Phoenix is Born

Starting with the base firmware Bored Individual and myself (Amiga) have worked on fixing updating and adding to the firmware at times we`ve been working into the small hours scratching our heads at what linksys had done. we have decided that it is now ready for release, so say hello to Phoenix V7.01.03

Whats Changed

-Added full telnet access
-Annex M modes are now selectable from the index page
-Annex M, and other modes are now detected properly by the status page
-Dnsmasq stability fix from TSM
-rc/network and pppd stability fix from TSM
-SNR 'fix' - show those ridiculously large numbers as negative
-Single port forwards have 'both' 'tcp' 'udp' selectable now from Amiga
-Static Addressing in UI changed to relative addressing
-Httpd Speedup Fix
-Busybox updated to v1.2.1
-Remote logging fix for busybox
-Status - DSL connection now has advanced status information Page (ADSLCStatus)
-Status - Gateway now has advanced gateway infomation page (SYSInfo)
-Status - Gateway now Shows Uptime (d/h/m/s) and dataflow (Rx/Tx)
-ADSLCStatus tweaks and fixes, now displays connection time (d/h/m/s)
-SYSLOG replaced - View by full or by levels category
-Firewall Replaced - view all,incoming,outgoing,blocked,aler­ t
-SYSLOG rotate and clear fix
-NTP Lasttime Bug fix
-Firewall log incoming fix
-Auto/Manual NTP servers, upto 2
-Cron log, Invalid Packet log Disable fixes
-Net Messages suppressed Disabled fix

So merry christmas from Bored Individual and myself.

Get your early Xmas present Here

www.hacknpatch.co.uk

The Phoenix rises in Pluto, AG241 V1 AU

Bored Individual ported back all the changes we made to the V2 to the V1 tree, so here is AG241-AU-v1.01.15. It's basically a merge of the firmware mentioned above, and the Pluto firmware you already know. He also added the 7.04 DSP from Acorp, which may or may not be better.

download.yousendit.com/C1E894D42518A9AB
rapidshare.de/files/3812...1.01.15.bin.html
bi.cockcrotch.com
www.hacknpatch.co.uk/BIFirmwareAU.htm

Source is available on request.

Amiga writes...

He also added the 7.04 DSP from Acorp, which may or may not be better.

Hi Amiga, great work on the info and files for the AG241_Ver1.

I have not really played with Linksys Router's before and have been trying to get a Ag241 to run more stable for a mate while I am in Sydney. Sync is fine, just disconnects PPPoA session every 30 to 45 minutes.

The firmware was standard 1.01.07 and downloading at almost 1.1meg a sec,fairly solid,across the board.

After putting Pluto 1.01.15 download throughput is anywhere from 0 to 85-90 kilobytes a second but erratic.

Have you guys had any feedback on a change to throughput with the change of DSP.

I haven't defaulted the AG241 after the upgrade to pluto , as I am reluctant to while mate is overseas as I won't be able to get straight back on the net. Is it possible this could cause the throughput issue?

Cheers

edit: spelling

mEDIEVIL writes...

After putting Pluto 1.01.15 download throughput is anywhere from 0 to 85-90 kilobytes a second but erratic.

Strange. Here's a version with the same drivers as the 1.01.07 firmware:

www.hacknpatch.co.uk/dow...1.01.15-DSP6.bin
1846ed206a7d50f13b05496c01954290 AG241-AU_v1.01.15-DSP6.bin

If that doesn't help, try the older 1.01.14 versions from here and let me know how they go:
forum-replies.cfm?t=511748&p=9#r179

Ok, thank you so much for the response, it is so appreciated. I have done the following to maybe try and help narrow it down.

1. Upgrade 1.01.07 to AG241-AU_v1.0­ 1.14-DSP7-static.bin -- Throughput appears to be the same as 1.01.07 1.3megabyte and stable

2 Upgrade 1.0.1.14-DSP7 to 1.01.15-Latest -- Throughput down to 85kB max a sec but all over the place

3 Roll back from 1.01.15-Latest to 1.0.1.14-DSP7 -- Just to make sure am not leading anyone astray. Throughput back to 1.3megabyte again

Now am not sure to put 1.01.15_DSP6 on it as am functioning on 1.01.14_DSP7.
Was there a change in the DSP 7 between 1.01.14 and 1.01.15?
Is their an advantage of DSP7 over DSP6?

Edit: To further help I have now tried 1.01.15_DSP6 and throughput drops to the erratic ramblings up to the 85kB a sec mark. So have rollbacked to 1.01.14_DSP7 and all is good.

Your help is once again welcome and if you need me to change or try anything, just let me know.

Cheers
Nayth

mEDIEVIL writes...

Your help is once again welcome and if you need me to change or try anything, just let me know.

No worries, you gave me enough details to narrow it down to a much smaller problem, which means it's easier for me to work on. Thanks.

In fact, I think I've already fixed it. Unfortunately, I can't get 1.3Mbit downloads, the best I can get is ~750kB/s, due to my line, but I was able to reproduce the same issue you got. Anyway, please give either of these a go:

7868dd674104d381debb19d1029da28f AG241-AU-1.01.15.1-DSP6.bin
d35fc9ea0a6a155dc25ab4a891da3af9 AG241-AU-1.01.15.1-DSP74.bin

www.hacknpatch.co.uk/dow...01.15.1-DSP6.bin
www.hacknpatch.co.uk/dow...1.15.1-DSP74.bin

(I've set the version to 1.01.15.1 - sorry if it's a bit confusing, but it's just 15, and a little bit more. The DSP6 one has the same drivers as the official firmware.)

Let me know how that goes.

OK, have loaded up "AG241-AU-1.01.15.1-DSP74.bin" and also tried the "AG241-AU-1.01.15.1-DSP6.bin" and throughput appears to be fixed with both giving me about 1.3 megabyte or so.

You are a fantastic help and your effort is greatly appreciated.

If I can be of any further help please just let me know.

Cheers
Nayth

Thanks. Don't forget to add release dates in future :)

AU Firmware updated at

www.hacknpatch.co.uk/BIFirmwareAU.htm

download is the AG241-AU-1.01.15.1-DSP74.bin

Works here tooooo :-) v2

Regards

Hi guys. Just a quick post to say how much I appreciate the work you guys have done on the firmware, and for taking the time to modify the firmware for other regions (such as EU).

I'm currently running one of the Bored Individual builds, and I'm going to flash my modem in a few. It's nice to see you have added uptime and dataflow, those were two things I was going to suggest :)

This will now be quite similar to my poor old ADSL2MUE which was running some strange Texas Instruments firmware I picked up somewhere (it was a nice firmware, but felt a bit hit and miss stability wise). For some unknown reason that modem kicked the bucket. Sometimes it would power on and load, sometimes it would get stuck with a red LED, however even when it did work, it was as though the DSL line was disconnected.

I'm glad to see people kindly using their time and sharing the results. Hopefully maybe more people will get on board and this modem will gain support something like that of the WRT54G.

Happy new year, and cheers.

Ok an update on whats been done so far for the next update to phoenix

1> IFstats modified to show Dataflow as b,KB,MB,GB only ie Tx: ## Rx:##
2> Port 12345 now stealthed in firewall
3> IPtables Updated (+ iptables compile fix)
4> NetFilter IPV4 Updated (well what can be) + IProute2 Update for compile
5> Disable Dsmark and Htb Debug output in log files ( Saves ram and not very useful)
6> Clean Up of the source files (Ongoing, CyberTan code is not the best)

If you find any problems with the current version please report them here and we`ll (well Bored with my help if i can will try to fix them).

Any improvements you would like added please post here as well, we will try to add features if possible. The current beta version i`m using at the moment with the above fixes is working really well and is a lot faster.

Watch this space for more......

Hi guys. Great work and amazing support. Thanks!!!
Can you publish the beta? Also I was wondering if DSP 7.4 code used in V1 Phoenix firmware is any better than 7.3 and if it could possibly be introduced into V2 routers? Maybe it would handle noisy lines a liitle better?

Sorry for double post. :(

I've not seen any Ti release notes for DSP7.4 but definitely could be used for other firmwares - I've used the Acorp version on my ADSL2MUe for example. But by looking at the atm code the biggest change between 7.2 and 7.3 was in extra "interoperability tweaks" - has anyone tried the Neo (of routertech)'s "router stats" program in "compatability mode" utility with Phoenix?

www.routertech.org/viewtopic.php?t=1501

mstombs writes...

I've not see any Ti release notes for DSP7.4

forum-replies.cfm?t=884111&p=1#r2

Amiga writes...

He also added the 7.04 DSP from Acorp, which may or may not be better.

mstombs writes...

has anyone tried the routertech "router stats" program

I Have now.

for anyone that wants to try it themselves get the zip of all needed files here:

www.hacknpatch.co.uk/sta...tats-logging.zip

how to use under xp

the zip contains:

MSWINSCK.OCX (this needs to be put in c:/windows/system32)
then start button -> run -> regsvr32 \windows\system32\MSWINSCK.OCX
vbrun60sp5.exe (double click to install - installs vbruntime 6 sp5 files)
stats_logger_1_0_17.zip (extract to desktop and run RouterStatsLogger.exe)

output and settings

img512.imageshack.us/img...reenstatsmt5.jpg

img81.imageshack.us/img8...nsettingspw9.jpg

img81.imageshack.us/img8...enreport1dj5.jpg

img81.imageshack.us/img8...enreport2mv2.jpg

img72.imageshack.us/img7...nexchangedu8.jpg

____________________________

fsjim writes...

Can you publish the beta?

AG241V2 Annex A
beta is here please post feedback.

www.hacknpatch.co.uk/bet...Network-beta.bin

NB: this Beta firmware does not contain the cleaned code yet and is not for general use but for Bug testing purposes

Amiga writes...

beta is here please post feedback

Thank you. So far no problems observed but (BTW) I've got one feature request: Administration/Reboot - to reset the device you need to confirm entering the password. Unfortunately you cannot use ENTER/RETURN key in requester; seems CR char is not ignored while doing strings comparison. This "bug" is not specific to the beta of course and is present in all fw versions so far (official and otherwise).

The atm/dsp drivers are Ti in origin, technical notes for 7.2 are here for example

mcmcc.bat.ru/acorp/ar7_d...ture_control.pdf

and we can see a few extras in source code for atm 7.3

Amiga writes...

4> NetFilter IPV4 Updated (well what can be) + IProute2 Update for compile

Further to that, I did some testing in Vista. With 2.01.03 official (+telnet), I was able to browse the web interface, but telnet stopped working as soon as I logged in. It worked fine when I disabled autotuning. This suggests there is probably still an issue with tcp window scaling in the official firmware.

After I updated most of netfilter, I tested again in Vista, enabled autotuning again, and telnet/etc worked fine. Just to make sure, I went back to 2.01.03 again and got the same issue with telnet.

So, make of that what you will.

EDIT: Bah, well, just to confuse me, the 2.01.03 firmware now decided it would work fine. Oh well, there's a bunch of shiny new options in /proc/sys/net/ipv4.

It is of course a major modification - old code thrown out left right and centre, so if you notice something not working right, do let us know.

fsjim writes...

Unfortunately you cannot use ENTER/RETURN key in requester

Fixed, tested in Firefox and Opera, not tested in IE yet, will do soon.

The problem wasn't that the return was being added to the pass, or wherever you pressed enter. It wasn't even getting to the validation stage, as the to_submit javascript wasn't run. I just changed the type of the submit button from "button" to "submit".

Hi Amiga and BI,
when putting links to new update, would you mind specify if it is EU, AU or else and also if it's for V1 or V2, I would be sad to brick my AG241, because I got the wrong firmware. :-)

You guys are doing an EXCELLENT job...

"AG241V2 Annex A
beta is here please post feedback.

www.hacknpatch.co.uk/bet...Network-beta.bin "

Is thsi also for annex M, V2, EU?

Regards

bimbobo writes...

Hi Amiga and BI,
when putting links to new update, would you mind specify if it is EU, AU or else and also if it's for V1 or V2, I would be sad to brick my AG241, because I got the wrong firmware. :-)

You guys are doing an EXCELLENT job...

"AG241V2 Annex A
beta is here please post feedback

Sorry it should of been

AG241V2 EU Annex A

i`ll make it clearer in future

NP... :-)

Still the question is, is it also for annex M...? I am in doubt if I understood this well.

Thanks again

Regards

This is going to sound very vague, and probably impossible to trace/answer but I'll throw it in for something to ponder.

I live quite a distance from my exchange. Before my ADSL2MUE kicked the bucket, it would usually sync around 3.5mbps, but my AG241 will only sync at 2.8mbps.

I heard that some routers/modems can extract better speeds on poor lines, is this correct? Also why might the ADSL2MUE have been getting a better speed compared to the AG241? It should be noted that during not having a modem, I tried the supplied BT homehub, and I've heard stories about it messing up a BRAS profile, so maybe this is what has happened and there would be no difference between the sync speed of both modems normally.

The ADSL2MUE I had was running (what I would call OEM) firmware from Texas Instruments (2.17ti). Could it possibly be this DSP thing I see mentioned around, for example if it had a faster/newer DSP in the ADSL2MUE TI firmware than in Linksys' AG241 firmware?

Also I wonder whether the difference is the ADSL2MUE was running at 250MHz and the AG241 is running at 200MHz? (I'm not sure what the ADSL2MUE was running at, but I'm wondering if the Texas Instruments firmware could have "overclocked" it)

Finally (and this really is a shot in the dark as I suspect you can't do anything about poor lines), are there any tweaks to the hardware that can be done in order to obtain higher speeds? I've heard about tweaking the modem with DMT tool; something about manually setting a target SNR margin. Could this target SNR margin be made available in the modem's firmware? I'm not sure if it's just me, but the DMT tool doesn't seem to work for me. Another thing may be tweaking of the output power (for example if you set it higher, might that help?)

Here's my current stats if it helps.

ADSL Line Type GDMT
ADSL Line Coding INTERLEAVED
Line Attenuation (Down):57 (Up):31
Noise Margin (Down):10 (Up):21
Total Output Power (Down):17 (Up):12
Current State DownStr: 2848(K) UpStr: 448(K)
ADSL Line Status UP
FEC(FAST) (Near):0 (Far):0
FEC(INTERL) (Near):68 (Far):31241
CRC(FAST) (Near):0 (Far):0
CRC(INTERL) (Near):1 (Far):101
HEC(FAST) (Near):0 (Far):0
HEC(INTERL) (Near):0 (Far):0
Connection Uptime 0 Days, 7 Hours, 18 Minutes, 46 Seconds.
ADSL FW Version 0.54
ADSL Board Type Annex A

One suggestion that should be pretty straight forward. How about putting a link in the firmware to hacknpatch.co.uk in the status tab under the firmware version or something?

Firmware Version: 7.01.03-Phoenix, 24. Dec 2007
Firmware Home Page: www.hacknpatch.co.uk
MAC Address: 00-00-00-00-00-00
Current Time: Sat 05 , Jan 2008 00:30:08
Uptime: 1 Day, 6 Hours, 38 Minutes, 14 Seconds.
Dataflow: Rx: 97636618 (93.1 Mb) Tx: 50407910 (48.0 Mb)

Something like that maybe?

Many thanks

The ADSL2MUE 2.17Ti firmware had V3 DSP drivers I recall - these were not ADSL2+ capable probably smaller/lighter than later ones. The later drivers may trade speed for stability - but your biggest problem appears to be distance from exchange/ attenuation, maybe re-adsl would help?

I believe the ADSL2MUE core is 150MHz vs the AG241 V2EU 212MHz, although I believe the DSP in the 2MUE is boosted to 250MHz by default - there are env vars no_boost and dsp_freq that control this I recall - also responsible for "mosquito noise".

[I'm still using my ADSL2MUEs - you would recognize the Routertech firmware coming from the same origin as 2.17Ti]

Thats very interesting. Thanks for sharing. Yes, IIRC the ADSL2MUE was only ADSL2 and not ADSL2+, however the Ti firmware did have some nice features which unfortunately I cannot recall or check because I have thrown it out since it was faulty and only had one ethernet port so I couldn't even use it for anything else.

Let me just check I'm on the right lines here. By boosting the DSP speed it is possible to gain higher speeds in a trade off for stability?

If this is the case then I wonder if this feature can be added to the firmware? Similar to how you can set your FSB and CPU multiplier in your PC's BIOS to increase/overclock your speed, I wonder if it would be possible to add the ability to select the DSP speed in the firmware? I was primarily thinking about suggesting this to squeeze some more speed out of bad lines, but it could work the other way so that people can underclock it to gain stability on shaky lines?

Unfortunately for me, I'm just on the brink of getting 2.5mbps. The sync rates are:
2272 - 2847 2 Mbps
2848 - 3423 2.5 Mbps

So the fact that my router is synced at 2848 doesn't give me much of a margin for error!

Thanks for suggesting RE-ADSL. It didn't occur to me at first, instead I went down the route of looking into ADSL2+ which, fingers crossed may give me 4mbps, but then again the estimate says 2-4mbps so it may be hardly any better. The only problem is that I can't seem to find a provider of RE-ADSL here in the UK, but I'll keep searching.

So to just outline my suggestions in a more readable format:
-Add a link to www.hacknpatch.co.uk in the modems status/general tab.
-Add the ability to set a target SNR margin
-Add the ability to set the modem/DSP speed. Suggest 150-250MHz in 10MHz increments (assuming 200MHz is default)?
-Add the ability to increase total output power (up and down)
-Add device name to login prompt. Currently in Internet Explorer and Firefox it just gives "The server 192.168.1.1 at null requires a username and password" whereas my WRT54G gives me "The server 192.168.1.1 at WRT54GL requires a username and password"

Naturally, I appreciate that this is something you guys do in your own time, and that some of my suggestions may not even be possible, but thank you anyway.

Cheers.

bimbobo writes...

Still the question is, is it also for annex M

Yes.

mstombs writes...

I believe the ADSL2MUE core is 150MHz vs the AG241 V2EU 212MHz, although I believe the DSP in the 2MUE is boosted to 250MHz by default - there are env vars no_boost and dsp_freq that control this I recall - also responsible for "mosquito noise".

Yes, there are 2 clock rates of concern. The CPU (TNETD7200/7300 running at 212/150Mhz respectively, and is not changed by running different firmware), and the DSP, which is boosted from 200Mhz to 250Mhz when supported, and by a tiatm.o that is new enough. Which means, I'm not sure if the ADSL2MUE is boosted. I believe support for boosting the DSP clock came in around tiatm 4.03/4 or so.

The AG241v2 has the TNETD7200 (212Mhz) and the ADSL2MUE has some form of TNETD7300 (150Mhz).

Zero1 writes...

are there any tweaks to the hardware that can be done in order to obtain higher speeds

No, most of the current tweaks will sacrifice speed to gain stability. Maybe an older DSP/Tiatm would suit you better, I'm not sure.

Could this target SNR margin be made available in the modem's firmware?

Nope, while it would be really really useful, it doesn't seem to be supported by the TI drivers/DSP.

Noise Margin (Down):10 (Up):21

Interesting that you have seemingly a fair bit of noise margin there, has your ISP set anything on their end?

If this is the case then I wonder if this feature can be added to the firmware?

I think you'll find it's already boosted in the AG241v2. Telnet in and run:
echo dspfreq > /proc/sys/dev/dslmod
tail /var/log/messages

Look for an entry at or near the bottom like:
user.warn kernel: dsp_freq = 250000000
(250Mhz).

If you add dsp_noboost 1 to /proc/ticfg/env, that dsp_freq will be 200000000 (200Mhz), ie.
echo "dsp_noboost 1" > /proc/ticfg/env
and reboot.

The really really old tiatm.o do not boost the dsp frequency at all.

To summarise:

Zero1 writes...

-Add a link to www.hacknpatch.co.uk in the modems status/general tab.
Should be easy, will probably be done eventually, unless someone has any issue with it.

-Add the ability to set a target SNR margin
Not currently possible, to my knowledge.

-Add the ability to set the modem/DSP speed. Suggest 150-250MHz in 10MHz increments (assuming 200MHz is default)?
Seems to be all or nothing - ie. 200Mhz or 250Mhz.

-Add the ability to increase total output power (up and down)
Don't think this is possible either.

-Add device name to login prompt.
Should be easy, as far as I remember the AG241v1 handled this properly, it's probably the realm (see the issue that prompted me to start modifying the AG241 firmware.)

Zero1 writes...

the Ti firmware did have some nice features which unfortunately I cannot recall or check

There are some Ti NSP userguides on the Russian Acorp mcmcc site, or a web simulator of an Italian mod of Routertech firmware (derived from Ti reference/Acorp) here

darkwolf.altervista.org/demo/login.htm

relevant to this thread due to similarity in Linux kernel, hardware and atm/dsp drivers - but Linksys/Phoenix use different gui and router control code.

The later atm version have a lot of "interoperability" tweaks 7.3 includes a specific bit for "BT qualification" I've tried them all but still have slow connect time and occasional dropouts in ADSL2+

I have experimented with different dsp frequencies, but found no effect on connect speed, I assume faster means more electrical power and hotter chips, but if enough processing can get done at lower clock speed the modem works. The ADSL2MUE has power supply issues - so maybe I should go back to the lower rate...

TALKTALK LLU in the UK do use RE-ADSL, no idea how this is selected I guess automatically by their HUAWEI DSLAM code - you wouldn't get far asking their customer service (when broadband is given away for free what support can you expect?)

I wonder if it is fair comparing old sync rates with current new - no doubt the chances of crosstalk interference are getting higher as more and more adsl connections made.

Not great news to you in this thread - but have you tried a different make of adsl modem? Some folk swear by Broadcom adsl chipsets as packaged by SpeedTouch for long lines! There is no simple answer its best to just try a few!

@Bored :- on my ADSL2MUE I can only see the responses from the dslmod queries on a serial mod cable - clearly my syslog not setup correctly or just another difference.

mstombs writes...

There are some Ti NSP userguides on the Russian Acorp mcmcc site, or a web simulator of an Italian mod of Routertech firmware (derived from Ti reference/Acorp) here

darkwolf.altervista.org/demo/login.htm
Bingo, that's almost exactly like the firmware that my ADSL2MUE was running, barring cosmetic changes. Heh, makes me miss it. That was a pretty nice modem.

mstombs writes...

The later atm version have a lot of "interoperability" tweaks 7.3 includes a specific bit for "BT qualification"
Do you happen to know anything else about this BT qualification? I'm currently with BT and I'm wondering if it's related to their BRAS profiles?

I think I'll hang on in there with my AG241, and perhaps look into this TalkTalk RE-ADSL. I wasn't aware anyone was operating this, so thanks for the heads up!

Also thanks, Bored Individual for answering those few questions, that helps me out now I know what can and cannot be done.

Zero1 writes...

Do you happen to know anything else about this BT qualification? I'm currently with BT and I'm wondering if it's related to their BRAS profiles?

All I see in partial source code leaked is

#define OAMFEATURE_FEATURE1_ENABLE_BRITISH­ _TELECOM_FIX 0x00000020 //CQ11307
03/02/07 Ram CQ11307: Added Interop_1 API bit5 for British Telecom qualification

BI may be able to see deeper, and advise how this can be set in which phoenix version. Its not yet in Neo's Router stats program, but I added it to my personal copy.

Note it may well make your connection worse if they are reducing power to reduce cross-talk for instance! It may also be a test bit never intended for us to see!

mstombs writes...

All I see in partial source code leaked

It's complete, it compiles and runs. I'm still looking for 7.04 source though.

BI may be able to see deeper, and advise how this can be set

Telnet in and run:
echo "DSL_FEATURE_CNTL_1 0x00000020" > /proc/ticfg/env

03/02/07 Ram CQ11307: Added Interop_1 API bit5 for British Telecom qualification

That's the only info I can see on it, and it really tells us nothing. So nope, no idea what it does either.

EDIT: Though it won't work on the AG241v2 - as it's only running 7.02 tiatm/dsp.

It may also be a test bit never intended for us to see!

Considering the feature control documents say Texas Instruments Confidential and proprietary on them, I don't think we were meant to see any of them. (It's a bit silly if you ask me, making the API confidential...oh no, people might be able to configure their routers.)

Bored Individual writes...

EDIT: Though it won't work on the AG241v2 - as it's only running 7.02 tiatm/dsp.


and that BT tweak won't be much use to you in the AU version with 7.04!

I installed the beta firmware, and it's working fine. :-)
When are you expecting the final version?

Regards

Ok next update on whats been done so far for the next version of phoenix

1> IFstats modified to show Dataflow as b,KB,MB,GB only ie Tx: ## Rx:##
2> Port 12345 now stealthed in firewall
3> IPtables Updated (+ iptables compile fix)
4> NetFilter IPV4 Updated (well what can be)
5> Disable Dsmark and Htb Debug output in log files ( Saves ram)
6> Clean Up of the source files
7> Reboot-Submit Patch (Enter/Return now working) Requested by fsjim
8> IPV6 core patch (Ready core for possible IPv6 usage)
9> Httpd Reduce Size patch
10> Iptables Reduce Size patch
11> Busybox updated 1.9
12> Realm Patch (Adds device name to login prompt) Requested by Zero1
13> IProute2 reverted (was causing QOS issues)

AG241 V2 EU Annex A
beta P10 is here please post feedback.

www.hacknpatch.co.uk/bet...Network-beta.bin

NB: this Beta firmware and is not for general use but for Bug testing purposes

----------------------------------­ ----------------------------------­

We are also working on getting Phoenix out for DE and AG241 v1 but hit a few unexpected glitches.....

Installed P10, not experiencing any issue..

From now on I will only install "the final", so I hope it will be soon released.

Regards

Hi Bimbobo

The final will have a new version number when released (Hopefully soon, we have a couple of things to finish first)

we are releasing the betas so as to get feedback from as many different setups as possible due to the amount of code changes that that have taken place, you`ll notice the betas are only released after the big changes (the ones that have the potential to have hidden bugs). The small changes we can check between ourselves.

Please bare with us as sometimes it can be hard to work on issues together as i`m in the uk and Bored Individual is in Australia 10.5 hours ahead

Thanks for the feedback,

As we say in Italy, "no news good news". They say the same here in Denmark too.

I run the beta 10 and I have no issue of any kind.
My test are very simple, I just use my PC and the connection as I always do.

No drops out, slow connection, or error in the way the AG 241 works, so I can only say it's stable afaik..

Regards

OK this is compile of the 2.01.03 source code with changes added for DE AnnexB

AG241 V2 AnnexB DE v2.01.03

www.hacknpatch.co.uk/bet...2.01.03-code.bin

Please note this is untested as i have no way of testing it here.

Please report feedback. thanks

Phoenix DE v7.01.03

Phoenix for AG241 V2 DE v7.01.03 is now up on the www.hacknpatch.co.uk website.

Enjoy.

Amiga / Bored Individual

Update on Phoenix v7.01.04, we are very near finalising just a couple last minute code changes.

Hopefully by the weekend the new version will be up (don`t hold me to it.)

www.hacknpatch.co.uk has been updated with new selection index page (currently AG241 EU/DE V2 and AG241 AU) AG241 EU V1 to be added once we get it ported back.

Thanks to Amiga and Bored Individual for the fantastic work.

Please announce clearly when the final V2 EU version is ready.
I will check this site everyday to receive the great news.

You guys are amazing ...
My truly respect and appreciation...

Regards

AG241V2 Phoenix EU AnnexA v7.01.04 , Phoenix DE AnnexB v7.01.04 and AG241 AU v1.01.18 Released

Phoenix v7.01.04 EU/DE Change log - Release: 16th January 2008

-IFstats modified to show Dataflow as b,KB,MB,GB only ie Tx: ## Rx:##
-Port 12345 now stealthed in firewall - Requested by Bimbobo
-IPtables Updated (+ iptables compile fix)
-NetFilter IPV4 Updated - Vista's tests now pass
-Disable Dsmark and Htb Debug output in log files - (Saves ram)
-Clean Up of the source files (CyberTAN code / Build) - (Saves ram)
-Reboot-Submit Patch (Enter/Return now working) - Requested by fsjim
-IPV6 core patch (Ready core for possible IPv6 usage)
-Httpd Reduce Size patch -Iptables Reduce Size patch
-Busybox updated 1.9.0 -Realm Patch (Adds device name to login prompt) - Requested by Zero1
-IProute2 reverted (was causing QOS issues)
-Homepage url link on Router status page added - Requested by Zero1
-Httpd Asp Parser - Rewrite ( Faster than its ever been)
-Make sure upnp-igd no longer overwrites manual port forwards
-Added a page to view upnp (auto) port forwards

Please note: The upnp forwards page can only be about as (in)accurate as the info the binary-only upnp-igd maintains.

AU v1.01.18 Change log - Release: 16th January 2008

-Make sure upnp-igd no longer overwrites manual port forwards
-Added a page to view upnp (auto) port forwards
-Pressing the enter key on the reboot pages should function correctly
-Rewrote asp parser yet again - slightly faster.

Please note: The upnp forwards page can only be about as (in)accurate
as the info the binary-only upnp-igd maintains.

Get them here www.hacknpatch.co.uk

Enjoy

Amiga / Bored Individual

Hi Amiga and BI,
the new firmware is running like a charm so far. (Annex A V2 EU)
Thanks a lot for the great work.
Regards

Amiga writes...

Please note: The upnp forwards page can only be about as (in)accurate
as the info the binary-only upnp-igd maintains.

If you would like to replace it I have recently built (to 100kb) and am using a version of miniupnpd from miniupnp.free.fr for my ADSL2MUE, advantage would be full source code availability, and given recent concerns over upnp vulnerabilities, controlled limited feature set with limited port range functionality. Miniupnp is already available in Openwrt. Acorp/Routertech includes source for the Ti version of upnpd (300kb uncompressed?) which has memory leaks - possibly due to old version of iptables libiptc. I've also noticed the WRT54GL etc Tomato's upnp, possibly similar to AG241 being based on Linksys/Broadcom is now also only 100kb and I believe has full source code and a neat option to not show the IGD icon in windows properties (but still do upnp forwarding functions). I've found just displaying this icon generates a lot of network traffic and router load.

Cool, I've just updated my modem and everything seems in great working order, and thanks for including my suggestions.

Also I had an idea, but then I realised it was probably a bit rubbish. Or if it wasn't, that perhaps it would bog down the modem too much or would be a pain to implement, or simply not benefit enough people to be viable. Basically I was thinking of a way of how something like Netlimiter could be implemented into the modem. I'm not sure if this would be possible or if it would be too resource consuming, but I think I'll outline it in case it is of interest at some point.

My general idea is to show the dataflow for individual IP addresses/devices connected to the modem under the status tab. In addition to this you would be able to click that IP address to expand it to show the dataflow for that IP address/device by year, month, day or hour. Obviously to save resources you could perhaps only keep the daily details for the current month (and hourly details for a week) with the other details just being totals for the previous months. Also if it was really tight on space, limit the stat keeping to 6 months, so as it rolls over to a new month, a previous month is deleted. This would preferably be written to internal memory rather than just kept in RAM in case of power outage. To keep things nice and clean, Dataflow could be moved to a tab of it's own (as when you get expanding things, the page could get very long).

Maybe add a current/average transfer rate to the dataflow section. I realise real time would not be possible, but I see the page does refresh periodically. Also if my idea was of any use and dataflow was moved to it's own tab, the refresh period for that page could probably be made shorter.

This is a very thrown together photoshop of what I had in mind.
img46.imageshack.us/my.p...=40447068dd2.png

I originally thought of this as a bit of fun and interest to see how much data my Wii is transferring while it is "asleep", but for those who are on capped ISP's I think it would be a good way to keep a check on their data usage. One of my thoughts (I am capped, but at 30GB so it's not often I go over it) was that you can install similar software on a PC, but these days so much stuff can use ADSL via a wireless router it's crazy, for example in our house we have 2 Wiis, Xbox360, mobile phone, DVD player and 2 laptops all connected to the net. Although you may be obeying your cap on your PC, you are never quite sure how much your Xbox is transferring especially if you are downloading demos, or netplay with the headset. Other devices do not get accounted for, so my idea was to account for them at the source of supply.

However I ran into a slight problem. If what I have described so far is within the limitations of the modem then it will run as intended, showing individual stats for each device, for the devices connected directly to it. However like many, I use a wireless router (WRT54GL) connected to the AG241. All my devices connect to the WRT54GL, which in turn connects to the AG241. If such a feature was implemented, I wonder if it would still be able to give stats for the individual devices, or would it only "look" as far as the WRT54GL and think that was the only device connected.

It's hard for me to explain because networking is not my strong point, but let's look at it like this:
First we have the AG241 at 192.168.2.1, laptop 1 at 192.168.2.100 and laptop 2 at 192.168.2.200. Both laptops have transferred 100MB each. Because this is a simple setup and they are connected to the modem directly, it should be possible for it to differentiate between data transferred to and from laptop 1 and laptop 2, showing 100MB transfer for 192.168.2.100 and 100MB for192.168.2.200

The next scenario is with the laptops connected to the router and the router connected to the modem. Now we have the AG241 at 192.168.2.1, laptop 1 at 192.168.1.100, laptop 2 at 192.168.1.200 and a WRT54GL at 192.168.1.1. Again, both laptops have transferred 100MB each. I'm wondering whether the modem will still be able to see what has been transferred individually, or if it will play dumb and think that there is one device connected at 192.168.1.1 and it has transferred 200MB (instead of seeing 2 laptops, it only sees the router).

Networking is by far my worst subject, so bear with me. If this is the case that it does not see the individual devices when a router is added, then I think there is another option, depending on how routers and modems work. If you can somehow pass on the IP addresses via port forwarding, or disabling a feature like NAT (which I assume is a modem only feature), then the current system could still work. If however the devices are referred by port numbers, then it could get interesting. What we could do would be to allow the listing of IP addresses and ports to monitor. For example if laptop 1 at 192.168.1.100 is passed on to the modem by the router as 192.168.1.1:100 (the router address followed by the last 3 digits of the devices IP address as the port number), then it could still work. Also, that means you could monitor how much data a device has transferred for a certain service, ie HTTP, torrents, voice on Xbox360 etc.

Other, more sane, useful suggestions follow. I was recently considering messing with QoS. I was wanting to set it up so my Xbox has priority over everything on my network, so other peoples downloading or browsing would not cause my games to lag. To that end, I'd like to request some more fields for user entry in the QoS tab. I see there are 3 fields for user defined port entry, which I believe you have already increased from 2. For the Xbox360 alone, I would need to add UDP 88, UDP 3074 and TCP 3074, so it would be greatly appreciated if there were some more fields so I could configure other devices on my network and also my Wii.

Also with the QoS, it doesn't specify whether it's UDP or TCP. Does it automatically configure both? Would there be a scenario where you would want to configure a specific port for TCP or UDP only? I'm assuming that if I add port 3074, then it will automatically set the priority for TCP and UDP.

Finally I wonder if QoS on a per device basis would be possible. For example, rather than specifying a port, you could specify a device's IP address so that all traffic on that device would get the same priority. As an example, if I was to set 192.168.1.125 (which we would just say the Xbox360 was on) to high priority, that would allow all ports for that device to be set to high priority, and if I set my laptop on 192.168.1.100 to low priority, that would affect all data, regardless of what port it was on (eg HTTP, FTP, torrent, messengers etc).

I had in mind something like 5 boxes. 4 sets of 3 for the IP address entry, and the last one for the port number. If the port number was left empty or set to 0, then all ports for the device are affected by the QoS priority.

Hope this makes sense, and thanks for sharing the new firmware. Great work guys, really appreciate it.

Zero1 writes...

All my devices connect to the WRT54GL, which in turn connects to the AG241. If such a feature was implemented, I wonder if it would still be able to give stats for the individual devices, or would it only "look" as far as the WRT54GL and think that was the only device connected.

or disabling a feature like NAT

The boxes connected to the WRT54GL are indeed only visible if you disable NAT. In the webpage of the WRT54GL under setup->Advanched routing you can change the Operating mode from Gateway to Router to disable NAT. In turn you also have to add a static route via the webpage of the AG241 to let the AG241 know where the other lan segment is.

If you use a WRT54GL behind an AG241, don't you preferably put the AG241 into full bridge mode and do everything on the WRT54GL? I use Tomato, but I recall it keeps logs only by interface not device. Is there any interest from AG241 users in emulating a Speedtouch style "Public IP address on the LAN" function, where one device on the LAN can use the external IP and others nat?

ziion writes...

The boxes connected to the WRT54GL are indeed only visible if you disable NAT. In the webpage of the WRT54GL under setup->Advanched routing you can change the Operating mode from Gateway to Router to disable NAT. In turn you also have to add a static route via the webpage of the AG241 to let the AG241 know where the other lan segment is.

Cool, it's good to know I was somewhat on the right lines. This is kinda embarrasing because I bet it's really obvious, but could someone help me with what I need to change and enter? I had a go at it earlier and the result was that I couldn't access my WRT54GL from wireless or ethernet and had to reset it. For reference my WRT54GL is at 192.168.1.1 and my AG241 is at 192.168.2.1.

Here is how my router looks: img245.imageshack.us/img245/1084/w­ rtow8.png
And here is the modem: img245.imageshack.us/img245/1413/agrk1.png

Also I wonder if I need to assign static IPs to my devices, or if I can still use DHCP in this non NAT mode?

mstombs writes...

If you use a WRT54GL behind an AG241, don't you preferably put the AG241 into full bridge mode and do everything on the WRT54GL?

Unfortunately, despite being savvy with most things when it comes to computers, networking just goes completely over my head for some weird reason. I've basically left the devices doing what they are doing from their stock setup, with only minor adjustments to the settings (stuff like transmit power, passwords, disabling PnP), or how they operate. I did just quickly have a look at doing as you suggested, but I couldn't see anything obvious in the WRT54GL that would allow me to connect via PPPoA, despite that it does have PPPoE. It may just be known as something else on the router and I don't appreciate it.

Also just having googled to read up about QoS, I found this:
blog.bananasplit.info/?cat=15

Well the most interesting thing is that he summed up my request in two sentances, "I have a Linksys AG241 which is a nice ADSL router. Nice that is except that the QoS on it doesn’t support prioritisation by address/subnet; a must with 3 children in the house." So that's basically what I was trying to suggest, QoS prioritisation by address/subnet, which in turn would allow, for example, setting game consoles to high priority, with other devices on the network to medium or low so as not to cause lag.

Also from the Linksys help text:
"Application-based QoS manages information as it is transmitted and received. Depending on the settings of the QoS screen, this feature will assign information a high (traffic on this queue shares 60% of the total bandwidth), medium (traffic on this queue shares 18% of the total bandwidth), or low priority (traffic on this queue shares 1% of the total bandwidth) for the five preset applications and three additional applications that you specify."
So it looks like hard coded priorities. It would perhaps be nice if more levels could be added, or rejigged the priorities, so instead of:
High = 60%
Medium = 18%
Low = 1%

We could now have:
Critical = 100% (depends on how QoS works... See below)
Very High = 80%
High = 60%
Medium = 40%
Low = 20%
Very Low = 1% (maybe 5%)

You may even consider scrapping predefined priority levels and let the user select the percentage (which may be beneficial for people on slower connections, or faster connections where 5% on a 24mbps connection is over 1mbps)

However, going on what that guys blog says, I have a question. How exactly does the QoS work? Say I have PC1 and PC2 connected to the AG241 with a 1000kbps connection (for the sake of simplicity). PC1 is used for gaming and is set to very high priority (80%) and PC2 is set to low (20%).

Is it a hard restriction, or does it depend on network traffic? In theory PC1 will get up to 800kbps and PC2 will get up to 200kbps. Let's say PC1 is watching streaming video at 500kbps. Will PC2 still be hard limited to 200kbps, or would it be able to make use of the "idle" data, in which case PC2 would be able to bump up to 500kbps? Then if PC1 becomes busy and requires 650kbps, PC2 falls back to 350kbps due to priority?

Zero1 writes...

I couldn't see anything obvious in the WRT54GL that would allow me to connect via PPPoA

You cannot full bridge PPPoA, you can only 'half-bridge', which I don't think is currently an option in the AG241 and is generally poorly implemented in most modems I have come across. The WRT54GL doesn't have an ADSL port so can't do PPPoA.

I support what you are trying to do to avoid 'double nat', but have no experience of the WRT54GL in this mode - and suspect you will have trouble with it passing everything through 'transparently'.

mstombs writes...

I support what you are trying to do to avoid 'double nat', but have no experience of the WRT54GL in this mode - and suspect you will have trouble with it passing everything through 'transparently'.

I have got a similar setup, my WRT54GL is connected as client of my AG241 which just works fine. The only trouble is services like upnp. As upnp uses broadcast message wich aren't forwarded to clients on the WRT54GL.

Cool, it's good to know I was somewhat on the right lines. This is kinda embarrasing because I bet it's really obvious, but could someone help me with what I need to change and enter? I had a go at it earlier and the result was that I couldn't access my WRT54GL from wireless or ethernet and had to reset it. For reference my WRT54GL is at 192.168.1.1 and my AG241 is at 192.168.2.1.

1. Assuming that you connect one of the ethernet ports of the AG241 with the WAN port of the WRT54GL.

2. You setup the WRT54GL to use a Static IP as Internet Connection Type (NOT the router ip) for example:
Internet IP: 192.168.2.85
Subnet Mask: 255.255.255.0
Gateway: 192.168.2.1

3. In the AG241 you setup a static route (Setup->Advanched routing).
- The Destination IP Address: 192.168.1.0
- Subnet Mask: 255.255.255.0
- Gateway: the static ip address you assigned to WRT54GL in the example 192.168.2.85

4. In the WRT54GL set the operating mode to Router.

It probably could also be setup with dynamic routing using RIP, but I have no experience with that.

Also I wonder if I need to assign static IPs to my devices, or if I can still use DHCP in this non NAT mode

You can still use DHCP to assign IPs to your devices.

Good news, we have managed to get the v2.01.03 firmware to boot on a AG241 v1 so Phoenix for Version 1 is now able to become reality :)

AG241V1 Phoenix EU AnnexA v7.01.04 , Phoenix DE AnnexB v7.01.04

Phoenix v7.01.04 for the AG241 V1 has been released, there are two versions EU AnnexA and DE AnnexB.

get it here:

www.hacknpatch.co.uk

Just a quick question, I don't own one of these devices but they seem nice. You posted on the first page that the new firmware has a full telnet blah. Will you give the modem a ssh (perhaps dropbear) server any time soon ?

dbmoodb writes...

Will you give the modem a ssh (perhaps dropbear) server any time soon ?

I did it once before, manually. The old gcc wont compile the current dropbear without a few simple modifications, but that's no big deal.

The main problem is the key generation. The options are
1. Have the same key for every device - not very good.
2. Generate a new key everytime it starts up - still not very good - any decent client will probably warn you that the key has changed, and it can get annoying for the end user.
3. Store the key somewhere. At the moment the only place it can be stored to survive a reboot is nvram. I suppose it could work - base64 encode it and all.

The other alternative is to take some of the flash memory for a new, small partition, and make it writeable I guess. I don't really want to have to do that, even if it would be a neater way of doing things - then all the config files would be user editable (would require a rewrite of a lot of things to make it useable in this way.)

I had a look at the AG310, which uses an old Dropbear (which didn't need modification to compile with gcc 2.95) - it seems they went with option 2 - generate a new key every time.

Anyone have any other ideas? Are there enough people interested in ssh support? It does add about 400K to the size before compression.

--> bored if you can get it running (if anyone other than me - the guy who doesn't have one of these units) just make it a download- the router has wget no ? . Just generate a ssh key each restart and force it to make one the first time- (script --> delete the key or blah). Uptime of the modems should be good enough to avoid the problem of ssh keys - for those who want to use ssh.

ziion writes...

1. Assuming that you connect one of the ethernet ports of the AG241 with the WAN port of the WRT54GL. etc
Thanks for the help, I got my router and modem configured correctly, but I could not get online. I think it's probably the case that my ISP is only allowing 1 IP address per customer, and so I must use NAT. Thanks anyway though, it will come in handy for when I change ISPs as soon as my contract ends.

Also, I'm not sure if anyone's noticed; it might even just be me but here goes. Did anyone notice that the Tx and Rx dataflows reset after approximately 4GB? My question is can this be improved to show more than 4GB? Only reason I ask is because I was intending to use it to gauge my internet usage.

Hi Zero1,
i had noticed the Tx Rx stats get to 4GB but i hard reboot my ag241 so often (testing) that i hadn`t exceded it. we`ll look into it and add any fix to 7.01.05 :)

Amiga writes...

Hi Zero1,
i had noticed the Tx Rx stats get to 4GB but i hard reboot my ag241 so often (testing) that i hadn`t exceded it. we`ll look into it and add any fix to 7.01.05 :)
Appreciate it. At first I thought maybe it was an addressing issue or limitation, but on second thoughts I think it's more likely the case that it was simply implemented the way it is (in that it is not a hardware limitation which causes 4GB limitations). At least that's what I think. My WRT54GL seems to be counting data as I would expect, and since I just realised that I can find my ADSL usage from the WRT54G, you might decide not to spend time on increasing the limit, unless it's useful to more people.

img153.imageshack.us/img...66/wrt54goj0.png
Thanks.

Zero1 writes...

you might decide not to spend time on increasing the limit, unless it's useful to more people.

I think it would be best to fix it, my isp account is unlimited and at times i`ve hammered it 30GB in less than a week, so just from my point of view it would be invaulable.

Amiga writes...

Tx Rx stats get to 4GB

That's a challenge, isn't 4GB the limit of 32 bit integer?

mstombs writes...

That's a challenge, isn't 4GB the limit of 32 bit integer

I think if we can find the source to the HyperWRT Thibor firmware, that will help as it doesn't seem affected by that, or if it is, it has already been worked around. DDWRT may even have a stats counter that is not limited at 4GB, but I haven't used it so I don't know if it will be of use. Also the HyperWRT firmware is very close to the original Linksys firmware, so it may be a case of some copy and paste with a few modifications.

Zero1 writes...

Did anyone notice that the Tx and Rx dataflows reset after approximately 4GB?

Well, the code I wrote is designed to output up to:

18446744073709551615 bytes, or around about 17179869184 Gb.

Indeed, compiling a test version, with tx_bytes and rx_bytes hard coded to that number, and running it on the AG241 gives:
# ./limit
17179869183.9 Gb, 17179869183.9 Gb

Unfortunately, the kernel stores it as an unsigned long (include/linux/netdevice.h), which is limited to:

4294967295 bytes, or around about 4Gb, as you saw.

Simply changing that comes with its own problems though:
lkml.org/lkml/2005/8/18/44

Zero1 writes...

I think if we can find the source to the HyperWRT Thibor firmware

Found it thanks to archive.org.

It looks like this is how it handles it:

rc/ifstat.c:
if (rx_bytes < ifsp->rx_prev) {
u = (0xFFFFFFFF - ifsp->rx_prev);
u += rx_bytes;
}
/* some stuff */
ifsp->rx_bytes += u;

Rx_bytes is the current data from /proc, rx_prev is the previous data, and ifsp->rx_bytes is the current total. (I'm mostly guessing, but it looks about right.)

So suppose rx_prev = 3Gb, ifsp->rx_bytes = 3Gb, and rx_bytes is now 1Gb, as it has overflowed. u = (Max - approx 4Gb) - 3Gb = 1Gb. Then add rx_bytes (1Gb) to u, which makes it 2Gb. ifsp->rx_bytes will now be 3Gb + 2Gb = 5Gb, which is correct.

Of course, there are many ways this could be inaccurate if it were only run when you load the status page, but it looks like it's run fairly regularly through cron:

rc/services.c:
if (!nvram_match("ifstat", "0")) buf_to_file("/tmp/cron.d/ifstat", ­ "*/2 * * * * root /sbin/ifstat\n");

Zero1 writes...

Thanks for the help, I got my router and modem configured correctly, but I could not get online. I think it's probably the case that my ISP is only allowing 1 IP address per customer, and so I must use NAT.

Using NAT shouldnt be the problem, the setup I described is using NAT but only on the first router (AG241) not on the second one (WRT54GL). So something else is probably not quite right.

-IPV6 core patch (Ready core for possible IPv6 usage)

Currenlty IPv6 is running happily on my AG241, altough not with the phoenix firmware. To improve the IPv6 support of the phoenix firmware, maybe it is a good idea to patch the kernel with the USAGI patches (www.linux-ipv6.org) and to include a routing deamon like radvd. The confiuration for radvd could be difficult as you need some way to change the configuration file or create the configuration file at boot time like udhcpd.

But probably there aren't much users using IPv6 so it might not be worth the effort.

First, thanks for the effort. The OEM firmware is really lacking and unstable, your firmware has been superior in every aspect so far.
I have 2 suggestions:
1. Remove the default telnet daemon account (tech) because this gives anyone access to the router. I commented the user in the "passwd" file but it can be omitted completely as is in this ag241.ickz.com/node/12
2. Is there any way to make logs sort by newest - first ?

Bored Individual writes...

Of course, there are many ways this could be inaccurate if it were only run when you load the status page, but it looks like it's run fairly regularly through cron:
Very interesting. If I had to suggest a way round the limitation, I would have suggested something like this. As I mentioned, coding isn't my strong point, but what with it being pretty basic, I can get the gist of what is happening.

It basically replaces rx_bytes with ifsp->rx_bytes, and runs a simple formula to keep count so that rather than resetting at 4GB, it starts counting again but adds it to the previous 4GB. I suppose in simple terms it's kind of like if you were to count to 25 on your fingers, you could use your left hand to keep track of how many sets of 5 you have counted, and your right hand to count the units. When you have counted 5, you extend one finger on your left hand to signify that 5 so you can start again with your other hand without losing track.

About the HyperWRT source; I did intend to find it but I had to leave for work. I didn't want to come across as suggesting something and leaving it to you. I did want to try and contribute some way :) Good find, it may come in handy as reference for future things.

Regarding the inaccuracy; realistically I don't think that is much of an issue. The options as I see them are keep it as current and reset every 4GB, or keep counting and be a little inaccurate. Assuming most people would use this feature to check monthly usage or something similar, I think the majority of people would prefer to have a less accurate, longer running counter than have it reset all the time. My general thoughts are that even if it's inaccurate, even if it's incredibly inaccurate, it would still provide a better idea of monthly usage than a counter that resets at 4GB, because you are unlikely to know how many times it has reset.

That might be another, slightly less polished way of doing it, keeping it resetting at 4GB, but have a reset counter to say how many times it has reset (reached 4GB). That way if the normal counter was at 2GB and had the reset counter at 3, I'd know it has transferred 14GB. However I suppose that implementing this and just making a straight >4GB counter are more or less the same or have the same implications, just a different way of displaying it.

ziion writes...

Using NAT shouldnt be the problem, the setup I described is using NAT but only on the first router (AG241) not on the second one (WRT54GL). So something else is probably not quite right.
It's probably just me doing something wrong. I may try again later and see if I have any success. Thanks for the help.

ziion writes...

But probably there aren't much users using IPv6 so it might not be worth the effort.

Nor the amount of size it increases the firmware by. I think I will leave it out, unless a whole lot of people start clamouring for it.

Chexara writes...

1. Remove the default telnet daemon account (tech)

Hmm, good point. It's actually changeable in the AG241-AU firmware - it's used for SSL remote access I believe. I'll chuck an #ifdef HTTPS_SUPPORT around that in mketc.c.

2. Is there any way to make logs sort by newest - first ?

Syslog outputs them in the order they appear in the web pages, ie. it's a simple append each time syslog gets a message. It's of course possible to display it in the reverse order, but figuring the best way to do it (in code) will take a bit of thinking. There's always tac (cat/tac - it's the cat utility backwards you see), then the same logic can be applied to a pre-processed tmp file. I think there's even a tac for busybox 1.82 patch that could be adapted for 1.90, so it shouldn't add too much to the size.

Zero1 writes...

source to the HyperWRT Thibor firmware

was freely available on www.thibor.co.uk as patch to standard Linksys firmware, now seems to have gone the same way as www.hyperwrt.org, but sources are still browsable through Google codesearch

www.google.com/codesearc...r15c&btnG=Search

I'll ask over on Linksysinfo.org where Thibor is a mod.

in any case probably now superceded by Tomato (www.polarcloud.com) also a patch on standard Linksys firmware, the same author produced Hyperwrt tofu.

Only just noticed "busybox 1.82 patch that could be adapted for 1.90", Tomato has just rolled back from the "unstable version" to busybox 1.2 in v1.15 because udhcpc and syslogd (at least) were not working properly with 1.90. I guess most people don't use a dhcp client in an AG241 - its function would be handled by pppd?

Ok Phoenix with dropbear added

AG241 V2 EU Annex A
beta P01 is here please post feedback.

www.hacknpatch.co.uk/bet...ropbear-beta.bin

NB: this Beta firmware and is not for general use but for Bug testing purposes

Amiga writes...

Ok Phoenix with dropbear added

To clarify, the key should be stored in nvram and only generated once. It will be wiped on a factory reset though.

The code was adapted from Hyperwrt, which has this comment at the top:
// thanks to Ewrt :)

Sshd can be disabled in the web interface, and you can set the port as well.

Hello,
I'm using AG241EU V2 devices with official Linksys firmware 2.01.03.
They are working with ADSL line (PPPoA) and connecting with Zyxel Zywall 35 (VPN Tunnel).

After upgrade to Phoenix v7.01.04 EU/DE I can't connect VPN tunnel.

With Linksys Official 2.01.03 firmware VPN works normally.

Is the VPN function broken or disabled in Phoenix v7.01.04 EU/DE firmware?

Regards,
Mario