[9fans] PPP error: cannot write IP - link not up

[9fans] PPP error: cannot write IP - link not up

[Stevie_Lancaster]

hi!
I have a problem with the direct PPPoE connection to my ISP.
I have a realtek 8139 NIC and wrote ether=type=rtl8139 in my
plan9.ini.
The kernel knows my NIC at startup. Until there everything is
allright.

There is no /net/ipifc/0 when the system just started and when I go on
and type:

ip/pppoe -dP
ndb/cs
ndb/dns -r

the error gets reproted in my ppp logfile.
PPPoE cannot write the IP. That is clear to me, because the interface
is not up.

If I bind the ether0 manually ( echo bind ether /net/ether0 > /net/
ipifc/clone) a "0" interface gets created. But PPPoE creates another
one "1". when I cat status of "1" there is no ether device.
Under "1" everything is there. Also my dynamic IP adress, there is
just no device for all that stuff...

What can I do to tell ip/pppoe to just use my formerly created "0"
interface?

Thanks for your help.

lancaster

Re: [9fans] PPP error: cannot write IP - link not up

[Vester Thacker]

On 3/5/07, Stevie_Lancaster <teenageriot@gmx.de> wrote:

> What can I do to tell ip/pppoe to just use my formerly created "0"
> interface?

Before running "ip/pppoe -dP", ensure "bind -b '#l0' /net" has been executed.

HTH,

Vester Thacker

[9fans] Re: PPP error: cannot write IP - link not up

[Stevie_Lancaster]

On 5 Mrz., 11:21, vester.thac...@gmail.com (Vester Thacker) wrote:
> On 3/5/07, Stevie_Lancaster <teenager...@gmx.de> wrote:
>
> > What can I do to tell ip/pppoe to just use my formerly created "0"
> > interface?
>
> Before running "ip/pppoe -dP", ensure "bind -b '#l0' /net" has been executed.
>
> HTH,
>
> Vester Thacker

Thanks for your reply. Under /net there is already a directory ether0.
So I guess the device got binded there. Am I right?
Are bind -a and bind -b in this case the same?

lancaster

Re: [9fans] Re: PPP error: cannot write IP - link not up

[Vester Thacker]

On 3/6/07, Stevie_Lancaster <teenageriot@gmx.de> wrote:

> Thanks for your reply. Under /net there is already a directory ether0.
> So I guess the device got binded there. Am I right?

I believe it to mean that the device was recognized, but it does not
mean that it was binded.

> Are bind -a and bind -b in this case the same?

Yes.

Regards,
Vester Thacker

Re: [9fans] Re: PPP error: cannot write IP - link not up

[erik quanstrom]

[-- Attachment #1: Type: text/plain, Size: 1093 bytes --]

On Tue Mar  6 04:55:44 EST 2007, vester.thacker@gmail.com wrote:
> On 3/6/07, Stevie_Lancaster <teenageriot@gmx.de> wrote:
> 
> > Thanks for your reply. Under /net there is already a directory ether0.
> > So I guess the device got binded there. Am I right?
> 
> I believe it to mean that the device was recognized, but it does not
> mean that it was binded.

i guess this is pedantic but, yes it does mean the device is bound
into the namespace and in most cases you can send and receive
raw ethernet frames.

i've attached an overly-complicated program which prints out the
ip configuration by interface.  output is like so:

	; ipifc
	0:/net/ether0 	00a0c973fa3f	1514
		205.185.197.2/120	205.185.197.0
	0:/net.alt/ether1 	000102ed3163	1514
		65.14.39.132/123	65.14.39.128

the first line is the ipifc index, a ':', the mountpoint, mac and mtu.
subsequent indented lines are  address, a '/', netmask and
network for each configured address.  you can get the same
information with

	; cat cat /net*/ipifc/?/status>[2=]

but it's a lot harder to read. ;-)

- erik

[-- Attachment #2: ipifc.c --]
[-- Type: text/plain, Size: 1141 bytes --]

#include <u.h>
#include <libc.h>
#include <ip.h>

char *mtab[10] = {"/net", "/net.alt"};
int nmtab = 2;

void 
ipifcprint(char *m, int idx)
{
	Ipifc 	*ifc;
	Iplifc	*l;
	uchar 	eth[6];

	for(ifc = readipifc(m, 0, idx); ifc; ifc = ifc->next){
		print("%d:%s ", ifc->index, ifc->dev);
		if(myetheraddr(eth, ifc->dev) == 0)
			print("\t"	"%E", eth);
		print("\t"	"%d\n", ifc->mtu);
		for(l = ifc->lifc; l; l = l->next)
			print("\t" "%I%M" "\t" "%I\n", l->ip, l->mask, l->net);
	}

//	freeipifc(ifc);
}

void
usage(void)
{
	fprint(2, "usage: ipfic [-m net mtpt] [interface number] ...\n");
	exits("usage");
}

void
main(int argc, char **argv){
	char* r;
	int i, m;

	m = 0;
	ARGBEGIN{
	case 'm':
		if(m++ == 0)
			nmtab = 0;
		if(++nmtab < nelem(mtab))
			mtab[nmtab-1] = EARGF(usage());
		break;
	}ARGEND;

	fmtinstall('I', eipfmt);
	fmtinstall('M', eipfmt);
	fmtinstall('E', eipfmt);
	if(*argv)
		for(; *argv; argv++){
			i = strtoul(*argv, &r, 0);
			if (*r)
				usage();
			ipifcprint(mtab[0], i);
		}
	else
		for(i = 0; i < nmtab; i++)
			ipifcprint(mtab[i], -1);

	exits(0);
}

Re: [9fans] Re: PPP error: cannot write IP - link not up

[Vester Thacker]

On 3/6/07, erik quanstrom <quanstro@coraid.com> wrote:
>
> i guess this is pedantic but, yes it does mean the device is bound
> into the namespace and in most cases you can send and receive
> raw ethernet frames.

I do not consider it pedantic at all. Thank you very much for the correction!
My apologies for the conceptual error.

--Vester