Asuswrt-Merlin - Project update

[RMerlin]

At least, the good news is that Asus now has an engineer actively taking care of networkmap, which is why they've been actively working on it over the past 18 months. Before that, it was mostly the same 5+ years old code which no one dared touch. That engineer actually visited SNBForums in the past.

 

[David B]

At least, the good news is that Asus now has an engineer actively taking care of networkmap, which is why they've been actively working on it over the past 18 months. Before that, it was mostly the same 5+ years old code which no one dared touch. That engineer actually visited SNBForums in the past.

Well considering that it's been re-written, that is to be expected, but the question is how long is maintenance going to last? We cannot be certain that Asus is going to continue to maintain it now that the bulk of the work on it is done. The 5+ year old code which no one dared touch was cared for at one point too.

 

[Xentrk]

AFAIK, there are already internal beta builds available for that model based on 382, so I expect they will eventually migrate that model to it.

The only device they specifically mentioned as being excluded when discussing it was the RT-N16.



I already squeezed everything I could out of openssl on these older ARM processors (including backporting optimizations from newer openssl releases at the time). For the RT-AC86U I don't know yet, optimization is far at the bottom of my list of things to do. Here are the current results using Asus's default optimizations on the RT-AC86U:

OpenSSL 1.0.2j  26 Sep 2016
built on: reproducible build, date unspecified
options:bn(64,32) rc4(ptr,char) des(idx,cisc,16,long) aes(partial) idea(int) blowfish(ptr)
compiler: /opt/toolchains/crosstools-arm-gcc-5.3-linux-4.1-glibc-2.22-binutils-2.25/usr/bin/arm-buildroot-linux-gnueabi-gcc -I. -I.. -I../include  -fPIC -DOPENSSL_PIC -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DOPENSSL_NO_HEARTBEATS -DL_ENDIAN -Os -march=armv7-a -fomit-frame-pointer -mabi=aapcs-linux -marm -ffixed-r8 -msoft-float -D__ARM_ARCH_7A__ -ffunction-sections -fdata-sections -O3 -Wall -DOPENSSL_BN_ASM_MONT -DOPENSSL_BN_ASM_GF2m -DSHA1_ASM -DSHA256_ASM -DSHA512_ASM -DAES_ASM -DBSAES_ASM -DGHASH_ASM

The 'numbers' are in 1000s of bytes per second processed.

type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
sha1             19452.99k    70705.77k  226198.86k   496435.88k   769111.38k
aes-128 cbc      67813.30k    74312.02k    77161.30k    78323.33k    78692.19k
aes-192 cbc      58153.93k    64428.71k    66564.44k    66911.77k    67012.75k
aes-256 cbc      52080.34k    56746.70k    58479.70k    58827.27k    59176.28k
sha256           43923.10k   151693.35k  383434.08k   621045.04k   762250.53k
sha512           11354.54k    45011.22k    70937.43k    99912.70k   113602.08k

I ran numbers for the AC88U. I appreciate the improvement.

type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
sha1             11352.24k    31662.34k    65645.34k    90957.67k   101701.76k
aes-128-cbc      41759.12k    48690.09k    50879.12k    51313.51k    50699.38k
aes-192-cbc      36538.26k    40220.68k    43558.09k    43825.43k    43813.36k
aes-256-cbc      32135.19k    35752.06k    38072.13k    38243.29k    37810.32k
sha256            8688.07k    22411.24k    43168.47k    56617.10k    62071.47k
sha512            3011.31k    11673.43k    17887.16k    25147.36k    28672.00k

 

[cybrnook]

I ran numbers for the AC88U. I appreciate the improvement.

type             16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
sha1             11352.24k    31662.34k    65645.34k    90957.67k   101701.76k
aes-128-cbc      41759.12k    48690.09k    50879.12k    51313.51k    50699.38k
aes-192-cbc      36538.26k    40220.68k    43558.09k    43825.43k    43813.36k
aes-256-cbc      32135.19k    35752.06k    38072.13k    38243.29k    37810.32k
sha256            8688.07k    22411.24k    43168.47k    56617.10k    62071.47k
sha512            3011.31k    11673.43k    17887.16k    25147.36k    28672.00k

Somewhat....

Look at aes-128-cbc 256kb for example.

 

[Xentrk]

Somewhat....

Look at aes-128-cbc 256kb for example.

Is the higher the number the better? In my test above, aes-128 cbc 256kb will process 50879.12k bytes/second while Merlin's aes-128-cbc 256kb result is 77161.30k bytes/second.

 

[cybrnook]

Sorry haven't had my coffee yet

 

[Xentrk]

Sorry haven't had my coffee yet

I have posted some dumb stuff on here a few times before I had my morning coffee. I now have a rule for myself not to post until I've had at least two cups of coffee in the morning.

 

[mstombs]

I don't suppose we'd be able to get Asus to consider open-sourcing networkmap by making lots of noise about them closing it? Just a thought...

Why would they? Surely with tight integration with Asus mobile tools they aim hook users into buying more Asus products. I assume Tomato's Device List is open-source (but gui licence not clear), just build a database linking arp table, dhcp static and dynamic lease table, wifi connection info, IEEE OUI lookup with user changeable icons and management methods...

 

[routers]

The RT-AC86U will be released today in the US (newegg link). Can't wait to read the reviews. I think it's gonna be a big deal for those in the market for a new Asus router. I hope to see it soon here in Europe as well. The fast CPU, the WiFi specs, the up to 110 MB/s USB transfer speed that almost maxes out 1Gbps. Wonder if a USB 3.1 Gen 1 device is really needed for that speed compared to 3.0. (when I connect my USB 3.0 enclosure with an SSD to my PC with USB 3.0 I get 4-500 MB/s read/write speeds). And the fact that Merlin already has the device makes me feel extra good about it. Only little downside initially IMO is the ROG design. I prefer the diamond design.

 

[RMerlin]

Wonder if a USB 3.1 Gen 1 device is really needed for that speed compared to 3.0.

USB 3.1 Gen 1 is the exact same thing as USB 3.0. It's just the USB working group that decided to confuse the hell out of end-users by renaming it.

USB 3.1 Gen 2 is the 10 Gbps stuff.

 

[maylyn]

USB 3.1 Gen 1 is the exact same thing as USB 3.0. It's just the USB working group that decided to confuse the hell out of end-users by renaming it.

USB 3.1 Gen 2 is the 10 Gbps stuff.

Second it

http://www.usb.org/developers/ssusb/USB_3_1_Language_Product_and_Packaging_Guidelines_FINAL.pdf

 

[Vexira]

what I don't understand is why asus never made the hard ware so obscure that, their firmware only works on their units and they don't need to close source the firmware, since it only works on certain hardware being theirs and no one elses so they wont have to close source it.

 

[RMerlin]

what I don't understand is why asus never made the hard ware so obscure that, their firmware only works on their units and they don't need to close source the firmware, since it only works on certain hardware being theirs and no one elses so they wont have to close source it.

Because at the root, the basic hardware design comes from the SoC manufacturer, so they're all very similar. Device manufacturers only make some tweaks around that design.

 

[huygens_25]

I understand that your 382 branch is currently private, so not accessible. But what about the official GPL 382 code from Asus? Is it public?

I'm interested in looking at the code to see if they bumped the Kernel version for the RT-AC68U. The reason is if they bump it, then I'm willing to keep around my Asus router for another couple of years. If not, I'm buying right now a Turris Omnia router. I want a modern OS on top of my router.

 

[RMerlin]

I understand that your 382 branch is currently private, so not accessible. But what about the official GPL 382 code from Asus? Is it public?

GPL drops are downloadable from Asus's support site.

They have not released a 382-based firmware for the RT-AC68U yet.

Kernels cannot be upgraded.

 

[Vexira]

Because at the root, the basic hardware design comes from the SoC manufacturer, so they're all very similar. Device manufacturers only make some tweaks around that design.

So in other words it's going to cost them a lot more for a fully custom solution. Where as it's more cost effective to get something already available but modifed since it's cheaper due to massive production.

 

[ColinTaylor]

So in other words it's going to cost them a lot more for a fully custom solution. Where as it's more cost effective to get something already available but modifed since it's cheaper due to massive production.

You see the same sort of thing with all sorts of electronic devices. Take graphics cards for instance, Nvidia will come out with a new GPU and publish a "reference design" for a PCI card using it. Then compare the resulting video cards from the likes of Asus, EVGA, Gigabyte, Palit, etc. They're all pretty much the same as the reference design with just little tweaks here and there.

 

[Vexira]

I understand that, about graphics cards I've seen it around for years, I was just suggesting a solution to the issue of close sourceing of the firmware.
If your remember agea and their phsyx and how it required a dedicated card to work or nvidia and their hardware phsyx requiring their gpu I order to work. I was suggesting propitiatory hardware to prevent the cloning issues then it can be open source.

 

[routers]

I asked ASUS Nordic by email about the release date of the RT-AC86U and they replied within a few hours (translated):
"Unfortunately we cannot inform here when it will be released. You can try online sellers or you can try asking local shops."

 

[huygens_25]

GPL drops are downloadable from Asus's support site.

They have not released a 382-based firmware for the RT-AC68U yet.

Kernels cannot be upgraded.

Thanks RMerlin, that's what I fear. But I had some hope that in order to offer the new WiFi meshing features, they had to go back to Broadcom and get some kind of updated SDK, hence newer Kernel.

OK, then bye bye Asus and hello Turris!