They sold enough wrt1900ac and made profit (at least broke even). I don't expect a "last word" in a positive way.

I think there be silence until TEH NEXT BIG THING has to be sold - like WRT2100AC which will definitely be "fully open source compatible" and openWRT firmware will be available "within a month".

This whole story kinda broke my heart...

After reading and verifying most facts and deducing some facts only said between the lines I am of the opinion that Linksys (or rather VEriK Systems) did a pretty good job on the WRT1900AC, but got mislead by one of their providers which have caused major headache for the Open Source aspect of the product.

The current progress with splitting the binary wlan driver is a major step forward, putting WRT1900AC at the same Open Source level as most Broadcom etc based router. It will allow the driver to be maintained with newer kernel releases and also fixing many of the kernel interface issues, but only to a limited extent. Very much improved situation compared to some weeks ago.

But it is not sufficient for an "Open Source" stamp on the product. But quite sufficient for making the product open source friendly. This combined with the other aspects of the product makes it a very promising alternative imho.

There is also some doubts on the stability of the new wlan driver, but it is unclear if that's firmware or driver issues, or simply wlan interference.

That's not "working", working is getting an email notification each time there's new post, or at least daily/weekly, are you saying this forum sw doesn't have that ability?

(Last edited by jalyst on 6 Aug 2014, 14:18)

If anything good comes of all this, perhaps it's that they've learnt their lesson in a big way, it'd be good to have more (healthy) competition, but, I won't hold my breath.

Whatever you like, but I like this feature the way it is. Go to PunBB Feature Request to talk about it. Please stop being off-topic.

+1

Afraid I'll have to remain OT for just tincy bit longer, as there's no way to PM users as far as I can see.

So, how do we stop being auto-logged out so that we can actually be emailed whilst we're "logged in"?
I'm not intimately familiar with the forum sw but, as I understand it now, it's very impractical...

I'm a member of MANY forums, & never have I found that I must log in every time I return, in order to start getting emails.*

Cheers.
*in the manner you described

(Last edited by jalyst on 6 Aug 2014, 15:08)

Check the "Remember me" when logging in. If that's not enough, check your browser config/extensions for something that cleans cookies. If that doesn't work, I am sorry, I can't help you. Further help can be provided to you on the PunBB site itself.

Thanks,

I wished I've read that page before buying the router.  Oh well, my 15 days exchange period is still good, may as well as grab the asus rt68u or netgear r7000 and put ddwrt or tomato on it.

Such a pity, it is a good beefy router too.  Just that not being able to put third-party firmware for openvpn and dnsmasq kinda ruin the experience.

Cheers

(Last edited by canucks on 6 Aug 2014, 17:13)

Actually my confidence in the quality of the programming in the marvell driver is very low.
I've been a professional embedded software developer for over 30 years and in that time have gotten pretty good at spotting familiar patterns.
Of course I haven't seen the driver source code itself, but most of the problems with the marvell driver that I've seen reported here strongly resemble in nature those that I've encountered on (only) other projects where the code looked like a dogs dinner.

(Last edited by justniz on 6 Aug 2014, 18:39)

If you have time I would encourage you to have a look at the driver that someone posted in the below link. AFAIK the user was fast and scooped the driver before it was removed from the Linksys GPL update.

https://github.com/TheDgtl/mrvl_wlan_v7drv

As a point of debate,  does a hardware specification exist for this marvel chipset... I've done some basic microcontroller stuff in the past but always worked from datasheets to get code written.  No doubt we have excellent software /embedded guys on this forum that could decode how this wireless chipset works?

(Last edited by guitarman on 6 Aug 2014, 19:05)

Not outside Marvell. Same for every other wireless chipset out there. Manufacturers sits tight on all documentation of the chips.

In best case there is documentation on the firmware interface, but usually that too is kept secret.

And in Marvell case even the code talking to the firmware is kept secret.. (almost), and implements most of the wireless stack not using the generic framework provided by Linux.

So my response yesterday to an email I sent about this issue, meh.. -

I am truly sorry for the issues you are having due to our device being an open source.
We take pride in our products and are willing to make this right for you.
If you choose to set up a refund I will be happy to assist you in this matter.

But first please take a look at the following links in regards to the open source.

We do have an SDK available and FW images for both Attitude Adjustment and Barrier Breaker available for download at GitHub.  The Links are:

Binary and OpenWrt patch at https://github.com/wrt1900ac/opensource/.   Here we have the ready to use image and the OpenWrt patch that we have submitted to OpenWrt.  The image is for someone who does not want to build code himself.   The patch is for folks who want to work with OpenWrt latest code where he would download code from OpenWrt.org, apply the patch and build the desired image.
Full source code is at https://github.com/jimmychungbelkin/Mamba.   Here includes our live source code for OpenWrt development in source code management repository.   Someone can do a code check-out from here and build the same image on #1 or create his own package.   For the developer convenience, this repository includes all necessary code from OpenWrt.org and our code.   This is exactly what we use for our own OpenWrt development.

Please let me know if these links help resolve your issues with the open source.
As I stated I will be more than willing to set you up for a refund if you choose to go that route.
Please advise and thank you for your patience sir,

George Rocha
Office of the President
Escalation Representative

I would not be at all suprised if they do what ASUS does with ASUSWrt, which is still not good enough.

I would say what happened is the marketing team got involved in "putting those ticks" on things, with a lack of technical knowledge/legal knowledge they made claims that they couldn't keep. I doubt they were mislead, companies like Belkin would be practiced enough at component acquisitions to be able to navigate their way around the legal issues.

If they were serious about releasing an open source router, (and not just marketing fluff) they would have actually made sure that partners/component makers were actually going to supply drivers with source and made that available on release.

I would not be at all surprised if Marvell would not sell them the source, or it would cost them dearly purchase it. The situation is very similar with Mobile chipsets (LTE/3G/UMTS etc)

I'm not surprised by this, unfortunately that is the case with most consumer hardware.

I think at this point Marvell is just fucking us around hoping we will go away and stop asking for the driver. I don't think Belkin has a legal agreement to force them to do so in place either. So much for their "open source router". I am really disappointed because I was going to buy this, thankfully I didn't. I really feel for those who got burned and I hope they all get refunds.

They first release ap8x.ko binary, then they release some source that includes a binary blob, they know full well people want the complete source that they can compile and patch it themselves so that the device may be supported when they decide to no longer support it. Eventually there's going to be a newer kernel their magical binary blob won't work with.

The simple fact is there's no financial gain for these companies to make a product so that it can be maintained in the future. They want you to throw it in the garbage and buy a new version when they drop support.

I think at this point for 802.11AC, something Atheros based is the best bet.

(Last edited by Euxauroo on 7 Aug 2014, 20:14)

I think at this point avoid buying future belkin/linksys/Marvell stuff will be my choice.

Agreed!

We have no sources for the Broadcom wifi and this won't change for quite some time. I didn't check the releases on github, but trunk runs flawlessly on the WRT1900AC (of course, without wifi).

You would be really surprised how little companies know about the FOSS world. Also, Marvell played the same game multiple times in the past with multiple vendors.

Well, they can simply tell them to either show up with mwl8k/mwifiex-pci support in a matter of weeks or they'll change the wifi card. The radio board connects via two standard miniPCIe slots, so replacing the cards and do a "v2" shouldn't be that bad for them - but would be for Marvell.

The Atheros radios are fine, but their MIPS cpus hardly have enough power to push them

Any reason the end-user can't do this today? Looks like it might be so easy to get the cards out?

If this is a good work-around, someone should tell Belkin about it.

Hit the ball back to Marvell (and get back in the game)

IMHO

(Last edited by gufus on 9 Aug 2014, 09:00)

That must be put to Belkin, if they don't try* despite all their bluster about F/OSS support.
Well it's official, they really don't have a clue...

*or haven't already

(Last edited by jalyst on 9 Aug 2014, 09:52)

Going from one vendor which at least tries to get into a reasonably acceptable workable situation while having to deal with binary objects to other who don't give a shit and lock you to a specific kernel release with specific config options (at least Netgear r7000 does, haven't looked in detail at rt68u blobs). Not sure how that's supposed to be a good thing.

Just make your own if there isn't one prebuilt. All the OpenWRT sources needed is there, just not in the main repository (yet).

Even a verbal agreement in a phone conversation can be held in court... If promises were made and Belkin wants to hold them to it they could fight this...

Of course not understanding or reading further in these forums before purchasing the router is my own problem. I will have to plod along with the process and hope that things improve. My 15/30 days is long over. And I doubt I can sell it to any one.

Other  than the wireless performance I'm pretty happy with the wan performance of this device. Many consumer routers can't support high speed high volume packet transfers. This router has the ram, CPU performance to keep up.

(Last edited by digitalgeek on 10 Aug 2014, 16:19)

Other then voiding your warranty, nothing stops you from doing it

And here you have it, contacting Linksys/Belkin about the open source issue and the best solution they have is to just immediately offer a refund for the router no matter how long you have had it since they can't offer any open source drivers to resolve the issue. Apparently after pointing them to this forum to have them review, the email I just got was to issue an immediate refund. Too bad, I like the router but...

-----

  We have gone over your case as a team and have come to the conclusion the best solution would be to offer you a refund for the WRT1900AC Linksys router.
I will also include a prepaid packing label to return the router to our Indiana warehouse.

Please reply with your phone number and the best time to call.

Do you still have your receipt?
Could you please reply with an attachment of the receipt in order to expedite this matter?

Thank you for your patience sir,

George Rocha
Office of the President
Escalation Representative
 
Belkin
12045 East Waterfront Drive
Playa Vista, California 90094