Asuswrt-Merlin 378.55 Beta is now available

[RMerlin]

OpenSSL scheduled a release (1.0.2d) for Thursday that fixes a high severity issue. It's probably worth including the updated release instead of 1.0.2c.

I know. Shouldn't be a problem to include it before the final 378.55 release, unless they break something else like they did with 1.0.2b.

 

[RMerlin]

OK no worries I'll revert back to 54 and wait for Asus to fix.

Incidentally I think that the parental time control (by Mac ID) may be broken too.

HB

Asus fixed it in 378_6975. I just backported it, and my test case confirmed the router was now generating the correct time array.

 

[IAAI]

Yeah or by the way OpenSSL is going, we can wait a few more days for 1.0.2e... or 1.0.2z.

Expect also other apps depending on OpenSSL such as OpenVPN also to release a new version soon, unless they are not affected which is highly unlikely.

Prepare for a high severity flaw, OpenSSL warns users

http://www.pcworld.com/article/2944...for-a-high-severity-flaw.html?linkId=15366487

 

[shooter40sw]

dnscrypt-proxy not working on this beta, at least for me.
I installed from last stable for some reason dnscrypt stopped working also on that version, when I upgraded anyway it deleted the jffs partition, so I installed entware and dnscrypt again, but dnscrypt fails to start, there is nothing in the logs that even mentions dnscrypt, those are my logs, so I have to disable the dnsmasq entries that are on the procedure for dnscrypt no I can have internet access to my pcs, but I can ping from ssh from the router, any ideas?
Thanks

ec 31 19:30:45 dnsmasq[650]: read /etc/hosts - 6 addresses
Dec 31 19:30:45 dnsmasq[650]: read /etc/hosts.dnsmasq - 4 addresses
Dec 31 19:30:45 dnsmasq-dhcp[650]: read /etc/ethers - 4 addresses
Dec 31 19:30:45 dnsmasq[650]: using nameserver 8.8.4.4#53 for domain local
Dec 31 19:30:45 dnsmasq[650]: using nameserver 8.8.8.8#53 for domain local
Dec 31 19:30:45 dnsmasq[650]: using nameserver 127.0.0.1#65053

-rwxrwxrwx 1 xxxxx root 80 Jul 8 05:43 post-mount
-rwxrwxrwx 1 xxxxx root 230 Jul 8 05:43 services-start
-rwxrwxrwx 1 xxxxx root 43 Jul 8 05:43 services-stop
-rwxrwxrwx 1 xxxxx root 1370 Jul 8 06:10 wan-start

Jul 8 06:16:26 ddns update: ez-ipupdate: starting...
Jul 8 06:16:56 ddns update: error connecting to updates.dnsomatic.com:80
Jul 8 06:16:56 ddns update: asusddns_update: 1

/tmp/home/root# ps |grep dns
273 nobody 1044 S dnsmasq --log-async
1363 Scorpio 1420 S grep dns

/tmp/home/root# opkg install dnscrypt-proxy
Package dnscrypt-proxy (1.4.3-1) installed in root is up to date.

The process is there but does not start automatically
-rwxr-xr-x 1 xxxxx root 256 Jul 8 06:07 S09dnscrypt-proxy


I need to start it manually
:/tmp/mnt/antares/entware/etc/init.d# ./S09dnscrypt-proxy start
Starting ... done.

nscrypt-proxy[1258]: Starting dnscrypt-proxy 1.4.3
Jul 8 06:38:17 dnscrypt-proxy[1258]: Initializing libsodium for optimal performance
Jul 8 06:38:17 dnscrypt-proxy[1258]: Generating a new key pair
Jul 8 06:38:17 dnscrypt-proxy[1258]: Done
Jul 8 06:38:17 dnscrypt-proxy[1258]: Server certificate #1435874751 received
Jul 8 06:38:17 dnscrypt-proxy[1258]: This certificate looks valid
Jul 8 06:38:17 dnscrypt-proxy[1258]: Chosen certificate #1435874751 is valid from [2015-07-03] to [2016-07-02]
Jul 8 06:38:17 dnscrypt-proxy[1258]: Server key fingerprint is ED19:BFBA:FAFC:9257FDC:68C7:69BF:AC24:94CD:743F:3C1D:4966:134D:FE2C:4BDC:F315
Jul 8 06:38:17 dnscrypt-proxy[1258]: Proxying from 127.0.0.1:65053 to 208.67.220.220:53

 

[RussellInCincinnati]

Prepare for a high severity flaw, OpenSSL warns users

http://www.pcworld.com/article/2944...for-a-high-severity-flaw.html?linkId=15366487

This is a Merlin _55 thread, maybe start another thread to bring up this topic (which is an interesting one, worth everyone seeing, not just people deep in a Merlin _55 thread).

 

[XzavieR]

Hey RMerlin - can't see it in your changelog but it might be one of those "miscellaneous" fixes: Did you find the issue that causes throughput to drop to a trickle every few days? As much as I have tried to check my logs, I can't find any error messages there, nor is there any CPU or memory hogging once the slowdown kicks in. It's only the throughput (both in and out) that slows to a crawl. It is definitely a router issue, as turning off/on my network cards doesn't fix the issue, only a router reboot fixes it.

 

[IAAI]

This is a Merlin _55 thread, maybe start another thread to bring up this topic (which is an interesting one, worth everyone seeing, not just people deep in a Merlin _55 thread).

sure thing , just adding additional reference to hggomes post
http://www.snbforums.com/threads/as...eta-is-now-available.25601/page-6#post-191321

----

Regarding 55 beta1

fw is great , just noticing these in log

Jul 7 22:00:41 disk_monitor: Got SIGALRM...
Jul 7 23:00:41 disk_monitor: Got SIGALRM...
Jul 7 23:58:26 miniupnpd[698]: upnp_event_recv: recv(): Connection reset by peer

no usb plugged in

 

[Blinkyz]

s

sure thing , just adding additional reference to hggomes post
http://www.snbforums.com/threads/as...eta-is-now-available.25601/page-6#post-191321

----

Regarding 55 beta1

fw is great , just noticing these in log

Jul 7 22:00:41 disk_monitor: Got SIGALRM...
Jul 7 23:00:41 disk_monitor: Got SIGALRM...
Jul 7 23:58:26 miniupnpd[698]: upnp_event_recv: recv(): Connection reset by peer

no usb plugged in

Jul 7 22:00:41 disk_monitor: Got SIGALRM=normal since .54

miniupnpd[698]: upnp_event_recv: recv(): Connection reset by peer= probly when closing Torrent programs (same here when i close Utorrent everytime)

 

[Wasteland]

I haven't noticed any particular change to that in 6117. Can you see if you have the same issue with Asus's 378_6117 beta firmware?

Hi Merlin. I tested with Asus' release Version 3.0.0.4.378.6975 and it has the same problem with only the Nexus 7 not able to re-join the WiFi network. So it looks like this is an Asus introduced issue.

 

[IAAI]

s
Jul 7 22:00:41 disk_monitor: Got SIGALRM=normal since .54

miniupnpd[698]: upnp_event_recv: recv(): Connection reset by peer= probly when closing Torrent programs (same here when i close Utorrent everytime)

No i didn't use torrents at that time at all

 

[Blinkyz]

No i didn't use torrents at that time at all

Ok im curious now, anyone can explain what is this line in the logs?:
upnp_event_recv: recv(): Connection reset by peer

I thought it was Utorrent the culprit

 

[ubimea]

Asus fixed it in 378_6975. I just backported it, and my test case confirmed the router was now generating the correct time array.

Hi Merlin; Thank you for all your excellent work on AsusWRT firmware. Just installed the latest 378.55 Beta on our three month old RT-AC3200 and can report no issues. Just Perfect.
Meanwhile, I thought about you today, as I read and heard several news sources report that the United Airlines system wide outage earlier today was caused by "router connectivity issues."
See here for example:
http://blogs.wsj.com/cio/2015/07/08/united-airlines-glitch-caused-by-router-issue/

Ha! United needs you!
RMerlin, the man that stamps out router issues!

Best regards, and again thanks for all that you do (and you will be missed when United comes offering that well deserved consulting contract.)

 

[RMerlin]

Hey RMerlin - can't see it in your changelog but it might be one of those "miscellaneous" fixes: Did you find the issue that causes throughput to drop to a trickle every few days? As much as I have tried to check my logs, I can't find any error messages there, nor is there any CPU or memory hogging once the slowdown kicks in. It's only the throughput (both in and out) that slows to a crawl. It is definitely a router issue, as turning off/on my network cards doesn't fix the issue, only a router reboot fixes it.

Anything related to wireless is entirely in Asus's camp. I have never experienced such issue myself, and I don't see it on their own changelog, so I can't say if it's something on your end, or something that they fixed, or simply never experienced either.

 

[RMerlin]

Meanwhile, I thought about you today, as I read and heard several news sources report that the United Airlines system wide outage earlier today was caused by "router connectivity issues."
See here for example:
http://blogs.wsj.com/cio/2015/07/08/united-airlines-glitch-caused-by-router-issue/

Ha! United needs you!
RMerlin, the man that stamps out router issues!

Best regards, and again thanks for all that you do (and you will be missed when United comes offering that well deserved consulting contract.)

Pretty sure they use something like Cisco or Juniper, not Asus. Completely different ball game, I have zero knowledge in ios

 

[RMerlin]

Ok im curious now, anyone can explain what is this line in the logs?:
upnp_event_recv: recv(): Connection reset by peer

I thought it was Utorrent the culprit

Could be another application that relies on UPNP, could be something on your LAN accessing the UPNP web server but not properly generating a UPNP request.

Just ignore it.

 

[IAAI]

getting these as well

Jul 8 23:59:57 kernel: htb: htb qdisc 14: is non-work-conserving?
Jul 9 00:08:09 kernel: htb: htb qdisc 12: is non-work-conserving?

 

[acale75]

I saw in the opening post the quote "ARM-routers will no longer periodically flush their memory caches".
I have a rt-ac68 router with this FW with an external usb 3 hd attached to it. If i reboot the router without the hd, the memory usage is completely normal, i say, about 50MB. When i plug in the hd, the memory usage increased very fast, up to 225MB. Only 26MB left. May i conclude thad the memory problem probably has something to do with de minidlna code?

 

[Chrysalis]

Running smooth. Less CPU it seems. Ping working, which was broken on previous fw.
AC87U+AC66U.

yep I am glad merlin added the buffer fix.

with buffers working properly again cpu usage will go down as wont be as many sys calls for i/o.

 

[RMerlin]

I saw in the opening post the quote "ARM-routers will no longer periodically flush their memory caches".
I have a rt-ac68 router with this FW with an external usb 3 hd attached to it. If i reboot the router without the hd, the memory usage is completely normal, i say, about 50MB. When i plug in the hd, the memory usage increased very fast, up to 225MB. Only 26MB left. May i conclude thad the memory problem probably has something to do with de minidlna code?

This is memory being allocated to disk buffers, and is perfectly normal.

 

[acale75]

This is memory being allocated to disk buffers, and is perfectly normal.

OK, thanks RMerlin