I don't think the Bssid being null is the only problem. I had fixed that problem but still couldn't get my mac to connect even though I could see the virtual SSID in the list
@jyavenard There is one user doing modified ASUS stockfirmware to RT-N66U and also port it to RT-N16 he seems have Multi ssid working in his program. Maby he have some tips, and on smallnets forum is a user from ASUS writing on messageboard. http://forums.smallnetbuilder.com/forumdisplay.php?f=37&order=desc https://github.com/RMerl/asuswrt-merlin. https://github.com/RMerl/asuswrt-merlin/downloads http://forums.smallnetbuilder.com/showpost.php?p=44154&postcount=21 http://www.lostrealm.ca/tower/node kthaddock
It works just fine with asus firmware. The issue isn't with the drivers per say, but how we are using them and the WL utilities.
Yes, I followed exactly the set-up guide. The signal of both bands are there, but I can't connect to it and when I see in Virtual Wireless tab, it said the driver error 00:00:00:00:00 although I saw there were connect MACs next to the error... What should I do ? Here is the picture of the error from others but it is similar to mine: http://imageshack.us/photo/my-images/835/homenetguest.png/
Yes I do as you can se on my signature, seems that it's 50/50 who has theirs routers working with Multi ssid. Maby that have to do with if you upgrade from earlier build without NVRAM resett or do a fresh install, If I remember right in the past there was if you set, Wireless Network Mode to AUTO or B, G, B/G or N mode. Eg set N-mode then reboot after that set auto again. kthaddock
I can assure you that no Toastman or shibby RT-N build are working for me with Multi-SSID. Clients can see the guest network, but can't connect to it. Knowing that Toastman's code there is identical to Shibby's one, I don't why there would be any difference between the two. There's also been no changes recently that would have made the code work better in newer build. As far as I can tell, Multi-SSID never worked properly with the newer drivers found in the RT-N branches. I can make it fail on all the asus and linksys routers I tried it on.
Guest Network is working for RT-N16 in rmerlins mod of unreleased Asus 155, assume the N66 one would work as well? RT-N16 wireless driver reports Code: admin@RT-N16:/www# wl ver 5.100 RC138.20 wl0: May 29 2012 08:58:33 version 5.100.138.20 Going to take some effort to port to Tomato, as all the interfaces seem to have changed name - wan port now a linux standard eth0, the guest wireless wl0.1 But interface name changes may also be due to fact that Asus firmware also has dual-wan code which is nearly usable!
THIS was the bug fix then: http://repo.or.cz/w/tomato.git/commit/d1b97fef19bec4e5714b0c01f2cc748dfa788e00
Confirm Shibby 097 MultiSSI working on RT-N16 and Belkin N300 Share Max. But only when br0 is DHCP enable. Does not work if br0 is disable ( on repeater)
Try this to get multissid working on a seperate bridge(br1), needed after reboot & if changes are made. Code: wl down wl up brctl show brctl addif br1 wl0.1 brctl show If on the Overview tab you don't see Interface Status: Up (LAN1) Code: ifconfig wl0.1 up
This fix is for the RT branch, not the RT-N with the newer drivers. I don't believe that there's an equivalent code in the newer drivers
Ok.... I have reliably working Multi-SSID now... will push after further testing... [removed as I've found an issue, for some reasons standard SSID only works if a virtual one is enabled]
@jyavenard: Cool! I started looking at the changes in the git, but then saw you pulled them. Why? BTW - what are these changes based on? Is the wl driver usage based on the Asus stock firmware, or just trial and error?
Yeah - the MAC address assignment is a mess. I had some posts about it in this thread. As far as I remember, the wlconf code did not just increment the last octet for each virtual network. It also incremented the first octet by 2 for all the virtual ones. On my E4200 RT-N, I get: 2.4GHz wl0 is C0:xx:xx:xx:62 5GHz wl1 is C0:xx:xx:xx:63 w0.1 is C2:xx:xx:xx:63 wl0.2 C2:xx:xx:xx:64 wl0.3 C2:xx:xx:xx:65 wl1.1 C2 :xx:xx:xx:64 wl 1.2 C2:xx:xx:xx:65 wl 1.3 C2:xx:xx:xx:66 So virtual and physical addresses don't conflict. But wl0.2 and wl1.1 do. In fact, I have wl0,wl1,wl0.1,wl0.2 all working at the same time (all with WPA2-PSK). One thing I wasn't able to do is operate wl1.1 (even without wl0.1), and also create more than two bridges and connect a vifs to them.
Hi All, I sucessfully setup Multissid in Rt-N16, and I'd isolated SAMBA disk using Shibby rules . Now I have one question, the second SSID I want to use it for guest service (no security), but I'm using "wireless filter" in my "normal" setup, could be possible to apply "wireless filter" only for SSID-1 and left SSID-2 without this rule? thanks in advanced
Ok. I've fixed most of the issues in the MultiSSID. It now works fine even with newer drivers, you can delete an entry after it's been created, don't have to just disable it anymore. I also fixed the issue where the new SSID wasn't working if using anything other than the main bridge/lan (br0)Changes on existing entries are properly saved too. Testet on RT-N66 (64k nvram) and RT-N16 (with both old old and new drivers) RT-N66 ONLY (64k NVRAM): http://www.avenard.org/wrt54-tomato....0501.2MIPSR2-jya-RT-N-VLAN-VPN-NOCAT-64K.trx RT-N16 and all other 8MB flash routers, with RT-N: http://www.avenard.org/wrt54-tomato...1.28.0501.2MIPSR2-jya-RT-N-VLAN-VPN-NOCAT.trx RT (older drivers) for all other MIPSR2: http://www.avenard.org/wrt54-tomato...B-1.28.7501.2MIPSR2-jya-RT-VLAN-VPN-NOCAT.trx All are based on Toastman-RT/RT-N Changes also in git.
IF it's brand new, install this one http://www.avenard.org/wrt54-tomato....0501.2MIPSR2-jya-RT-N-VLAN-VPN-NOCAT-64K.trx using the firmware recovery utility (there are a few guides available on how to do it). When done, make sure you reset the unit with either the reset button, or going into http://192.168.1.1/admin-config.asp and select Erase All date in NVRAM memory (thorough)
This image was done to test and debug Multi-SSID which is what this thread is about. It's based on Toastman images as that's what I use usually. I don't need all the new stuff Shibby keeps adding, I prefer something with less new features but more tested. Shibby will take my fixes and put it in his version like he usually does. So if you're aim is to have Multi-SSID properly working on your RT-N66 then use the image above, otherwise wait for a few days ....
I tried your RT-N fix on my RT-N16 and noticed that after about 18 hours the free memory was down to under 26M. Looks like there is a memory leak in your patch. Thought you would want to know.
Running well so far for about 3+ hours on an RT-N66U. Does not seem to have any memory leak problem - so far. Time Sun, 29 Jul 2012 20:19:56 -0400 Uptime 03:34:37 CPU Load (1 / 5 / 15 mins) 0.00 / 0.00 / 0.00 CPU Usage 0.99% Total / Free Memory 249.96 MB / 232.16 MB
If you looked at the patch and what it does, it's impossible that it introduces a leak. The patch modifies how the wireless interface is turned off and then back on. It does ifconfig eth1 down, wlconf down, re-assign the mac address, wlconf up, ifconfig up, wlconf start It's very similar to how it was done earlier, just in a different order. Say you have a leak there, as soon as the process has finished running, the memory would be freed. So whatever leak you are seeing is unrelated. Plus it must be a continuously running process, not a driver being stopped/started.
BTW, if you wan to see what uses the ram, run a ps or top command and look at the culprit. No point saying the memory is low with no extra details.
BTW, I'd like to thank tvlz hint on what got it to work for him... This helped me find where to look.. Was a key in finding what was wrong
I just flasht RT-N66 ONLY (64k NVRAM): http://www.avenard.org/wrt54-tomato....0501.2MIPSR2-jya-RT-N-VLAN-VPN-NOCAT-64K.trx But the router does not boot (flasht with firmware restore tool) Lucky i got i flasht back to original firmware,any idea what i am doing wrong? I have the asus rt-n66u
Did you use a 64K nvram firmware before? If not, you must do a full reset by holding the reset button until the pwr LED blinks. Or use the same as the RT-N16 one (no 64K nvram)
No,i did not use the 64k version before,i used the normal merlin build. So best way is to first reset the router? and then in recovery mode,flash your firmware,and after that clear the nvram?
Merlin version 144 and later use 64K nvram Exactly as if you had flashed tomato the first time, after flashing from official Asus (Merlin build falls under this category), then you *must* perform a full reset. You do the reset after flashing of course; not before! This has nothing to do with this thread, your questions/problems are about how to flash tomato. Can we keep this on topic?
@jyavenard did you try compile K24? After commits of MultiSSID K24 seems to be broken make[4]: Wejście do katalogu `/root/openwrt/tomato.git-compile/src/router/rc' [rc] CC rc.o [rc] CC init.o [rc] CC interface.o [rc] CC network.o network.c: In function 'stop_lan_wl': network.c:367: warning: unused variable 'subunit' network.c:367: warning: unused variable 'unit' network.c: In function 'start_lan_wl': network.c:415: warning: unused variable 'ifr' network.c: At top level: network.c:501: error: expected identifier or '(' before '}' token make[4]: *** [network.o] Błąd 1 make[4]: Opuszczenie katalogu `/root/openwrt/tomato.git-compile/src/router/rc' make[3]: *** [rc] Błąd 2 make[3]: Opuszczenie katalogu `/root/openwrt/tomato.git-compile/src/router' make[2]: *** [all] Błąd 2 make[2]: Opuszczenie katalogu `/root/openwrt/tomato.git-compile/src' make[1]: *** [bin] Błąd 2 make[1]: Opuszczenie katalogu `/root/openwrt/tomato.git-compile/src' make: *** [e] Błąd 2
I'll have a look. What make option are you using? I haven't a single K24 build that actually compile properly. So even fixing the problem there isn't going to help much Edit: there's no way tomato-shibby would have built a K24 build as it is today... There's plenty of references to K26 header fine, in particular in router/iptables
Do you use a different /opt/brcm link to compile? Now iptables compile with your latest change, but router/rc/mtd.c fail sounds to me that the toolchain used to compile is using K26 headers , and so that fail
i propably found problem and solution. In function start_lan_wl we have #ifdef CONFIG_BCMWL5 <------><------><------><------><------>if (strncmp(ifname, "wl", 2) == 0 && strchr(ifname, '.')) { #endif <------><------><------><------><------><------>char nv[40]; <------><------><------><------><------><------>snprintf(nv, sizeof(nv) - 1, "%s_bss_enabled", ifname); <------><------><------><------><------><------>if (!nvram_get_int(nv)) <------><------><------><------><------><------><------>continue; #ifdef CONFIG_BCMWL5 <------><------><------><------><------><------>if (get_ifname_unit(ifname, &unit, &subunit) < 0) <------><------><------><------><------><------><------>continue; <------><------><------><------><------><------>set_wlmac(0, unit, subunit, NULL); #endif <------><------><------><------><------>} // <-- this is our problem #ifdef CONFIG_BCMWL5 <------><------><------><------><------>else <------><------><------><------><------><------>wl_ioctl(ifname, WLC_GET_INSTANCE, &unit, sizeof(unit)); #endif IMO should be #ifdef CONFIG_BCMWL5 <------><------><------><------><------>if (strncmp(ifname, "wl", 2) == 0 && strchr(ifname, '.')) { #endif <------><------><------><------><------><------>char nv[40]; <------><------><------><------><------><------>snprintf(nv, sizeof(nv) - 1, "%s_bss_enabled", ifname); <------><------><------><------><------><------>if (!nvram_get_int(nv)) <------><------><------><------><------><------><------>continue; #ifdef CONFIG_BCMWL5 <------><------><------><------><------><------>if (get_ifname_unit(ifname, &unit, &subunit) < 0) <------><------><------><------><------><------><------>continue; <------><------><------><------><------><------>set_wlmac(0, unit, subunit, NULL); <------><------><------><------><------>} <------><------><------><------><------>else <------><------><------><------><------><------>wl_ioctl(ifname, WLC_GET_INSTANCE, &unit, sizeof(unit)); #endif Am i right?
I have a fix for the compilation not building, however I want to test it first as what it does is quite significant, I need to test how it behaves with my WRT54. Last I try to flash a K26 build on it, I got a nasty brick...
Ok.. The issue was the mipsel toolchain was using a K26 link... The last update you just did on iptables, give me compilation error, reverting it and it compiles
when you want compile K24 you have to use toolchain for K24 http://repo.or.cz/w/tomato.git/tree/refs/heads/tomato-shibby:/tools/brcm you have to link hndtools-mipsel-uclibc -> K24/hndtools-mipsel-uclibc-4.2.4 hndtools-mipsel-linux -> K24/hndtools-mipsel-uclibc-4.2.4 i`m still fighting with IPP2P on K24. I have kernel compilation error because in K24 file include/net/netfilter/nf_conntrack_core.h does not exist. I changed #include <net/netfilter/nf_conntrack_core.h> to #include <linux/netfilter_ipv4/ip_conntrack.h> and now i have: ipt_ipp2p.c:32: error: syntax error before "int" ipt_ipp2p.c:32: warning: type defaults to `int' in declaration of `module_param' ipt_ipp2p.c:32: warning: function declaration isn't a prototype ipt_ipp2p.c:32: warning: data definition has no type or storage class ipt_ipp2p.c: In function `match': ipt_ipp2p.c:836: warning: implicit declaration of function `nf_ct_get' ipt_ipp2p.c:836: warning: assignment makes pointer from integer without a cast ipt_ipp2p.c:841: error: dereferencing pointer to incomplete type ipt_ipp2p.c:841: error: dereferencing pointer to incomplete type ipt_ipp2p.c:842: error: dereferencing pointer to incomplete type ipt_ipp2p.c:844: error: dereferencing pointer to incomplete type ipt_ipp2p.c:845: error: dereferencing pointer to incomplete type ipt_ipp2p.c:847: error: dereferencing pointer to incomplete type ipt_ipp2p.c:847: error: dereferencing pointer to incomplete type ipt_ipp2p.c:847: error: dereferencing pointer to incomplete type ipt_ipp2p.c:878: error: dereferencing pointer to incomplete type ipt_ipp2p.c:900: error: dereferencing pointer to incomplete type ipt_ipp2p.c: In function `init': ipt_ipp2p.c:962: warning: implicit declaration of function `need_conntrack' make[6]: *** [ipt_ipp2p.o] Błąd 1
I managed to compile after reverting your last change to iptable. I have noticed a linking error compiling the K24 kernel when some options are enabled... I didn't touch any of that part of the code, hence why I asked when it last built for you, because it seems to have been broken for a little while... BTW, Toastman-RT branch build just fine (after the rc/network.c fix) I am doing some more testing and I'll be pushing the changes. Unfortunately, I can't set if MultiSSID works on K24 as I have a WRT54G v2 which doesn't support BSSID. So what I did is make the old code conditional, so when building a K24 image, it will use the same code before my MultiSSID code. I figured if it worked before, it will work now too
It works with a caveat: There is an issue with the way the MAC address for the virtual interfaces are calculated by the Broadcom drivers. As such, the MAC address between 2.4Ghz and 5Ghz overlap. After declaring the 5Ghz virtual interface, you'll have to manually adjust the MAC address of each adapter so they don't overlap. It will not survive a reboot due to how the drivers initialise things.
You seem so good, why don't you have a go at it... Oh wait, do you have the source code of the wireless drivers to fix the problem? No? Oh, of course, it's not open source... Duh! And for your information, before posting stupid unhelpful suggestion, the Asus source code doesn't let you create a guest network in 5GHz mode, only 2.4...
Which part of "the Asus source code doesn't let you create a guest network in 5GHz mode, only 2.4..." don't you get? The broadcom driver is a binary blob, same with the Asus code and any broadcom based wireless router. Again, 5GHz virtual wireless works, with caveats. If you only set a 5GHz one and not a 2.4GHz one, it will work. 2.4GHz and 5GHz virtual wireless at the same time, you have to manually adjust the MAC address... don't see the point in answering any more of your posts. Luckily there are people appreciative of the hard work done FOR FREE...
(Wow, my account here is still open LOL) These are the files jyavenard is referring to: https://github.com/RMerl/asuswrt-merlin/tree/master/release/src-rt/wl/sysdeps/RT-N66U/linux Those files are binary-only. They are linked with other code to form wl.ko, the actual driver.
I successfully got 2- SSIDs working on a used e2000 with the Toastman 7500 build. I didn't play with it long because when I got around to speed testing the e2000, my speed test showed a terrible 9Mb when my cable has 15Mb to supply. I returned to the Linksys firmware and got the same result, so the issue is probably hardware related. From what I did attemt, I had one question on using qoS with mult-SSIDs..... I tried to place a bandwidth limit on the guest network by using the pseudo-MAC address generated that was found on the status page. I found the bandwidth could not be limited using this MAC. Any suggestions how I could trap this guest network in the qoS bandwidth settings? I know I could target the MACs of all the connected devices to this network, but since they change frequently, I prefer a global solution. I am probably missing an easy solution becasue I am fairly new to this game. Thanks.
Finally had a chance to check the new functionality on me E4200v1. This is a great improvement!! I have successfully set up multiSSID with two SSIDs both for the 2.4Ghz and 5Ghz bands (simultanuously). I have eth1 and wl0.1 on 2.4ghz, and eth2 and wl1.1 on 5Ghz. There are no MAC address conflicts. All interfaces working with WPA2. As for bridges, the following configurations worked: 1) all interfaces on LAN. 2) eth1, eth2 on LAN, wl0.1 on LAN1, wl1.1 on LAN 3) eth1, eth2 on LAN, wl0.1 on LAN1, wl1.1 on LAN2 The following did NOT work: 4) eth1 and eth2 on LAN, wl0.1 and wl1.1 on LAN1. In this configuration, clients could not connect to wl1.1.
In teaman's 1.28.0024-SDHC build, he included the following fix: "fix WL MAC filter to include all WL VIFs (Basic: Wireless Filter, basic-wfilter.asp" However this fix is detrimental to me, because my virtual SSID is for guests, and while I want my MAIN SSID to have wfilters, I do not want my guest SSID to have mac filtering. With this fix, wfilters are applied to main SSID and virtual SSID. My guest SSID is on a seperate vlan, so I want all MACs to be able to connect to it, since it can't access my main LAN at all, I don't care as much who connects (as long as they have the WPA password I provide them) My question is, could we make the wfilters for the WL VIF's seperate? This "fix" is also in the latest shiibby builds, I'm assuming this is because the changes in teaman's builds are incorporated into shibby's builds as well.
OK, I have problem with multiSSID on ASUS RT-16n... I use Shibby RT build 100 AIO. I set multiSSID according to the teaman manual... I see guest wifi on my devices but I cant connect to it...
Hi, I was interested in using MultiSSID for a guest network, and installed this build Tomato Firmware 1.28.0005 101 ND VPN (tomato-K26USB-1.28.RT-MIPSR1-101-Nocat-VPN on two routers. The first, a Linksys WRT610N flashed to become an e3000, works for me, and I have two networks connected to my LAN (2.4 Ghz and 5Ghz) and two guest networks that route to the internet. I occasionally get 'blips' when the network seems slow - but that could be my broadband provider, not the e3000 The second is aWRT54GL 1.1 - This does not work. It appears to Work fine until I define the virtual wireless interface. Then the router will not reboot. A Power cycle brings router up. The Lan port interfaces work, the non virtual wireless interface works but guest network doesn't. As the wrt54gl is for my daughter who's rushing off to uni, I'm going to reflash, and not bother with the guest network, it was a nice to have.
Is QOS supposed to work when using multi SSID? It works fine for the main lan but the extra SSID lan has its traffic classified but is shown under the wan IP instead of the lan IP for all devices. The bandwidth limiter seems to be ineffective also. Tried this but does not help: http://tomatousb.org/forum/t-436502/multi-ssid-qos I am running : v1.28.7500 MIPSR2Toastman-VLAN-RT K26 USB VPN-NOCAT Thanks.
Have you set bandwidth limits for br1, br2 and br3? I haven't played around with the multi SSID's yet, but I think that's where you'd set the limits for them.
Saarge, I found the same behaviour > wrong IP classification see my previous post: http://www.linksysinfo.org/index.php?threads/qos-and-vlan.53308/ But for me bandwidth limiter is working. I'm using it on the default class level br0, br1
Thanks guys. I thought the idea of switching the main lan with the virtual one to be clever. I have a fairly big pipe to the interwebs so I don't require QOS absolutely. I setup the second SSID for the neighbours a couple of months ago. It's just that in the past few weeks they have been eating quite a bit more bandwidth than expected. A simple e-mail to them should solve things. I can always pull the plug on them if they continue abusing Low tech solutions work best sometimes. I might spend some time reading a bit on iptables and maybe just write my own script and skip the GUI. I suspect there are a few here that do that. I hope that skipping the GUI won't affect the stats being gathered too much. As long as the bandwidth stats are correct, I would be fine.
Hi Huys, I am using a Cisco E2000 wireless router (Tomato Firmware v1.28.2025 MIPSR2Teaman-RT-VLAN-PPTPD K26 USB Ext) and want to accomplish this setup: Current situation: 1. Cisco E2000 with Uplink on port 1 2. DHCP on this router is Disabled 3. Connected users will get IP address etc. from domain 4. Users get IP address within 10.x.x.x. range (br0) 5. this is working fine for employees 6. The router is basically working as an accesspoint. Now there is a question for implementing another SSID for a Guest network on this device Guests need to get their DHCP address from another range (i.e. 192.168.2.x (br1)) A new LAN bridge has been created with DHCP enabled (192.168.2.100-150) DHCP is working fine when connected to this Guestnet (wireless) VLAN on port 3 also got DHCP (wired within the 192.168.2.x subnet) I can ping the virtual gateway on 192.168.2.1 DNS requests are resolved, but there is no route to the outside. Here is some other info: ---------------------------------------- System (Click here to hide) Name wlanrtr1 Model Linksys E2000 Chipset Broadcom BCM4716 chip rev 1 pkg 9 CPU Freq 354 MHz Flash RAM Size 8 MB Time Tue, 11 Dec 2012 20:10:01 +0100 Uptime 0 days, 03:09:23 CPU Load (1 / 5 / 15 mins) 0.02 / 0.02 / 0.00 CPU Usage 0.95% Total / Free Memory 28.10 MB / 20.61 MB (73.34%) LAN (Click here to hide) Router MAC Address c0:c1:xx:xx:xx:80 Router IP Addresses 10.0.0.25/24 on LAN (br0), 192.168.2.1/24 on LAN1 (br1) Gateway 10.0.0.1 DNS 10.0.0.10:53 DHCP Disabled on LAN (br0), 192.168.2.100 - 192.168.2.150 on LAN1 (br1) Wireless (eth1) (Click here to hide) MAC Address C0:C1:xx:xx:xx:82 Wireless Mode Access Point + WDS Wireless Network Mode G Only Interface Status Up (LAN) Radio Enabled SSID WLAN Broadcast Enabled Security WPA Personal (PSK) + TKIP / AES Channel 8 - 2.447 GHz Channel Width 20 MHz Interference Level Acceptable Rate 24 Mbps Wireless (wl0.1) (Click here to hide) MAC Address C0:C1:xx:xx:xx:83 Wireless Mode Access Point Interface Status Up (LAN1) SSID WLAN-Guestnet Broadcast Enabled Security - ---------------------------------------- Is it possible to have a guest net working in this manner or do i need to enter some extra information elsewhere? Thank in advance for all your answers and if the information provided is incomplete, please let me know.
