Kernel 2.2.x: PPP routing problem - solved!
To those who were (or weren't) having the same problem as me (PPP freezes in kernel 2.2), I found the answer after talking with Clifford Kite on comp.os.linux.networking (he didn't suggest it, though, he just prompted me to get on with the diagnosing): The modem UART type wasn't set right. On boot, the kernel (2.2) reported ttyS2 to be a 16550A type, which was correct. However, the 0setserial script decided to set the UART type to whatever it was just underneath the manual configuration area (can't remember at present). So, if you've got a problem with PPP freezing under a 2.2 kernel, check your UARTS!! To see if it works for you, do a 'setserial /dev/ttySx uart 16550A', with 'x' being whatever your modem is. Everything will go back to normal on reboot. Anyway, I'm off to reinstall; I kinda messed everything up while I was messing about with everything :-) Yours, A very happy uebergeek Kris :-)
Re: Kernel 2.2.x: PPP routing problem?
I had the same problem.. IP masquerade isn't re-reading the firewall rules
after the link goes down. Creating a null /etc/{wherever}ipmasqueradeislocatedinthisdirectory/ppp
file.
touch /etc/ipmasq/ppp .. That should solve your problem.. ;)
If that doesn't work, uninstall ipmasq because you probably don't need
it anyway.
BTW: STOP REINSTALLING.. :)
Kris wrote:
Problem: PPP works fine under kernel 2.0.36; under
2.2.x it connects, but
it seems like no packets go in or out. tcpdump only shows what looks
like
things trying to get out, but not getting anything back (really technical,
huh?). I can ping myself (localhost local IP), but not _any_
remote IPs
(including my nameservers). All IRQs, nameservers, so on are
set correctly.
PPP is compiled into the kernel (of 2.2.9), and I'm using the latest
version of PPP. It's slink, but I've tried it with apt-get update
upgrade
using the unstable area; makes no difference. Re-installed many times.
Booting from floppy due to a resistant HD/BIOS. I've spent about two
weeks
on this problem, scouring searchlinux.com, deja(news), the HOW-TOs,
Google,
and so on. And I reckon I've spent about 5 on phone bills trying
things.
Here are various outputs while connected (my apologies if everything
gets
munged by Eudora):
# route -n
Destination Gateway
Genmask Flags Metric Ref
Use Iface
195.147.160.5 0.0.0.0 255.255.255.255
UH 0 0
0 ppp0
127.0.0.1 0.0.0.0
255.255.255.255 UH 0 0
0 lo
0.0.0.0 195.147.160.5 0.0.0.0
UG 0 0
0 ppp0
^- I only get loopback when I do 'route add -host 127.0.0.1'
# ifconfig
lo Link encap:Local Loopback
inet addr:127.0.0.1
Mask:255.0.0.0
UP LOOPBACK
RUNNING MTU:3924 Metric:1
RX packets:0
errors:0 dropped:0 overruns:0 frame:0
TX packets:0
errors:0 dropped:0 overruns:0 carrier:0
Collisions:0
ppp0 Link encap:Point-to-Point Protocol
inet addr:195.147.165.134
P-t-P:195.147.160.5 Mask:255.255.255.255
UP POINTOPOINT
RUNNING NOARP MULTICAST MTU:1500 Metric:1
RX packets:13
errors:0 dropped:0 overruns:0 frame:0
TX packets:15
errors:0 dropped:0 overruns:0 carrier:0
Collisions:0
# tcpdump -i ppp0 (single telnet to my POP3 mailbox (there are no problems
at the ISP end: it's fine within Windows))
195.147.165.134.1024 > 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk.
(38)
195.147.165.134.1024 > 194.126.86.9.domain: 55880+ A? mail.globalnet.co.uk.
(38)
195.147.165.134.1024 > 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk.
(38)
195.147.165.134.1024 > 194.126.86.9.domain: 55880+ A? mail.globalnet.co.uk.
(38)
195.147.165.134.1024 > 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk.
(38)
# egrep -v '#|^ *$' /etc/ppp/options
asyncmap 0 auth crtscts
lock hide-password modem
noipdefault -am -vj
debug lcp-echo-interval 30
lcp-echo-failure 4 noipx
-am and -vj make no difference.
# /var/log/ppp.log (start and end trimmed, only showing pppd)
Serial connection established.
Using interface ppp0
Connect: ppp0 --> /dev/ttyS2
sent [LCP ConfReq id=0x1 magic 0x277274c6> pcomp> accomp>]
rcvd [LCP ConfReq id=0x5b asyncmap 0xa> magic 0x1921c20>
pcomp>
accomp>]
sent [LCP ConfRej id=0x5b asyncmap 0xa>]
rcvd [LCP ConfReq id=0x5c magic 0x1921c20> pcomp> accomp>]
sent [LCP ConfAck id=0x5c magic 0x1921c20> pcomp> accomp>]
rcvd [IPCP ConfReq id=0xc6 compress VJ 0f 00> addr 195.147.160.5>]
sent [LCP ConfReq id=0x1 magic 0x277274c6> pcomp> accomp>]
rcvd [LCP ConfReq id=0x5d asyncmap 0xa> magic 0x19227e9>
pcomp>
accomp>]
sent [LCP ConfRej id=0x5d asyncmap 0xa>]
rcvd [LCP ConfAck id=0x1 magic 0x277274c6> pcomp> accomp>]
rcvd [LCP ConfReq id=0x5e magic 0x19227e9> pcomp> accomp>]
sent [LCP ConfAck id=0x5e magic 0x19227e9> pcomp> accomp>]
sent [LCP EchoReq id=0x0 magic=0x277274c6]
sent [IPCP ConfReq id=0x1 addr 0.0.0.0>]
sent [CCP ConfReq id=0x1 deflate 15> deflate(old#) 15> bsd
v1 15>]
rcvd [IPCP ConfReq id=0xc7 compress VJ 0f 00> addr 195.147.160.5>]
sent [IPCP ConfRej id=0xc7 compress VJ 0f 00>]
rcvd [LCP EchoRep id=0x0 magic=0x19227e9]
rcvd [LCP ProtRej id=0x5f 80 fd 01 01 00 0f 1a 04 78 00 18 04 78 00
15 03 2f]
rcvd [IPCP ConfReq id=0xc8 addr 195.147.160.5>]
sent [IPCP ConfAck id=0xc8 addr 195.147.160.5>]
rcvd [IPCP ConfReq id=0xc9 addr 195.147.160.5>]
sent [IPCP ConfAck id=0xc9 addr 195.147.160.5>]
sent [IPCP ConfReq id=0x1 addr 0.0.0.0>]
rcvd [IPCP ConfNak id=0x1 addr 195.147.165.134>]
sent [IPCP ConfReq id=0x2 addr 195.147.165.134>]
rcvd [IPCP ConfAck id=0x2 addr 195.147.165.134>]
local IP address 195.147.165.134
remote IP address 195.147.160.5
Script /etc/ppp/ip-up started; pid = 159
Script /etc/ppp/ip-up finished (pid 159), status = 0x0
#
That's it. If you need more info then just say. I'm really lost with
this,
so if any of you live near me and know how to sort this out, then you
can
have a couple of free pints someday.
Finally, I'm _really_ sorry if this is all a waste of bandwith. (Just
remember to snip plenty if you're kind enough to reply).
Thanks,
Kris
--
Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED]
/dev/null
Re: Kernel 2.2.x: PPP routing problem?
Like a dolt, I didn't even recognize this. I had the exact same problem
that Kris has described, and after a reinstall, I noticed it went away
automagically. Of course, I don't have ipmasq installed this time.
*doh!*
BTW, HTML-mail SUCKS. :)
Sean
Johnny Thompson wrote:
I had the same problem.. IP masquerade isn't re-reading the firewall
rules after the
link goes down. Creating a null
/etc/{wherever}ipmasqueradeislocatedinthisdirectory/ppp file.
touch /etc/ipmasq/ppp .. That should solve your problem.. ;) If that
doesn't work,
uninstall ipmasq because you probably don't need it anyway.
BTW: STOP REINSTALLING.. :)
Re: Kernel 2.2.x: PPP routing problem?
In reply to Johnny Thompson:
I had the same problem.. IP masquerade isn't re-reading
the firewall rules after the link goes down. Creating a
null /etc/{wherever}ipmasqueradeislocatedinthisdirectory/ppp
.file.
touch /etc/ipmasq/ppp .. That should solve your problem.. ;)
If that doesn't work, uninstall ipmasq because you probably
don't need it anyway.
I don't have IP masquerading installed/set up... I always do a base install
(using the basic 28 megs option), and add what I need after that -- which
fits neatly into less than 100Mb (~78 now I think).
I've done a dpkg -l .gz and whacked it up on my web space, just in case you
don't believe me :-)
http://www.users.globalnet.co.uk/~dufas/dpkg.gz (a massive 3.12k)
ipchains/SSH will be next on my list when I sort this thing out.
Good idea, but... nah.
BTW: STOP REINSTALLING.. :)
Sorry, can't help it :-) I like having a neat filesystem. Compared to what
my room looks like right now that's actually bit strange ;-)
Au revoir...
Kris
PS. Don't blame me if anything gets HTMLised in these e-mails -- Eudora
sucks when it doesn't ask me ONLY plain text *ever*?
Re: Kernel 2.2.x: PPP routing problem?
{blush} You can tell I follow the digest now... darn. Apologies for the
waste of bandwidth, but it's for archive continuity (and so people actually
know wts(meg) I was going on about). Re-send, in reply to Sean
[EMAIL PROTECTED], with the correct subject:
snip suggestion about killing inetd, as it worked for Sean
Nope. inetd has been well and truly slaughtered, and nothing changes (even
renamed it to make sure it didn't get restarted... call me strange).
Output of 'ps ax' while connected (after I've killed gpm, portmap inetd):
init
[kflushd]
[kpiod]
[kswapd]
update
/sbin/syslogd
/sbin/klogd
/usr/sbin/atd
/usr/sbin/cron
-bash
-bash
/sbin/getty 38400 tty3
/sbin/getty 38400 tty4
/sbin/getty 38400 tty5
/sbin/getty 38400 tty6
tail -f /var/log/debug
/usr/sbin/pppd call provider
sh /etc/ppp/ip-up ppp0 /dev/ttyS2 115200 195.147.135.
run-parts /etc/ppp/ip-up.d
sh /etc/ppp/ip-up.d/exim
/usr/sbin/exim -qf
ps ax
Makes no difference whether I kill gpm/portmap or not. I do notice that the
last digits are missing from the IP, though... but that's probably just
running off the end of the screen or something (80 chars in total,
including all the other columns, just as it would on-screen). I think Exim
is waiting for my smarthost domain name to resolve (or I just didn't stay
on long enough for it to finish anyway).
I also let the 0setserial script do my serial ports by itself -- it got my
IRQs wrong, and set the I/0 to 0x03e8 (which is right), whereas I set it to
0x3e8 (or the other way round; can't remember). That doesn't matter, though.
My HD is getting seriously warped with all this rebooting...
Anyone got any other plausible ideas?
Thanks,
Kris
(Oh, and I changed /etc/hosts - splodge.org. Dunno why I said that.)
Re: Kernel 2.2.x: PPP routing problem?
In reply to John Pearson: These [Exim things in ps ax output] should run briefly when you start your session, and then be gone. If they are still running it may be that there is a problem with exim's setup that is preventing it from quitting in a timely manner, or it may just be a further symptom of the fact that your PPP connection isn't working. What happens if you kill /usr/sbin/exim-qf? I have now re-installed Slink (Cheapbytes Official 2.1 CDs) completely, and I'm using a minimum install (basic + gpm, make, patch, lynx, etc, with no X) using my 2.2.9 kernel from floppy with no updates/upgrades. I've changed: fstab, 0setserial (ttyS2 - IRQ 5), and /etc/ppp/options (added novjccomp, nodeflate, nobsdcomp. Removed asyncmap other stuff (no difference before and after, so it's not any of that)). Also commented out all lines /etc/ppp/ip-up.d/exim and removed the 15-min exim cron job. Also tried it with killing inetd, portmap, and gpm. So exim doesn't run unless I tell it to. No difference. Theories: pppd bug or kernel 2.2 bug. If anyone wants me to use various kdebug options (I'm guessing that will reveal a grand total of zero bytes) and upload the output to my web space in a .gz for easy zless viewing, I shall. Just ask. I honestly think of anything else :-( Next step: comp.os.linux.networking tomorrow, and/or the linux-ppp mailing list after that. I feel as though I'm interrupting everyone... Thanks for trying, and goodnight... Kris
Re: Kernel 2.2.x: PPP routing problem?
At 19:35 16/05/1999 +, you wrote: After I got everything installed, I immediately downloaded, installed and ran apt-get, and promptly upgraded to Slink, just as I had done when Slink first went stable. And now I can't use any of the 2.0.x kernels AND have ppp work. I've tried everything that has been mentioned in this thread, recompiling, moving back to the 2.2.7 kernel, which had just been working fine on my previous Slink installation, etc all to no avail. This is really driving me batty, but I thought you might like to know (Kris) that it isn't just your machine. I've just about decided to try a complete reformat/reinstall, as I've run out of just about all other options. Nice to know I'm not alone :-) I also thought about just using the 2.0.36 kernel, but as I have a dual PPro system... Git ;-) I'm on a lowly Cyrix 166 :-( Oh yes, something else I forgot to mention. I also put the proposed-updates link in my sources.list file, and upgraded to the packages present at that location. As I had not done this previously, I suspect this is where the problem is, so I'm thinking about reinstalling, upgrading, and then NOT upgrading again with the proposed-updates. I've tried it with a virgin Slink install, proposed-updates, and then did an apt-get update; apt-get install with libc6, netbase so on using proposed-updates and the potato unstable directories. Nothing changed (tried at all stages). My next guess: incorrect serial port settings (mouse ttyS1 modem ttyS2) them manually for my modem IRQ). I'll see what letting it set them automatically does. 95% chance it'll do nothing (it works in 2.0.36 as I said), but I can try. Otherwise, it's to the newsgroups/linux-ppp. But I can't try it now. 02.24am. C++ assigments suck. I'll get back to everyone tomorrow on what the serial port autodetect does. Thanks, Kris
Re: Kernel 2.2.x: PPP routing problem?
A quick kludge that might fix your problem (it fixed it on my machine under hamm, and after I get done upgrading to slink, I'm going to try it there) is to kill inetd. I think the problem revolves around the netbase script in the /etc/init.d folder trying to use ifwadm(or something like that) instead of ifchains(or something like that). I still don't understand why this works or why the problem has never happened to me before now with the 2.2 kernels under both hamm and slink, but thought this info might be of use to you. Sean My next guess: incorrect serial port settings (mouse ttyS1 modem ttyS2) them manually for my modem IRQ). I'll see what letting it set them automatically does. 95% chance it'll do nothing (it works in 2.0.36 as I said), but I can try. Otherwise, it's to the newsgroups/linux-ppp. But I can't try it now. 02.24am. C++ assigments suck. I'll get back to everyone tomorrow on what the serial port autodetect does. Thanks, Kris
Re: Kernel 2.2.x: PPP routing problem?
On 17-May-99 Sean wrote: A quick kludge that might fix your problem (it fixed it on my machine under hamm, and after I get done upgrading to slink, I'm going to try it there) is to kill inetd. I think the problem revolves around the netbase script in the /etc/init.d folder trying to use ifwadm(or something like that) instead of ifchains(or something like that). I still don't understand why this works or why the problem has never happened to me before now with the 2.2 kernels under both hamm and slink, but thought this info might be of use to you. Interesting, because when it happened to me, I thought I had set up ipchains incorrectly, until diald started to work even with the same ipchains rules. -- Andrew Sean My next guess: incorrect serial port settings (mouse ttyS1 modem ttyS2) them manually for my modem IRQ). I'll see what letting it set them automatically does. 95% chance it'll do nothing (it works in 2.0.36 as I said), but I can try. Otherwise, it's to the newsgroups/linux-ppp. But I can't try it now. 02.24am. C++ assigments suck. I'll get back to everyone tomorrow on what the serial port autodetect does. Thanks, Kris -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Kernel 2.2.x: PPP routing problem?
Subject: Kernel 2.2.x: PPP routing problem? Date: Sat, May 15, 1999 at 05:46:14PM +0100 In reply to:Kris Quoting Kris([EMAIL PROTECTED]): Problem: PPP works fine under kernel 2.0.36; under 2.2.x it connects, but it seems like no packets go in or out. tcpdump only shows what looks like things trying to get out, but not getting anything back (really technical, huh?). I can ping myself (localhost local IP), but not _any_ remote IPs (including my nameservers). All IRQs, nameservers, so on are set correctly. PPP is compiled into the kernel (of 2.2.9), and I'm using the latest version of PPP. It's slink, but I've tried it with apt-get update upgrade using the unstable area; makes no difference. Re-installed many times. Booting from floppy due to a resistant HD/BIOS. I've spent about two weeks on this problem, scouring searchlinux.com, deja(news), the HOW-TOs, Google, and so on. And I reckon I've spent about £5 on phone bills trying things. Here are various outputs while connected (my apologies if everything gets munged by Eudora): [ snip ] # egrep -v '#|^ *$' /etc/ppp/options asyncmap 0 authcrtscts lock hide-password modem noipdefault-am -vj debug lcp-echo-interval 30 lcp-echo-failure 4 noipx I have a working ppp on kernel 2.2.9. my pppd command is exec /usr/sbin/pppd /dev/ttyS0 115200 lock modem crtscts \ asyncmap 0 defaultroute connect $DIALER_SCRIPT pppd --version (pppd version 2.3 patch level 5) egrep -v '#|^ *$' /etc/ppp/options domain mindspring.com mru 1500 mtu 1500 name login-name I don't recall making any changes to the above when going to the 2.2.x kernel. The only differences I can see are what i show above. HTH Wayne -- Bringing computers into the home won't change either one, but may revitalize the corner saloon. ___ Wayne T. Topa [EMAIL PROTECTED]
Re: Kernel 2.2.x: PPP routing problem?
At 20:21 15/05/1999 +, you wrote: Ppp is known to have problems with 2.2.X. Indeed it does. *thwap* How about the far end of the ppp link? Nope; the only thing I can ping or otherwise communicate with is localhost the dynamic IP which I'm assigned (from /var/log/ppp.log). Option names have changes and the old ones may not be supported anymore. To turn off compression use 'nobsdcomp', 'nodeflate', and 'novj'. I no longer recall what '-am' did. Check the man page. As a long shot, try turning off all compression. I've tried nobsdcomp, nodeflate, novjc, novjccomp, and probably a few other things; nothing cures it. I reckon that even if the compression stuff wasn't set right, _some_ packets should make their way in/out. Incidentaly, I earlier tried upgrading to potato (libc6 so on) and recompiling the kernel to include anything which looks even remotely route-looking; no luck (made the man output have a Perl error though; nothing serious). I'm starting to think that I've either got a curse put on me my by an evil goat, or I'm just dmed to failure. Any other ideas, anyone? (Pretty please?)
Re: Kernel 2.2.x: PPP routing problem?
At 08:32 16/05/1999 +, you wrote: I have a working ppp on kernel 2.2.9. my pppd command is exec /usr/sbin/pppd /dev/ttyS0 115200 lock modem crtscts \ asyncmap 0 defaultroute connect $DIALER_SCRIPT pppd --version (pppd version 2.3 patch level 5) egrep -v '#|^ *$' /etc/ppp/options domain mindspring.com mru 1500 mtu 1500 name login-name I don't recall making any changes to the above when going to the 2.2.x kernel. The only differences I can see are what i show above. Nope. None of that makes any difference. I've now tried downgrading to PPP(d) 2.3.5, and upgrading to 2.3.8. I don't think it's anything to do with DNS, but here are a few things: # cat /etc/resolv.conf domain globalnet.co.uk # - Tried with, without, search globalnet.co.uk # - and combinations of. nameserver 194.126.82.5 nameserver 194.126.86.9 # /etc/hosts 127.0.0.1 splodge localhost loopback Minus loopback, no change. Splodge is current hostname. # /etc/inetd.conf discard daytime timetalk ntalk shell login exec smtpident inetd.conf shouldn't matter if it works in 2.0.36. Someone might need a good LARTing over this but it's probably just me. Still seeking help on this... Thanks, Kris
Re: Kernel 2.2.x: PPP routing problem?
It's not just you. A couple of days ago I decided that I had pretty much butchered my Debian installation (which I had installed close to a year previous), so I decided to do the whole reformat/reinstall maneuver (after backing up my /home directory, of course). As I now have the Debian 2.0 CDs, the process was pretty painless. After I got everything installed, I immediately downloaded, installed and ran apt-get, and promptly upgraded to Slink, just as I had done when Slink first went stable. And now I can't use any of the 2.0.x kernels AND have ppp work. I've tried everything that has been mentioned in this thread, recompiling, moving back to the 2.2.7 kernel, which had just been working fine on my previous Slink installation, etc all to no avail. This is really driving me batty, but I thought you might like to know (Kris) that it isn't just your machine. I've just about decided to try a complete reformat/reinstall, as I've run out of just about all other options. I also thought about just using the 2.0.36 kernel, but as I have a dual PPro system the SMP performance just isn't there. Oh yes, something else I forgot to mention. I also put the proposed-updates link in my sources.list file, and upgraded to the packages present at that location. As I had not done this previously, I suspect this is where the problem is, so I'm thinking about reinstalling, upgrading, and then NOT upgrading again with the proposed-updates. Sean -Original Message- From: Kris [EMAIL PROTECTED] To: debian-user@lists.debian.org debian-user@lists.debian.org Date: Sunday, May 16, 1999 10:58 AM Subject: Re: Kernel 2.2.x: PPP routing problem? At 08:32 16/05/1999 +, you wrote: I have a working ppp on kernel 2.2.9. my pppd command is exec /usr/sbin/pppd /dev/ttyS0 115200 lock modem crtscts \ asyncmap 0 defaultroute connect $DIALER_SCRIPT pppd --version (pppd version 2.3 patch level 5) egrep -v '#|^ *$' /etc/ppp/options domain mindspring.com mru 1500 mtu 1500 name login-name I don't recall making any changes to the above when going to the 2.2.x kernel. The only differences I can see are what i show above. Nope. None of that makes any difference. I've now tried downgrading to PPP(d) 2.3.5, and upgrading to 2.3.8. I don't think it's anything to do with DNS, but here are a few things: # cat /etc/resolv.conf domain globalnet.co.uk # - Tried with, without, search globalnet.co.uk # - and combinations of. nameserver 194.126.82.5 nameserver 194.126.86.9 # /etc/hosts 127.0.0.1 splodge localhost loopback Minus loopback, no change. Splodge is current hostname. # /etc/inetd.conf discard daytime time talk ntalk shell login exec smtp ident inetd.conf shouldn't matter if it works in 2.0.36. Someone might need a good LARTing over this but it's probably just me. Still seeking help on this... Thanks, Kris -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Kernel 2.2.x: PPP routing problem?
Kris writes: # cat /etc/resolv.conf domain globalnet.co.uk # - Tried with, without, search globalnet.co.uk # - and combinations of. nameserver 194.126.82.5 nameserver 194.126.86.9 The 'domain' and 'search' directives won't affect your problem (In fact they are rarely needed at all). # /etc/hosts 127.0.0.1 splodge localhost loopback Minus loopback, no change. Splodge is current hostname. 'loopback' just adds a nickname for localhost. You don't need it. You do want an fqdn in hosts, though. I use: 127.0.0.1 hasler.dhh hasler localhost This won't affect your problem either, though. inetd.conf shouldn't matter if it works in 2.0.36. inetd.conf shouldn't matter at all for this problem. Still seeking help on this... Wish I could be more help. -- John HaslerThis posting is in the public domain. [EMAIL PROTECTED] Do with it what you will. Dancing Horse Hill Make money from it if you can; I don't mind. Elmwood, Wisconsin Do not send email advertisements to this address.
Kernel 2.2.x: PPP routing problem?
Problem: PPP works fine under kernel 2.0.36; under 2.2.x it connects, but it seems like no packets go in or out. tcpdump only shows what looks like things trying to get out, but not getting anything back (really technical, huh?). I can ping myself (localhost local IP), but not _any_ remote IPs (including my nameservers). All IRQs, nameservers, so on are set correctly. PPP is compiled into the kernel (of 2.2.9), and I'm using the latest version of PPP. It's slink, but I've tried it with apt-get update upgrade using the unstable area; makes no difference. Re-installed many times. Booting from floppy due to a resistant HD/BIOS. I've spent about two weeks on this problem, scouring searchlinux.com, deja(news), the HOW-TOs, Google, and so on. And I reckon I've spent about £5 on phone bills trying things. Here are various outputs while connected (my apologies if everything gets munged by Eudora): # route -n Destination Gateway Genmask Flags Metric Ref Use Iface 195.147.160.5 0.0.0.0 255.255.255.255 UH0 0 0 ppp0 127.0.0.1 0.0.0.0 255.255.255.255 UH0 0 0 lo 0.0.0.0 195.147.160.5 0.0.0.0 UG0 0 0 ppp0 ^- I only get loopback when I do 'route add -host 127.0.0.1' # ifconfig loLink encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:3924 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 Collisions:0 ppp0 Link encap:Point-to-Point Protocol inet addr:195.147.165.134 P-t-P:195.147.160.5 Mask:255.255.255.255 UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1500 Metric:1 RX packets:13 errors:0 dropped:0 overruns:0 frame:0 TX packets:15 errors:0 dropped:0 overruns:0 carrier:0 Collisions:0 # tcpdump -i ppp0 (single telnet to my POP3 mailbox (there are no problems at the ISP end: it's fine within Windows)) 195.147.165.134.1024 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk. (38) 195.147.165.134.1024 194.126.86.9.domain: 55880+ A? mail.globalnet.co.uk. (38) 195.147.165.134.1024 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk. (38) 195.147.165.134.1024 194.126.86.9.domain: 55880+ A? mail.globalnet.co.uk. (38) 195.147.165.134.1024 194.126.82.5.domain: 55880+ A? mail.globalnet.co.uk. (38) # egrep -v '#|^ *$' /etc/ppp/options asyncmap 0 authcrtscts lock hide-password modem noipdefault-am -vj debug lcp-echo-interval 30 lcp-echo-failure 4 noipx -am and -vj make no difference. # /var/log/ppp.log (start and end trimmed, only showing pppd) Serial connection established. Using interface ppp0 Connect: ppp0 -- /dev/ttyS2 sent [LCP ConfReq id=0x1 magic 0x277274c6 pcomp accomp] rcvd [LCP ConfReq id=0x5b asyncmap 0xa magic 0x1921c20 pcomp accomp] sent [LCP ConfRej id=0x5b asyncmap 0xa] rcvd [LCP ConfReq id=0x5c magic 0x1921c20 pcomp accomp] sent [LCP ConfAck id=0x5c magic 0x1921c20 pcomp accomp] rcvd [IPCP ConfReq id=0xc6 compress VJ 0f 00 addr 195.147.160.5] sent [LCP ConfReq id=0x1 magic 0x277274c6 pcomp accomp] rcvd [LCP ConfReq id=0x5d asyncmap 0xa magic 0x19227e9 pcomp accomp] sent [LCP ConfRej id=0x5d asyncmap 0xa] rcvd [LCP ConfAck id=0x1 magic 0x277274c6 pcomp accomp] rcvd [LCP ConfReq id=0x5e magic 0x19227e9 pcomp accomp] sent [LCP ConfAck id=0x5e magic 0x19227e9 pcomp accomp] sent [LCP EchoReq id=0x0 magic=0x277274c6] sent [IPCP ConfReq id=0x1 addr 0.0.0.0] sent [CCP ConfReq id=0x1 deflate 15 deflate(old#) 15 bsd v1 15] rcvd [IPCP ConfReq id=0xc7 compress VJ 0f 00 addr 195.147.160.5] sent [IPCP ConfRej id=0xc7 compress VJ 0f 00] rcvd [LCP EchoRep id=0x0 magic=0x19227e9] rcvd [LCP ProtRej id=0x5f 80 fd 01 01 00 0f 1a 04 78 00 18 04 78 00 15 03 2f] rcvd [IPCP ConfReq id=0xc8 addr 195.147.160.5] sent [IPCP ConfAck id=0xc8 addr 195.147.160.5] rcvd [IPCP ConfReq id=0xc9 addr 195.147.160.5] sent [IPCP ConfAck id=0xc9 addr 195.147.160.5] sent [IPCP ConfReq id=0x1 addr 0.0.0.0] rcvd [IPCP ConfNak id=0x1 addr 195.147.165.134] sent [IPCP ConfReq id=0x2 addr 195.147.165.134] rcvd [IPCP ConfAck id=0x2 addr 195.147.165.134] local IP address 195.147.165.134 remote IP address 195.147.160.5 Script /etc/ppp/ip-up started; pid = 159 Script /etc/ppp/ip-up finished (pid 159), status = 0x0 # That's it. If you need more info then just say. I'm really lost with this, so if any of you live near me and know how to sort this out, then you can have a couple of free pints someday. Finally, I'm _really_ sorry if this is all a waste of bandwith. (Just remember to snip plenty if you're kind enough to reply). Thanks, Kris
Re: Kernel 2.2.x: PPP routing problem?
Kris writes: Problem: PPP works fine under kernel 2.0.36; under 2.2.x it connects, but it seems like no packets go in or out. Ppp is known to have problems with 2.2.X. ...not _any_ remote IPs (including my nameservers). How about the far end of the ppp link? -am and -vj make no difference. Option names have changes and the old ones may not be supported anymore. To turn off compression use 'nobsdcomp', 'nodeflate', and 'novj'. I no longer recall what '-am' did. Check the man page. As a long shot, try turning off all compression. -- John HaslerThis posting is in the public domain. [EMAIL PROTECTED] Do with it what you will. Dancing Horse Hill Make money from it if you can; I don't mind. Elmwood, Wisconsin Do not send email advertisements to this address.
