Ghm... I believe it should use external toolchain (the one which is provided by ASUS and me). Probably Rod could clarify this, as I've not used asusoe yet.Quote:
Originally Posted by Jean-Fabrice
Printable View
Ghm... I believe it should use external toolchain (the one which is provided by ASUS and me). Probably Rod could clarify this, as I've not used asusoe yet.Quote:
Originally Posted by Jean-Fabrice
Yes, AsusOE uses either the Asus external toolchain, or the one that Oleg distributes (depending on where your /opt/bcm/hndtools-mipsel-uclibc symlink points).Quote:
Originally Posted by Oleg
-- Rod
For Unslung packages: http://www.nslu2-linux.org/wiki/Unsl...ckageToUnslungQuote:
Originally Posted by Jean-Fabrice
For AsusOE: http://www.nslu2-linux.org/wiki/OpenEmbedded/HomePage
Note that these pages are written for the NSLU2, but you should be able to work out the changes required to build things for the wl500g if I say the following:
1/ For Unslung, "TARGET = wl500g"
2/ For AsusOE, "MACHINE = wl500g, DISTRO = asusoe"
-- Rod
Thanks a lot Rod.
CVS and packages development are really new for me, so I'm learning...
I have checkout unslung CVS tree in my colinux and now trying to build a package. Seems to work pretty fine for now...
After building my package in builds/, what should I do ? 'make mypackage-ipk' ? Don't I need a CVS write access before ?
Thanks again.
Yes. "make <foo>-ipk" will build the ipk file, which you can then scp up to the wl500g and "ipkg install <foo>.ipk" (install from a local file, rather than from an external feed - just give ipkg the full pathname to the local file). You can also setup a local feed - see the "upload" target in the top-level Makefile.Quote:
Originally Posted by Jean-Fabrice
You only need CVS write access to write <foo>.mk and any other files back to the repository. If you attach the make/<foo>.mk and any other sources/<foo>/... files here, I can add them to the Unslung CVS reposoitory for you. Note that the Unslung CVS repository is meant for the NSLU2, and we are re-using it for the WL500g, so I need to check new wl500g packages to make sure they compile on the nslu2 before I add them to the repository.
Most of the users and developers for the Unslung CVS repository can be found in the #nslu2-linux IRC channel on irc.freenode.net - if you need help on developing Unslung packages feel free to ask there. That channel is *not* for end-user help about using pre-compiled Unslung packages on the wl500g - it is a developer-only channel for those who are creating packages (like you are doing).
-- Rod
BTW, I use CoLinux too :-)Quote:
Originally Posted by Jean-Fabrice
-- Rod
Here is the output from hdden admin page:
cat /tmp/ppp/options.wan0:
noauth refuse-eap
user '*****'
password '*****'
plugin rp-pppoe.so eth0
mru 1492
mtu 1492
idle demand
defaultroute
persist
ipcp-accept-remote ipcp-accept-local noipdefault
ktune
default-asyncmap nopcomp noaccomp
novj nobsdcomp nodeflate
lcp-echo-interval 10
lcp-echo-failure 6
unit 0
nvram get wan_pppoe_idletime:
it is blank
nvram get wan0_pppoe_idletime:
it is blank
And the system log:
Jan 1 01:00:09 pppd[79]: Plugin rp-pppoe.so loaded.
Jan 1 01:00:09 pppd[79]: RP-PPPoE plugin version 3.3 compiled against pppd 2.4.2
Jan 1 01:00:09 pppd[79]: invalid numeric parameter 'demand' for idle option
Have you entered anything to the PPP Idle time box? According to this output it's empty...Quote:
Originally Posted by dejay
Thanks, I noticed this slightly improved my browsing with the original firmware but performance was still lousy with the custom firmware. Even downloading a file was about 10 times as slow. :mad: Luckily i figured out why. I had the following settings in Bandwidth Management / Download Policy List: Port 80, Max.(kbps) 128, Min. 96 (kbps). Of course "known bug n. 10" (Bandwidth management page treat rate values as kilobytes, not kilobits) limited the bandwidth heavily on the custom firmware with these settings. Btw, i've put a minimum of 192 and no maximum using the custom firmware and surfing has never been faster, really happy now with the new firmware. :) Thanks!Quote:
Originally Posted by tomilius
The Unslung Makefile has been updated with new entries for WL500G_PACKAGES and also another variable for packages that compile but just need a simple (but tedious) fix to emit the correct ipk file.Quote:
Originally Posted by rwhitby
I have started a separate thread at http://wl500g.info/showthread.php?t=2036 for discussion about the Unslung/WL500g packages, as from now on they should work with all future versions of Oleg's firmware, not just the version that this thread describes.
-- Rod
I've been looking everywhere, but I'm curious: Has anybody got a working NFS share working using the Windows services for Unix toolkit (SFU)?
I can see the shared directory's in 'My Network Places' -> 'Entire Network' -> 'Favourite LAN'. When I want to mount it (or just access it), it asks me for a username and password but every username/password I enter is wrong (and yes, I tried the admin user as well)...
NFS does not require username/password. The checks and access control are performed by client itself (which is why nfs is considered insecure).
Well that seems to be the problem. I did enter a 0, but for some reason it won't save it. This isn't a proflem for me, I enterd 65000 and it should be online all the time :) , but it is still interesting.Quote:
Originally Posted by Oleg
Hmm, yes it seems to be: http://wl500g.info/attachment.php?at...ntid=398&stc=1Quote:
Originally Posted by Oleg
But when I want to access it, Windows says: http://wl500g.info/attachment.php?at...ntid=399&stc=1
I've got not a clue why... (I am administrator on my computer).
EDIT: Forget this question please, I found out I need to setup some usermapping for SFU.... I'll report later :)
Summary of problem:
Asus is setup as WiFi client,attaches to ISP’s AP, but is not able to send or receive any IP packets on WiFi interface.
LAN ethernet ports are isolated from router inside Asus.
----------------------
Hi,
I tried to setup Asus 500g Deluxe as Wifi client to connect my home LAN to ISP using WiFi. It looks like there is either some bug in gx version or (more probably) there is problem in my setup (typical problem between keyboard and chair). I have some networking experience (Cisco CCIE,CCNP,CCDP) but Asus is new for me. I tested real life connection to ISP and later (after long hours of fight with ASUS) I built testing lab environment to limit ISP's complications (MAC filtering, security filters etc..). My ASUS worked in lab same way as in real life - WiFi connection has been built (my ASUS attached to AP) but there is no possibility to get DHCP reply, there is 0 packets received on WiFi interface from air. Another (very similar) behaviour is on LAN (eth 0.1) ports, it seems that they are only unidirectional ports.
I am sorry if I made some stupid error :confused: in setup but I spent long hours with Asus and I am desperate now. I tried almost everything (different setups, tried all HOWTOs I was able to find) and it does not work as expected.
I described my setup (pictures, screenshots, listings) and my tests as a Word file at http://www.tenzor.cz/surfer/asus/ASUSproblem.doc and I am attaching it to this post.
Could you give me some hint how to make this great router with Oleg's firmware (I like it, really) working in my setup?
Thanks
Ivan
i've not read your .doc file yetQuote:
Originally Posted by Aiwan
and I'm also not sure if WL500g.Deluxe as Wireless Client works
I will try this asap when I find some time..
(I have the same setup with WL-HDD/WL-500g and 1.9.2.7-3c firmware which works flawless only no encryption is used, so please try 1.9.2.7-3c too if you use no encryption)
Where are you from? If you're from Russia, then PM me. To be honest, client mode was never tested with wl500gx, so LAN port problems could be there.Quote:
Originally Posted by Aiwan
As for WLAN problem - do you use hex wep keys or passphrases? Passphrases are know to cause problems (as they're mostly vendor specific, unlike WPA-PSK keys), use hex keys only. Also, have you tried using it without WEP at all?
Also, dependening on the peer software it could require either "Shared Key" authentication (like you've specified) or "Open System / Shared Key" (this way key is used for data encryption only).
Ok, I will check on this. Thanks.Quote:
Originally Posted by dejay
To get back on the Perl discussion: Why is it so difficult to get Perl compiled for a non x86 CPU? It's not that I want to equalize PHP with Perl, but PHP did compile quite easilly for WL-500g as well as NSLU2 AFAIK...
Perl on the WL-500g or any other non x86 would be great because of the enormous amount of scripts available for it (like Webmin ;)).
I have the same problem as AIWAN but then with a 500g. I have an 500gx that is set in AP mode and an 500g in client mode.
With a PC on a 500gx-LAN-port I can logon in the 500g but I cannot connect to a PC on a 500g-LAN-port. The other way around gives also a problem: an PC on a 500g-LAN-port can login in the 500g but cannot login in the 500gx nor connect to an PC on an 500gx-LAN-port.
So there is no connection between the wireless part of the 500g and the LAN-ports on that box. In ethernet-bridge-mode this problem doesnot exists. On both boxes I use fw 1.9.2.7-4.
We had to natively compile Perl - we couldn't easily cross-compile it. So a lot of the effort was getting a native compilation environment working.Quote:
Originally Posted by Styno
I believe PHP was cross-compiled without many issues.
In fact microperl could be cross compiled, at least openwrt has it.
Just a small question in between. Is this 'range problem' solved in the latest CR4 release? I searched through the forum but couldn't find any info.
regards,
Theo.
wi-fi part is still the same, as this firmware is based on the 1.9.2.7.Quote:
Originally Posted by Theo
Thanks for reply Oleg, I am from Czech Republic (I have studied russian language 11 years but my technichal russian is bad :( ). I am using passphrase, but I assume that if wl status shows that I am connected, it means that WEP encryption works well and data should go through. I'll test all possibilities recommended and let you know.Quote:
Originally Posted by Oleg
LAN ports (eth0.1) is solvable by hack using crossover cable between WAN and LAN ethernet port (3 remaining ports are enough for now). This is low priority issue.
WiFi isssue is much more serious.
Thanks for hints Oleg. I appreciate it.
Hi,Quote:
Originally Posted by Hoorna
your network behaviour is same as if PC on 500g LAN port has no route to other subnet than connected LAN. Check if you have default route on PC on 500g LAN port with next hop 500g's LAN interface. Of course I can not be sure only from your decription, checking route tables from all PCs and Asuses might help. Also firewall settings may confuse things.
Hope it may help
Aiwan
Hi all,
recently I bought the Asus 500g to connect another remote LAN to my wireless network. I wanted to use the WDS function but so far i am not succesfull in it.
I loaded the newest release of the firmware (1.9.2.7-4) and configured the 500g via the web panels. (IP-address, WEP-key, MAC filtering etc.)
Computers connected to the already existing LAN are now able to ping/ configre the new Asus 500g via the WDS (so no cables!) It's even possible to ping from a telnet session on the asus over WDS via the other AP to the already existing LAN and even Internet. (e.g. www.chupa.nl)
But now the problem:
Computers connected to the Asus 500g AP. Can ping to the Asus 500g via Wireless connection and via LAN connection but they are not allowed to ping the remote Access Point (already existing LAN) at all . When i take al look at the log via telnet tot the 500g I notice the message duplicate IP-address which comes from the eth0 and eth2 combination (same HW address for wireless and LAN connection)
First I thought it was a routing problem but somewhere I read the WirelessLAN adapter and LAN adapter ought to be in the same subnet. This is why they are bridged together....
I tried to change the ip address with ifconfig but I lose my connection several times and gave up.
(where to store the settings via telnet session?)
Is someone succesfull in connecting two LANS via ASUS 500g with this firmware?
Keywords: bridge,stp,gateway,wds,route,metric, midnight commander mc,ifconfig
additional question: Has someone a compiled binary for midnight commander to use on the asus?
Is anyone else experiencing instabilities with the latest 1.9.2.7-4 release of Oleg's firmware?
It seems my unit regurarly looses wireless (wl down; wl up re-enables it). Also it seems to do spontaneous reboots. I usually have a USB stick mounted on /mnt but it gone after a while (seems to indicate reboot). Also the logfile (/tmp/syslog.log) indicates a reboot. Ofcourse there is no clue why this happens because the old log is gone. Usually the issue occurs somewhere between 12 and 24 hours after boot.
I am using an IPV6 tunnel, that may make a difference. Maybe I should try without it for a while (also not loading ipv6 module).
Regarding IPV6, the firmware seems to be imcomplete with regard to busybox. There is no ping6 and the available ping only understands ipv4. The same is true for traceroute. Anyway, IPV6 addresses result in 'host unknown'.
I have the same issue, and I don't run IPV6. I have only experienced the reboots (but maybe the wl down occurs just before that and I don't notice).Quote:
Originally Posted by Feico de Boer
I haven't taken the time to characterise the problem yet, so I have no better data on it than you do.
-- Rod
Do you use WPA? This could keying problem - try to isolate this by using WEP. If this still occures - try playing with other wl options.Quote:
Originally Posted by Feico de Boer
The reboots problem is seems to be wl.o module related and it's common for 1.9.2.7 series. 1.9.3.6 includes an updated wl.o, so once ASUS release sources for the 1.9.3.x I will switch to it.Quote:
Also it seems to do spontaneous reboots. I usually have a USB stick mounted on /mnt but it gone after a while (seems to indicate reboot). Also the logfile (/tmp/syslog.log) indicates a reboot. Ofcourse there is no clue why this happens because the old log is gone. Usually the issue occurs somewhere between 12 and 24 hours after boot.
I will add ping6 and traceroute6. As for dns problem - have no idea, I do not use ipv6 at all. Ask phedny - I've included his work to the firmware, but it seems to be incomplete at the moment.Quote:
I am using an IPV6 tunnel, that may make a difference. Maybe I should try without it for a while (also not loading ipv6 module).
Regarding IPV6, the firmware seems to be imcomplete with regard to busybox. There is no ping6 and the available ping only understands ipv4. The same is true for traceroute. Anyway, IPV6 addresses result in 'host unknown'.
I have the same issue, and I do not use IPV6.Quote:
Originally Posted by rwhitby
I use the WPA-TKIP kryptering
I have down gradet to 1.9.2.7-3c it works fine.
/Esben S. Nielsen
As far as I know I do use the WEP stuff but I will check when I have access to the router again. That won't be until I'm back home. I have not changed anything when I upgraded from 1.9.2.7-3c and that version was stable as a rock. Maybe I should hard reset the unit once and try again with clean settings.Quote:
Originally Posted by Oleg
As I mentioned, I never had problems with the previous firmwares. When I was not playing around with it, it was simple up day after day 24/7. Got home and had wireless, now I have to walk upstairs first and reboot the thing (every day) to get connection. Not really the idea behind wireless.Quote:
Originally Posted by Oleg
Actually I think there is not a DNS issue. I have been tracing the stuff before with ping6 on older firmware (1.9.2.7-3c) with the not ipv6 enable ulibc. There it was a library thing (if I recall correctly ntoh() like stuff), so just enabling IPV6 in busybox might fix things. I guess ping and traceroute just don't understand the ipv6 format and interprete these as hostnames instead of 128 bit ipv6 addresses.Quote:
Originally Posted by Oleg
Still, thanks for the efforts Oleg, I really appreciate your work.
My guess is that somebody near you just installed wireless equipment. Nothing changed in the wireless stuff in the -4, as it's available as precompiled binary only. The problem is that I've never experienced reboots, etc., so I could not even try to catch this.Quote:
Originally Posted by Feico de Boer
As for stability - this one runs in the AP mode - no reboots, etc:
Code:[root@wl500g-1 root]$ uptime
18:08:19 up 21 days, 25 min, load average: 0.08, 0.01, 0.00
[root@wl500g-1 root]$ cat /.version
1.9.2.7-4
Ok, I see now.Quote:
Actually I think there is not a DNS issue. I have been tracing the stuff before with ping6 on older firmware (1.9.2.7-3c) with the not ipv6 enable ulibc. There it was a library thing (if I recall correctly ntoh() like stuff), so just enabling IPV6 in busybox might fix things. I guess ping and traceroute just don't understand the ipv6 format and interprete these as hostnames instead of 128 bit ipv6 addresses.
Still, thanks for the efforts Oleg, I really appreciate your work.
500gx seems to work OK as WiFi client with 1.9.2.7-4 with no encryption.Quote:
Originally Posted by Oleg
500gx seems to work OK as WiFi client with 1.9.2.7-4 with shared key WEP encryption, but key must be input in Hexadecimal form, not passphrase.
Bug in firmware is, that Asus wrongly (or at least differently from Cisco Aironet 350 PCMCIA card, Orinoco AP, Intel Centrino chipset WiFi card and my ISP's D-link AP) translates passphrase WEP key into Hexadecimal form (which is probably used for encryption process). When I used table below (from Orinoco AP Help) and translated passphrase manualy into Hexadecimal key, it started to work.
There was no difference between "Open System / Shared Key" and "Shared Key" option, after translation of passphrase, both worked well with my ISP's D-link AP.
My fault was that I missunderstood wl status output
[admin@ASUS500GX root]$ wl status
SSID: "pokus"
Mode: Managed RSSI: -28 dBm noise: -99 dBm Channel: 1
BSSID: 00:02:2D:2D:10:AD Capability: ESS WEP
Supported Rates: [ 1(b) 2(b) 5.5 11 ]
I understood that WEP keys were compared and encryption works well. It is not true. The output above is probably radio interface status with no connection with encryption.
Problem with strange behaviour of LAN ports is not solved and I will have to investigate it. :confused:
Ivan
----------------------------------------
Translation from passphrase to hexadecimal equivalent (thanks Orinoco AP for intellectual property, I am not breaching their copyright too much :o ).
Encryption Key strings are case-sensitive!
Example: If your Encryption Key reads: "Key2Z" the hexadecimal equivalent would be: "4B657932A"
Alphanumeric Hex Alphanumeric Hex
A 41 a 61
B 42 b 62
C 43 c 63
D 44 d 64
E 45 e 65
F 46 f 66
G 47 g 67
H 48 h 68
I 49 i 69
J 4A j 6A
K 4B k 6B
L 4C l 6C
M 4D m 6D
N 4E n 6E
O 4F o 6F
P 50 p 70
Q 51 q 71
R 52 r 72
S 53 s 73
T 54 t 74
U 55 u 75
V 56 v 76
W 57 w 77
X 58 x 78
Y 59 y 79
Z A z 7A
Spacebar 20 0 30
! 21 1 31
" 22 2 32
# 23 3 33
$ 24 4 34
% 25 5 35
& 26 6 36
' 27 7 37
( 28 8 38
) 29 9 39
* 2A : 3A
+ 2B ; 3B
, 2C < 3C
- 2D = 3D
. 2E > 3E
/ 2F ? 3F
[ 5B @ 40
5C
] 5D
^ 5E
_ 5F
` 60
7B
| 7C
7D
~ 7E
Del. 7F
In fact there is no WEP hex key - passphrase translation schemes, so every vendor uses it's own implementation. Agere/Orinoco drivers supports really weak scheme, which only uses ASCII translation (this reduces key strength), also at some point I had problems interacting Orinoco Gold Card with Orinoco AP - they've translated passphrases differently (this was driver bug).
So, as I've suggested several times - do not use WEP passphrases, use hex keys.
Oleg, is there a chance to get a shutdown-function for the WL-HDD implemented in a further version of your excellent firmware ?
Well, as I do not have wlhdd, I could not even check how this should work. Does it really turns off (i.e. all LEDs are turned off and HDD is stoped)?Quote:
Originally Posted by ManfredK2
To test WebTools on WL-500g I dug up an old 40 Gb HDD and created 2 partitions on it:
- 64 Mb Swap ( /dev/scsi/host0/bus0/target0/lun0/part5)
- the rest ext3 ( /dev/scsi/host0/bus0/target0/lun0/part6)
When I boot the router there's a `/opt` directory, but I can't write to it: If I do `mkdir /opt/test` I get:this is some config on my WL-500g:Code:mkdir: Cannot create directory `/opt/test': Read-only file system
proc:
mount:Code:cd /dev/scsi/host0/bus0/target0/lun0/
$ ls
disc part1 part5 part6
How can I get /proc to be read-write?Code:mount
/dev/root on / type squashfs (ro)
none on /dev type devfs (rw)
proc on /proc type proc (rw)
ramfs on /tmp type ramfs (rw)
usbfs on /proc/bus/usb type usbfs (rw)
/dev/discs/disc0/part6 on /tmp/harddisk type ext3 (rw,noatime)
Styno, I'm wonder: have you read this thread from the begining? Also, try using search, as read-only /opt is discussed several times already. :)
Hint: you need to mount /opt first or to switch to the external rootfs. :)
Dear Oleg, thank you for replying so quickly.
- Yes, I've read the entire thread -> It has only hints for people using FAT, rootfs or NFS. I'm using native ext3 partition without rootfs.
- Yes, I've searched (the search button is a big friend of mine remember? :p) on `/opt` and read all topics (about 14 of them) but could only find pointers to this topic, no solution.
Please can you help me?
Edit: Ok, I've tried every option in the mount man page and voila, the following works:
Thanks for your patience.Code:mount -obind /tmp/harddisk /opt