Re: [Cerowrt-devel] WNDR alternative for higher capacity

[Dave Taht <dave.taht@gmail.com>]

Rich Brown had contacted linksys. I don't know the current status.

So far we're not impressed with their follow through, or their
announcement, or their hardware. They seemed to think that merely
announcing they'd work with openwrt precluded things like, actually
contacting the openwrt devs to talk about their plans. This is the
rough equivalent of someone announcing a product that Apple would
support, without Apple;s actual endorsement.

Certainly linksys was a powerful brand and many folk have warm
feelings about their breakthrough new products in the early 00s, but
the company was a gutted shell long before it was sold off.

Over the past year...

I have had plenty of "early hardware" from various vendors to
evaluate, all rejected thus far, primarily due to the binary blob
problem. Some had pretty good kernel support, but rejected (like the
mirabox) for heat and ancient kernel issues as well, and a couple got
rejected because they couldn't push even half a gigabit.

It's a rather long list of unacceptable hardware at this point. I can
try to sort through what I can and cannot publish and get it out here
at some  point.

Currently on the top of the "not so-horrible-I'm going-to-barf list"
is the latest arm based atheros CPU chipset and mildly below that is
the mindspeed stuff. Mindspeed got bought recently and that's taken it
out of the running until they get their internal processes sorted out
or sold again.

The atheros chipset is still hotly smoking off the fab... I've held a
dev board in my hands but they took it away from me real fast. :)

The octeon was a contender for a while but with the death of mips it's
time to leave that arch. The armada 370s are not bad, the imx6 stuff
is not bad (there's been some good changes on the imx6 I should look
at it again)

That's just for cpu.

In all cases from all vendors the 802.11ac radio code is in binary
blobs, and barely works in the first place. The ath10k work is
proceeding in public at least, but they are still battling with
blowing up on the PCIe bus and other
severely low level problems. Ath10k support HAS landed in openwrt....

At the moment what I think I'm trying to do is divide the problem in half,
find a decent X86 based box with mini-pcie support, and solve the
gigE problem that way, and the wifi problem separately.

The cost goes way up... The closest I've found so far to what I wanted
was the latest nuc with sata support. Still want two hardwired
ethernet ports which it doesn't have....



On Sun, Feb 9, 2014 at 10:51 AM, Toke Høiland-Jørgensen <toke@toke.dk> wrote:
> Bill Merriam <lists@billmerriam.com> writes:
>
>> The announced but not yet available Linksys WRT1900AC might meet your
>> needs. The press release says they are providing advance copies of the
>> hardware to the openwrt developers and linksys expects openwrt to
>> ready when the hardware is released in "Spring 2014".
>
> I think it might, and I considered that; but got stuck on the "announced
> but not yet available" part...
>
>> It seems cerowrt developers need early hardware also.
>
> Well, maybe I'll try to convince someone at linksys that this is the
> case; any ideas about who I should email?
>
> -Toke
>
> _______________________________________________
> Cerowrt-devel mailing list
> Cerowrt-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cerowrt-devel
>



-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html