OpenWrt Forum Archive

Topic: davidc502 1900ac 3200acm builds

The content of this topic has been archived between 26 Feb 2018 and 7 May 2018. Unfortunately there are posts – most likely complete pages – missing.

Good information -- Thanks

Hi David,

Does your build use this wrt1200ac repo?

github.com/zhouruixi/WRT1200AC_LEDE_source

eliaspizarro wrote:

Hi David,

Does your build use this wrt1200ac repo?

github.com/zhouruixi/WRT1200AC_LEDE_source

The above is a fork from the lede project, so no. davidc502 builds are from the lede project sources.

hnyman wrote:

LEDE firewall includes time-based per-host restrictions by default (allowed/forbidden times per host, but no restrictions on durations are available). And there is even LuCI support for that as "traffic rules", so there is no explicit need for that add-on from k-szuster.

For blocking certain sites/domains, just install adblock (and luci-app-adblock).

While I'm certain that you could accomplish what I'm after using the built-in "traffic rules" option under Firewall settings, the luci-access-control-package certainly seems to be a more novice/user-friendly way to do it.

I understand not wanting to include it in David's base build, but could it at least be added to the available repository so that it could be manually added if desired? Is there any downside to that approach?

Thanks...Scott

(Last edited by zabolots on 27 Jun 2017, 15:58)

starcms wrote:

I don't believe I have seen anyone mention these in any openwrt or LEDE thread ever.  Honestly I didn't know they existed until I saw them on a one day sale on Amazon last week.

I got a great deal on them ($20 normally $45 for 2, $65 for 4), and while they might look ridiculously big, they certainly work well.  Linksys claims 7dBi (5GHz) and 4dBi (2.4GHz) vs 3.5dBi and 2.5dBi, respectively (which translates to 2x antenna gain at 5GHz and 1.5x gain at 2.4GHz), compared to the stock antennas.  Model number: WRT002ANT (or if you have any model besides the 1200AC, WRT004ANT (for 4 antennas)).

Hey, great info!! I'm sure I'm suffering from gain loss due to the construction techniques used down here in Chile ( we build everyhing in concrete due to being a heartquake active country sad ) ... so I guess going with this linksys high gain antennas will help!

I have a WRT1900ACS v2 ... should I replace 2 of the antennas , or the 4 of them?

I found this on amazon: https://www.amazon.com/s/ref=nb_sb_noss … WRT004ANT+

Thanks in advance for your advice.

starcms wrote:

Attention: Public Service Advisory:

To those who have changed their distros to point to the LEDE snapshot server in order to get the latest version of packages between @david's releases, please be sure to enter the following command line before upgrading any more packages.  The updates for these were released yesterday I believe; they are pre-reqs for one another; and if you update any or all of them, the router gets stuck in a boot-loop and/or won't boot fully (trust me, I learned from experience.  Luckily I was able to download the original versions from @david's distro using my cell-phone as a wifi hotspot, put the router into recovery mode, use winscp to push them onto the router, and install them using opkg --force-downgrade).

opkg flag hold netifd procd libubox libjson-script jshn

This way if you are using the script or command which updates all available packages, it won't update these and soft-brick your router.

That is all.

@starcms ... i saw libmbedtls appearing in the list of upgradable packages and I holded it just in case... should I?

Looks like we have a couple of kernel bumps, so will start on a build.  Of course as soon as it is finished, the new wifi driver will be out... lol

mariano.silva wrote:

Hey, great info!! I'm sure I'm suffering from gain loss due to the construction techniques used down here in Chile ( we build everyhing in concrete due to being a heartquake active country sad ) ... so I guess going with this linksys high gain antennas will help!

I have a WRT1900ACS v2 ... should I replace 2 of the antennas , or the 4 of them?

I found this on amazon: https://www.amazon.com/s/ref=nb_sb_noss … WRT004ANT+

Thanks in advance for your advice.

I've experiment with antennas for a while, and seems like you need all 4 set. If you change only 2 of them, you have higher signal level but significant performance drop. And forget about beamforming in this case.

mariano.silva wrote:
starcms wrote:

Attention: Public Service Advisory:

To those who have changed their distros to point to the LEDE snapshot server in order to get the latest version of packages between @david's releases, please be sure to enter the following command line before upgrading any more packages.  The updates for these were released yesterday I believe; they are pre-reqs for one another; and if you update any or all of them, the router gets stuck in a boot-loop and/or won't boot fully (trust me, I learned from experience.  Luckily I was able to download the original versions from @david's distro using my cell-phone as a wifi hotspot, put the router into recovery mode, use winscp to push them onto the router, and install them using opkg --force-downgrade).

opkg flag hold netifd procd libubox libjson-script jshn

This way if you are using the script or command which updates all available packages, it won't update these and soft-brick your router.

That is all.

@starcms ... i saw libmbedtls appearing in the list of upgradable packages and I holded it just in case... should I?

No, that one is fine to upgrade.  Just hold the ones I had mentioned.

(Last edited by starcms on 28 Jun 2017, 11:24)

mariano.silva wrote:
starcms wrote:

I don't believe I have seen anyone mention these in any openwrt or LEDE thread ever.  Honestly I didn't know they existed until I saw them on a one day sale on Amazon last week.

I got a great deal on them ($20 normally $45 for 2, $65 for 4), and while they might look ridiculously big, they certainly work well.  Linksys claims 7dBi (5GHz) and 4dBi (2.4GHz) vs 3.5dBi and 2.5dBi, respectively (which translates to 2x antenna gain at 5GHz and 1.5x gain at 2.4GHz), compared to the stock antennas.  Model number: WRT002ANT (or if you have any model besides the 1200AC, WRT004ANT (for 4 antennas)).

Hey, great info!! I'm sure I'm suffering from gain loss due to the construction techniques used down here in Chile ( we build everyhing in concrete due to being a heartquake active country sad ) ... so I guess going with this linksys high gain antennas will help!

I have a WRT1900ACS v2 ... should I replace 2 of the antennas , or the 4 of them?

I found this on amazon: https://www.amazon.com/s/ref=nb_sb_noss … WRT004ANT+

Thanks in advance for your advice.

You need the 4 pack (WRT004ANT).  Only the 1200AC uses the 2 pack because it only has 2 antennae smile

Edit: And remember what I said in my first post.  The high-gain antennae will only work properly if you only need coverage on one floor of a house (so if you have live in a one story house or apartment, you're fine).  The higher the gain, the more horizontal they broadcast (see the third pic I had attached to my original post).  You should also have all 4 pointing straight up.

(Last edited by starcms on 28 Jun 2017, 11:31)

starcms wrote:
mariano.silva wrote:
starcms wrote:

I don't believe I have seen anyone mention these in any openwrt or LEDE thread ever.  Honestly I didn't know they existed until I saw them on a one day sale on Amazon last week.

Hey, great info!! I'm sure I'm suffering from gain loss due to the construction techniques used down here in Chile ( we build everyhing in concrete due to being a heartquake active country sad ) ... so I guess going with this linksys high gain antennas will help!

I have a WRT1900ACS v2 ... should I replace 2 of the antennas , or the 4 of them?

I found this on amazon: https://www.amazon.com/s/ref=nb_sb_noss … WRT004ANT+

Thanks in advance for your advice.

You need the 4 pack (WRT004ANT).  Only the 1200AC uses the 2 pack because it only has 2 antennae smile

Edit: And remember what I said in my first post.  The high-gain antennae will only work properly if you only need coverage on one floor of a house (so if you have live in a one story house or apartment, you're fine).  The higher the gain, the more horizontal they broadcast (see the third pic I had attached to my original post).  You should also have all 4 pointing straight up.

Got it, yeap.... what if i just purchase 2 and leave 2 "low gain" and 2 "high gain" antennas installed? Will I have "some" better coverage in the same floor, and a little less on the upper floors? Sorry, but my software engineering degree didn't go into too much depth on electromagnetic radiations theory smile

With the 3200acm power being weaker, high gain antenna might be a must for people..

BTW -- I have the new build done. Will push it out to the server tonight.

New build r4470 on Kernel 4.9.34 has just been uploaded to the server.

This build was built with a new AMD Ryzen system which so far has been pretty sweet with how fast it can crank out a new build.... This last build for everything was 18 minutes... which is completely insane, considering my AMD 1090T 6 core processor would do it in 5 hours... lol 

I guess I'll go ahead and do the 4.4.x kernel now. I have time wink

*EDIT*

I'm thinking the last build didn't have to make everything because the 4.4.x is taking a lot longer.  So, 18 minutes must not be accurate. It was accurate in a sense of how long it too, but there must have been packages already compiled.

(Last edited by davidc502 on 28 Jun 2017, 23:42)

Oh, I have a kernel compilation with modules on wrt3200 takes 45 minutes. I'm afraid to see how many hours will take to compile Openwrt on wrt3200. Sorry about the offtop.

mariano.silva wrote:
starcms wrote:
mariano.silva wrote:

Hey, great info!! I'm sure I'm suffering from gain loss due to the construction techniques used down here in Chile ( we build everyhing in concrete due to being a heartquake active country sad ) ... so I guess going with this linksys high gain antennas will help!

I have a WRT1900ACS v2 ... should I replace 2 of the antennas , or the 4 of them?

I found this on amazon: https://www.amazon.com/s/ref=nb_sb_noss … WRT004ANT+

Thanks in advance for your advice.

You need the 4 pack (WRT004ANT).  Only the 1200AC uses the 2 pack because it only has 2 antennae smile

Edit: And remember what I said in my first post.  The high-gain antennae will only work properly if you only need coverage on one floor of a house (so if you have live in a one story house or apartment, you're fine).  The higher the gain, the more horizontal they broadcast (see the third pic I had attached to my original post).  You should also have all 4 pointing straight up.

Got it, yeap.... what if i just purchase 2 and leave 2 "low gain" and 2 "high gain" antennas installed? Will I have "some" better coverage in the same floor, and a little less on the upper floors? Sorry, but my software engineering degree didn't go into too much depth on electromagnetic radiations theory smile

You'd have to experiment.  Theoretically, if you put two high-gain in the back pointing up (for the lower floor) and two regular/low-gain in the front/side pointing diagonally (for the upper floor), it should work decently.  But you would lose beam-forming and on the upper floor you most likely would only get 2 spacial streams.  Theoretically you'd have better coverage on the bottom floor and possibly the same or a little less coverage on the upper floor, but only 2 spacial streams available.  Or it could bomb out completely and work like crap smile  You'd really have to try and see.  But the arrangement I just said has the best chance of working.  And the router would have to be located on the bottom floor.

(Last edited by starcms on 29 Jun 2017, 05:01)

davidc502 wrote:

New build r4470 on Kernel 4.9.34 has just been uploaded to the server.

This build was built with a new AMD Ryzen system which so far has been pretty sweet with how fast it can crank out a new build.... This last build for everything was 18 minutes... which is completely insane, considering my AMD 1090T 6 core processor would do it in 5 hours... lol 

I guess I'll go ahead and do the 4.4.x kernel now. I have time wink

*EDIT*

I'm thinking the last build didn't have to make everything because the 4.4.x is taking a lot longer.  So, 18 minutes must not be accurate. It was accurate in a sense of how long it too, but there must have been packages already compiled.

That's incredible about the AMD Ryzen.  How many cores?  I've only read about them; haven't gotten a chance to use one yet. 

I'll try the new build shortly and report back.

(Last edited by starcms on 29 Jun 2017, 02:10)

starcms wrote:
davidc502 wrote:

New build r4470 on Kernel 4.9.34 has just been uploaded to the server.

This build was built with a new AMD Ryzen system which so far has been pretty sweet with how fast it can crank out a new build.... This last build for everything was 18 minutes... which is completely insane, considering my AMD 1090T 6 core processor would do it in 5 hours... lol 

I guess I'll go ahead and do the 4.4.x kernel now. I have time wink

*EDIT*

I'm thinking the last build didn't have to make everything because the 4.4.x is taking a lot longer.  So, 18 minutes must not be accurate. It was accurate in a sense of how long it too, but there must have been packages already compiled.

That's incredible about the AMD Ryzen.  How many cores?  I've only read about them; haven't gotten a chance to use one yet. 

I'll try the new build shortly and report back.

This was built with the Ryzen 1700x which is a 8 core 16 Logical Processors. It was pretty much built to compete with Intel's high end processors. It doesn't even have a APU, so that means on-board video will not work as you have to have a separate video card.

So far I've been very pleased with the performance considering the price.

If you want to know more, let me know, and we can take it to anther thread.

(Last edited by davidc502 on 29 Jun 2017, 03:18)

@david, new build (r4470) is working great.  However, when you built dnscrypt-proxy, I believe you forgot to enable plugins.  I am seeing this in log:

Wed Jun 28 23:34:57 2017 user.warn dnscrypt-proxy: dnscrypt-proxy plugins support not present, ignoring block_ipv6 parameter...
Wed Jun 28 23:34:57 2017 user.warn dnscrypt-proxy: dnscrypt-proxy plugins support not present, ignoring blacklist parameter...
Wed Jun 28 23:34:57 2017 daemon.notice dnscrypt-proxy[4880]: dnscrypt-proxy Starting dnscrypt-proxy 1.9.5
Wed Jun 28 23:34:57 2017 daemon.notice dnscrypt-proxy[4880]: dnscrypt-proxy Proxying from 127.0.0.1:5353 to 107.181.168.52:443

It honestly doesn't affect me, but I know some of your users do use some of the dnscrypt plugins.



Edit: And the .ipk, which is the same version as in version r4323 (your last build), has shrunk from 78K to 75K.

If anyone needs plugins, they can simply install the .ipk from @david's repo from version r4323:

opkg install http://davidc502sis.dynamic-dns.net/snapshots/r4323/packages/arm_cortex-a9_vfpv3/packages/dnscrypt-proxy_1.9.5-4_arm_cortex-a9_vfpv3.ipk --force-reinstall

(Last edited by starcms on 29 Jun 2017, 05:50)

@david,

I'm also seeing this oddness in my log with the new build:

Wed Jun 28 23:33:43 2017 user.err kernel: [    8.076638] procd: cannot set group tty for /dev/ttyS11
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.082165] procd: cannot set group tty for /dev/ttyS8
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.087581] procd: cannot set group tty for /dev/ttyS6
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.092929] procd: cannot set group tty for /dev/ttyS4
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.098251] procd: cannot set group tty for /dev/ttyS2
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.103621] procd: cannot set group tty for /dev/ttyS14
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.109034] procd: cannot set group tty for /dev/ttyS0
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.114369] procd: cannot set group tty for /dev/ttyS12
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.119787] procd: cannot set group tty for /dev/ttyS9
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.125124] procd: cannot set group tty for /dev/ttyS10
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.130531] procd: cannot set group tty for /dev/ttyS7
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.135866] procd: cannot set group tty for /dev/ttyS5
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.141221] procd: cannot set group tty for /dev/ttyS3
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.146554] procd: cannot set group tty for /dev/ttyS15
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.151975] procd: cannot set group tty for /dev/ttyS1
Wed Jun 28 23:33:43 2017 user.err kernel: [    8.157296] procd: cannot set group tty for /dev/ttyS13

Never seen it before; no idea what it means.  I know we got a new kernel version and those are kernel errors...

Everything seems to be working fine though.

(Last edited by starcms on 29 Jun 2017, 05:58)

starcms wrote:

[Or it could bomb out completely and work like crap smile

This one. Any 2-kind-of-antenna (with different gain) config (front, back, diagonal) provide better signal level, but transfer rate divided by 4.
I check it before and roll back to ones from package.

starcms wrote:

I'm also seeing this oddness in my log with the new build:

Wed Jun 28 23:33:43 2017 user.err kernel: [    8.076638] procd: cannot set group tty for /dev/ttyS11

Likely due to https://git.lede-project.org/?p=source. … 9122dfa242

starcms wrote:

@david, new build (r4470) is working great.  However, when you built dnscrypt-proxy, I believe you forgot to enable plugins.  I am seeing this in log:

Wed Jun 28 23:34:57 2017 user.warn dnscrypt-proxy: dnscrypt-proxy plugins support not present, ignoring block_ipv6 parameter...
Wed Jun 28 23:34:57 2017 user.warn dnscrypt-proxy: dnscrypt-proxy plugins support not present, ignoring blacklist parameter...
Wed Jun 28 23:34:57 2017 daemon.notice dnscrypt-proxy[4880]: dnscrypt-proxy Starting dnscrypt-proxy 1.9.5
Wed Jun 28 23:34:57 2017 daemon.notice dnscrypt-proxy[4880]: dnscrypt-proxy Proxying from 127.0.0.1:5353 to 107.181.168.52:443

It honestly doesn't affect me, but I know some of your users do use some of the dnscrypt plugins.



Edit: And the .ipk, which is the same version as in version r4323 (your last build), has shrunk from 78K to 75K.

If anyone needs plugins, they can simply install the .ipk from @david's repo from version r4323:

opkg install http://davidc502sis.dynamic-dns.net/snapshots/r4323/packages/arm_cortex-a9_vfpv3/packages/dnscrypt-proxy_1.9.5-4_arm_cortex-a9_vfpv3.ipk --force-reinstall

Thank you --- I will look at this.

Short story:  WRT3200ACM running newest build r4470, getting dropouts on 2.4Ghz.
System log shows many
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.775567] ieee80211 phy1: buffer is NULL for tx done ring
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.784368] ieee80211 phy1: buffer is NULL for tx done ring
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.789972] ieee80211 phy1: buffer is NULL for tx done ring

Disabling phy1/2.4Ghz and re-enabling does not resolve, it gets same errors before and after.
Power off/back on does resolve it.

Long story:  I was on stock Linksys firmware FW_WRT3200ACM_1.0.6.181063_prod.img until about a week ago and was getting intermittent router reboots.  Tried a factory reset, and that seemed to cure the reboots, but would still get intermittent wifi dropout.
I installed LEDE r4323 a week ago, only ran it for about an hour because it gave frequent wifi dropouts.  Went back to stock, and ran that for about a week, still getting intermittent wifi dropouts, a few a day but it would recover so family not too disgruntled.
I saw new LEDE build r4470 yesterday so I popped that on last night, and the frequent wifi dropouts are back, now I have the problem mentioned above.
My changes from stock LEDE config:
Change LAN IP
change timezone
change phy0 and phy1 SSID, channel, security

rhauff wrote:

Short story:  WRT3200ACM running newest build r4470, getting dropouts on 2.4Ghz.
System log shows many
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.775567] ieee80211 phy1: buffer is NULL for tx done ring
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.784368] ieee80211 phy1: buffer is NULL for tx done ring
Thu Jun 29 12:43:14 2017 kern.err kernel: [31847.789972] ieee80211 phy1: buffer is NULL for tx done ring

Disabling phy1/2.4Ghz and re-enabling does not resolve, it gets same errors before and after.
Power off/back on does resolve it.

Long story:  I was on stock Linksys firmware FW_WRT3200ACM_1.0.6.181063_prod.img until about a week ago and was getting intermittent router reboots.  Tried a factory reset, and that seemed to cure the reboots, but would still get intermittent wifi dropout.
I installed LEDE r4323 a week ago, only ran it for about an hour because it gave frequent wifi dropouts.  Went back to stock, and ran that for about a week, still getting intermittent wifi dropouts, a few a day but it would recover so family not too disgruntled.
I saw new LEDE build r4470 yesterday so I popped that on last night, and the frequent wifi dropouts are back, now I have the problem mentioned above.
My changes from stock LEDE config:
Change LAN IP
change timezone
change phy0 and phy1 SSID, channel, security

Is this happening on all or just some devices connected via 2.4Ghz?  what are these devices?
If you used the same SSID for both stock and lede images, please configure your device(s) to "forget" the SSID completely, and then re-join it.

2.4Ghz drops out for ALL devices, mix of Samsung, Apple, Android, HP, phones, tablets laptops.

Yes, SSID is the same.  I will try deleting the 2.4 Ghz SSID on the devices, will be a challenge to get them all :-)