new firmware on asus site 374.257

[RMerlin]

I have 3 Xbox 360's and all stream fine on all .27X firmwares. This is very frustrating. Do you know when Asus will finally fix the 5 GHz band for all wireless adapters?

I don't, and they probably don't either. Debugging isn't something you can say: "I'll sit in front of the code, and have it fixed in two hours". All I know (and probably all they know themselves as well) is they are looking into it.

 

[hfm]

I was using the .270 firmware with the only driver that would work with my 6235 in this Samsung Gamer 7 I have which was the 15.5.6.48 version of Intel's driver. I was getting solid 300mbit N connections.

I really want to upgrade my firmware for the security fixes, so I tried the latest one on Friday and it was a complete disaster. Neither my notebook (I even tried upgrading drivers on the 6235) or my Nexus 4 would be able to connect to the network. I've downgraded to the first revision that had the security fixes and downgraded my 6235 driver. I can now connect again to 5Ghz with the Nexus 4 but my laptop still wont. I'm just using 2.4Ghz again. Only getting 144mbit now, the spectrum is crowded in my building there's at least 20 networks in range. Hilariously I am the ONLY 5Ghz network.

My wife hasn't been complaining about it since I was messing around with firmwares, usually she's the first one to go nuclear about poor wifi stability. So it's working fine on her MacBook Air late-2011 (scratch that she just told me she still has to fiddle with it sometimes during the day..) So upgrading firmware covertly didn't help..

It's pretty frustrating to say the least..

 

[wouterv]

The new Broadcom SDK they switched to implements HW acceleration in PPPoE, which is required to be able to use this router on various high speed ADSL connections. It also implements a fix for XBox360 compatibility.

I am not really getting this: PPPoE is at the WAN side of the router, the 5 GHz WiFi issues are at the LAN side of the router. Are drivers for those combined in one piece of software or hardware? I would think it are totally separated parts.

 

[RMerlin]

I am not really getting this: PPPoE is at the WAN side of the router, the 5 GHz WiFi issues are at the LAN side of the router. Are drivers for those combined in one piece of software or hardware? I would think it are totally separated parts.

The chips used in the router are manufactured by Broadcom. Broadcom provides router manufacturers with an SDK which contains (closed source) drivers, kernel patches, (closed-source) libraries to handle encryption, closed-source module to handle HW acceleration, and basic code for using their chips.

Router manufacturers will use that SDK to write their router firmware. Components of the SDK are tied together. You can't just take the wireless driver from one SDK version, and use it with the HW acceleration module taken from another SDK version.

That's why downgrading to the older driver was so difficult. I had to figure out which components were directly tied to the wireless driver, and also downgrade these components.

 

[wouterv]

Thanks for the explanation Merlin, that makes it clear.
Did you as insider got any confirmation from Asus that they are aware of ongoing issues with 5 GHz in the 374.257 build?

To jump onto network interface drivers, as also discussed here: I am also very reluctant to update the wireless drivers on the different PC's. They currently work well (with the RT-N66U / 270 build, and on other locations with various accesspoints). Experiences from the past learn that usual the drivers from the OEM are the best (obviously tested in the specific PC hardware) but often behind what the wireless chip manufacturer offers. I have seen very different odd behaviour with NIC driver updates....or old versions. I recently upgraded a Ralink NIC driver to the latest of the manufacturer and after it, it didnt see the 5 GHz of the RT-N66U / 270 build, so I reverted to the old driver. My advise is to not touch them as long as they dont cause problems.
In this case the Asus (Broadcom) firmware seems to be the culprit, as it looks like a single point of problems for various NICs. I can imagine though that the solution lies in a magic combination of router upgrade and NIC upgrade.

 

[RMerlin]

Thanks for the explanation Merlin, that makes it clear.
Did you as insider got any confirmation from Asus that they are aware of ongoing issues with 5 GHz in the 374.257 build?

I was told that the information has been relayed to development - they had seen the posted feedback.

 

[starkiller]

It's apparently not just the 5G band having issues with this latest build. I recently posted about drop outs on the 5G band and then flashed Merlins firmware with older drivers, and boom, no drop outs on 5G band. So just for laughs, I flashed back the latest stock Asus firmware and used my phone on the 2.4G band for the extra range as suggested by many that I wouldn't notice the speed difference on a cell phone . Worked great for 2 days speeds in the 40's... Suddenly today, speeds have dropped from the 40's to the teens... I tried changing all the channels, changing it to 20mhz, nothing helped. I figured perhaps some local interference... But I just now flashed back Merlin's firmware and wouldn't you know it, speeds on 2.4G band back into the 40's. What on Earth is Asus doing with these awful firmware updates? If it weren't for Merlin's firmware, I think I would have trashed this router as it seems Asus just can't get it right.

 

[RogerSC]

Okay, after reporting my problems with my Ralink-based wireless adapters to Asus, including their own USB-N66, Asus just sent me a new Beta version to test.

I've done so, and it works great with the Ralink wireless adapters that I have (Linksys AE3000 and the Asus one). Connects first time every time, good wireless speed, seems just fine.

So the fixes for these wireless adapters are coming, I'm hoping in the next official release version, since they have this fixed in Beta now.

I won't be running the new Beta for very long, since I've already gotten the "Neighbour table overflow" errors, and don't have an IPv6 firewall (which I've been working around with boot time scripts that I don't think I can run easily with Asus firmware), I may just turn IPv6 off soon, since it seems to be fine.

I'm sorry, I don't know about the Broadcom chipset wireless adapters, I don't have one of those to test with.

 

[starkiller]

Good to hear that they are finally getting close. Hopefully they will fix the issues with broadcom chips. They have had plenty of time to get this issue ironed out.

Sent from my SPH-L720 using Tapatalk 4

 

[bluepoint]

That's good to know that they fixed your Ralink clients. It's also good to know that Asus responds to bug reports from customers. It only shows when you report problems to asus they respond. Thanks for sharing.

 

[bluepoint]

Good to hear that they are finally getting close. Hopefully they will fix the issues with broadcom chips.

Not all Broadcom chips have issues though. I hope you reported your Broadcom chip's wireless model to asus.

 

[RMerlin]

Okay, after reporting my problems with my Ralink-based wireless adapters to Asus, including their own USB-N66, Asus just sent me a new Beta version to test.

I've done so, and it works great with the Ralink wireless adapters that I have (Linksys AE3000 and the Asus one). Connects first time every time, good wireless speed, seems just fine.

Out of curiosity, was that with build 374_334-g686a713?

I won't be running the new Beta for very long, since I've already gotten the "Neighbour table overflow" errors, and don't have an IPv6 firewall (which I've been working around with boot time scripts that I don't think I can run easily with Asus firmware), I may just turn IPv6 off soon, since it seems to be fine.

I've been trying to pitch the IPv6 firewall at Asus, since I believe such a thing is kinda essential if they are to provide usable IPv6 support. Hopefully their devs will either adopt my implementation, or implement at least some basic default rules on their end.

 

[RogerSC]

Out of curiosity, was that with build 374_334-g686a713?

I've been trying to pitch the IPv6 firewall at Asus, since I believe such a thing is kinda essential if they are to provide usable IPv6 support. Hopefully their devs will either adopt my implementation, or implement at least some basic default rules on their end.

The build version was actually pretty odd:

RT-AC66U_3.0.0.4_372_1191-g5da0c47_0801.trx

They assured me that it work on the rt-n66u despite the name, so I took the jump of faith and installed it. Incognito firmware *smile*. No one would suspect that it's really rt-n66u firmware *smile*.

I also mentioned the IPv6 firewall that you've implemented...so you got another, independent mention.

Also mentioned the /jffs stuff, since I've found it very useful, especially in avoiding the Comcast "neighbour table overflow" thing.

 

[RMerlin]

The build version was actually pretty odd:

RT-AC66U_3.0.0.4_372_1191-g5da0c47_0801.trx

They assured me that it work on the rt-n66u despite the name, so I took the jump of faith and installed it. Incognito firmware *smile*. No one would suspect that it's really rt-n66u firmware *smile*.

That is quite odd indeed Can you see if it's actually using the RT-N66U or RT-AC66U driver? Telnet in, and run:

wl ver

5.xxx is N66. 6.xx is AC66.

Maybe they are considering switching the N66 to the 6.x SDK in an attempt to resolve all the lingering wireless issues, and having one single driver to maintain for all those similar routers (N16/N66/AC66).

Also mentioned the /jffs stuff, since I've found it very useful, especially in avoiding the Comcast "neighbour table overflow" thing.

I'd be surprised to see Asus implement this on all their routers, as this could become quite difficult for their tech support to properly support (imagine a level one tech support being asked "I wrote a shell script and it doesn't work! HALP!", or "I flashed my router and all my scripts are gone! This is a bug, fix it!"

I do know that they are implementing JFFS support on one specific model (I forgot which, but it ain't one of the four that I support), as they intend to keep a copy of syslog there. Assuming it's not just an experiment they are toying with. Like the dual firmware support that's been silently popping in the source code about a year ago...

 

[RogerSC]

That is quite odd indeed Can you see if it's actually using the RT-N66U or RT-AC66U driver? Telnet in, and run:

wl ver

5.xxx is N66. 6.xx is AC66.

Unfortunately I've already gone back to the future (your beta3).

Sorry. I could go back to it, but it was kind of a pain to flash. If I do, I'll get the version to you.

 

[dsring]

After reporting my 5 GHz issues to Asus, they also sent me a link to this Beta firmware. I was told that the beta features a new SDK for the RT-N66. You have to put the router in recovery mode to flash it. It also changes the GUI to the AC66U. It is working well, and the 5G issues seem to be resolved.

 

[RMerlin]

Unfortunately I've already gone back to the future (your beta3).

Sorry. I could go back to it, but it was kind of a pain to flash. If I do, I'll get the version to you.

So it was that special test build that required to be flashed through recovery. I heard about that one.

Don't sweat it then.

 

[chief_os]

If I do, I'll get the version to you.

can you share to me this FW ?

 

[RogerSC]

can you share to me this FW ?

If you're having problems with your wireless network adapters, I'd suggest not going to this firmware unless Asus sends it to you to test. I'm back on RMerlin's new Beta3 firmware myself, and it works great with my wireless network adapters, having the old wireless driver. And it has the IPv6 firewall, which you should use if you're using IPv6.

On the other hand, you could write to:

cl-jeremy@asus.com

and mention to Jeremy that you're having problems with such-and-such wireless network adapters and want to test their latest firmware for your router to verify that it solves your problems.

I don't think that I should propagate firmware that was sent to me just to test, especially given the version string that it comes labeled with.

 

[jlake]

can you share to me this FW ?

Please make sure you tell him that you have tested and observed a performance issue with your Broadcom BCM43224. I know you had noted that in one of your posts. So it's not only a link speed issue, but also a throughput issue. I guess what I'm trying to say is....it's just not an erroneous link speed. Performance is involved too.

My Broadcom BCM43236 suffers the same 54-108 Mbps link speed issue.

It seems like a couple other people with centrino were still having issues as well.