[Openvas-devel] segfault : openvas-nasl 5.0.1

[Sebastien Aucouturier]

Jan,
i may wrong,
but here is what i see in code,

the NVTI values are add  in fonction of OPENVAS-LIBRARIES:
   - 'store_load_plugin'  (libraries/misc/store.c)

'store_load_plugin' is only used in OPENVAS-SCANNER :
   - 'nasl_plugin_add'    (scanner/src/nasl_plugins.c)
   were the NVTI are add to plugin_args and use by 'exec_nasl_script'

it means the online openvas-nasl tools in OPENVAS-LIBRARIES, does not 
initiate NVTI ...
so plugin using nvti went wrong (or crash) when used through CLI 
openvas-nasl  ;-)

can you confirm ?


Le 15/06/2012 10:13, Sebastien Aucouturier a écrit :
> Hi jan,
> i check,
> the nvti cache folder is full, the nmap.nasl.nvti is in the directory
> and fully accessible.
> So i will also check why the nvti pointer remains null.
>
> BR
> Seb
>
>
> Le 15/06/2012 08:17, Jan-Oliver Wagner a écrit :
>> Hi Sebastien,
>>
>> thanks for the hint. I fixed this and several other
>> places with r13600.
>>
>> But this may only solve the segfault and now you may
>> see other interesting problems.
>>
>> The source of the actual problem is somewhere else.
>> The question to follow is: Why is there no NVTI object?
>>
>> Can you check the NVTI cache?
>>
>> Best
>>
>> 	Jan
>>
>> On Dienstag, 12. Juni 2012, Sebastien Aucouturier wrote:
>>> the faulty is in get_plugin_preference function  (misc/plugutils.c)
>>>
>>> with the following code :
>>> char *plug_name = nvti_name (arg_get_value (desc, "NVTI"));
>>>
>>> arg_get_value (desc, "NVTI") return null ...
>>>
>>> so nvti_name segfault ;-)
>>> because :
>>> gchar *
>>> nvti_name (const nvti_t * n)
>>> {
>>>    return (n->name);
>>> }

-- 
     | Sébastien AUCOUTURIER | Software Design Engineer Lead
     | ITrust | 55 rue l'Occitane BP 67303 31673 LABEGE CEDEX
     | Email: s.aucouturier at itrust.fr
     | Fixe Sdt. 05.67.34.67.80 | Fax. 09.80.08.37.23
     | IT Security Services & SaaS Editor