PPTPD works fine unencrypted, I use NETBIOS and FTP to move files over the tunnel, but when I use MPPE-128 it seems to be an buffer overflow when I send data more then a couple of KB. So I can browse a folder via NETBIOS with 10 files/folders but when I try to browse a folder with 50 files/folders it just hangs. Same goes for FTP, so it's probably not a NETBIOS problem.
Here is some log from an session when the datastream hangs on trying to browse a folder with 50 files
pptpd[11224]: CTRL: Client 217.215.208.3 control connection started
pptpd[11224]: CTRL: Starting call (launching pppd, opening GRE)
pppd[11225]: Plugin /usr/lib/pptpd/pptpd-logwtmp.so loaded.
pptpd[11224]: GRE: Bad checksum from pppd.
pppd[11225]: pppd 2.4.3 started by root, uid 0
pppd[11225]: Using interface ppp0
pppd[11225]: Connect: ppp0 <--> /dev/pts/2
/etc/hotplug/net.agent: assuming ppp0 is already up
pptpd[11224]: CTRL: Ignored a SET LINK INFO packet with real ACCMs!
pppd[11225]: MPPE 128-bit stateless compression enabled
pppd[11225]: found interface eth0 for proxy arp
pppd[11225]: local IP address 192.168.5.4
pppd[11225]: remote IP address 192.168.5.5
pptpd[11224]: GRE: Discarding duplicate packet
last message repeated 4 times
pptpd[11224]: GRE: Discarding duplicate packet
last message repeated 8 times
pppd[11225]: LCP terminated by peer (cM-(^A^A^@<M-Mt^@^@^@^@)
pppd[11225]: Connect time 1.5 minutes.
pppd[11225]: Sent 32043 bytes, received 26070 bytes.
kernel: application bug: pptpctrl(11224) has SIGCHLD set to SIG_IGN but calls wait().
kernel: (see the NOTES section of 'man 2 wait'). Workaround activated.
pppd[11225]: Modem hangup
pppd[11225]: Connection terminated.
pppd[11225]: Exit.
pptpd[11224]: CTRL: Client 217.215.208.3 control connection finished
/etc/hotplug/net.agent: NET unregister event not supported
I run redhat 9 and I've installed like all different versions of mppe, ppp,and pptpd, mostley from rpm and src.rpm, they all give me the same bug. One thing I haven't tried is recompiling kernel.
Hope someone has any idea where the problem is and what the solution is.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The difference when connecting with Win2k from WinXP was the MTU of the pppX interface. When connection with WinXP it was set to 1396 and with Win2K to 1496.
For some reason it doesn't work setting the mtu in /etc/ppp/options.pptpd so I added the following to /etc/ppp/ip-up.local
!/bin/bash
PATH=/sbin:/usr/sbin:/bin:/usr/bin
export PATH
ifconfig $1 mtu 1500
which set the mtu on the starting pppX interface. this will affect all ppp devices, but as i'm not using dialup this works for me.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
PPTPD works fine unencrypted, I use NETBIOS and FTP to move files over the tunnel, but when I use MPPE-128 it seems to be an buffer overflow when I send data more then a couple of KB. So I can browse a folder via NETBIOS with 10 files/folders but when I try to browse a folder with 50 files/folders it just hangs. Same goes for FTP, so it's probably not a NETBIOS problem.
Here is some log from an session when the datastream hangs on trying to browse a folder with 50 files
pptpd[11224]: CTRL: Client 217.215.208.3 control connection started
pptpd[11224]: CTRL: Starting call (launching pppd, opening GRE)
pppd[11225]: Plugin /usr/lib/pptpd/pptpd-logwtmp.so loaded.
pptpd[11224]: GRE: Bad checksum from pppd.
pppd[11225]: pppd 2.4.3 started by root, uid 0
pppd[11225]: Using interface ppp0
pppd[11225]: Connect: ppp0 <--> /dev/pts/2
/etc/hotplug/net.agent: assuming ppp0 is already up
pptpd[11224]: CTRL: Ignored a SET LINK INFO packet with real ACCMs!
pppd[11225]: MPPE 128-bit stateless compression enabled
pppd[11225]: found interface eth0 for proxy arp
pppd[11225]: local IP address 192.168.5.4
pppd[11225]: remote IP address 192.168.5.5
pptpd[11224]: GRE: Discarding duplicate packet
last message repeated 4 times
pptpd[11224]: GRE: Discarding duplicate packet
last message repeated 8 times
pppd[11225]: LCP terminated by peer (cM-(^A^A^@<M-Mt^@^@^@^@)
pppd[11225]: Connect time 1.5 minutes.
pppd[11225]: Sent 32043 bytes, received 26070 bytes.
kernel: application bug: pptpctrl(11224) has SIGCHLD set to SIG_IGN but calls wait().
kernel: (see the NOTES section of 'man 2 wait'). Workaround activated.
pppd[11225]: Modem hangup
pppd[11225]: Connection terminated.
pppd[11225]: Exit.
pptpd[11224]: CTRL: Client 217.215.208.3 control connection finished
/etc/hotplug/net.agent: NET unregister event not supported
I run redhat 9 and I've installed like all different versions of mppe, ppp,and pptpd, mostley from rpm and src.rpm, they all give me the same bug. One thing I haven't tried is recompiling kernel.
Hope someone has any idea where the problem is and what the solution is.
I realized it's a problem with WinXP client. When i tried with Win2K it worked fine with MPPE-128.
Anyone knows if there is a patch for XP or for pptpd to fix this problem ?
The difference when connecting with Win2k from WinXP was the MTU of the pppX interface. When connection with WinXP it was set to 1396 and with Win2K to 1496.
For some reason it doesn't work setting the mtu in /etc/ppp/options.pptpd so I added the following to /etc/ppp/ip-up.local
!/bin/bash
PATH=/sbin:/usr/sbin:/bin:/usr/bin
export PATH
ifconfig $1 mtu 1500
which set the mtu on the starting pppX interface. this will affect all ppp devices, but as i'm not using dialup this works for me.