cipher
   fun:SSL_library_init
   fun:pn_ssl_domain
   fun:pn_messenger_tsync
   fun:pn_messenger_sync
   fun:pn_messenger_recv
   fun:main
}


On 4 April 2013 15:07, Ken Giusti <kgiu...@redhat.com> wrote:
> Hi Keith,
>
> This patch:  http://pastebin.com/bSGjm0nN
>
> *should* get valgrind working for older versions.  Can you try it against 
> your 3.2.1 valgrind and let me know (my version of valgrind is 3.5).
>
> If running the above patch fails to suppress any errors on your system, it 
> will prompt you to print out a new suppression.  If that happens, can you 
> print the suppression and send it to me?
>
> thanks in advance
>
> -K
>
> ----- Original Message -----
>> From: "Ken Giusti" <kgiu...@redhat.com>
>> To: proton@qpid.apache.org, "keith wall" <keith.w...@gmail.com>
>> Sent: Thursday, April 4, 2013 8:44:17 AM
>> Subject: Re: Valgrind based soak tests fail against earlier versions of 
>> valgrind (3.2.1)
>>
>> Hi Keith,
>>
>> Ah, so sorry - I'll take a look.
>>
>> My first instinct is to simply disable the valgrind tests for older versions,
>> but let me check if I can fix the suppression in some backward compatible
>> way.
>>
>> /me valgrind noob.
>>
>>
>> ----- Original Message -----
>> > From: "Keith W" <keith.w...@gmail.com>
>> > To: proton@qpid.apache.org
>> > Sent: Thursday, April 4, 2013 7:26:29 AM
>> > Subject: Valgrind based soak tests fail against earlier versions of
>> > valgrind (3.2.1)
>> >
>> > Hello
>> >
>> > We are seeing a the valgrind based soak tests fail on boxes with
>> > earlier versions of valgrind (3.2.1).
>> >
>> >  proton_tests.soak.MessengerTests.test_star_topology_valgrind
>> > ...............................................................................location
>> > should start with fun: or obj:
>> > ==1403== FATAL: in suppressions file
>> > '/home/keith/src/live/qpid-proton/tests/python/proton_tests/valgrind.supp':
>> > location should start with 'fun:' or 'obj:'
>> > ==1403== exiting now.
>> >
>> > Is it possible to avoid the dependency on the newer version, or get
>> > config.sh to check the version of valgrind if it requires a specific
>> > version?
>> >
>> > Thanks in advance, Keith
>> >
>>
>> --
>> -K
>>
>
> --
> -K

Reply via email to