Asuswrt-Merlin 3.0.0.4.374.32 is out

[RMerlin]

I just tested AP mode, and the router doesn't assign any IP to the eth0 interface, so it's not that the mode check was backward. I suspect it was a mistake as they removed the check for router mode when they also removed that nvram var that allows you to force WAN binding.

I'll put back the former mode check so the router won't bind Samba to the WAN interface again while in router or AP mode.

If you need an immediate fix, a temporary workaround would be to copy your smb.conf to /jffs/configs/ and then edit the "interfaces =" parameter to remove eth0 from it. After restarting Samba, it will start using that custom config file. That means you will have to re-create that custom config if you ever do any change to your plugged USB shares.

 

[cmillar6]

I just tested AP mode, and the router doesn't assign any IP to the eth0 interface, so it's not that the mode check was backward. I suspect it was a mistake as they removed the check for router mode when they also removed that nvram var that allows you to force WAN binding.

I'll put back the former mode check so the router won't bind Samba to the WAN interface again while in router or AP mode.

If you need an immediate fix, a temporary workaround would be to copy your smb.conf to /jffs/configs/ and then edit the "interfaces =" parameter to remove eth0 from it. After restarting Samba, it will start using that custom config file. That means you will have to re-create that custom config if you ever do any change to your plugged USB shares.

Amazing how you leaf through all that code and spot the problem right away, don't know how you do it. I'll wait for your fix as you've previously convinced me its not a security risk having Samba bind to WAN.

 

[RMerlin]

Amazing how you leaf through all that code and spot the problem right away, don't know how you do it. I'll wait for your fix as you've previously convinced me its not a security risk having Samba bind to WAN.

Every time a new GPL is released, the first thing I do is I diff the code with the previous release, so I can have a look at everything that Asus has changed between both versions. It can be a bit time consuming, but it's essential as I will later on need to merge that code into mine. I have to ensure that none of the Asus changes would conflict with my own code.

I remembered they had done that change to the Samba config generation, so a single git command showed be the exact change.

This should hopefully resolve the issue (as I can't reproduce it easily here).

 

[saintdev]

@@ -271,8 +274,12 @@ int main(int argc, char *argv[])
        fprintf(fp, "host msdfs = no\n");               // ASUS add
        fprintf(fp, "strict allocate = No\n");          // ASUS add
 //     fprintf(fp, "mangling method = hash2\n");       // ASUS add
+#ifndef RTCONFIG_BCMARM
        fprintf(fp, "bind interfaces only = yes\n");    // ASUS add
-       fprintf(fp, "interfaces = lo br0 %s\n", ( ( (!nvram_match("sw_mode", "3") && !nvram_match("sw_mode", "1")) || (!strcmp(nvram_safe_get("smbd_bind_wan"), 
+       fprintf(fp, "interfaces = lo br0 %s\n", (!nvram_match("sw_mode", "3") ? nvram_safe_get("wan0_ifname") : ""));
+#else
+       fprintf(fp, "interfaces = br0 %s\n", (!nvram_match("sw_mode", "3") ? nvram_safe_get("wan0_ifname") : ""));
+#endif
 #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,36)
        fprintf(fp, "use sendfile = no\n");
 #else

Makes me wonder if their router mode check might not just be backward. Mode 3 corresponds to Access Point mode. The current code states that Samba should only bind to WAN port if the router is NOT in access point mode. Previously, it also required a certain nvram setting to be set.

This means I will have to test in access point mode to see if it would make any sense for Samba to bind to WAN port. If yes, then I will assume they simply got their logic backward, and reverse it. Otherwise, I might just bind to the LAN bridge - this will (again) require further testing to ensure it doesn't break anything else.

Assuming nvram_match returns 1 for a match, the old line would be:

if sw_mode IS 1 or 3, then strcmp() smb_bind_wan (to whatever is on the rest of that line)

the new line is: if sw_mode is not 3

So they messed up somewhere, either when they changed the old line, or the old line itself.

 

[tom2415]

minidlna

since the version 3.0.0.4.372.31_0 find my device the router no longer in the network! I could NOT access my music from the tv, pc and from android smartphone.

I recently tried 3.0.0.4.372.31_0 and 3.0.0.4.374_32_0 and am now because of problems with the stream returned to 3.0.0.4_372.30_2.

3.0.0.4.372.30_2 appeared on the router and my router connected hard drive in network. I could access my music from the tv, pc and from android smartphone.

 

[aircoreboy]

I'll put back the former mode check so the router won't bind Samba to the WAN interface again while in router or AP mode.

Is this now a permanent solution in the F\W?

If you need an immediate fix, a temporary workaround would be to copy your smb.conf to /jffs/configs/ and then edit the "interfaces =" parameter to remove eth0 from it. After restarting Samba, it will start using that custom config file. That means you will have to re-create that custom config if you ever do any change to your plugged USB shares.

, or is this still the current solution?

 

[Martineau]

RMerlin,

Just installed 374.32_0-dwrpyd onto RT-N66U (not yet confident to try 374.32_0-SDK6 assuming it is the same as 374.32_0-dwrpyd with the new wireless drivers?) and as usual looks very stable.

P.S. Many thanks for including my DAR to diffentiate the WAN-STARTevent...

README-merlin.txt

<snip>

- wan-start: WAN interface just come up (includes if it went down and back up)
The WAN unit number will be passed as argument (0 = primary WAN)

Regards,

 

[primitivo]

Is it normal that DHCP IP assignment is completely random? E.g. 192.1.68.1.121, 192.168.1.72 etc (RTC-AC56U), usually my previous routers (e.g. Draytek) assigned them gradually from the IP pool count start.

 

[cmillar6]

Is it normal that DHCP IP assignment is completely random? E.g. 192.1.68.1.121, 192.168.1.72 etc (RTC-AC56U), usually my previous routers (e.g. Draytek) assigned them gradually from the IP pool count start.

Normal, my AC66u does the same thing.

 

[saintdev]

Is it normal that DHCP IP assignment is completely random? E.g. 192.1.68.1.121, 192.168.1.72 etc (RTC-AC56U), usually my previous routers (e.g. Draytek) assigned them gradually from the IP pool count start.

It's not actually random, just looks that way. It uses a hash of the client's MAC address to determine which IP address to use. This means it is more likely a client will receive the same IP every time, instead of being forced to change IPs because of a conflict.

I have not tested this with the dnsmasq version in included in Merlin's firmware, but you should be able to add dhcp-sequential-ip to /jffs/configs/dnsmasq.conf.add if you would like them to be sequential (not recommended).

Look at the entry under dhcp-sequential-ip for more info.

 

[tom2415]

RMerlin,

my tv and smartphone can not find my wireless router (rt-AC66U) by lan or wlan. previously with older ... 372.30_2 worked the stream to the devices with dlna correctly.

has it to do with the updated version of minidlna that I can suddenly not stream my music on the net for the two latest versions?
thanks for hints!

 

[Fab30]

Hi, i have problem with transmission and Lighttpd, service won't start after reboot... Any solution?

 

[RMerlin]

Is this now a permanent solution in the F\W?

This is what will be included in the next release.

 

[RMerlin]

RMerlin,

my tv and smartphone can not find my wireless router (rt-AC66U) by lan or wlan. previously with older ... 372.30_2 worked the stream to the devices with dlna correctly.

has it to do with the updated version of minidlna that I can suddenly not stream my music on the net for the two latest versions?
thanks for hints!

Check the location of your DLNA database, under USB Services.

 

[verbovet]

RMerlin,

I use rt-ac56u with the Russian ISP Beeline-Moscow, which uses l2tp with dual access.
Versions 3.0.0.4.374.32 and 3.0.0.4.372.31_2 don't work, log reads:
Unable to connect PPPoL2TP socket

Version 3.0.0.4.372.30_2 and current stock version 3.0.0.4.374.37 work perfectly.

 

[tom2415]

Check the location of your DLNA database, under USB Services.

Many thanks! Thats it!

p.s.: by the way. thank you for this great alternative firmware!

 

[jerry6]

I have noticed no difference in signal strength on either band with this new build. Both my Roku boxes show the same and same with the wireless computers. Anybody else notice any decrease in signal with this build ??

Not so much the build as the 5ghz band is weaker on the AC 66 U compared to the RT N 66U . Just changed routers as the RT N 66 U stopped working last weak . 2.4 band has a much stronger signal on the AC 66 compared to the N 66 .
Also the AC 66 RUNS A BIT HOTTER on the 5.0 band, 63 C compared to 52 C on the N 66 , on the 2.4 band both run at 51c

 

[RMerlin]

RMerlin,

I use rt-ac56u with the Russian ISP Beeline-Moscow, which uses l2tp with dual access.
Versions 3.0.0.4.374.32 and 3.0.0.4.372.31_2 don't work, log reads:
Unable to connect PPPoL2TP socket

Version 3.0.0.4.372.30_2 and current stock version 3.0.0.4.374.37 work perfectly.

I'll try to reproduce a similar test setup next week to see if I can reproduce the problem on an RT-AC56U (last time I tested it was on an RT-N66U).

 

[primitivo]

Hi, i have problem with transmission and Lighttpd, service won't start after reboot... Any solution?

I sent RMerlin pm regarding Lighttpd. According to some devs, there is a bug, however RMerlin has yet to confirm whether it's true or not.

 

[SjoerdNLD]

The Client List relies on ARP entries to build its list. If the client hasn't sent any traffic to the router yet or is behind a switch that means it hasn't sent traffic to the router itself yet but only on other devices on the same switch, then the client might not appear on the Networkmap yet.

Still 5 clients, and I'm working from ip 10.0.0.164, which has traffic for sure.
I tried finding javascript error but using firefox I could not find any. Pls see screenshot. I checked arp on the router, 10.0.0.164 is listed, pls see screenshot 2. Is there any other place to look? (RT-AC66U)