Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[banger]

+1

[johnson]

So it seemed like a good time for a consolidation post.

These are firmwares based of the latest release from Zyxel, AAJK.16 for the VMG1312-B10A & AAKL.24 for the VMG8924/8324-B10A.

Each folder has several firmware with naming as follows:


-jumboframes
Patches applied to allow 1508 byte packets to pass from a router to the modem in bridge mode. See this thread for further info.

-oldtelnet
Including a slightly older version of the telnetd binary that allows multiple concurrent telnet sessions. This is useful if you want to be able to log into your modem while running monitoring software such as DSLStats. Basic comparisons of it with newer versions indicate that the only change made by zyxel was the disallowing of multiple sessions.

-customcmds
A small change to add the execution of a user modifiable script located in the persistent /data partition during boot. This is useful if you wish to add a cap to your line to possibly remove banding/reduce retransmission, in a persistent way that survives reboots. See this post and thread for more info on reasons to cap. It is also useful if you have an ADSL line and wish to have persistent storage of changes to your target SNR. Please see this post for how to add the commands you want.

-x1
These firmwares contain a newer version of part of the broadcom DSL driver - adsl_phy.bin. It has been tested by several users and does not seem to offer any speed or stability improvements, however it does seem to fix issues I have personally encountered on a line with vectoring. If you have vectoring (unlikely unless you live in a BDUK area) then these versions are worth trying.

-x6
As above but including the newer A2pv6F039x6 adsl_phy.

VMG1312-B10A
https://www.dropbox.com/sh/c5no9h8ckyp2i4i/AADwAR6oVbzwFjtE-bLdy81Za?dl=0&lst=

VMG8924/8324-B10A
https://www.dropbox.com/sh/4xfif7ssq8pdq6u/AAChWy_N8UIVmpqQizmI2CUra?dl=0&lst=


They have all been at a minimum tested to boot have the correct changes on my devices, but as always you are flashing at your own risk and having a backup modem is highly recommended.

Edit: Change dropbox view to list

[Moderator edited to add new information from the OP.]

[johnson]

It is I hope clear just how much we all appreciate your hard work.

Its just nice to be able to provide something that people find useful.

If there are any other combinations of the options from the above post required that are not in the folders, just post here and I can make, test and upload them.

[Weaver]

regarding the graphical stats web server build, a feature request: How about a couple of links, maybe inside a <ul><li> etc which just point to /data/whatever, the textual stats report pages, for convenience. It would be a finger saver.

Also musing about another ui convenience request, could there be any way of arranging for a preceding call to/getdata to happen automatically before these individual per-command requests? If asking for such a thing even makes sense.

--

I was thinking about the on-boot custom script execution feature. There's nowhere appropriate to store the custom boot commands in writeable flash. Could we pull in a boot script by fetching it from the internet somewhere somehow ? Pull its content from a download from some fixed url and then run it, or perhaps even better fetch a well-known dns TXT record and execute its contents ? But what would the 'standard' well-known external reference be ? Could such an external reference be made variable by defining it in a config xml entry somehow, perhaps re-purposing an entry.

Or go further and simply put the whole custom script content into the config xml somewhere?

[7cfm]

I have just got a VMG1312-B10A from Ebay and tried to upgrade from 1.00(AAJZ.17)C0 to any of the johnson firmwares, I also tried to revert to stock 1.00(AAJZ.16)C0 but the router won't play ball, any ideas?

[vultura]

I have just got a VMG1312-B10A from Ebay and tried to upgrade from 1.00(AAJZ.17)C0 to any of the johnson firmwares, I also tried to revert to stock 1.00(AAJZ.16)C0 but the router won't play ball, any ideas?

Welcome to the forum.

Have you factory reset the unit prior to attempting firmware upgrade?

[7cfm]

Yep, no difference, I suspect this might have something to do with the router not letting me downgrade to an older firmware?

[vultura]

Yep, no difference, I suspect this might have something to do with the router not letting me downgrade to an older firmware?

I think it is possible to downgrade, but it may well require the .rom file from the firmware archive file you downloaded to be applied first using the restore configuration option.  It is simply a standard configuration file.

Others may be able to advise better than I.

[Weaver]

There’s an emergency boot mode that you can force it into with the appropriate hole-poking procedure and magic incantation. I don’t know if that would be relevant with this.

Burakkucat or Johnson may be able to help with that.

There is a procedure for killing the .rom file and indeed as vultura said that might be the way to go. High priests can advise with .rom-killing.

[Weaver]

Welcome, 7cfm!

[Weaver]

I must have been talking utter rubbish in this earlier post of mine when I thought there was no writeable flash area, no? I was thinking that /data/ was ramdisk? Is that wrong?

The stats code writes to /tmp and that obviously is writeable, and it survives a reboot? I haven’t tried pulling the power - ?

[Weaver]

The convenience links, urls I mentioned earlier, would be a nice tiny finger-saver. Could I even do it myself by modifying the html file, by code in the custom boot cmds thing? It would probably be too late and the web server would have loaded that html into ram and be running that copy permanently so it wouldn’t make sense, I presume. I would have added some

Code: [Select]

<a /> elements into the html with

Code: [Select]

sed perhaps.

More wish list:

2. I wonder if the graph could have a range limit. Sometimes, when things go wrong, huge SNRM values get introduced as giant spikes, and dominate the scale making it impossible to read the normal items as they are so tiny by comparison. I don’t know what to suggest about the details of sorting that out and I don’t know where to set range limit parameters, or just to hard-code some seemingly sensible range limits since we know what SNRM values make sense, from experience.

3. What about logging attenuation ? What do you think?

— Note: I am finding the stats logging and graphing thing so incredibly useful. I’m using it all the time and it has already helped me with a recent intermittent high-resistance fault on my line #3 which is now sorted.

— The easy access to the raw ZyXEL current stats at http /data/varStats is fantastic. I’ve used it in a program in my iPad which queries the upstream sync speeds from all the modems, calculates downrated IP data rates derived from these numbers them and puts these rates into a dynamically modified Firebrick router config file which is uploaded into my router.  This ensures that the router drives each modem at the correct max speed, and the total upstream traffic gets split between the links in the right ratios according to the links’ differing speeds.

So thank you so very much indeed for your work!

[johnson]

I must have been talking utter rubbish in this earlier post of mine when I thought there was no writeable flash area, no? I was thinking that /data/ was ramdisk? Is that wrong?

The stats code writes to /tmp and that obviously is writeable, and it survives a reboot? I haven’t tried pulling the power - ?

Indeed /data is writeable and persistent, not sure of its size though. There is about 155K of certificates and things there, but I wouldnt want the the log file written there every 30 seconds. The html/js/logfile etc is all copied to /var/tmp (in ram) at boot, so a reboot will unfortunately lose the days log.

The convenience links, urls I mentioned earlier, would be a nice tiny finger-saver. Could I even do it myself by modifying the html file, by code in the custom boot cmds thing? It would probably be too late and the web server would have loaded that html into ram and be running that copy permanently so it wouldn’t make sense, I presume. I would have added some

Have a look at the current state of the interface I have running on my VMG1312:
https://imgur.com/a/DzvYY9E

Buttons for each graph, options to scale linear or log and for tone range data either full or scaled to the top tone with bits used.

Also a button that goes to the data folder where you can select each raw stats file. They all update on access now so no /getdata call required.

Playing with having a side bar with useful info next to the graphs, will look something like this:
https://imgur.com/a/Iz3UsGY

Does this sort of interface appeal?

2. I wonder if the graph could have a range limit. Sometimes, when things go wrong, huge SNRM values get introduced as giant spikes, and dominate the scale making it impossible to read the normal items as they are so tiny by comparison. I don’t know what to suggest about the details of sorting that out and I don’t know where to set range limit parameters, or just to hard-code some seemingly sensible range limits since we know what SNRM values make sense, from experience.

3. What about logging attenuation ? What do you think?

Are there other giant spikes than the ones caused by the bugged upstream SNRM (going to 100 for negative values)? The log/lin button helps with visualising graphs with spikes in, but certainly clipping would be good.

Can certainly add attenuation, I imagine the graphs will be pretty boring on most lines but maybe they will offer some insight on lines with problems like yours.

Note: I am finding the stats logging and graphing thing so incredibly useful. I’m using it all the time and it has already helped me with a recent intermittent high-resistance fault on my line #3 which is now sorted.

Really glad its useful!  Apologies for the glacial speed with which I'm updating it.

One more thing I have got working on the testing version is in browser zipping of files, so one click to get an archive of the log file and all stats, should be useful for capturing events easily.

[Weaver]

> Does this sort of interface appeal?

Wow, just a bit. That looks fantastic. You have been working hard. Very appealing.

I had a problem where the attenuation jumped up by 8-10dB and so I’m thinking that that could possibly be a common pattern where there is some fault condition? Of course as you say, deathly boring for the most part.

> Are there any other giant spikes

No, I am certain that’s the only source of them. I’m still running that version, can’t remember where we are now.

Regarding zipping, does the small webserver used for stats support gzip encoding for http?

[johnson]

I had a problem where the attenuation jumped up by 8-10dB and so I’m thinking that that could possibly be a common pattern where there is some fault condition? Of course as you say, deathly boring for the most part.

> Are there any other giant spikes

No, I am certain that’s the only source of them. I’m still running that version, can’t remember where we are now.

Regarding zipping, does the small webserver used for stats support gzip encoding for http?

Ok will add attenuation. The 100+db spikes should be sorted by the fudge of subtracting upstream SNRM values from 100 when > 100, sure there are no cases when SNRM would be that high else where so no problems.

Good idea. I dont think the server supports dynamic compressing of files, but maybe with the right headers set it could serve pre-gziped ones - should save some ram! But the zipping I was talking about is in browser using a javascript library (96K - 28K gzipped  ) mostly just to take a snapshot of the various stats files and logs and present a single archive to download. Good for debugging as with the 100db spikes you had and with the possibility of being able to "upload" (not really because its all in browser) archived stats to view at a later date.