Re: [mpls-linux-devel] Quagga compile error

[Marco S. <mar...@s2...>]

The command-line is the same Miguel tried:
./configure --enable-vtysh --enable-mpls --enable-netlink
and a make afterwards -> leads to those declaration errors below.

I have to add, that I changed the compiler-flag in the configure
script from std=c99 to std=gnu99 because of an earlier compiling-error:
http://www.gossamer-threads.com/lists/quagga/dev/13305#13305

----------
Now to the stability of zebra:
The network-topology I use, is the same of mpls-linux-labs (very
helpful, indeed). I've experimented with
LDP-enabeled-Link-Protection and experienced a few strange things.
http://www.elcom.pub.ro/~adrian.popa/mpls-linux/mpls-linux-docs/4-1-link_protection.html

First of all, today LSR E1 made the same problems Gonzalo mentions.
All Routes are learned during OSPF-Phase, and although label mappings
are recieved via LDP (can be seen via vtysh), E1 didn't send any labeled
packets... All other Routers are doing almost ;-) fine.
I'll try to help isolating this error, but I had no time to investigate
this any further, because I wanted to show you that zebra crashes.

All zebra debug options are enabled, these are the last few lines
(complete log is appended):
2007/03/16 15:04:43 ZEBRA: netlink_talk: mpls-netlink-cmd type (29), seq=13
2007/03/16 15:04:43 ZEBRA: netlink_parse_info: mpls-netlink-cmd ACK: type=(29), seq=13, pid=0
2007/03/16 15:04:43 ZEBRA: netlink_parse_info: mpls-netlink-listen type (29), seq=0, pid=0
2007/03/16 15:04:43 ZEBRA: zebra message comes from socket [16]
2007/03/16 15:04:43 ZEBRA: zebra message received [ZEBRA_MPLS_FTN_ADD] 11
2007/03/16 15:04:43 ZEBRA: zebra message comes from socket [16]
2007/03/16 15:04:43 ZEBRA: zebra message received [ZEBRA_MPLS_XC_ADD] 11
ZEBRA: Received signal 11 at 1174053883 (si_addr 0x64, PC 0x80537cd); aborting...
Program counter: /usr/sbin/zebra[0x80537cd]
Backtrace for 9 stack frames:
/usr/lib/quagga/libzebra.so.0(zlog_backtrace_sigsafe+0x47)[0xd3c267]
/usr/lib/quagga/libzebra.so.0(zlog_signal+0x272)[0xd3c882]
/usr/lib/quagga/libzebra.so.0[0xd49caa]
[0x8a8440]
/usr/sbin/zebra[0x80537cd]
/usr/lib/quagga/libzebra.so.0(thread_call+0x3f)[0xd2d3ff]
/usr/sbin/zebra(main+0x3ab)[0x80543cb]
/lib/libc.so.6(__libc_start_main+0xdc)[0xa7a4e4]
/usr/sbin/zebra[0x8050721]

This happened at LSR E2 after (I'm not completely sure about the
moment) disconnecting a link from E4 to E3. Pings are being sent from
A2 to A3 all the time, so the packets would have to find the other
possible way. I'll try it again on monday without enabling LDP.

How do I configure the LSRs? I've taken all advice from the mailing
list. The daemons are started with a little script in the following
order:
modprobe mpls4
service ... start (zebra, ospfd, ldpd)
-> the conf files are chown'ed with quagga.quagga, they only contain
-> information about the interfaces and ospf subnets, ldpd.conf is
-> almost empty
vtysh

In vtysh I type:
conf t -> mpls ldp -> ex
int ethx -> mpls l 0 -> mpls ip -> end
.

Regards,
Marco

> On Thu, Mar 15, 2007 at 07:49:03PM +0100, Marco Schmidt wrote:
> <snip>
>> 
>> This happens, if I try to compile the patched quagga-sources from sf
>> on my FC6 machine. I've got no other MPLS Packages installed on this
>> machine (it is used for other work), perhaps you forgot to install the
>> glibc-kernheaders from sf?
>>
>> I'm also trying to compile Quagga, because the zebra-daemon is not as
>> stable as I wished... Which snapshot has been used for the RPM at sf?
>> I encounter the following errors on my FC5 MPLS machines:

> I'm using http://quagga.net/download/quagga-0.99.6.tar.gz as the base,
> with my nexthop and MPLS framework changes which are significant.

> Can you elaborate on "zebra-daemon is not as stable"?

> What configure command line are you using before building quagga that results
> in the errors below?

>> 
>>     In file included from /usr/include/sys/uio.h:24,
>>                     from /usr/include/sys/socket.h:27,
>>                     from /usr/include/linux/mpls.h:32,
>>                     from mpls_netlink.c:1:
>>     /usr/include/sys/types.h:62: error: conflicting types for 'dev_t'
>>     /usr/include/linux/types.h:10: error: previous declaration of 'dev_t' was here
>>     /usr/include/sys/types.h:67: error: conflicting types for 'gid_t'
>>     /usr/include/linux/types.h:21: error: previous declaration of 'gid_t' was here
>>     /usr/include/sys/types.h:72: error: conflicting types for 'mode_t'
>>     /usr/include/linux/types.h:12: error: previous declaration of 'mode_t' was here
>>     /usr/include/sys/types.h:77: error: conflicting types for 'nlink_t'
>>     /usr/include/linux/types.h:13: error: previous declaration of 'nlink_t' was here
>>     /usr/include/sys/types.h:82: error: conflicting types for 'uid_t'
>>     /usr/include/linux/types.h:20: error: previous declaration of 'uid_t' was here
>>     In file included from /usr/include/sys/types.h:220,
>>                      from /usr/include/sys/uio.h:24,
>>                      from /usr/include/sys/socket.h:27,
>>                      from /usr/include/linux/mpls.h:32,
>>                      from mpls_netlink.c:1:
>>     /usr/include/sys/select.h:78: error: conflicting types for 'fd_set'
>>     /usr/include/linux/types.h:9: error: previous declaration of 'fd_set' was here
>>     make[2]: *** [mpls_netlink.o] Error 1
>>     make[2]: Leaving directory `/usr/src/redhat/BUILD/quagga-0.99.6/zebra'
>>     make[1]: *** [all-recursive] Error 1
>>     make[1]: Leaving directory `/usr/src/redhat/BUILD/quagga-0.99.6'
>>     make: *** [all] Error 2
>> 
>> Does anyone know, how to solve these declaration-problems?
>> 
>> Regards,
>> Marco