- Bird-users - Mailing Lists

Re: BIRD doesn't compile
by David Rohleder 06 Apr '01

06 Apr '01

"James A. Crippen" <james(a)UnLambda.COM> writes: > On Thu, 5 Apr 2001, Florian Lohoff wrote: > > > On Thu, Apr 05, 2001 at 03:48:34PM +0200, David Rohleder wrote: > > > > > > if [ -S /var/bird.ctl ] ; then > > > echo already running > > > fi > > > > > > does the same thing. > > > > > And off course many daemons are using start-stop-daemon > > > > Just by staring at it i saw a couple of ugly things - Must have been > > very early when i actually did it. > > > > The above would only imply that the pipe exists but not actually > > tell you if bird is running. > > > > "birdc" exits with an error code > 0 if bird is not reachable which > > i try to detect. > > > > The problem with the start-stop-daemon is that it needs a pid file > > which when i remember correctly is not written by bird. One > > solution would have been to make a patch to bird (Which i had > > done if i remember) but i dropped it because of lazyness. > > I have my own init system which uses a C program similar to debian's > start-stop-daemon. I've written it such that given a command line flag > the daemon spawning program will write a pid file if you know the daemon > you're running doesn't write one (such as with bird). Once I did this I > had no trouble starting and stopping bird using the pid file > mechanism. Perhaps you might patch debian's start-stop-daemon to do a > similar thing? It's a thought. start-stop-daemon already does this: START-STOP-DAEMON(8) Debian GNU/Linux START-STOP-DAEMON(8) --make-pidfile Used when starting a program that does not create its own pid file. This option will make start-stop- daemon create the file referenced with --pidfile and place the pid into it just before executing the process. Note, it will not be removed when stopping the program. NOTE: This feature may not work in all cases. Most notably when the program being exe cuted forks from it's main process. Because of this it is usually only useful when combined with the --background option. -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

3 2

Re: BIRD doesn't compile
by David Rohleder 05 Apr '01

05 Apr '01

Florian Lohoff <flo(a)rfc822.org> writes: > On Thu, Apr 05, 2001 at 02:29:52PM +0200, David Rohleder wrote: > > > > I was looking for debian package, but i was not able to find it. I was > > quite happy with that, because I wanted to become a debian > > developper. Bad luck this time. I will have to look for another > > package. > > > You are using interesting init.d script :-) I have never seen this before. > > You mean this ? > > if [ -d /proc -a -z "`cat /proc/ksyms | grep rtnetlink`" ]; then > > I wanted to give a resonable output if people dont compile > in "rtnetlink" - I mean if they even skip proc there is no hope. No, I mean this kind of starting and other operations with bird ---- echo quit | birdc >/dev/null 2>&1 if [ $? -eq 0 ]; then echo already running fi ---- if [ -S /var/bird.ctl ] ; then echo already running fi does the same thing. --- And off course many daemons are using start-stop-daemon -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

3 2

Re: BIRD doesn't compile
by David Rohleder 05 Apr '01

05 Apr '01

Florian Lohoff <flo(a)rfc822.org> writes: > On Tue, Mar 06, 2001 at 10:49:18AM +0100, David Rohleder wrote: > > Hello, > > > > i have got a problem compiling BIRD 1.0.4 on newly installed debian > > box with gcc 2.95.3 > > apt-get install bird ? Doesn't work with OSPF area other than backbone. (I posted a patch month ago). I was looking for debian package, but i was not able to find it. I was quite happy with that, because I wanted to become a debian developper. Bad luck this time. I will have to look for another package. You are using interesting init.d script :-) I have never seen this before. > Flo > -- > Florian Lohoff flo(a)rfc822.org +49-5201-669912 > Why is it called "common sense" when nobody seems to have any? -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

3 2

Re: BIRD doesn't compile
by David Rohleder 05 Apr '01

05 Apr '01

Ondrej Feela Filip <feela(a)ipex.cz> writes: > > I was looking for debian package, but i was not able to find it. I was > > quite happy with that, because I wanted to become a debian > > developper. Bad luck this time. I will have to look for another > > package. > > You can maintain a debian package of Bird. Superb, but bird is not orphaned package and Florian Lohoff is still it's maintainer. -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

1 0

IS-IS ?
by Florian Lohoff 05 Apr '01

05 Apr '01

Hi, some people at work think that replacing eigrp by IS-IS is "a good think" (tm). Is there anyone working on IS-IS support for Bird ? Flo -- Florian Lohoff flo(a)rfc822.org +49-5201-669912 Why is it called "common sense" when nobody seems to have any?

1 0

BIRD doesn't compile
by David Rohleder 05 Apr '01

05 Apr '01

Hello, i have got a problem compiling BIRD 1.0.4 on newly installed debian box with gcc 2.95.3 gcc 2.95.2 compiles without errors (but with many warnings). Here is what i have got (seems that #include <time.h> is missing) on 2.95.3 io.c: At top level: io.c:438: warning: missing initializer io.c:438: warning: (near initialization for `sk_class.lookup') io.c: In function `io_init': io.c:1021: warning: implicit declaration of function `time' make[2]: *** [io.o] Error 1 make[2]: Leaving directory `/home/davro/source/bird-1.0.4/obj/lib' make[1]: *** [subdir] Error 2 make[1]: Leaving directory `/home/davro/source/bird-1.0.4/obj' make: *** [all] Error 2 when i added #include <time.h> to several sources, it has passed without errors. But i don't know if this is "the right way" to do it. -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

3 2

Is this an active project?
by Ralph Jennings 02 Apr '01

02 Apr '01

Is this an active project? There seems to be no changes since September 2000? Are all the developers on vacation? Permanently? -Ralph

4 4

About func in md5.c
by Zheng Yuan {ZYUAN1} 07 Mar '01

07 Mar '01

Hi, All maybe i totally dont figure this right, in the func void byteReverse(unsigned char *buf, unsigned longs) { u32 t; do { t = (u32) ((unsigned) buf[3] << 8 | buf[2]) << 16 | ((unsigned) buf[1] << 8 | buf[0]); *(u32 *) buf = t; buf += 4; } while (--longs); } isnt that buf[3] should shift 24?

1 0

patch to OSPF area problem
by David Rohleder 07 Mar '01

07 Mar '01

Hello, I have got a patch, seems, that it works --- packet.c.orig Sun Sep 3 08:53:15 2000 +++ packet.c Wed Mar 7 16:53:36 2001 @@ -131,7 +131,7 @@ return(1); } - if(ps->areaid!=ifa->an) + if(ntohl(ps->areaid)!=ifa->an) { log("%s: Bad OSPF packet received: other area %ld", p->name, ps->areaid); log("%s: Discarding",p->name); -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

1 0

OSPF and strange area ID
by David Rohleder 07 Mar '01

07 Mar '01

Hello, i have got strange problem running OSPF: R ----- F ----- R is Bay Networks VNR F is Debian Linux 2.2 (with kernel 2.4.2) when i try to configure OSPF: protocol ospf MyOSPF { # rfc1583compat yes; area 0.0.0.2 { # stub no; # tick 1; interface "eth0" { cost 500; dead count 4; type broadcast; }; interface "eth1" { cost 550; dead count 4; type broadcast; }; interface "eth2" { stub yes; cost 10; }; interface "eth3" { stub yes; cost 10; }; }; } i am getting strange debug messages: Mar 7 16:32:53 fw5 bird: MyOSPF: Bad OSPF packet received: other area 33554432 Mar 7 16:32:53 fw5 bird: MyOSPF: Discarding and tcpdump shows: 16:29:36.338278 fw5.ics.muni.cz > OSPF-ALL.MCAST.net: OSPFv2-hello 44: rtrid 147.251.8.17 area 0.0.0.2 [|ospf] [tos 0xc0] [ttl 1] 16:29:43.932693 routics-are.brno.cesnet.cz > OSPF-ALL.MCAST.net: OSPFv2-hello 52: rtrid routics-are4.brno.cesnet.cz area 0.0.0.2 [|ospf] [tos 0xc0] [ttl 1] i don't see any neighbour. when i change area to 2.0.0.0 (33554432), bird doesn't drop packets and goes to state init/other with R, but tcpdump shows, that bird sends packets with area 2.0.0.0 which collides with R config. It seems to me, that bird does't do translation from net byte order to host byte order when receiving packets (I am not sure, but area 0.0.0.0 could work :-) -- ------------------------------------------------------------------------- David Rohleder davro(a)ics.muni.cz Institute of Computer Science, Masaryk University Brno, Czech Republic -------------------------------------------------------------------------

1 0