[libvirt] [PATCH 1/2] tests: commandtest: handle tcmalloc hacking environment

Nikolay Shirokovskiy nshirokovskiy at virtuozzo.com
Fri Nov 17 13:31:13 UTC 2017



On 17.11.2017 16:24, Daniel P. Berrange wrote:
> On Fri, Nov 17, 2017 at 04:17:37PM +0300, Nikolay Shirokovskiy wrote:
>> If one of the libraries is compiled with tcmalloc then
>> the latter will add GLIBCPP_FORCE_NEW and GLIBCXX_FORCE_NEW to
>> environment at startup and thus break commandtest.
> 
> How are they getting those envs into our environment after we clean it
> out ? We strongly aim to prevent any non-whitelisted env variable
> leakage into children we spawn, so I would really like to kill these
> env vars instead of changin the test.

They inserted at process startup I guess [1]. They are cleared out by commandtest
but visible in commandhelper.

[1] https://github.com/gperftools/gperftools/blob/6e3a702fb9c86eb450f22b326ecbceef4b0d6604/src/malloc_extension.cc#L85

> 
>> ---
>>  tests/commandhelper.c | 11 +++++++++--
>>  1 file changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/commandhelper.c b/tests/commandhelper.c
>> index 1da2834..0f6ce07 100644
>> --- a/tests/commandhelper.c
>> +++ b/tests/commandhelper.c
>> @@ -94,8 +94,15 @@ int main(int argc, char **argv) {
>>      for (i = 0; i < n; i++) {
>>          /* Ignore the variables used to instruct the loader into
>>           * behaving differently, as they could throw the tests off. */
>> -        if (!STRPREFIX(newenv[i], "LD_"))
>> -            fprintf(log, "ENV:%s\n", newenv[i]);
>> +        if (STRPREFIX(newenv[i], "LD_"))
>> +            continue;
>> +
>> +        /* Fix tests if tcmalloc is used in libraries */
>> +        if (STRPREFIX(newenv[i], "GLIBCPP_FORCE_NEW=") ||
>> +            STRPREFIX(newenv[i], "GLIBCXX_FORCE_NEW="))
>> +            continue;
>> +
>> +        fprintf(log, "ENV:%s\n", newenv[i]);
>>      }
>>  
>>      open_max = sysconf(_SC_OPEN_MAX);
>> -- 
>> 1.8.3.1
>>
>> --
>> libvir-list mailing list
>> libvir-list at redhat.com
>> https://www.redhat.com/mailman/listinfo/libvir-list
> 
> Regards,
> Daniel
> 




More information about the libvir-list mailing list