LWN: Comments on "The new Linksys WRT1900ACS router" https://lwn.net/Articles/660229/ This is a special feed containing comments posted to the individual LWN article titled "The new Linksys WRT1900ACS router". en-us Thu, 30 Oct 2025 05:13:02 +0000 Thu, 30 Oct 2025 05:13:02 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net The new Linksys WRT1900ACS router https://lwn.net/Articles/660620/ https://lwn.net/Articles/660620/ JGR <div class="FormattedComment"> <font class="QuotedText">&gt; Exactly like scenario 1 but I pot openwrt on it. Now I can't get it replaced.</font><br> You could flash it back to the stock firmware, verify that the port is still broken, and then send it back to them. The manufacturer does not need to know what firmware was on it, only that their approved firmware is currently on it, and it doesn't work.<br> </div> Wed, 14 Oct 2015 11:28:38 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660617/ https://lwn.net/Articles/660617/ pboddie <div class="FormattedComment"> I don't think it should surprise anyone that, given observations of certain mailing lists and the temptation to simplify product hardware, various smartphone producers are interested in adopting things like L4 implementations. I think it's inevitable that many devices will end up with Linux as some kind of task cohabiting with some real-time stuff and all sorts of other things besides.<br> </div> Wed, 14 Oct 2015 10:49:40 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660616/ https://lwn.net/Articles/660616/ sorpigal <div class="FormattedComment"> <font class="QuotedText">&gt; And if 'warranty' doesn't mean the vendor guarantees the router will work as advertised, what does it mean?</font><br> <p> "Warranty" should be read here as "The hardware won't die unexpectedly."<br> <p> Scenario 1: I buy one of these and a month later one of the wired ports stops working. Hardware defect. No problem, I can get it replaced under warranty.<br> <p> Scenario 2: Exactly like scenario 1 but I pot openwrt on it. Now I can't get it replaced.<br> <p> This isn't reasonable. If the flashing had failed that should void the warranty, but the buyer should still have some protection against unrelated defects when using a custom rom.<br> </div> Wed, 14 Oct 2015 10:41:31 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660592/ https://lwn.net/Articles/660592/ flussence <div class="FormattedComment"> PCEngines boards have first-class support for Coreboot too: in fact it's the only manufacturer to have *any* "cooperation score" on their supported hardware list:<br> <p> <a href="http://www.coreboot.org/Supported_Motherboards#Embedded_.2F_PC.2F104_.2F_Half-size_boards">http://www.coreboot.org/Supported_Motherboards#Embedded_....</a><br> <p> It's quite a bit more freedom-respecting (and customer-respecting!) than Linksys' bait-and-switch announcement here.<br> </div> Wed, 14 Oct 2015 04:38:11 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660577/ https://lwn.net/Articles/660577/ dlang <div class="FormattedComment"> some devices are like the Raspberry Pi where they are really a video chip that happens to have an ARM core that runs Linux. But these are pretty much the low end devices.<br> <p> The higher end multi-core devices aren't that way.<br> <p> It was never the case that the radio processor ran Linux. If it was a separate chip, it always ran it's own software (and usually nothing as sophisticated as an "operating system" or even QNX)<br> </div> Wed, 14 Oct 2015 02:37:06 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660556/ https://lwn.net/Articles/660556/ rahvin <div class="FormattedComment"> Honestly this is way beyond my skill, I'm just going by what I've read on XDA-developers from the people that do the actual hacking on phone ROMs. <br> <p> My understanding of the phones where the radio is integrated with the CPU is the CPU boots the radio OS and code first and the User facing OS (typically a version of android) is booted as a VM on top of the radio OS. This is supposedly a very stripped down VM where everything pretty much runs directly without translation to avoid the VM penalty. Again the point is to prevent any modification of the radio. I'm paraphrasing this, probably badly, but this is my understanding of how it typically works. In the past both OS's were Linux but there has been a recent trend to make the radio OS proprietary so they don't have to release the code for it and can prevent redistribution. They use memory and CPU locks that will only allowed signed code to run. <br> <p> This is typically why when you obtain hardware access on locked down Android phones you need to turn off the hardware security (typically referred to as S-Off on HTC phones), then flash the new ROM. There are also typically radio ROM's available to update the radio OS portion to newer released versions because the phone maker will typically update the radio a couple times after release to fix reception issues. These ROM's are flashed separately because they run completely segregated in separate memory but IIRC the radio OS has DMA access to the User facing OS. Because the manufacturer will typically apply a new radio ROM at the same time they deploy a new Android ROM the first thing the modding community does is try to decompress the update package into the separate updates so those that don't want to update the Android ROM but want the Radio fixes can obtain them. <br> <p> Anyway I'm going off track. If you want to understand more about how this works I would suggest visiting the developer forums on xda-developers.com and reading through some of the various forums for the device you are interested in and some of the more general android forums. IMO these are ugly hacks to try to enforce this goal of keeping the radio from being modified. <br> </div> Tue, 13 Oct 2015 23:51:19 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660555/ https://lwn.net/Articles/660555/ rahvin <div class="FormattedComment"> It is my understanding that the radio CPU has DMA access and boots the user OS on the second CPU in a VM at Ring 2. It could be secure, it could also be an exploit waiting to happen for those with the resources to reverse engineer it. <br> </div> Tue, 13 Oct 2015 23:20:08 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660543/ https://lwn.net/Articles/660543/ dlang <div class="FormattedComment"> Dell has a reputation of switching out the components without changing the model number.<br> </div> Tue, 13 Oct 2015 20:38:13 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660542/ https://lwn.net/Articles/660542/ mathstuf <div class="FormattedComment"> That's hopeful. Unfortunately, all of the ones I've seen are "preconfigured" where you don't actually get to change much of anything, so if the model you want is on a Broadcom line, all you can do is wait for a different line to make it that week :( .<br> </div> Tue, 13 Oct 2015 20:36:08 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660541/ https://lwn.net/Articles/660541/ mathstuf <div class="FormattedComment"> Yeah, I do something similar. Though I run 2 wireless networks (well, 4; both run on 2.4GHz and 5GHz), one guest and another which is no different than wired as far as network topo goes. I haven't separated the wifi from the wired internal network though; no reason to AFAIK (same as vasvir). Separating that again would be possible. The two holes poked in the firewall between the networks for guests (but not WAN) are for DHCP and DNS. Everything else is (well, should be) blocked coming in except for the forwarded ports through the router to the WAN (which is the same way guests access local services).<br> <p> The way it's set up is that it is agnostic to what is providing the Internet connection (currently a netbook which gets tries to reconnect to the tether every 15 seconds when it can't ping out 3 times in a row).<br> </div> Tue, 13 Oct 2015 20:35:00 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660532/ https://lwn.net/Articles/660532/ pak9rabid <div class="FormattedComment"> Personally, I've been using x86-based embedded systems for my routing/firewall needs. Currently I'm running one of these with an 802.11ac (ath10k) wireless card in it. It works pretty well:<br> <p> <a rel="nofollow" href="http://www.pcengines.ch/apu.htm">http://www.pcengines.ch/apu.htm</a><br> </div> Tue, 13 Oct 2015 18:14:55 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660527/ https://lwn.net/Articles/660527/ zlynx <div class="FormattedComment"> Bad software can actually destroy hardware and in those cases I don't think the manufacturer should have to fix it. <br> <p> For example, AMD X2 CPUs didn't have Intel's hard-coded temperature management. So if you modified the BIOS code to ignore the temperature sensors and overclock the CPU it was entirely your own fault if it melted down.<br> <p> Some GPUs have writable clocks and temp sensors as well. If you modify the driver to overclock without regard to the temperature it will melt down.<br> <p> I lost an AMD laptop to a bug in the Linux kernel and Xorg. It did something that locked the bus and put the CPU in a tight loop while I was away from my desk. When I came back I could hardly touch the case it was that hot. And it never booted ever again. Intel chips won't do that, and for all I know AMD won't either these days.<br> <p> A lot of embedded hardware does not have the same kind of hardware protection that x86 systems have. Because people expect to be able to install anything on desktop PCs, so desktop PCs have all kinds of firmware coding that you aren't allowed to touch.<br> <p> You can have software freedom over all of the embedded code but that means you can break everything too, and its your fault.<br> </div> Tue, 13 Oct 2015 17:19:55 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660519/ https://lwn.net/Articles/660519/ brunowolff <div class="FormattedComment"> This won't work with older devices with less memory. I tried doing that with a couple of routers that had only 16 MB of memory and they would constantly crash under load. I had to run them in bridge mode for most traffic. (Though they did NAT for wireless connections.) Since getting a couple of refurbished WNDR3800s I can route each port to do traffic isolation.<br> </div> Tue, 13 Oct 2015 15:15:50 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660499/ https://lwn.net/Articles/660499/ lsl <div class="FormattedComment"> Johill is right. Using OpenWRT you can run as many networks as you want, routing (and firewalling) between them.<br> <p> As an example you might look at the CeroWRT default configuration, that puts Ethernet and Wifi into different subnets, and also adds separate subnets for so-called guest wireless networks with differing firewall rules.<br> <p> See:<br> <a href="http://www.bufferbloat.net/projects/cerowrt/wiki/Default_network_numbering">http://www.bufferbloat.net/projects/cerowrt/wiki/Default_...</a><br> </div> Tue, 13 Oct 2015 14:44:57 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660490/ https://lwn.net/Articles/660490/ pinkpony <div class="FormattedComment"> How is anyone supposed to give you warranty the thing will work as advertised, when you flash it with a different firmware? When you flash it with something that doesn't work properly and you go back to the vendor you router works badly or not at all, what you expect? And if 'warranty' doesn't mean the vendor guarantees the router will work as advertised, what does it mean?<br> <p> Still, this shouldn't (and very often it legally *can't*) mean you are not getting any warranty on anything, if you bring the router back because the HW was obviously faulty (say, the power supply short-circuits and kills the entire router), the warranty in this particular case should be honored and the faulty HW replaced or repaired. But you cannot say you can flash the router and still expected it to have everything working properly as advertised on the box and have that covered by the warranty for the device...<br> </div> Tue, 13 Oct 2015 13:31:14 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660478/ https://lwn.net/Articles/660478/ tajyrink <div class="FormattedComment"> As a sidenote, Dell XPS 13 Ubuntu edition is available with Intel wifi in at least certain countries/configurations. Mine has Intel 7265.<br> </div> Tue, 13 Oct 2015 10:46:06 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660476/ https://lwn.net/Articles/660476/ ewan <div class="FormattedComment"> There's a difference between a hand-holding support contract for the software and a hardware warranty. I can buy an x86 server with a warranty and install any OS I like - the hardware vendor won't help be configure my web server, but they will replace dead disks. This should be the same.<br> </div> Tue, 13 Oct 2015 10:06:58 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660475/ https://lwn.net/Articles/660475/ arnd <div class="FormattedComment"> Interesting! Do you have an example which ones do that? A lot of the modern high-end SoCs (Broadcom bcm4709, Mediatek MT7621A, Marvell Armada, Qualcomm IPQ8064, ...) don't even have a radio and instead rely on PCIe wireless, so I assume you are not talking about those.<br> <p> Another weird setup that seemed to get more popular for a while was to have multiple SoCs that each run Linux in one box, I think I've seen one case that had three of them (DSL, router/switch, and Quantenna WiFi), though usually there are just two of them.<br> </div> Tue, 13 Oct 2015 08:12:07 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660474/ https://lwn.net/Articles/660474/ johill <div class="FormattedComment"> Anything you can install OpenWRT on should be able to do that (unless I'm not understanding it correctly)<br> <p> I have a few different networks on my systems - some have VPN access, some don't, etc.<br> </div> Tue, 13 Oct 2015 07:42:38 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660473/ https://lwn.net/Articles/660473/ vasvir <div class="FormattedComment"> Thanks for the heads up. I can totally see that happening to me also.<br> <p> In my office we have separated WIFI from DSL devices simply because DSL routers are bridging the WIFI with the LAN while I need a separate network with firewall rules for proper separation of the WAN, the internal LAN and random WIFI guests. So what you are saying makes total sense not only for decoupling and future upgradability but also for better security.<br> <p> In my home I don't need to decouple the LAN from the WIFI... Or maybe I should and I just don't know it yet?<br> <p> I wonder is there any product which can be configured to have 3 different networks, WAN, WIFI, LAN and have a gigabit switch?<br> <p> <p> <br> </div> Tue, 13 Oct 2015 07:18:11 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660472/ https://lwn.net/Articles/660472/ Cyberax <div class="FormattedComment"> <font class="QuotedText">&gt; I personally find this rather scary and a dramatic step back in freedom. That radio OS has ring 0 control of everything</font><br> Not really. Usually it has its own RAM and the host&lt;-&gt;controller protocol is fairly simple. If anything, it's even better than the current model from the security standpoint.<br> <p> Of course, a compromise in the controller OS will still be problematic, since it can do a lot of mischief (like trying to exploit the host driver by using malicious replies or by diverting traffic to attacker's host).<br> </div> Tue, 13 Oct 2015 02:48:38 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660470/ https://lwn.net/Articles/660470/ luto <div class="FormattedComment"> On the Motorola X Developer Edition (mine, at least), Motorola doesn't warranty that anything works if I reflash, but they go warranty that I'll be able to flash back, and they provide full support if I flash back to the factory images.<br> <p> I'm sure I could come up with a malicious image that would brick the device (by destroying the flash chip, for example), but that would be a silly thing to do.<br> </div> Tue, 13 Oct 2015 00:44:38 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660466/ https://lwn.net/Articles/660466/ rahvin <div class="FormattedComment"> I doubt anyone expects support, but the automatic voiding of a warranty where a bad flash didn't take place is just bad. If you are advertising this feature you shouldn't be saying that using said feature voids your warranty. I can certainly expect someone to have to flash back to the "supported" firmware to troubleshoot but just flashing a different OS should not be grounds for canceling a warranty. <br> <p> Frankly given OpenWRT is an advertised feature there are a few jurisdictions that such a cancellation of warranty clause would not be legally valid. <br> </div> Mon, 12 Oct 2015 23:14:09 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660463/ https://lwn.net/Articles/660463/ rahvin <div class="FormattedComment"> The newer routers behave just like the cellular has for awhile. There is the general OS CPU, and underneath that is the radio cpu and OS. The Radio CPU boots first then loads the OS controlling the radio. The radio OS then loads the second CPU and calls the User OS. The User OS communicates with the radio through commands much like the old modem ATA commands. <br> <p> I personally believe the FCC has been steering companies this direction because the radio OS is often a proprietary realtime OS that looks like a binary blob and is unalterable. It makes it pretty impossible to adjust any settings or power levels on the radio which is an FCC goal as we recently saw with the recent FCC request for comment. <br> <p> I personally find this rather scary and a dramatic step back in freedom. That radio OS has ring 0 control of everything, with the host OS running at a higher level. A compromise on the radio OS compromises the entire system and all data stored on it and all data passing through it. It's an NSA wet dream and a hackers gold mine. Sure it's harder to break but given the time and resources I doubt it's an impossible task. Frankly it's just bad design trying to route around something that's AFAIK never been a problem and as a side benefit opens the whole OS to exploit regardless of User OS. <br> </div> Mon, 12 Oct 2015 23:05:03 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660455/ https://lwn.net/Articles/660455/ roc <div class="FormattedComment"> How so?<br> <p> Providing support for random hacked firmware is impossible. They'd be crazy to promise to do that.<br> </div> Mon, 12 Oct 2015 22:28:11 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660444/ https://lwn.net/Articles/660444/ sdalley <div class="FormattedComment"> From the press release:<br> <font class="QuotedText">&gt; Linksys has collaborated with OpenWrt and Marvell to provide full open source</font><br> <font class="QuotedText">&gt; support for the WRT1900ACS in OpenWrt's stable and development branches.</font><br> <p> From the small print at the bottom of the sales page:<br> <font class="QuotedText">&gt; **While Linksys fully embraces the open source community</font><br> <font class="QuotedText">&gt; and is providing open source use capabilities,</font><br> <font class="QuotedText">&gt; they do not offer technical support on using open source firmware.</font><br> <font class="QuotedText">&gt; Installing 3rd party firmware is done at your own risk and replacing</font><br> <font class="QuotedText">&gt; factory-installed firmware with OpenWRT firmware will void your warranty.</font><br> <p> What the large print giveth, the small print taketh away...<br> </div> Mon, 12 Oct 2015 19:42:42 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660441/ https://lwn.net/Articles/660441/ Wol <div class="FormattedComment"> PC World it's £150.<br> <p> My problem is I'd like to get one and stick a raid array on it - an enclosure for the disks costs as much as a disk, if not more :-( A 3TB Red costs £100, a four disk enclosure about £300 :-(<br> <p> Cheers,<br> Wol<br> </div> Mon, 12 Oct 2015 17:40:55 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660434/ https://lwn.net/Articles/660434/ Pc5Y9sbv <div class="FormattedComment"> I understand your position with ADSL, as had a similar attitude a decade ago. I scoured the earth to find an ADSL router that worked with my ISP and with OpenWRT, and I spent time building and installing pre-release firmware to keep it working pretty well for my home office. But, you can fall into a trap that way. I couldn't move to newer ADSL standards from my old ISP while continuing to use my unicorn of a router that supported ADSL and OpenWRT. When I moved to a cable internet provider, it didn't even seem like an option anymore.<br> <p> If you care about the features and flexibility of OpenWRT you may opt for a separation between the modem that works with your ISP and the gateway router that runs OpenWRT on your behalf. In my more recent deployment, I first switched a cable router into modem mode. It stopped trying to provide a NAT gateway and just exposed DHCP to allow my real gateway to get the WAN-facing IP issued by the ISP. Later, I upgraded to a new cable modem without having to change my OpenWRT gateway router at all.<br> </div> Mon, 12 Oct 2015 17:19:16 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660430/ https://lwn.net/Articles/660430/ mathstuf <div class="FormattedComment"> Agreed. But it could be the differentiator between two alternatives. For example the Dell XPS 13 works great…except for the fact you need deal with the Broadcom wireless driver. It works fine once you get it, but you need to tiptoe around kernel updates.<br> <p> As for "viable" I mean "works without external modules and crashes". Even staging would be fine since that has a better hope of getting cleaned up.<br> </div> Mon, 12 Oct 2015 16:51:06 +0000 Is it certified by FSF's RYF (Respect your Freedom) program? https://lwn.net/Articles/660418/ https://lwn.net/Articles/660418/ drag <div class="FormattedComment"> Personally I am only going to be interested in running openwrt on a router of this sort, unless it has something extra-special going on or if it is a 'enterprise' grade switch running something like 'cumulus linux'. <br> <p> Does OpenWRT or Cumulus Linux themselves even meet the definition of 'free software that respects your freedoms' under the FSF's definitions? I really doubt it. <br> <p> So is RYF going to be something that I consider highly important or relevant? No, of course not. And I don't think most people here will consider it that relevant. It's a 'nice to have' thing, not a requirement. RYF is a nice thing, but it is too narrow in scope and divergent goals from most people means that it's not really that useful as a certification for most people.<br> <p> If you are a hard-core FSF-type then, yeah sure, it matters to you. But for most people it is probably never going to be that important despite the effort FSF puts into it.<br> <p> <p> <p> <br> </div> Mon, 12 Oct 2015 16:01:19 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660403/ https://lwn.net/Articles/660403/ TD-Linux <div class="FormattedComment"> That comparison would work better if you could actually use the laptop as a wireless router, but it doesn't even have an Ethernet port.<br> </div> Mon, 12 Oct 2015 14:43:25 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660398/ https://lwn.net/Articles/660398/ dlang <div class="FormattedComment"> It is disappointing that while they are bragging about it being 'open source'. they don't list Linux as a certified/supported OS<br> </div> Mon, 12 Oct 2015 13:46:47 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660397/ https://lwn.net/Articles/660397/ dlang <div class="FormattedComment"> 802.11ac routers are all in that price range. This is middle of the road to low cost for the specs.<br> </div> Mon, 12 Oct 2015 13:45:47 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660395/ https://lwn.net/Articles/660395/ dlang <div class="FormattedComment"> I think part of the question is what you define as "viable" :-)<br> <p> I have seen a couple recommendations, but I'd have to dig for them . The reality is that most people don't pick one, they get whatever's in the device they buy.<br> </div> Mon, 12 Oct 2015 13:32:55 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660394/ https://lwn.net/Articles/660394/ mathstuf <div class="FormattedComment"> Is there a list of viable -ac cards that ship in laptops?<br> </div> Mon, 12 Oct 2015 13:08:16 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660380/ https://lwn.net/Articles/660380/ niner <div class="FormattedComment"> If all you need is a WiFi router, the Archer C7 will probably suffice. If you need more, then it's no comparison to the WRT1900ACS as the latter has 4 times the RAM, 8 times the Flash, USB 3 (instead of 2), eSATA, and multiple times the CPU power. It also draws 4 times as much electricity (which still is not that much).<br> </div> Mon, 12 Oct 2015 08:13:00 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660378/ https://lwn.net/Articles/660378/ arnd <div class="FormattedComment"> They use the ath9k wireless driver, which is a softmac implementation that does everything in the kernel, and there is no other CPU in the wireless module. Newer Atheros hardware (including all 802.11ac devices) uses the ath10k driver, which has large portions of the code in a binary blob running on another CPU.<br> <p> The same seems to be true for almost everyone: As you get to higher data rates, the chip vendors slap another CPU core in front of the wireless hardware to offload the host CPU. The WRT1900ACS case is particularly interesting, because it's even the same kind of CPU core (Cortex-A9) in both the host SoC and in the wireless chip, and a lot of the code (from looking at strings in the binary blob) running on the wireless side is also the same as what we have in upstream drivers for older Marvell chips.<br> </div> Mon, 12 Oct 2015 07:48:49 +0000 Commodity distribution? https://lwn.net/Articles/660377/ https://lwn.net/Articles/660377/ arnd <div class="FormattedComment"> 128 MB of flash is really tight for modern distros. I'd probably install Debian to a nano USB stick with f2fs and just put the boot partition on the internal flash.<br> Another advantage of that setup is that you can boot the same drive with qemu-sytem-arm on a PC for testing.<br> </div> Mon, 12 Oct 2015 07:22:00 +0000 The new Linksys WRT1900ACS router https://lwn.net/Articles/660375/ https://lwn.net/Articles/660375/ tao <div class="FormattedComment"> Do they not need binary blobs, or are those binary blobs a (for users) non-upgradable part?<br> <p> Whenever there's a choice between "I can upgrade the firmware, but I don't have the source code" vs "The firmware cannot be upgraded at all" (or "The firmware can be upgraded, but only by a service centre" or "The firmware can be upgraded, but only using a Windows-only tool"), I prefer the former. The FSF seems to encourage the latter -- and consider it more free. I still have not understood why.<br> </div> Mon, 12 Oct 2015 06:42:14 +0000 "Turris" Open Hardware Router from nic.cz https://lwn.net/Articles/660291/ https://lwn.net/Articles/660291/ bnewbold <div class="FormattedComment"> I think it's great that Linksys is publicly framing open source support as a feature.<br> <p> This is sort of a late an tangential reference, but I recently found a (relatively) high performance Open Hardware (CERN OHL) home router which claims to run an open OpenWrt stack.<br> <p> nic.cz's Turris router: <a href="https://www.turris.cz/en/hardware">https://www.turris.cz/en/hardware</a><br> <p> It has PCIe slots for wireless, so even if the default hardware doesn't have appropriately libre drivers once could swap out for a different card. This doesn't seem to be for sale, it's a project of a Czech Repulic entity (business association? non-profit?) which distributed hardware for research, but they seem to be doing a follow-on product:<br> <p> <a href="https://omnia.turris.cz/en/">https://omnia.turris.cz/en/</a><br> </div> Mon, 12 Oct 2015 02:34:36 +0000