Cisco ATA-186/188 reset / upgrade

As I have a couple old, locked down Cisco ATAs, from time to time I have to reset one of them to put into use, but have to play hours figuring out how to reset and how to upgrade them correctly, here is a howto. As DHCP, I’ve used isc-dhcpd

– Press “function” key on the ATA188
– Enter 322873738# (FACTRESET)
– Enter the reset PIN, then #
– Check the dhcpd server’s log on the ATA’s IP address
– If using SIP 3.1.1 firmware, open http://ip_address/dev/
– If using SIP 3.2.1 already, open http://ip_address
– Download ata18x_pack.tgz package, which includes the latest 3.2.1 SIP firmware and the associated Linux tools
– If You want to upgrade the firmware via the built-in Linux tool, simply run sata186us.linux with the required parameter, and dial in the *100*etc keys on the connected phone
– If You want to upgrade the firmware via DHCP, see below. This is also useful if You want to configure / use your ATA via automatic configuration via TFTP.

Required dhcpd.conf parameters:

option tftp-server-name "";

Filename is not required. The ATA will automatically look for ata.cfg or atadefault.cfg, in this order.

Required files:
test.txt – this is the example configuration file
ptag.dat – this is the descriptor listing
cfgfmt.linux – this is the actual binary that creates the binary configuration file for the ATA

You can configure the SIP parameters also in ‘test.txt’.

If You’ll be using the ATA behind NAT, it is advisable to set SipRegInterval=60 for keepalives. This puts a small burden on the PBX, but it won’t really hurt.


Nokia E90 + Ubuntu + 3G with data cable

After succeeding with Dapper to kick my GPRS connection alive with my 9300/E90 via BT 2yrs ago, I haven’t had much joy in configuring BT+E90+3G on Edgy, Feisty, Hardy, and so on. A friend has written a good howto on connecting up via data cable (if you want an english translation, just poke him), so this afternoon I decided to try it.

My E90 is running v210.34.75. The difference is that using the AT+CGDCONT command returns an ERROR. Another howto states that the user has to set the default access point on the phone, and dial out without the CGDCONT – this worked for me. So here is the final deal:

Set up the default AP on your E90

Tools / Settings / Connection / Packet data / Access point
Add (3G) or (plain old GPRS) depending on what you want to use.

Do an lsusb do check the USB vendor/product ID:

Bus 005 Device 001: ID 0000:0000
Bus 004 Device 001: ID 0000:0000
Bus 003 Device 001: ID 0000:0000
Bus 002 Device 003: ID 0421:04cf Nokia Mobile Phones
Bus 002 Device 002: ID 08ff:2580 AuthenTec, Inc.
Bus 002 Device 001: ID 0000:0000
Bus 001 Device 004: ID 03f0:171d Hewlett-Packard Bus 001 Device 001: ID 0000:0000

Load the cdc_acm and the usbserial module with parameters:

# modprobe cdc_acm
# modprobe usbserial vendor=0ร—0421 product=0ร—04cf

(This will probably screw your other modules if you try to use other USB-serial converters, but decide what you want to use… ๐Ÿ™‚ )

Connect up your E90 in PC Suite mode.

(I wanted to use the default pon/poff config, but I finally stuck with copy-pasting the wvdial conf).

Insert these into your wvdial.conf

[Dialer Defaults]
Init1 = ATZ
Init2 = ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
Init3 = ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
Modem = /dev/ttyACM0
Phone = *99#
Idle Seconds = 300
Password = vodafone
Modem Type = USB Modem
Stupid Mode = 1
Compuserve = 0
Baud = 460800
Auto DNS = 1
Dial Command = ATDT
Ask Password = 0
ISDN = 0
Username = vodafone

Start up wvdial:

# wvdial
–> WvDial: Internet dialer version 1.60
–> Cannot get information for serial port.
–> Initializing modem.
–> Sending: ATZ
–> Sending: ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
–> Sending: ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
ATQ0 V1 E1 S0=0 &C1 &D2 +FCLASS=0
–> Modem initialized.
–> Idle Seconds = 300, disabling automatic reconnect.
–> Sending: ATDT*99#
–> Waiting for carrier.
~[7f]}#@!}!} } }2}#}$@#}!}$}%}”}&} }*} } g}%~
–> Carrier detected. Starting PPP immediately.
–> Starting pppd at Sun Feb 15 17:23:06 2009
–> Pid of pppd: 29057
–> Using interface ppp0
–> pppd: [1b][7f]
–> pppd: [1b][7f]
–> pppd: [1b][7f]
–> pppd: [1b][7f]
–> local IP address x.x.x.x
–> pppd: [1b][7f]
–> remote IP address y.y.y.y
–> pppd: [1b][7f]
–> primary DNS address
–> pppd: [1b][7f]
–> secondary DNS address
–> pppd: [1b][7f]

Say hi to Mr. Pkerese for writing the original howto.

Logging into SQL with apache2 or nginx

I’ve been playing around this weekend with my webservers because some PHP coders want to see the accesslog of a few virtualhosts in SQL. In different databases, to be on the fun side. Apache2 has two options:

  • mod_log_sql does not support logging into different database servers per virtualhost, and also sports a static table layout, which is not necessarily nice. Could be used in other configs, but *trash*.
  • mod_log_mysql seems to be nice on the first spot, but then it gets out that one has to rebuild the Apache binaries with the modified mod_log_config the project provides. Hell wants an if-you-want-an-uprade-rebuild-everything nightmare, even in chroots. *trash*
  • Nginx has a patchset which allows the server to log into syslog, then in turn into syslog-ng, but since it has no configuration options – hardcoded syslog facility, no per-vhost config -, it seems promising, but *trash*

Now I seem to end up in using a script which parses the accesslog with tail -f and inserts the entries into SQL. Fuck what?

Preparing SSD and Sun x2250 test…

A few days ago I’ve got an OCZ 64Gb SSD, and apart from doing bonnie++ and sql-bench on it, I want to know what would _YOU_ like to see tested on that. A Sun x2250 is also in the queue, maybe I’ll pack the SSD into it, question is the same on that also.

Comment, ask, see results. ๐Ÿ™‚

vodafone meg az asterisk

Kellemeset csalodtam ma a Vodafone ugyfelszolgakban. Az alapfelallas az, hogy van egy asteriskem, par helyre bedrotozva, es ezt van, amikor napkozben is el akarom erni a telefonomrol, nem feltetlenul wifirol, hanem GPRS/3G-rol, a 100Mb keretbe ugyis boven belefer. E90, SIP konfig, beallit, stb, se kep, se hang. Sniff a gepen, UDP/5060-on se kep, se hang. Oke, UDP/5061-re atrak, ott is ugyanez. Akkor mar kezdett gyanus lenni a dolog, megneztem TCP-n, hogy nem az van-e, hogy csak az alap 22, 25, 80, stb portokat engedik at, nem, TCP/5060 siman atjon. Oke, menjunk UDP-n lejjebb, 54 mondjuk, az atmegy, 1025 (1024+1 hogy szep kerek legyen) mar nem. Ugyfelszolgalat, ASZF-nezegetes, WTF-eles. jo kis szaftos anyazasra szamitas…

Aztan felhiv egy technikainak tuno ugyfelszolga, ket perces telefonalassal lezavartuk, hogy van egy masik APN, a **, probaljam meg azt. clicketyclick mukodik. Meg ilyet… ๐Ÿ™‚

Azota amugy belottem a localphone-os londoni szamomat is, ugyhogy a +44 203 129 6287-en is elerheto vagyok csittcsattcsittcsatt – illetve a vojszmailboxom. Ku’l.

elmentem aproert dolgozni

21:43 <@V1pr> amรบgy meg mit gondoltรกl? elmรฉsz aprรณรฉrt dolgozni, รฉs meg se kรฉrdidk, hogy milyen?

Nem csak varjal, bontok egy sort, meg atnezem a mailjeimet, merthogy SSH kifele az nincs bentrol…

Szoval a szokasos elso nap, delelott uldogeles, doksiolvasas, elso login mindenhova, jelszavakkal szopas. Aztan confcall, szokni az angolok akcentusat, amit valahogy ugy kell elkepzelni, mintha minden szo veget elharapnad es hadarnal es mindezen a voip codec is sokat javit am. Aztan haromoras eloadas par rendszer strukturajarol, “folytkov holnap”, arra raebredes h az egyik huzosabb rendszer padawanja mostantol en leszek…unatkozni nem fogok, az ziher. Egy 2-3 hetes londoni ut kinez, varhatoan szeptemberig, abban azert bizom h mar Gaab-bal egyutt fogunk kimenni iszonyatosan bekurni azon a hetvegen.

Amugy jo helynek tunik, arc kollegak, tobbi csoportban is, kilatas a dunapartra, kajalda a kozelben nincs, de van ingyenkave meg csocsoasztal meg ket sakktabla is, meglattyuk…

Holnap meg nem megyunk melozni, talaltak egy laza kettonnas bombat, valszin ilyet: , pont az epulettel szemben. ๐Ÿ˜€

Ja, Morgan Stanley ๐Ÿ˜›


Compaq / HP Smartarray and raw disks

For the record: SmartArray controllers (probably all, but tested with a 5i) are storing 512+32Kbytes of data at the beginning of the disk. You have to get rid of this with dd if you want to read a disk directly that has been configured as an array member with these cards – for example for data rescue from a completely dead controller.

Nokia E90 part 1

After way too many months of waiting, we finally got our new Nokia E90. Apart from the design and manufacturing bugs, which pushed back the street dates, and apart from Vodafone’s cock-sucking mentality over the firmwares, which caused a few days of headaches, here is a litle debate over the firmware versions and the phone’s WiFi/WPA-capability.

Vodafone provided the phone with another product ID than the usual unlocked phones are. Failed to provide the latest firmware not immediately, but weeks after the release date of the normal phones’ firmware, making the NSU (Nokia Software Update) unable to update to the currently latest 7.40 firmware.

By some *pooof* *magic*, we were able to update the firmware to this version – TBH, we just wanted to see how AGPS and the new kernel performs -, but this unveiled a strange bug when connecting to our multi-SSID WPA network based on Merakis and OpenWrt. The phone rebooted. The bug was reproducible. It didn’t happen with the earlier 7.24 firmware.

After an evening of tweaking and almost fucking up my Meraki, I found out that we might have ran into a bug which happens when combining OpenWrt’s hostapd-0.5.8 and E90 7.40. Using hostapd-0.5.7 from a stable Kamikaze 7.09, the error did not occur, even with multi-SSID test networks.

I have yet to try to some other combinations, but it seems that I have to bug around with some sniffing with 7.24 and 7.40 versions, and hostapd-0.5.7 and 0.5.8. Bug or feature? ๐Ÿ˜›

Testing Linux network appliances

It seems that one of my customer would like to have some kind of a network appliance in front of its windows server (good idea!), and I have some strange animals^W^Whardware lying around, I decided to do some performance testing. Following will be used:

  • Gateway 7001 AP – Intel XScale IXP422-based AP with 2 ethernet ports and 1 miniPCI.
  • Atmel NGW100 – AVR32-based development board with 2 ethernet ports
  • PeeCee with various hardware – “pick your favourite mobo brand and build some crap”

Testing will include NAT, some IPsec with various crypto engines measured with iperf, and also some power consumption measurement if I can get a good meter.

Results will be posted here, so check back later.

phpadsnew 2001

peklapattal basznam agyon azt a bandat, akinek a 2001-ben felrakott phpadsnew-jan (bocsanat, php banner management system) keresztul toltak meg a gepemet, uszkve 10mbit kimeno mailforgalmat generalva. persze regglobals kellett a bandanak, meg az is h egy kanyi buznyakot se fizessenek az elmult 6 evben.
it’s 2007 you fat fuck!!!111111

ostoba clamav-freshclam

nyuszika baszd meg a funyirodat

+ su clamav -p -s /bin/sh -c ‘. /lib/lsb/init-functions && start_daemon /usr/bin/freshclam /var/run/clamav/ -d –quiet’
ERROR: Can’t open /var/log/clamav/freshclam.log in append mode (check permissions!).
ERROR: Problem with internal logger (UpdateLogFile = /var/log/clamav/freshclam.log).

hardcodeolt usernev initscriptben, mi, nyuszika?

it’s alive, it’s alive!

RedBoot(tm) debug environment – built 11:11:58, Jan 22 2001
Platform: IQ80310 (XScale)
Copyright (C) 2000, Red Hat, Inc.

RAM: 0xa0000000-0xa2000000
FLASH: 0x00000000 – 0x00800000, 64 blocks of 0x00020000 bytes each.
IP:, Default server:

You’ll never find out what this is. It’s an Intel IQ80310 XScale development board, which was given to me as a late christmas gift by a good old friend ๐Ÿ™‚ It’s not a Johnnie Walker Blue Label, it’s something more pervert ๐Ÿ˜› Nevertheless, this will be a good toy to play with OpenWrt on in the next few weeks, I’ll try to get some newer RedBoot onto it after I get a basic kernel booting.

(These develboards are cheap on eBay, I’ll get some more when I finish with this. Stay tuned, I’ll have to make a garage sale soon, or else these will flow out from my room like a river ๐Ÿ˜› )

UPDATE: hahahaha, this will be a funny ride ๐Ÿ˜›
“Inte XScale IOP310 removal”

AVM Fritz!Classic A1 card


I was a bit disappointed in AVM, since they released a half-binary-half-opensource driver for their Fritz A1 card, and it’s also a bit buggy. It needs to be patched to run with upstream 2.6.x kernels, so here is an easy-to-use download-patch-compile-and-enjoy patch for it. Since AVM did a bad job on this, and everyone ran into it, and AVM seemed to not to care about it, I do not care now about re-distributing issues, leave me alone with it.

Download link


Ah, my work with Kaloz got committed into trunk. ๐Ÿ˜›

Changeset 7533

BTW, looking at freenode/#avr32, it seems that Atmel reached its aim – to have a cheap AVR32 development for everyone who likes strange anima^Warchitectures. I’ll try to get an LCD panel – the same what the STK1000 has – to make this small board a … settopbox ๐Ÿ˜›


Yes, I’m building a port for my alphapc just to be perverse.
Yes, I was stopped with a uclibc error as follows:

Note to self: uClibc- alpha bug

No, I’m not ready with the SH4 port yet, the userland has to be fixed yet, but my kernel boots fine with the generic openwrt patches and the Landisk fixups. ๐Ÿ™‚

SH4 bootlog

10th floor test


“A machine part of an infrastructure should be able to pass “The 10th floor test”, a term first used by Steve Traugott from which refers to the ability to take a random machine in your infrastructure , drop it from the 10th floor, and be able to restore your infrastructure to working within 5-10 minutes.”


CHDK firmware

EZ jopofa cucc, csak kene DIgic I-hez is, upgradelnem az A85-omet ๐Ÿ™‚ Amugy erre az elborult cuccra is jellemzo, h oroszok csinaltak ๐Ÿ˜›