1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Tomato Mod 1.20.8000

Discussion in 'Tomato Firmware' started by Victek, Jul 17, 2008.

Thread Status:
Not open for further replies.
  1. guillaumy

    guillaumy LI Guru Member

    Did you mention brownouts? Ah, your location has unstable voltage?

    It could even be that that has something to do with the Intel connection problems (I read somewhere I think where some said it had to do with some power issue).

    It is my practise to power ALL my Linksys routers with universal transformers (switching transformers) that can handle variations in input voltage (e.g. 100-240V) and spikes in current uptake (e.g. up to 2 amps).

    Maybe that's why I'm seeing none of those problems?
     
  2. Toastman

    Toastman Super Moderator Staff Member Member

    I think you're seeing none of these problems because you don't have 25 users with Intel laptops, is all :biggrin:
     
  3. guillaumy

    guillaumy LI Guru Member

    Surely it would not hurt for you to try ... :wink:
     
  4. mstombs

    mstombs Network Guru Member

    Another suggestion for Victek. I'm trying to debug an issue in adsl router firmware, and have found that the tcpdump (symlinked to etherdump) in Woytekm's WAG200G firmware (for Ti AR7 linux 2.4.17 mips le) seems to run OK on Tomato (2.4.20 mips le).
    It is much smaller than a full tcpdump I have been given for Ti AR7 which also appears to work on Tomato - but not on my adsl router!

    Code:
    # ls -laF
    -rwxr-xr-x    1 root     root        13760 Dec 16  2007 etherdump*
    -rwxr-xr-x    1 root     root       787076 Aug 14 13:59 tcpdump*
    drwxr-xr-x    1 root     root            0 Jan  1  1970 wwwext/
    Woytekm's utility http://sourceforge.net/projects/openwag200/ (its a binary in the source) is clearly from

    http://freshmeat.net/projects/etherdump/

    I can't see if he has modded it or the build scripts, but it looks an interesting toy ie a snip of my ssh connection to my router:-

    Code:
    # ./etherdump -i br0
    07-15 00:29:14  TCP:     192.168.0.1 : 22    ->    192.168.0.99 : 4358     92
    07-15 00:29:14  TCP:     192.168.0.1 : 22    ->    192.168.0.99 : 4358    172
    07-15 00:29:14  TCP:    192.168.0.99 : 4358  ->     192.168.0.1 : 22       46
    07-15 00:29:14  TCP:     192.168.0.1 : 22    ->    192.168.0.99 : 4358    236
    07-15 00:29:14  TCP:     192.168.0.1 : 22    ->    192.168.0.99 : 4358    172
    07-15 00:29:14  TCP:    192.168.0.99 : 4358  ->     192.168.0.1 : 22       46
    07-15 00:29:14  TCP:     192.168.0.1 : 22    ->    192.168.0.99 : 4358    236
    To ensure full compatability with Tomato it should really be compiled/built with Tomato toolchain and kernel headers, but looks small enough to be included in the firmware?
     
  5. TexasFlood

    TexasFlood Network Guru Member

    Cool, mstombs, I grabbed etherdump and threw it out on a cifs share and runs fine on my router. Ran it for a while on br0, eth1 and vlan1. Nice to have such a tool available for diagnostics or just satisfying curiosity, :grin:

    Thanks
     
  6. Victek

    Victek Network Guru Member

    Thanks mstombs, run and the most important, is smaller. :biggrin:
     
  7. NewAtair

    NewAtair Network Guru Member

    Show nvram doesn't work

    Hello,

    the Function "Tools -> NVRam show" doesn't work correctly, because it always show a quite empty window with:
    "
    Tomato
    Version 1.21

    Thor
    NVRam contents
    "
    Nothing more. (See Attachment)
     

    Attached Files:

  8. Toastman

    Toastman Super Moderator Staff Member Member

    Hi Vic

    I have a router running at another residence nearby, this one is interesting because almost all of the port forwards were from one guy, using Messenger. He isn't on line right now so all of these connections by .110 should have been closed.

    It does always seem to be Messenger not closing the ports.

    On another subject, I put ND firmware on an AP in the worst location next to several guys with Intel laptops, set to b only, 11 Mbps. It has been up for 8 days, they do seem able to connect - at least nobody's complained to the desk staff :smile:


    mstombs: I can't find much about miniUPnP - have you a link?

    ** OK - found it. You've been busy, huh ? :biggrin:
     

    Attached Files:

  9. Toastman

    Toastman Super Moderator Staff Member Member

    I spoke too soon. I changed all firmware to ND driver. Several Intel's remained connected, three at least cannot associate no matter what I do to the machines. Gone back to normal driver :frown:
     
  10. guillaumy

    guillaumy LI Guru Member

    Why don't you try one of those switching transformers. Put it on one unit and see how it goes?
     
  11. Toastman

    Toastman Super Moderator Staff Member Member

    Read my post - routers are already on UPS (uninterruptible power supply).

    guillamy, just go google "Intel 2200BG problems" - the association problems are due to some as yet unknown problem, perhaps in the wireless driver setup. I'm sure it will be fixed son ...
     
  12. mstombs

    mstombs Network Guru Member

    This is a truly international collaboration! Miniupnp homepage is

    http://miniupnp.free.fr/

    And yes the documentation is not great, but its in active development, and the developer responds quickly to any query!

    With recent scares over upnp security I believe this is the most secure implementation around, it only supports reporting and port forward setting. It supports NAT-PMP as well as or instead of upnp. NAT-PMP http://en.wikipedia.org/wiki/NAT_Port_Mapping_Protocol is more restrictive than upnp which, in a full implementation, can be used to configure everything including DNS servers.

    It has a secure mode where clients can only set port-forwards to their own IP (try it - even Windows XP provides a UI where you can manage other users upnp port-forwards in most implementations!)

    It also has rules which restrict which clients can forward which ports - as well as the auto rule cleaning. An example conf file is

    Code:
    # WAN network interface
    #ext_ifname=eth0
    ext_ifname=xl1
    # if the WAN interface has several IP addresses, you
    # can specify the one to use below
    #ext_ip=
    
    # LAN network interfaces IPs / networks
    # there can be multiple listening ips for SSDP traffic.
    # should be under the form nnn.nnn.nnn.nnn/nn
    # HTTP is available on all interfaces
    listening_ip=192.168.0.1/24
    listening_ip=192.168.1.1/24
    #listening_ip=
    # port for HTTP (descriptions and SOAP) traffic
    port=5555
    
    # enable NAT-PMP support (default is no)
    enable_natpmp=yes
    
    # enable UPNP support (default is yes)
    enable_upnp=yes
    
    # lease file location
    #lease_file=/var/log/upnp.leases
    
    # bitrates reported by daemon in bits per second
    bitrate_up=1000000
    bitrate_down=10000000
    
    # "secure" mode : UPnP client are allowed to add mappings only
    # to their IP
    secure_mode=yes
    
    # default presentation url is http address on port 80
    #presentation_url=http://www.mylan/index.php
    
    # report system uptime instead of daemon uptime
    system_uptime=yes
    
    # notify interval in seconds. default is 30 seconds.
    #notify_interval=240
    notify_interval=60
    
    # unused rules cleaning.
    # never remove any rule before this threshold for the number
    # of redirections is exceeded. default to 20
    #clean_ruleset_threshold=10
    # clean process work interval in seconds. default to 0 (disabled).
    # a 600 seconds (10 minutes) interval makes sense
    clean_ruleset_interval=600
    
    # log packets in pf
    #packet_log=no
    
    # ALTQ queue in pf
    # filter rules must be used for this to be used.
    # compile with PF_ENABLE_FILTER_RULES (see config.h file)
    #queue=queue_name1
    
    # tag name in pf
    #tag=tag_name1
    
    # make filter rules in pf quick or not. default is yes
    # active when compiled with PF_ENABLE_FILTER_RULES (see config.h file)
    #quickrules=no
    
    # uuid : generate your own with "make genuuid"
    uuid=fc4ec57e-b051-11db-88f8-0060085db3f6
    
    # serial and model number the daemon will report to clients
    # in its XML description
    serial=12345678
    model_number=1
    
    # UPnP permission rules
    # (allow|deny) (external port range) ip/mask (internal port range)
    # A port range is <min port>-<max port> or <port> if there is only
    # one port in the range.
    # ip/mask format must be nn.nn.nn.nn/nn
    # it is advised to only allow redirection of port above 1024
    # and to finish the rule set with "deny 0-65535 0.0.0.0/0 0-65535"
    allow 1024-65535 192.168.0.0/24 1024-65535
    allow 1024-65535 192.168.1.0/24 1024-65535
    allow 1024-65535 192.168.0.0/23 22
    allow 12345 192.168.7.113/32 54321
    deny 0-65535 0.0.0.0/0 0-65535
    Its easy to compile and run on a Tomato router (about same size as Tomato upnp), the harder part is the integration into the GUI and rc control code.
     
  13. Toastman

    Toastman Super Moderator Staff Member Member

    mstombs

    - I found that link just a while ago, just been reading up on it. That is a very responsive guy. Seems like a useful tool. You mentioned a configurable UPnP timeout/reset of unused forwards? I didn't find any mention of that, probably something you find out when you run it. Are any of the major firmware upgrades using it yet in earnest?
     
  14. mstombs

    mstombs Network Guru Member

    The rule cleaning is controlled by this section of the conf

    Code:
    # unused rules cleaning.
    # never remove any rule before this threshold for the number
    # of redirections is exceeded. default to 20
    #clean_ruleset_threshold=10
    # clean process work interval in seconds. default to 0 (disabled).
    # a 600 seconds (10 minutes) interval makes sense
    clean_ruleset_interval=600
    but I will have to look at the code to be sure exactly what it does, and how it decides what ones to remove. Not sure about major firmwares, but we've been using it in RouterTech (AR7 adsl router) firmware for a couple of revs, with a few thousand downloads and no complaints - but not hard to be better than original upnp daemon that didn't clean rules on WAN IP change and had a memory leak which meant the recommendation was not to use it! Tomato upnp is pretty good to start with, and it may be possible for Jon to add extra functionality?
     
  15. Toastman

    Toastman Super Moderator Staff Member Member

    The author believes that cleanup every 10 minutes makes sense. The "threshold" might be the interesting bit.

    I've been looking in Tomato code to see if there's an insertion point for a "cleanup" and I think I found one, I'll have to go find it again.

    I think you are about 1,000 light - years ahead of me in this game :smile:

    Yes, I can't find fault with Tomato's UPnP implementation, only that it doesn't address situations like people pulling out the plug without letting their PC shut down the connections.
     
  16. Toastman

    Toastman Super Moderator Staff Member Member

    Remember the discussion on White Russian forum, in which Soltec reckoned he found the problem with 2200BG association was due to the "h" power management not being disabled while in "g" mode? There was no functioning variable to turn this off, he had to patch the code to enable it:-

    QUOTE

    "Solution: disable 802.11h in the WRT. The problem is that it doesn't exists a nvram variable for that, command for wl is:
    wl spect 0
    and for wlc:
    wlc 802.11h 0"

    Just for curiosity's sake I tried telnet to an AP here - entering "wl spect" returns "Off". "wl spect 1" or "wl spect 2" doesn't do anything, so has someone already tried this, as it seems to be switched off already, with no function to turn it back on ?

    So was that really a cure?

    OpenWRT's Kamikaze was supposed to have fixed it, but nobody seems to have posted anything that verifies if that is true or not. Kamikaze release date is behind schedule, so we'll have to wait and see I guess.

    I've been looking to see if there is any WRT54GL firmware at all which does not suffer from the driver bugs, to use as a AP's, but I can't find one that's actually been acknowledged to work. [The last Linksys WRT54GL I bought, 2 weeks ago, also didn't work with 2200BG with the original software.]
     
  17. mstombs

    mstombs Network Guru Member

    Re upnp

    miniupnpd cleans unused rules when it checks on its periodic timer and there are more than the minimum. I can't see where it resets the rule use counters so I will ask on the miniupnp forum.

    Re wireless 2200BG

    I have never had a problem with my WRT54GSv1.1 with Tomato old drivers and a Dell laptop with this Intel chip. Is it only newer Broadcom chipsets?
     
  18. Toastman

    Toastman Super Moderator Staff Member Member

    I don't know, I never had anything but the WRT54GL, perhaps older stuff was OK. Also, it seems to me that it isn't only Broadcom based routers that have problems.

    I'm getting a little fed up with Intel stuff, because everybody else seems to be able to make wireless adaptors that work. All of the residents that I've spoken to about it are really pi**ed at having bought expensive laptops that don't work at home, at work, or even in most hotspots! I've just tried half a dozen different driver versions on one guy's laptop, which he left here for the day, but can't get the damn thing to associate. I've tried everything I've read in every forum on the planet...

    Edit...

    The problem is slowly going away as the laptops containing most of the offending Intel cards (Compaq/HP) are overheating and burning out. There are a lot of very unhappy purchasers here :biggrin:
     
  19. hogrod

    hogrod Network Guru Member

    I have a dell B130 laptop with a intel 2200bg wifi card. My laptop originally came with a dell 1350 wifi card, but I wanted to get the intel card so combined with my Pentium M processor I would have a Centrino laptop for slightly better battery life. I have used the intel card with my WRT54G V3 for more than a couple years without any major issues.

    I also use this laptop on many different access points when traveling without any problem.

    A few firmware updates I did to my home router in the past(older versions on tomato) my wireless wouldn't work after resuming from standby, but a system reboot would always fix that. When it wasn't working it would always be connected to my access point but just not have a working internet connection.

    I always use the newest drivers from intel, when I was running XP, and now running vista business. I always let windows manage my wireless, and don't use or install the intel wireless software.
     
  20. gingernut

    gingernut LI Guru Member

    Toastman,

    I've had some of the same problems with the 2200BG Intel card. The New Drivers in Tomato was impossible to connect and get the DHCP server to hand out an IP. I decided to give DD-WRT a shot on one AP, a Buffalo HP unit, and at the moment seems to have the solved my connection problems. I've got a few units in WDS, two with DD-WRT 24 SP1 and main AP Tomato ND Victek Mod and all seems to work well, fingers crossed.
     
  21. Toastman

    Toastman Super Moderator Staff Member Member

    Thanks for that, Hogrod.

    It is a real can of worms, isn't it? Searching google for information, it appears that there are several companies making Intel wireless cards and some people say they behave differently. Some of them work fine all the time, which really miffs the guys who do have problems! This afternoon I had another guy bring me a laptop to look at, he had bought two extra wireless interfaces but couldn't set them up. The intel original sure enough was a no-no. I tried his USB adaptor, a Netgear WG111v3 which worked perfectly. Next I swapped out his Intel card for a Broadcom based one - it worked perfectly - so that was what I left in the machine.

    There is a repairman in the local computer store at Seri Center who has a couple hundred Intel cards he's swapped out for people - a veritable graveyard.
     
  22. Toastman

    Toastman Super Moderator Staff Member Member

    gingernut, what version DD-WRT did you flash, and can you see from the logs what version of the wireless driver it is running? I'll give it a shot on one of my AP's too. Last time I tried DD-WRT it just wasn't stable.
     
  23. TexasFlood

    TexasFlood Network Guru Member

    First let me state that I'm not recommending or not recommending DD-WRT, just providing information. I feel Tomato is easy to use, stable, and has good QOS. However DD-WRT has some features Tomato does not and supports hardware that Tomato does not. Do your homework and make your own choice. I suspect that in most cases if you fell in love with Tomato for it's strong points, it will be hard to switch. Let me add that I run Tomato on some of my routers but some are not supported so I run DD-WRT as well. So I am grateful to both Tomato and DD-WRT for giving me better choices than the stock firmware.

    Having said that, with recent DD-WRT releases (v24 vs the older v23), your choices depend somewhat on your hardware, specifically does it support the "new driver" builds or require the "vintage driver" builds. There is some guidance on this in the DD-WRT wiki. This is the official v24 SP1 download link.

    The rest of this is FYI as the above is all you need to know if you're sticking with official releases. The hardware requiring vintage builds -should- link to the latest official vintage build here as required.

    There are some newer builds of both the vintage and new driver builds in the "TNG" (The Next Generation") directory.

    Note that the vintage builds have recently fallen behind the new driver builds, I believe due to an as yet unresolved issue with the new code in the nas module. Those of us who need the vintage builds hope this is resolved soon but the existing builds are fine for the time being. So v24 final is the latest official vintage build and v24 sp1 the latest official new driver build. And as mentioned above, there are some slightly newer builds of both out in the TNG directory.

    If you want more info on the official releases vs. the newer TNG builds, suggest you consult the DD-WRT forums and the DD-WRT timeline for more information about specific builds. Searching by release number is a good way to start.
     
  24. gingernut

    gingernut LI Guru Member

    Toastman,

    On a Buffalo HP I'm using the latest 24 sp1 Eko TNG build, which is the AP the Intel 2200BG normaly connects to, and at the moment I've not had any problems. I do prefer Tomato on my routers where possible but in this case I wanted the better performing ND which in Tomato, at the time being, cause a few connection problem with some wireless adapters. Maybe Victek will be able to get some better drivers to work in Tomato we will just have to wait and see.
    I'll have to get back to you on the actual driver version used as I'm not sure at the moment.
     
  25. Toastman

    Toastman Super Moderator Staff Member Member

    Righto, thanks guys.

    I've put the v24sp1 generic nokaid version on an 2 AP's that are next to several Intel users, let's see what happens. I can't find any feedback on the dd-wrt forums, did you see any, Tex?

    [going completely off-track, a resident brought me a laptop to look at tonight, a Compaq V3000. Just out of warranty. No boot or display on startup, wireless not found by BIOS, been using a USB adaptor for three months instead. A quick Google reveals that one in 7 have faulty motherboards! What a dog of a laptop! I can't believe she never pressed the warranty claim...]
     
  26. TexasFlood

    TexasFlood Network Guru Member

    You mean about Intel 2200BG? If so, I found a few hits on that in the dd-wrt forums, this one, that one, and yet one more. Not sure if any of they are what you're looking for.
     
  27. Toastman

    Toastman Super Moderator Staff Member Member

    What I have been looking for is some feedback on whether the DD-WRT's latest driver has actually a) Stopped the rebooting of the router and at the same time b) Still allows Intel cards like 2200BG and later to associate, and get an IP address allocated by DHCP.

    The forum is very big and disorganised, and most of the posts rambling and uninformative, so I gave up after the first 300 or so - but I couldn't find any comment on the subject at all! I must be wrong, I'd have though this was an important enough topic for someone to comment on!

    What I am getting from the OpenWRT and other forums is that the Broadcom driver is still closed-source and essentially not much can be done with it. There is a new open-source driver being built (somewhere) by (someone) from scratch, but I haven't discovered who or where he is :biggrin:
     
  28. TexasFlood

    TexasFlood Network Guru Member

    I can't argue with you about the dd-wrt forum. The threads range from seemingly pointless bickering to genuine pearls of wisdom and inside help from developers. So, ya know, it's pretty much like any other forum, just bigger and more active, :grin: . I do get the feeling over there that is possible to get the ear of a developer, like eko, but it really depends on the issue, YMMV

    OpenWRT forums can be interesting as OpenWRT by definition tends to scare off neophytes and attract a more technical crowd. It scared me off. Well, I was more lazy that scared. I loaded up white russian back in the day but HyperWRT then Tomato was so much easier, :wink:

    I have read that the newer driver in DD-WRT works better with Intel wireless cards but don't remember any detail so maybe some googling is in order. I did see in one of those threads I listed earlier about DD-WRT users resolving the issues they were having by upgrading drivers and there was a link included. Although it sounds like you've already been down that road to no avail.
     
  29. Toastman

    Toastman Super Moderator Staff Member Member

    All comments noted! One big critcism I have of the DD-WRT setup is that it is actually quite difficult to find out what is in a particular released Binary.

    I find that the openWRT crowd seem to be more knowledgeable and keen to get things going. Not so much of " did you try polishing the power cord?"

    More power to the guys from x-wrt!

    As for the DD-WRT AP's I set up this morning, they are still working, however, not a single person has connected to them yet. Which seems odd. I've heard that the driver does the same (effects a cure) , but can't actually remember where I read that. I suspect it hasn't changed, so this evening when it gets busy, around 8-11 p.m. will tell.
     
  30. TexasFlood

    TexasFlood Network Guru Member

    Yes, the lack of release notes for DD-WRT is a frequent, question/complaint. The general answer is to read the comments in the dd-wrt timeline, which is helpful but not nearly as easy as nice concise release notes.
     
  31. Toastman

    Toastman Super Moderator Staff Member Member

    Well, after last evening, when it got busier, both AP's running DD-WRT had rebooted, I cannot be sure how many times but one certainly 3 times and the other I don't know. It did seem that most people could access them with Intel machines, but the problem machine I spoke of still could not associate. My conclusion is that nothing much has changed with DD-WRT!

    *** update: Persevering with it, I have now found a stable version of DD-WRT and have it on all the AP's. See later.
     
  32. Victek

    Victek Network Guru Member

    Toastman, I think you need a real router to handle all the clients and the embedded firmwares in Linksys routers or AP's are not enough to support the stress of connections (honestly thought).

    Do you have any spare PC (P3-256MB-10GBHDD) to give a try to x86 version of dd-wrt or Openwrt? I installed one in my holiday little town for a free wireless offered by the local authorities and be sure, no problems, up to 320 users connected in peak hours without any glitch from the unit using dd-wrt x86. :smile:
     
  33. Toastman

    Toastman Super Moderator Staff Member Member

    Hi Victek

    Ok on your thoughts.

    Firstly, let me say I'm actually very happy indeed with the progress that we've made.

    The test build you made for me (v1.21.8800 dated 10 August 01:53:26) copes very well with the number of users, the average number online in an evening is up to half of the clients - 35 to 40 people.

    Currently there are 37 users online using approx 3.7Mbps. DHCP leasetime 2 hours. UPnP forwards 62. Processor/memory 0.07/0.04/0.00 4,108 free. No. of connections varying between 280 and about 600.

    On weekends when it is busier, up to 70 people, there has still been no sign of instability, and the router has not rebooted nor has dnsmasq given up under the strain. It isn't the router that I am having trouble with, it's the "simple" AP's !

    Let's look at the progress that we've made with the firmware of the router:

    1) The increase in support to 100 users in SDHCP & Access Restrictions has made it possible for me to use the router effectively in my buildings, instead of using several routers and ADSL lines. The staff could not maintain that system because it was too complicated for them. Now the desk staff add and delete residents by themselves, much simpler with only one router and Tomato's excellent GUI interface.

    2) P2P has now been controlled and isn't excessive, and I haven't seen the number of connections in Conntrack rise much over 2000 for a few weeks now. The higher conntrack limit with SpeedMod is effective if and when it rises higher.

    3) The router is stable with high load now. There has been no sign of high processor useage nor low memory for several weeks.

    4) The increase to 100 forwarded ports from 25 has meant that there are no longer people who can't use the service they want to because they couldn't open ports. UPnP port starvation is no longer a problem, because the list has stabilised at around the 70 mark, even though many of those entries were not being cleared. This mod is also stable.

    So, as it is, the system works well.

    The main router has stayed up for over a week with the wireless switched off, proving that there's nothing wrong with the firmware.

    So, the last remaining "real" problem is the wireless driver. And that manifests itself in the AP's - which aren't running any extras in the firmware, and don't have to do anything much at all. I have just borrowed and tested a new WAP54G with original Linksys software, which also has the same problem.

    There has been some change in the numbers of residents this month, but I still have 22 people with Intel cards. Some of these have no problems, some of them can associate with a "b" only AP, a few cannot associate at all.

    But there is still a certain number of them which cause problems to the AP's when they do associate. It's very easy to prove, because it is AP's next to the bulk of the Intel users that have problems. If I relocate the same AP to a different point which has no Intel machines nearby, they often stay up for a week and more. That particular AP has 5 very regular users, 4 of them are Intel, one is D-Link. The 24 hour chart shows they use very little bandwidth, about 200kbps, for a few hours in the evening. That isn't much of a load for an AP!

    So please don't be discouraged, because we've actually accomplished the original goal! There are many people here who have commented on how the service works properly now, compared with three months ago when it was almost useless. :smile: As I mentioned before, owners of other apartment blocks are coming to see how it works and to ask questions about it, because the residents here have told their friends.

    I'm confident that it isn't the inability of the firmware to work with this number of users, it's really now just the wireless which the one remaining problem area. I know that the ultimate answer is to replace all my AP's with a make that doesn't have the association problem - but that would be an expensive cure :) However, I would still keep the main router with Tomato, even if I had to do that.

    Thanks Victek, for all of your help and encouragement, without which I'd still be living in the dark ages :biggrin:

    Toast
     
  34. Toastman

    Toastman Super Moderator Staff Member Member

    I forgot to mention, I found the reason the router rebooted a week ago when we were testing the UPnP stability. One guy here has an email virus. His machine sometimes starts up and sends several thousand DNS lookups simultaneously, followed shortly afterwards by a corresponding several thousand connections to remote SMTP clients on port 25.

    Now that's something I'd not thought of before! How on earth do I stop that but still allow normal SMTP to work with a high priority? I can't get an individual limit to work well with Tomato's normal QOS rules.
     
  35. Victek

    Victek Network Guru Member

    Hi Toastman, thanks for the follow up, your findings are highly appreciated, well, I'm not 100% operative yet to test all the features and work I'm doing during my holidays but miniupnpd, etherdump and finally igmp-proxy are compiled, I have to install in one router as soon I'll back at home. I have some doubts because the size of generated files has decreased (good signal) but I think some function is lost.
    I think we'll have time enough before the new Tomato release will appear to test it.

    Thanks :biggrin:
     
  36. Toastman

    Toastman Super Moderator Staff Member Member

    4.150.10.5 Broadcom Driver cures Intel Wireless problems

    Hi Vic

    So you compiled mini upnp? I look forward to try it out !

    I am desperately trying to test out a 4.150.10.5 driver. (I believe that DD-WRT v24sp1 has the 4.150.10.5 version of the driver according to TexasFlood). I decided to use the mini compile of DD-WRT v24sp1 because some of the earlier versions of v24 were unstable or just didn't work - some people have found they bricked their routers. The later versions seem stable - if anyone else is going to try DD-WRT be very careful as the "dodgy" versions have not been removed from the upload site.

    This version flashed OK, responded well, and appears stable. It has been up 24 hours so far, no complaints from anybody.

    http://www.dd-wrt.com/dd-wrtv2/down...nksys/WRT54GL_1.1/dd-wrt.v24_mini_generic.bin

    Apart from that, not been doing much - I gave up trying to compile your code because I think I am missing some vital element - (like a brain) :wall: But now the environment is fine and I am still very impressed with OpenSUS. I really need to get another machine, this one is always busy ....

    **** Update - Saturday Midday - DD-WRT AP up 24 hours, still OK. I have now loaded it on 16 of my AP's. 7 of my 20 Intel clients are associated, so far no complaints to the reception desk from anybody.

    **** Update - Monday evening - all AP's now up 3 days. 17 of my Intel users have all associated OK. I believe this confirms that the 4.150.10.5 driver is a positive cure for the reboot problem, and so far it also seems that Intel cards can associate with the AP's. This is good news indeed!

    **** Final Edit - October 1st - all AP's have been for well over a month. No complaints, everyone can associate. QED ! Now I look forward to see this driver in Tomato!
     
  37. Toastman

    Toastman Super Moderator Staff Member Member

    Power Save Polling in Wifi Cards

    At the risk of hogging the thread, I found some info which I have never seen published on these forums, this is the first article I've seen that really explains what the problem is with the Intel wireless:

    HOW POWER SAVE POLLING IS SUPPOSED TO WORK:

    "Timing is important when using power-save polling mode (PSPM), where the Wi-Fi client goes into Sleep/Standby mode for a predetermined length of time in order to save power. Makes sense, but what happens to the traffic destined for the sleeping Wi-Fi client? PSPM has that covered. The access point acts as a storage buffer for the Wi-Fi client whenever the Wi-Fi client is in reduced power mode. Here are the details:

    The Wi-Fi client announces its intention to enter PSPM by transmitting an 802.11 frame with the power-save bit in the packet header toggled on. This frame informs the access point that the Wi-Fi client wants to power down everything except a simple timer. The timer’s main purpose is to control when the Wi-Fi client powers back up to receive traffic from the access point. Upon receiving the PSPM frame, the access point and Wi-Fi client synchronize their timers and agree upon a beacon interval, which is the allocated time between beacon frames.

    The PSPM frame also informs the access point of the need to begin saving any traffic earmarked for that Wi-Fi client. With everything understood the Wi-Fi client powers down and the access point starts buffering traffic destined for the sleeping Wi-Fi client. At the end of a beacon interval, the Wi-Fi client powers up opening what’s called the Announcement Traffic Indication Message (ATIM) window. The access point then takes advantage of the ATIM window to send a beacon frame.

    Because the Wi-Fi client is in PSPM, the access point uses a special beacon frame called Delivery Traffic Indication Message (DTIM). The DTIM beacon frame synchronizes the timers and notifies the Wi-Fi client on whether it has any buffered unicast traffic or not.

    Right after notifying the Wi-Fi client via the DTIM beacon frame, the access point will forward any multicast or broadcast traffic that it was holding for the client. Then if no unicast traffic exists, the client immediately goes back into powered-down state.

    If there’s buffered unicast traffic, the access point will transmit the stored data using one of two methods:

    PSP method: The Wi-Fi client will send a Power Save Poll (PSP) message to the access point, telling it to release a set amount of traffic. If the total amount of traffic stored by the access point exceeds the amount requested by the original PSP message, additional PSP handshakes take place until all the data is transferred to the Wi-Fi client. At that time, the access point indicates there’s no more data and the Wi-Fi client powers down.

    CAM method: The Wi-Fi client advises the access point that it will enter Continually Aware Mode (CAM) for a certain period of time and is capable of receiving data traffic at the discretion of the access point. This approach eliminates overhead created by having multiple PSP handshakes, but at the cost of using more power."

    So what is wrong with Intel cards, which have problems associating with many routers, especially those which use Broadcom wireless devices such as the Linksys WRT54 series routers? We can only deduce that there is some incompatibilty in the handshaking between them. Other manufacturers, however, implement Power Save Polling on their wireless cards which does appear to work successfully with Linksys routers. It seems strange that the problem has existed now for several years, but we are still no closer to a solution.

    ***Note*** - See info on latest 4.150.10.5 version of Broadcom Driver, which appears to be a solution for some).

    There is a big question as to whether this fragile technology is even worth the effort, especially considering the frustration caused to hundreds of thousands of owners who cannot connect to an access point. There’s very little elapsed battery time gained in the modern laptop by using PSPM. The power saving just isn't significant!

    (http://www.networkworld.com/research/2008/051208-wireless-power-savings.html)

    "The 802.11 standard was initially developed during a time when processor clocks were in the 100MHz to 200MHz range, and initial WLAN designs involved a significant number of power-hungry components. Today, however, Wi-Fi adapters are highly integrated — meaning fewer chips are required to implement a Wi-Fi solution — and designs are more power-efficient. While the notebooks’ other components — most notably the processor (because of higher clock rates) and display and backlighting (due to much higher resolutions) — often consume more energy than in the past.

    Notebook designers have compensated with larger batteries and a continual emphasis on power-conservative designs and provisions for a high degree of end-user control over power conservation settings in many cases, but the proportion of energy consumed between the computer and the WLAN adapter has clearly flipped."


    Intel have apparently given this explanation:

    Are they saying that the cards associate, but then get kicked off after 60 seconds? That's not what I see here. Some cards **appear** to go to sleep immediately after associating, and before even being issued an IP. One fix is to assign an IP manually in your OS, but this doesn't always work.

    EDIT January 2009

    There is a cure for the problem, see here:

    http://www.linksysinfo.org/forums/showpost.php?p=338606&postcount=1
     
  38. phykris

    phykris Addicted to LI Member

    Hi Victek,

    Thanks for your great software.

    I installed it on a network with 4 wireless access points and 1 wireless router, all based on the linksys WRT54GL. The access points run tomato 1.19. The router runs your mod (stable version).

    My downstream bandwidth is 1700 kbit
    My upstream bandwidth is 375 kbit

    I have 17 users. For every user I use the folllowing settings:
    DLRate = 100 kbit
    DLciel = 500 kbit
    ULRate = 22kbit
    ULCeil = 375 kbit
    Priority: normal
    TCP Limit : no limit
    UDP Limit : no limit

    When saving the settings, I get the following warnings in the log file:

    Aug 24 17:15:46 myrouter user.warn kernel: HTB: quantum of class 10010 is small. Consider r2q change.
    Aug 24 17:15:46 myrouter user.warn kernel: HTB: quantum of class 10011 is small. Consider r2q change.
    Aug 24 17:15:46 myrouter user.warn kernel: HTB: quantum of class 10012 is small. Consider r2q change.
    Aug 24 17:15:47 myrouter user.warn kernel: HTB: quantum of class 10013 is small. Consider r2q change.
    Aug 24 17:15:47 myrouter user.warn kernel: HTB: quantum of class 10014 is small. Consider r2q change.
    Aug 24 17:15:48 myrouter user.warn kernel: HTB: quantum of class 10015 is small. Consider r2q change.
    Aug 24 17:15:48 myrouter user.warn kernel: HTB: quantum of class 10016 is small. Consider r2q change.
    Aug 24 17:15:48 myrouter user.warn kernel: HTB: quantum of class 10017 is small. Consider r2q change.
    Aug 24 17:15:49 myrouter user.warn kernel: HTB: quantum of class 10018 is small. Consider r2q change.
    Aug 24 17:15:49 myrouter user.warn kernel: HTB: quantum of class 10019 is small. Consider r2q change.
    Aug 24 17:15:50 myrouter user.warn kernel: HTB: quantum of class 10020 is small. Consider r2q change.
    Aug 24 17:15:50 myrouter user.warn kernel: HTB: quantum of class 10021 is small. Consider r2q change.
    Aug 24 17:15:50 myrouter user.warn kernel: HTB: quantum of class 10022 is small. Consider r2q change.
    Aug 24 17:15:51 myrouter user.warn kernel: HTB: quantum of class 10023 is small. Consider r2q change.
    Aug 24 17:15:51 myrouter user.warn kernel: HTB: quantum of class 10024 is small. Consider r2q change.
    Aug 24 17:15:52 myrouter user.warn kernel: HTB: quantum of class 10025 is small. Consider r2q change.
    Aug 24 17:15:52 myrouter user.warn kernel: HTB: quantum of class 10026 is small. Consider r2q change.

    Everything seems to work, but I am a bit worried about these warnings. What do they mean? What are the consequences?
     
  39. Toastman

    Toastman Super Moderator Staff Member Member

    I'm not an expert on code-related stuff, but searching for information to satisfy my own curiosity, I found this - it seems to be the only answer to a similar question that didn't assume you already knew everything :hmm:

    "Quantum is the amount of bytes a class may send when 2 classes are fighting
    for excess bandwidth. When quantum is too big, it can create extra bursts.
    When quantum is too small (smaller then 1 packet) it can will/disturb the htb
    calculations.

    quantum = rate / r2q
    with r2q = 10 (can be overruled when you add a qdisc)

    Solution :
    r2q = smallest_rate_you_have / 1500

    Stef"

    Try changing your limits upwards/downwards a little, see if it goes away. Victek may not reply until he gets home from his holidays.
     
  40. phykris

    phykris Addicted to LI Member

    Thanks. I also found this.

    It's a manual that describes how the linux outbound traffic throttling works.

    http://luxik.cdi.cz/~devik/qos/htb/manual/userg.htm

    It rather hard to understand. I'll need a while for this.

    I think this part is important:

    It might be good time to touch concept of quantums now. In fact when more classes want to borrow bandwidth they are each given some number of bytes before serving other competing class. This number is called quantum. You should see that if several classes are competing for parent's bandwidth then they get it in proportion of their quantums. It is important to know that for precise operation quantums need to be as small as possible and larger than MTU.
    Normaly you don't need to specify quantums manualy as HTB chooses precomputed values. It computes classe's quantum (when you add or change it) as its rate divided by r2q global parameter. Its default value is 10 and because typical MTU is 1500 the default is good for rates from 15 kBps (120 kbit). For smaller minimal rates specify r2q 1 when creating qdisc - it is good from 12 kbit which should be enough. If you will need you can specify quantum manualy when adding or changing the class. You can avoid warnings in log if precomputed value would be bad. When you specify quantum on command line the r2q is ignored for that class.

    How do I set r2q?

    It would also be nice if r2q could be filled in in the web interface.

    I think the creation of a qdisc (with child classes) stops the traffic flow. So we can't change r2q while there's traffic running without interrupting the traffic. There should an option in the web interface the fill in r2q.

    If all outgoing rates are higher that 120bit, I think we can keep the default r2q (10).

    If there's an outgoing rate <120bit the user can calculate r2q as
    r2q < bitrate /(1500*8)

    For instance in my case: r2q < 22000 /(1500*8) or r2q < 1.83
    And because r2q is an integer, it would need to be 1.

    With the current software things go wrong (no fair sharing of the excess bandwidth) if there's a birrate <120 kbit/s.
     
  41. phykris

    phykris Addicted to LI Member

    I found another problem. Sometimes I get the folllowing messages in my log:

    Aug 26 00:56:27 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:28 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:30 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:32 myrouter user.warn kernel: NET: 4 messages suppressed.
    Aug 26 00:56:32 myrouter user.warn kernel: dst cache overflow
    Aug 26 00:56:37 myrouter user.warn kernel: NET: 4 messages suppressed.
    Aug 26 00:56:37 myrouter user.warn kernel: dst cache overflow
    Aug 26 00:56:37 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!

    I think it's realated to what a specific users does. When he's doing it I get these messages about every 30 minutes.

    I didn't get any customer that came to complain, so I think the consequences are not that big.

    I never saw this in version 1.19 (original).
     
  42. Toastman

    Toastman Super Moderator Staff Member Member

    Saw that myself once or twice, it's a memory problem, happened to me when too many connections opened by a couple of P2P guys - usually at that point the router gives up and reboots. Maybe you can knock that guy down in his useage of resources somehow? I've throttled P2P to the point where connections usually limit at around 2000, only rarely does it rise much above this. It still allows people to max out the connection, and nobody's even noticed any restrictions at all. But I'd like to be able to set a definite limit on everybody in a more controlled way.

    It's possible that something is different in this compile, in which case Victek might be able to fix it.

    ** Re QOS issue - That manual is interesting, but can't say I understood much of it :frown: I never saw any router that had an adjustment for r2q...

    "With the current software things go wrong (no fair sharing of the excess bandwidth) if there's a birrate <120 kbit/s." - which is to say, quite often - however, I don't see those messages here.

    "for precise operation quantums need to be as small as possible...." implies that it probably is still acceptable but not "precise". What do you think, does it still appear to be working reasonably OK?
     
  43. phykris

    phykris Addicted to LI Member

    The QoS is working OK. But I don't have that many users. And when I use the old QoS system in Tomato 1.19 (not based on IP addresses / individual users) everything worked fine too.

    About the other error messages: I limit the number of TCP connections to 300 per user and I am still getting those messages, although not that often any more. I almost never have more than 700 connections. The router doesn't reboot and I think the software is stable, but we shouldn't be getting these messages and I never saw them in tomato 1.19.

    I am also very, very interested in the changes Victek made too solve your problems or to add the features requested by you. I think they are all excellent changes and I hope they will become part of the new release. If the router is technically capable of supporting 100 users (but limited/low bitrates), software shouldn't the limiting factor.

    I am also taking care of the internet in a condo in Thailand.
     
  44. phykris

    phykris Addicted to LI Member

    There are a few strange things about the QoS:

    - I get the the warnings in the log when the ULrate is set to a value lower than 80kbit. This is a bit strange because I would expect it would be when the ULrate is lower than 120kbit/s (= 1500*8*10, if r2q has the default value of 10).

    When I login with telnet and I ask for r2q of the qdisc, I see it's set to its default value of 10.

    I really don't get it.

    - When I type "tc -d qdisc" after logging in. I get this:

    qdisc sfq 29: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 28: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 27: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 26: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 25: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 24: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 23: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 22: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 21: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 20: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 19: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 18: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 17: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 16: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 15: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 14: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 13: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 12: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc sfq 11: dev br0 limit 128p quantum 1514b flows 128/1024 perturb 10sec
    qdisc htb 1: dev br0 r2q 10 default 0 direct_packets_stat 41072 ver 3.17
    qdisc sfq 29: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 28: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 27: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 26: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 25: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 24: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 23: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 22: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 21: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 20: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 19: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 18: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 17: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 16: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 15: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 14: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 13: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 12: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc sfq 11: dev imq0 limit 128p quantum 1500b flows 128/1024 perturb 10sec
    qdisc htb 1: dev imq0 r2q 10 default 0 direct_packets_stat 75153 ver 3.17

    This seems to suggest that the quantum is set to its minimum value when the calculated value is too low. With these settings there's a fair distribution of the bandwidth and there's no problem. But then I don't understand why I get the warnings??

    I am slowly beginning to understand how the mod works. I found the scripts it generates but I still have a long way to go.
     
  45. Toastman

    Toastman Super Moderator Staff Member Member

    Have you seen this?

    "BTW, for those of you who use the QoS in Tomato, there is a new undocumented "feature". By default, Tomato QoS uses SFQ qdiscs which should give 'fair queueing'. But it takes more CPU power, so I recommended to Jon to replace SFQ with simple PFIFO qdisc which is faster. It is in the Tomato 1.18 firmware already (all versions, including my SpeedMod one), but by default its not activated.

    To activate the PFIFO qdisc, telnet into the router and create a new NVRAM variable called "qos_pfifo" and set the value to 1:

    #nvram set qos_pfifo=1
    #nvram commit

    Reboot or restart QoS (go to QoS > Basic Settings page, click Save) to make it take effect."


    see also

    http://www.linksysinfo.org/forums/showthread.php?t=47720

    searching this site for "pfifo" turned up a lot more stuff on QOS operation
     
  46. bripab007

    bripab007 Network Guru Member

    Yes, I'm curious to know more about PFIFO vs. SFQ qdiscs in Tomato.

    Who here is using them, and what testing has been done?
     
  47. Toastman

    Toastman Super Moderator Staff Member Member

    Haven't seen many comparisons - in truth, unless done in a lab they tend to be pretty subjective.

    I have a lot of users, many of them P2P. I changed it a month or so ago to see what would happen. It seems to have reduced the number of connections somewhat, my own P2P (emule, uTorrent) seems unaffected. The firmware appears a little more stable. BUT - hard to say....

    Try enabling it and see what you think!
     
  48. Victek

    Victek Network Guru Member

    I did not see any variation in terms of connection_stability_speed BUT the CPU load increased and from 97% iddle time dropped to 83%. I tested 6 months ago after I reviewed my logbook with all changes and test done with Tomato.

    Thanks for the information, already tested. :smile:
     
  49. Victek

    Victek Network Guru Member

    Well, some modification are in the way waiting for the next oficial release but meanwhile you can download stable release:

    Tomato RAF 1.21.8005

    Changelog from previous version.

    UPnP up to 100 user
    IGMProxy for IPTV (testing)

    Link to download as usual http://victek.is-a-geek.com/tomato.html :biggrin:
     
  50. Toink

    Toink Network Guru Member

    Thanks for build 8005, Señor Victek :) Currently loaded on my WRT54GL.

    I'm not sure if build 8002 (or the base Tomato 1.21) has a problem re-connecting via PPPoE. But I really noticed my DSL won't re-connect after a while. I would wake up in the morning without DSL connection, I had to manually re-connect.

    Tomato 1.19 would reconnect a-ok.
     
  51. Victek

    Victek Network Guru Member

    Hi Toink, thanks for the info, I integrated a few mods in the build that wouldn't affect the issue you mention. As usual and sorry to bother you :rolleyes:, please do a nvram erase and reconfigure again. The sync for a new connection to the modem and IP WAN renewal takes no more than 5 seconds when you disconnect-connect manually?
     
  52. Toink

    Toink Network Guru Member

    Thanks for the prompt reply, Victek :)

    Yeah. It's been my personal preference to clear the NVRAM before and after each upgrade - via the http://192.168.1.1/admin-config.asp and then manually - Over kill, I know! lolz :biggrin: But I have O.C.D. he-he. :biggrin:

    I don't mind starting the config from scratch....

    Well, it could be that it's a modem issue (WAG54GX2). Or it could be that it's my ISP... So nothing is definite - probably just co-incidence :)

    I usually set the keep alive setting in Tomato to 5 or 10 secs.

    If anyone is also experiencing a PPPoE re-connect problem (Base build or Victek's), kindly confirm.

    Thanks!
     
  53. a9988cd

    a9988cd LI Guru Member

    Victek, Thanks for the new version.(v1.21.8005)

    whether "busybox" downgraded to 1.2.2 ?
     
  54. Victek

    Victek Network Guru Member

    Hi a9988cd,

    I downgraded to 1.2.2 because I'm now creating modules for each function that allows me to run partially the modifications without affecting the base firmware and find any bug easily confirming that if a problem is arising I have to look in the module/function instead of the whole firmware, modules like snmp, samba, miniupnpd don't need to run with BB higher version.

    Of course I had the opportunity to go in the other direction :biggrin: and integrate functions now in modules directly into BB like udhcpd. It's a question of skills and commodity for each modder and in my case I prefer to have clean modules to be ported to other firmwares.

    This is the root of the change.
     
  55. mstombs

    mstombs Network Guru Member

    I thought BB 1.11.x had much smaller footprint and seems faster/more efficient? Used to have 0.61_pre in our adsl routers so maybe anything above 1.x is better?

    You're not thinking of getting rid of dnsmasq for udhcpd are you? I recall Tomato uses a modded version of udhcpc inside BB (which broke when Official Tomato tried 1.10.).
     
  56. Victek

    Victek Network Guru Member

    Hi mstombs, nice to hear from you,

    The only difference I noticed compiling from BB_1.11 was 22-44KB size increase, never decreased even when I kicked out new functions. I was unable to build an smaller busybox (also using shared libraries to reduce size and using mips -03/02, no success.

    About dhcpc, yes, the service broke in Tomato official release, Jon got the patch to bypass the problem in /src/router/rc/dhcp.c file and modification in /src/router/Makefile, roadkill solved already in release 1.16 with BB_1.9.x.

    Thanks :biggrin:
     
  57. mstombs

    mstombs Network Guru Member

    When I get a spare router (i was thinking of WRT54G-tm) I'll have a go myself (BB 1.11 has send_mail...) - but for Samba you must be working with an Asus?
     
  58. Victek

    Victek Network Guru Member

    Yep, WL500 GP v1.
    Let me know if you need the BB_1.11 and collateral files mod, I'll send to you.
     
  59. bagu

    bagu Network Guru Member

    Can we expect a v1.21.8505 with new drivers ? or not ?
     
  60. Victek

    Victek Network Guru Member

    Yes, but not with the existing ND drivers, I'm working on it since one month ago. :biggrin:
     
  61. phykris

    phykris Addicted to LI Member

    Hi Victek,

    Thanks for release 1.21.8005. It's working fine on my local router and the next step is to go to test it in the condo. Just a small remark : I don't find the show NVRAM option any more (under tools) in the WEB interface.

    I am looking forward to your new version. It looks very promising ... I really appreciate the way you work on the stability of the software and the increase of the number supported users (up to 100 now, which is a fair amount). This is exactly what I was looking for!

    Edit: Ok, I've installed it in the condo. Seems to be working fine. The most terrible users wake up at night... so the real test still has to come. I've only 19 users now, I can't test the changes made to support 100 users.

    Edit: It's night time now. I am getting my first error messages.
    Aug 29 21:17:56 myrouter user.info kernel: ip_conntrack_rtsp.c: help_out: ip_conntrack_expect_related failed (-17)
    I saw them only twice. The consequences seem to be minimal (no complaints until now).

    Edit: Morning time now. Software still running fine. No new error messages.

    I tried to build the software myself. It complained about missing bison and flex, so I installed those. After that the build failed as follows:

    make[2]: Entering directory `/home/kris/tomato/release/src/linux/linux'
    touch: cannot touch `/tomato/release/src/linux/linux/include/asm/param.h': No such file or directory
    make[2]: *** [/tomato/release/src/linux/linux/include/asm/param.h] Error 1
    make[2]: Leaving directory `/home/kris/tomato/release/src/linux/linux'
    make[1]: *** [kernel] Error 2
    make[1]: Leaving directory `/home/kris/tomato/release/src/router'
    make: *** [all] Error 2

    Any idea what might be missing?
    The file param.h exists in tomato/release/src/linux/linux/include/asm. But tomato is not under the root folder, it's under /home/kris.

    Edit: ok, fixed that problem already. The dependencies were not correct. I removed the dependencies as follows:
    rm -rf $(find . -name '.depend')

    Edit: The software has been running for 2 days now. Since the upgrade to 1.21.8005 I don't get the following error messages any more:

    Aug 26 00:56:27 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:28 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:30 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!
    Aug 26 00:56:32 myrouter user.warn kernel: NET: 4 messages suppressed.
    Aug 26 00:56:32 myrouter user.warn kernel: dst cache overflow
    Aug 26 00:56:37 myrouter user.warn kernel: NET: 4 messages suppressed.
    Aug 26 00:56:37 myrouter user.warn kernel: dst cache overflow
    Aug 26 00:56:37 myrouter user.warn kernel: MASQUERADE: No route: Rusty's brain broke!

    This might just be a coincidence, but I think it's not, because those messages appeared rather systematic with the previous version.

    Thanks!
     
  62. Victek

    Victek Network Guru Member

    Nice to see this release resist your terrible users!! :biggrin:
     
  63. Toastman

    Toastman Super Moderator Staff Member Member

    Want some of mine?
    :banned:


    I sometimes see the "user.info kernel: ip_conntrack_rtsp.c: help_out: ip_conntrack_expect_related failed" message. Not often though, I think the problem mostly went away when I put rtsp in QOS rules with a high priority.
     
  64. phykris

    phykris Addicted to LI Member

    Hi Victek,

    Remember the error/warning when using small ulrate or dlrate?

    I fixed that error in the new_qoslimit.c file.

    I added a function "void check_for_small_ulrate_dlrate (int * ul_r2q, int * dl_r2q)" that calculates the most optimal r2q values and I changed the "void new_qoslimit_start(void)" function a little bit so it will use these optimal values instead of the default value of 10 when using "tc add qdisc".

    The warning/errors are gone now.

    I attached the file. What do you think about it?
     

    Attached Files:

  65. Victek

    Victek Network Guru Member

    That's a very good contribution, you know what?...

    You fixed also the HTB problem with dd-wrtx86 since r2q will be always smaller than the maximal MTU, see the error message that appears in dd-wrt:

    HTB: quantum of class 11127 is big. Consider r2q change.

    This is the advantage when I published the source code, everybody can contribute to improve the firmware.

    Attached here is the diff between original file and modified file, I'll integrate in the next release.

    Thank You!
     

    Attached Files:

  66. phykris

    phykris Addicted to LI Member

    You're welcome Victek, that's the spirit of open source.

    This error:

    HTB: quantum of class 11127 is big. Consider r2q change.

    occurs when the quantum is big. The quantum is big when r2q is too small. This warning is much less serious than the warning about the quantum that is too small because a quantum that is big leads to unnecessary delays while a quantum that is too small leads to inaccuracy.

    Some messages on the internet suggest to take r2q =1 for bitrates smaller than 120kbit. Doing so would lead to this warning (quantum is big). It's better to take the biggest possible r2q that will make the quantum bigger than the MTU.

    My implementation is not perfect. When changing the MTU size the "quamtum is big" messages might still occur. However, the more serious "quamtum is too small" messages should be gone.
     
  67. Toastman

    Toastman Super Moderator Staff Member Member

    Nice Mod, Kris!

    I want to try ipt_ROUTE and ipt_random to split traffic between two gateways.

    They are Netfilter Extensions which are normally installed by "Rusty" Russel's "patch-o-matic" module installer.

    http://www.netfilter.org/documentation/HOWTO//netfilter-extensions-HOWTO-4.html

    I believe I found references in Tomato source code - they may in fact already be added, but they don't work. Can anyone confirm?
     
  68. y2kboy23

    y2kboy23 Network Guru Member

    I've been testing version 1.21.8005 for about a week now and I have to say it's very stable. While browsing the UI tonight I noticed a type on the menu. On the 'IP/Mac Bandwidth Limiter' item, its spelled 'Bandwidht.' Not sure if you spotted it.
     
  69. Victek

    Victek Network Guru Member

    Thanks I'll fix it in next release
     
  70. phykris

    phykris Addicted to LI Member

    I also consider version 1.21.8005 to be an improvement over the previous version. On 4 days I got only 2 warning in my log (which I already mentioned).

    But today the most strange thing happened:

    The WAN(ppp0) TAB just disappeared from my "Last 24 hours" bandwidth graph, while it's still there in the real time bandwidth graph. And I got a new imq1 TAB in the "last 24 hours" graph. I might have done something wrong myself. I'll reset the router and see if it happens again.

    The show NVRAM option is also not available in the latest version (which I already mentioned).

    Anyway, I like version 1.21.8005 :)
     
  71. Victek

    Victek Network Guru Member

    :biggrin:, Yes I think you did something in the IMQ parameter .. remember that you can show the activity up to 16 IP's/machines conected to the router (only available in the Tomato RAF version).

    About NVRam Show... it's a pretty option but I have to review an type all the new nvram variables for each new version, I'll try to loose 10 minutes in my next mod and build it.

    Thanks!
     
  72. Toastman

    Toastman Super Moderator Staff Member Member

    Sounds like you need an exorcist :biggrin:

    Victek: all AP's still up after 2 weeks.
     
  73. nobugme

    nobugme Network Guru Registered

    How can we do that? Would be great to see who's really hogging the pipe! :)

    Thanks.
     
  74. Toastman

    Toastman Super Moderator Staff Member Member

  75. Kent_Diego

    Kent_Diego LI Guru Member

    Any idea when a new ND, New Driver, version will be released? My version 4 WRT54G seems to work better with the ND firmware.

    Thanks,
    Kent
     
  76. Victek

    Victek Network Guru Member

    No, I'm working on 4.150.10.16 wireless drivers, as repported problems free with Centrino 2200 b/g. Patience Diego :biggrin:
     
  77. luk-a-sz

    luk-a-sz Addicted to LI Member

    hi,

    I have question.

    I can't set group of IP in IP/MAC BW Limiter.

    In the old rom (conanxu) i can set:

    ex. 192.168.1.114-192.168.1.115

    In the newest rom by Victek i can't do that.

    Could you help me ?
     
  78. Victek

    Victek Network Guru Member

    No, you can't. It makes no sense to fix the same parameter for a group of IP's cause then the HTB is overkilled.

    Thanks
     
  79. luk-a-sz

    luk-a-sz Addicted to LI Member

    I don't speak very vell in eglish but I try explein my problem.


    I have in LAN 20 users. Many of them have few computer.

    ex. Mr.1 have 3 PC: 192.168.1.5; 192.168.1.6; 192.168.1.7
    I make group of IP 192.168.1.5-192.168.1.7.
    Mr.1 use mostly only one PC in the same time.
    I give Mr.1
    DLRate 250 DLCeil 1000 ULRate 200 ULCeil 500

    WHEN

    I set each IP separately I use 3 lines and when Mr.1 use two PC in the same time take MAX 2Mb/s transfer.


    ------------------------------------------

    When I set each IP separately 20 record I can't set next IP :/ I need 25 lines in IP/MAC BW Limiter if i wont write separately.
     
  80. keasley

    keasley Guest

    I'm the new kid on the block. Been using stock tomato V1.21 - (love it). Currently I am deployed to the Middle East. There is a group of us (25 or so) using Satellite Internet service. Currently we have 512down & 2048up (10to1 ratio). Got rid of the Dlink 604 router (crap router) and purchased 2 Asus WL-G500Gp V2 routers. One is currently doing it's job and the QOS feature is wonderful (except when the guys are using P2P) even with IPP2P filters enabled. I am actively working on getting P2P under control.

    Gotta to give the creator a shout-out for the wonderful job he did on the stock firmware.

    The 2nd Asus router, just loaded Victec's firmware Tomato_RAF_1.20.8501ND I got from the repos to test and look at the features. Man, I love it. Thanks Victec, I believe with your firmware I can get the P2P users under control. Also be able to utilize the bandwith for it's true purposes. (communication with love ones back at home). The service is not cheap and we all pay for it, plus the additional equipment we had to purchase to wire the network together.

    Just wondering and here is the question; when will your version of the ND firmware come out for version 1.21? Tried upgrading to Tomato_RAF_1.21.8005 fm Tomato_RAF_1.20.8501ND and it a no go. Tried restoring the backup config and it was a no go.

    Reading this thread completely and the improvements you have made with the community's contributions, I am considering adding more people to the network (up towards 100) and increasing the bandwidth to accommodate increase number of users.

    I am still reading about the IP/MAC bandwidth limits, speedmod and the others. I am quick learner. Able to move around linux CLI to see what is going on with the help of forums/threads like this one.

    So, again thanks everyone for the hard work and the wonderful product. It is true what they say - "Turn a $60 home router to a $600 network appliance"
     
  81. Victek

    Victek Network Guru Member

    As I said some post ago, I'm working in new wireless driver to avoid Centrino 2200 b/g issue, this is the reason for a non Tomato RAF 1.21.8505 ND version yet ...
     
  82. jnappert

    jnappert LI Guru Member

    At Victeks Site there is a note that he plans to implement an IGMProxy for IPTV connections.

    Can this Proxy be used für RTP connections (VOIP) too? I had to switch to DD-WRT because of the possibility to run an outbound proxy (milkfish) for the VOIP-lines.
     
  83. Victek

    Victek Network Guru Member

    Yes, the first attempt was failed after test done with DarEinda, I'm testing why. The module used is the same as dd-wrt is using (from Openwrt).

    :biggrin:
     
  84. jnappert

    jnappert LI Guru Member

    Multiple SSIDs with the ND Version AND milkfish? That would be a dream ;-)
    Looking forward to it. If you need a tester....
     
  85. phykris

    phykris Addicted to LI Member

    okay, I can confirm I am not getting mad. :biggrin:

    The following problem has occurred again without me touching anything:

    The WAN(ppp0) TAB just disappeared from my "Last 24 hours" bandwidth graph, while it's still there in the real time bandwidth graph. And I got a new imq1 TAB in the "last 24 hours" graph.

    This time I am sure didn't do something wrong myself. Today I only changed the DLceil from 500kbit to 600 kbit for every user, for the rest I didn't touch the router. This time it happened after 3.5 days.

    Edit: I think the problem is not related to the number of days the router is on, because my home router, which runs the same software, is already on for 10 days and doesn't show the problem. So, it might be related to heavy use. The people in the condominium download around 10 GB per day (all together). When the WAN(ppp0) TAB-sheet disappears the daily/weekly and monthly traffic counters also don't increase anymore.
     
  86. Kiwi8

    Kiwi8 LI Guru Member

    Victek, can u specify the changelog between 1.21.8002 and 1.21.8005? Thanks.
     
  87. Victek

    Victek Network Guru Member

    Yes, it was the official implementation of the last patch for DNS 2.45, nothing else... but.. we are now in the RAF 1.21.8005 :rolleyes:

    Thanks!
     
  88. luk-a-sz

    luk-a-sz Addicted to LI Member

    Hi,

    Victek send me for test firmware.

    Change:
    - we can set more than 20 record in BW limit.

    On first look all works good. I set 24 record.
    I'm still test it.
     
  89. Kiwi8

    Kiwi8 LI Guru Member

    Yes, I meant 1.21.8002 and 1.21.8005 actually. :redface:
     
  90. m007

    m007 LI Guru Member

    I have read 29 pages and I don't see any of answer to my question. Also, I can't find anything of README in your source tarball. I am wondering what process I should do with your source tarball? Extract WRT54GL source tarball, then follow orignal tomato's README.txt for remove some stuff first before extract tomato source tarball. Is it same as with your? Thanks!
     
  91. Victek

    Victek Network Guru Member

    Just extract and make, all files (linksys sources+ Original Tomato source+ Victek mod) are in.
     
  92. bagu

    bagu Network Guru Member

  93. Victek

    Victek Network Guru Member

    Yes, it was...
     
  94. bagu

    bagu Network Guru Member

    ok, so i don't have to change my glasses ^^
     
  95. DerEineDa

    DerEineDa Network Guru Member

    Well, at the moment the igmpproxy found in Victek's current firmware is exactly the same as the previous igmprt, just renamed. It's bit-identical to the previous igmprt. I don't know what went wrong there.

    Maybe Victek got an outdated module from Openwrt? Because the igmpproxy found in the newer builds of DD-WRT is exactly this one: http://sourceforge.net/projects/igmpproxy . And this is certainly not the module Victek uses right now, unfortunately.
     
  96. m007

    m007 LI Guru Member

    Thanks!
     
  97. Slimey

    Slimey Network Guru Member

    I love tomato and all the work that's been done on making it the best, just spotted a small type-o "IP/MAC Bandwidht Limiter" :)
     
  98. GhaladReam

    GhaladReam Network Guru Member

    Victek is aware of this typo and it will be corrected in the next release :)
     
  99. jnappert

    jnappert LI Guru Member

    I found the manual for IP/MAC Bandwidht Limiter.

    I studied it several times but still need a hint: I only want to give my VOIP-ATA a garantied bandwith from 512d/128u.

    All other connections should be handled via "normal" qos. What settings must be done and do they work together?
     
  100. zoz2000

    zoz2000 Addicted to LI Member

    Hi Everyone!

    I have a problem with my WRT54GL. I have Victek's last firmware.
    My router works perfect with Utorrent then the speed goes down approx half of the full(down and up too). When i limit the upload the download goes up. I use QOS so usually the upload doesn't slow down the download. When i turn off the QOS and limit upload the download speed goes up but it doesn't reach the usually maximum. I tried it with Ftp, http and it affects too.
    The only thing that helps to restart the cable modem. It solves the problem awhile.(Approx a day):frown:
     
Thread Status:
Not open for further replies.

Share This Page