New oleg firmware version

[DrChair]

since the libk8055 doesn't work with libusb-compat and doesnt work with libusb-0.1.12 on new oleg, I ported it to libusb1.0.

I tested the ported source on i368 ubuntu -> here it works with libusb-1.0

However on the asus it doesn't work. Enabling debug mode, gives me the error: [calculate_timeout] failed to read monotonic clock.

So I patched io.c from libusb-1.0 and changed all occurences of usbi_backend->clock_gettime in to clock get_time. (like it used to be)

This solves the error. However now running the program again cold boots my router.

Reverting to old oleg and the program now gives the error:
libusb:warning [handle_bulk_completion] unrecognised urb status -130
libusb:warning [do_sync_bulk_transfer] unrecognised status code 1
intr error -99

I'm starting to get lost on this one

[greengreen]

HI I'm using lastest oleg's firmware and I tried to run passive ftp over vsftpd. Everythink is O.K. but I have to open one specific port for this connestion forever. But I think that firewall could open this port automaticly without hard wite rule in iptables. i think it depend on any module
modprobe ip_tables
modprobe ip_conntrack
modprobe ip_conntrack_ftp
modprobe ip_conntrack_irc
modprobe iptable_nat
modprobe ip_nat_ftp

Do you know how to run this?Is possible to run it?
thanks a lot

[lly]

since the libk8055 doesn't work with libusb-compat and doesnt work with libusb-0.1.12 on new oleg, I ported it to libusb1.0.

I tested the ported source on i368 ubuntu -> here it works with libusb-1.0

This should sounds like "works with libusb-1.0 + kernel 2.6"!

However on the asus it doesn't work. Enabling debug mode, gives me the error: [calculate_timeout] failed to read monotonic clock.

Since r379 bundled libusb-1.0 already has similar patch

However now running the program again cold boots my router.

Probably, bug in kernel. Which firmware build you use? Can problem be reproduced without specific hardware? If yes, please provide modified sources of libk8055 & simple test case, we will try to fix it.

[Lesiuk]

You enabled IMQ in newest revision.

Is it possible to do some QoS with it?

[lly]

You enabled IMQ in newest revision.

Is it possible to do some QoS with it?

It is under tests now, but I hope, yes. Unfortunately, there will no web-GUI until someone contribute it.

[wpte]

is there any chance that a ppp_mppe module can be added to the kernel?

[Lesiuk]

It is under tests now, but I hope, yes. Unfortunately, there will no web-GUI until someone contribute it.

I do not need webui, i like do everything in console.

It's great news.

Maybe you need some help?
Do you use jabber/xmpp ?

[lly]

is there any chance that a ppp_mppe module can be added to the kernel?

CONFIG_PPP_MPPE_MPPC=y is set for years in Oleg-based firmwares...
Maybe I understand your wrong?

[wpte]

CONFIG_PPP_MPPE_MPPC=y is set for years in Oleg-based firmwares...
Maybe I understand your wrong?

sorry, I didn't knew it was enabled already

[DrChair]

Probably, bug in kernel. Which firmware build you use? Can problem be reproduced without specific hardware? If yes, please provide modified sources of libk8055 & simple test case, we will try to fix it.

Oke, for now I assume libusb-1.0 (and compat-libusb) isn't full compatible with the current kernel.
So let's try to find out why the the libusb-0.1 version crashes the router.

If I remove the USB device, the program doesn't crash the router. So I guess you need the specific hardware.
However I can reproduce the problem with an usb-mouse (with also uses interrupt transfers)

I'll PM you the details...

Anyway, I spend the whole day compiling different revisions of the firmware.

R334 is the last one that works (with the libusb-0.1 version)
R335 and R336 don't compile
R337 has the problem.

So the problem is caused by "Add missed urb->interval initialization for usbdevfs"

[lly]

R334 is the last one that works (with the libusb-0.1 version)
R335 and R336 don't compile
R337 has the problem.

So the problem is caused by "Add missed urb->interval initialization for usbdevfs"

Problem reproduced, thanks to report.
As workaround you simply can exclude kernel/242-usb-devio.patch from custom build process. This patch repairs EHCI(2.0), but broke UHCI(1.1) drivers interface via usbdevfs ioctl calls. It will take some time to find proper solution.

[cichy]

Great FW with great options
I wonder if there's a possibility to add a file called version in /etc/ directory with actual firmware version and a name of OS the router is running (i know there's a .version in / dir but it's read-only). I'm asking that coz i found something like that http://phpsysinfo.sourceforge.net/ and it gives great info about router (i added nslu icon and it's being recognized as Oleg's Firmware after finding /etc/version file). I know that this is something stupid but i think it's quite useful

[lly]

I wonder if there's a possibility to add a file called version in /etc/ directory with actual firmware version and a name of OS the router is running (i know there's a .version in / dir but it's read-only).

/etc directory is writeable, as in all Oleg FW. File "/.version" must be read-only. OS version you can get from uname utility or from /proc/version

[lly]

In r461 we add experimental accel-pptp kernel driver, which should significantly improve PPTP speed. Please help us to test it.

To turn it on you have to do:

nvram set wan0_pptp_kernel=1

nvram commit is superfluous, since there is not needed that driver will start after reboot or crash.
After this, push "Disconnect", wait 10 sec, push "Connect" in web-GUI.

Syslog should have lines like below:

Aug 18 08:39:26 wan: connected manually
Aug 18 08:39:29 kernel: vlan1: Setting MAC address to  xx xx xx xx xx xx.
Aug 18 08:39:29 post-firewall: started [ppp0 0.0.0.0 br0 192.168.0.1 vlan1 xxx.xxx.xxx.xxx]
Aug 18 08:39:30 pptp[258]: Plugin pptp.so loaded.
Aug 18 08:39:30 pptp[258]: PPTP plugin version 0.8.3 compiled for pppd-2.4.5, linux-2.4.37.5
Aug 18 08:39:30 pptp[259]: pppd 2.4.5 started by admin, uid 0
Aug 18 08:39:30 pptp[259]: PPTP: failed to setsockopt PPTP_SO_TIMEOUT (Operation not supported)
Aug 18 08:39:30 pptp[260]: anon log[callmgr_main:pptp_callmgr.c:132]: IP: xxx.xxx.xxx.xxx
Aug 18 08:39:30 pptp[260]: anon log[callmgr_main:pptp_callmgr.c:136]: control connection
Aug 18 08:39:30 pptp[260]: anon log[callmgr_main:pptp_callmgr.c:140]: unix_sock
Aug 18 08:39:30 pptp[261]: anon log[ctrlp_rep:pptp_ctrl.c:251]: Sent control packet type is 1 'Start-Control-Connection-Request'
Aug 18 08:39:30 pptp[261]: anon log[ctrlp_disp:pptp_ctrl.c:737]: Received Start Control Connection Reply
Aug 18 08:39:30 pptp[261]: anon log[ctrlp_disp:pptp_ctrl.c:771]: Client connection established.
Aug 18 08:39:31 pptp[261]: anon log[ctrlp_rep:pptp_ctrl.c:251]: Sent control packet type is 7 'Outgoing-Call-Request'
Aug 18 08:39:31 pptp[261]: anon log[ctrlp_disp:pptp_ctrl.c:856]: Received Outgoing Call Reply.
Aug 18 08:39:31 pptp[261]: anon log[ctrlp_disp:pptp_ctrl.c:895]: Outgoing call established (call ID 2, peer's call ID 22850).
Aug 18 08:39:31 pptp[259]: Using interface ppp0
Aug 18 08:39:31 pptp[259]: Connect: ppp0 <--> pptp (xxx.xxx.xxx.xxx)
Aug 18 08:39:33 pptp[259]: CHAP authentication succeeded
Aug 18 08:39:33 pptp[259]: local  IP address xxx.xxx.xxx.xxx
Aug 18 08:39:33 pptp[259]: remote IP address xxx.xxx.xxx.xxx
Aug 18 08:39:33 post-firewall: started [ppp0 xxx.xxx.xxx.xxx br0 192.168.0.1 vlan1 xxx.xxx.xxx.xxx]
Aug 18 08:39:33 ddns: IP address has not changed since the last update
Aug 18 08:39:33 PPTP: connected to ISP

p.s automatic reconnect don't work for now.

[cichy]

New firmware already in place but now i'm getting this error:

Aug 20 21:07:28 ddns update: ez-ipupdate Version 3.0.11b7, Copyright (C) 1998-2001 Angus Mackay.
Aug 20 21:07:28 ddns update: ez-ipupdate started for interface vlan1 host cichy.ath.cx using server members.dyndns.org and service dyndns-static
Aug 20 21:07:28 ddns update: connected to members.dyndns.org (204.13.248.112) on port 80.
Aug 20 21:07:28 ddns update: authentication failure
Aug 20 21:07:28 ddns update: failure to update vlan1->xx.xx.xx.xx (domain.ath.cx)
Aug 20 21:07:28 ddns update: shutting down updater for domain.ath.cx due to fatal error

Although settings were the same as in earlier version and worked fine. What could be the problem?