[pisa] Re: Bug in hipd
- From: Dongsu Park <Dongsu.Park@xxxxxxxxxxxxxx>
- To: pisa@xxxxxxxxxxxxx
- Date: Tue, 16 Sep 2008 14:09:58 +0200
Hi Wolfram,
sorry, there has been an another bug in my patch 151. I resolved it and
committed as patch-152. Try to test it with '--disable-privsep'. I'm
sure you'll get no problem any more.
By the way I'll report the first bug about capability, where
'--disable-privsep' is sometimes needed, to HIPL Bugzilla system.
Cheers,
Dongsu
Dongsu Park wrote:
> First, for userspace branch, you can avoid this problem by giving a
> '--disable-privsep' option for configure script. That works for every
> revision in the userspace branch.
>
> However, for midauth branch, that way does not solve the problem.
> hipconf hangs at some point in any case, even if you configure with
> '--disable-privsep' option.
>
> This is a subtle aspect of the problem. There could be two possible
> reasons. Either because there is an another bug only for midauth branch,
> or because some part is not imported from userspace to midauth.
>
> Dongsu Park wrote:
>> Hi,
>>
>> hipl--userspace--2.6--patch-1610 works, while patch-1611 does not work.
>> But its fix would not be trivial. Let's see...
>>
>> Dongsu Park wrote:
>>> Hi Wolfram,
>>>
>>> patch-141 does not work, while patch-140 works.
>>> The thing is however, that patch-141 is a set of huge amount of patches,
>>> which makes it hard to find out the real reason.
>>>
>>> Dongsu
>>>
>>> Wolfram Fischer wrote:
>>>> Hi Dongsu,
>>>>
>>>> I just checked the problem with the hipd. As tcpdump doesn't show any
>>>> sign of traffic over the interfaces, this seems to be a local error.
>>>> This behavior must have been introduced between patch-140 and
>>>> patch-current, because 140 works fine without problems and current
>>>> doesn't work and ended at this point in my two runs with gdb.
>>>>
>>>> Perhaps you could use patch-current and add some debug message before
>>>> and after the call to recv in message.c:37?
>>>>
>>>> At the moment I can't add any entries to Bugzilla (already wrote Miika
>>>> a mail about it) - I'll add this as soon as I get a login.
>>>>
>>>> Following the backtrace, after stopping the hipd with CTRL-C.
>>>>
>>>> Cheers,
>>>> Wolf
>>>>
>>>> [New Thread 0xb7b5b6f0 (LWP 3095)]
>>>>
>>>> Program received signal SIGINT, Interrupt.
>>>> [Switching to Thread 0xb7b5b6f0 (LWP 3095)]
>>>> 0xffffe402 in __kernel_vsyscall ()
>>>> (gdb) bt
>>>> #0 0xffffe402 in __kernel_vsyscall ()
>>>> #1 0xb7f60c91 in recv () from /lib/tls/i686/cmov/libpthread.so.0
>>>> #2 0x080afa6a in hip_peek_recv_total_len (socket=13, encap_hdr_size=0)
>>>> at message.c:37
>>>> #3 0x080b0252 in hip_send_recv_daemon_info (msg=0x80f9608) at
>>>> message.c:228
>>>> #4 0x080aed17 in hip_conf_handle_handover (msg=0x80f9608, action=16,
>>>> opt=0xbfa4342c, optc=0) at hipconf.c:1689
>>>> #5 0x080ae535 in hip_do_hipconf (argc=3, argv=0xbfa43424, send_only=1)
>>>> at hipconf.c:1568
>>>> #6 0x0809c157 in hip_conf_handle_load (msg=0x0, action=11,
>>>> opt=0xbfa436e4,
>>>> optc=1) at getendpointinfo.c:2253
>>>> #7 0x08085f66 in hip_load_configuration () at init.c:165
>>>> #8 0x080866ff in hipd_init (flush_ipsec=1, killold=0) at init.c:364
>>>> #9 0x08057655 in hipd_main (argc=1, argv=0xbfa43b94) at hipd.c:393
>>>> #10 0x08058a95 in main (argc=1, argv=0xbfa43b94) at hipd.c:717
>>>>
>>>
>>
>>
>
>
Other related posts:
