Nope, I'm just in testing mode at the moment so there is no traffic other than my browser trying to access the config pages. Only 1 wireless client and no other traffic.
ok, thanks for the feedback. I will try some more experiments then. Maybe something on my end. I already performed "mtd erase nvram;reboot" but maybe that was part of the problem because my first task was to figure out the WAN problem, which ended up being fixed by Eko's posted firmware and possibly in that process I broke some WLAN stuff.
BTW I don't get any wireless errors, it is just very slow to respond. Wired connection is fine, both within the LAN and LAN->WAN. WLAN is very choppy even reading DD-WRT config pages and WLAN->WAN.
I'm using Dell 1390 internal WLAN BCM4311/BCM2050.
The WLAN card works fine with my Linksys WRT-54G v4 and WRT-54GS v3. I set up the Linksys and Belkin routers the same day on v23 SP2 so I was running through the same procedures, installing firmware, clearing NVRAM, verifying WAN connectivity. I have yet to start on my Buffalos because I'm playing with the F5D7230-4 V1112 right now and will start on F5D7130 v1010 next but maybe if I get too frustrated I'll do the Bufallos first.
It is just a little strange to me and a puzzle to solve especially now that I have feedback others are not having WLAN troubles.
@Eko
Is there a way to set WAN port in full duplex mode ?
Right now it seems to be in half duplex (100Mb/s) , what I can see on linksys router leds.
If I connect my PC to WAN port it also connect with 100 Mbit/s half duples, at least that is what "mii-tools" reports in Fedora 6.
@sfhub
Browsing on labtop seems to be ok.
I get about 11Mbit/s RX and TX at the same time, what is about 22Mbit/s total. Not bad for a little 125MHz CPU. this is messured with sending ping flood from my PC (Fedora 6) to an older IBM labtop (Unbutu 6.10). I have a Ralink RT2500 chipset card in the labtop.
ping -f -s16192 192.168.1.147
The size value for "-s" is a multible of 1472+n*1480 , this will give you the biggest packets possible. I have "iptraf" running in a second terminal to messure throughput.
For just pinging PC to Belkin I get about 18 Mbit/s Rx and Tx , what is 36 Mbit/s per connection.
ping -f -s 25152 192.168.1.1 _________________ SB5120 motorola cablemodem
F5D7230-4 V1232 Belkin, DD-WRT v23 SP3beta std - build 4788
AMD Sempron 2600+ running Win98SE or FedoraCore6
Joined: 07 Jun 2006 Posts: 23 Location: New York, NY
Posted: Mon Nov 13, 2006 13:34 Post subject:
Build 4538 successfully enables the WAN port on my F5D7230-4 v1111tt. Since I'm using it as a wireless bridge, the WAN port is only useful to me bridged to the LAN ports. It appears that this model (Broadcom BCM4710 chip rev 0) has no VLAN support, so the only way to do this is using
Code:
ifconfig eth1 0.0.0.0
brctl addif br0 eth1
which I put in the startup commands so it will survive reboots.
max2010 wrote:
Eko,
thanks for your answers.
I just reportet what I found different from Belkin firmware in case somebody else wants to compare it with his results.
It is just a little strange to me and a puzzle to solve especially now that I have feedback others are not having WLAN troubles.
Armed with the knowledge others had WLAN working, I tried mini-4204 again. This time everything just worked. As mentioned earlier WAN was already working the first time around with Eko's build (and future builds). Only WLAN was giving me troubles.
I don't know what changed. The only thing I could think of was the first time I went from an older Belkin 3.00.05->micro-sp2final->mini-4204. The second time I went from newer Belkin 4.05.03->mini-4204->std-4538.
Incidentally I didn't need to go from Belkin->micro(2mb) as the guide suggested as the second case showed. Perhaps my earlier shenanigans set things up so the router could take > 2mb flash directly.
Thanks for the feedback that WLAN was working for you guys. It gave me encouragement to continue trying. Now on to the F5D7130 v1010, which is rumored to only have 2mb flash. I will try micro first to see what it has then proceed from there if it has 4mb flash.
I loaded std-4538 onto the 7130 v1010 and everything was working fine. I should have left it alone. The I got the bright idea that I would clean out NVRAM using "mtd erase nvram;reboot"
Well it rebooted and everything worked, but I had reset my 7130 into a generic RT210W with the default MAC addresses instead of the ones programmed into CFE.
Then I figured I would TFTP the original Belkin 4.05.03 firmware and have it reset everything before I did the upgrade again. TFTP went fine, couldn't get to web config pages. I was a little panicked at this point. I think the problem was because I reset to RT210W values, the Belkin firmware couldn't run to completion.
Luckily boot load was still enabled, so I was able to TFTP std-4538 back again. However even then I could not reach the DD-WRT console. Luckily I found that I could telnet into the 7130. So I did that and on the theory that I messed up the NVRAM with RT210W values, I diff'd the RT210W nvram values against the ones I had just after I upgraded to std-4538. I hand entered each value and did "nvram commit"
After this I was able to load DD-WRT console. I think the critical NVRAM setting was some bridging parameter. Anyway, I saw with nvram show that even more nvram was being used than before so growing bold, I TFTP back to Belkin 4.05.03 in the hopes I could do factory defaults, then go through and upgrade to DD-WRT for a cleaner NVRAM.
Belkin firmware started up fine. Now when I went to do factory defaults, that is when the big trouble started. I guess I forgot I was using 7230-4 4.05.03 on the 7130. Now when I reboot, there is no boot load pause for TFTP. The router never responds to any pings at 192.168.2.1, 192.168.2.100, 192.168.2.254, 192.168.1.1, 192.168.1.100, 192.168.1.254
I've tried using crossover cables, using a 100baseT hub, and various attempts to TFTP flash. I think it is not hopeful I will be able to TFTP flash because I opened up 6 command prompts and "ping -t -w 1" the six addresses above and went through lots of different reboot sequences with the reset button. I never got a single ping response so I'm doubtful I would ever be able to TFTP in the current state.
I went ahead and opened up the unit, which in itself was a little confusing. The 2 screws are hidden under the product label at opposite corners. I removed those but still the case wasn't that easy to open. After a some squeezing of the sides the first clips came loose and then it was obvious what needed to be done to get the whole thing open.
Anyway, I now have the board sitting in front of me. There is a heat shield covering most of the components. I removed the mini-PCI and rebooted because someone mentioned earlier that this might cause the system to think something was wrong and enter TFTP boot load wait state. No dice, the router happily booted without the mini-PCI card inserted but still not response.
I have 2 options at this point. I can try to short some pins either on the 20-pin header or somewhere else. The other option is to go the JTAG route. The board looks very similar to the pictures Eko posted for the 7230-4 v1000.
I've tried shorting pins 1&2 on the 20-pin header because someone mentioned that might get boot wait working, but it didn't seem to do anything or make any difference in the boot sequence.
I started looking at the components under the heat shield. I'm too lazy at this moment to remove the heat shield so it is hard to read the component labels underneath the shield.
I found my unit is using
ESMT M12L64164A 8MB RAM (2 pieces for total of 16MB)
and
Fujitsu 29DL323BE-90PFTN 4MB ROM
The Fujitsu ROM didn't seem to match anything in the readme for the wrt54g flash JTAG tool. I noticed Eko put /fc:29 on his command line for the 7230-4 v1000 firmware flash which forced selection of
MX29LV320B 2Mx16 BotB (4MB) Flash Chip
I'm wondering if I should use the same or something else. Will there be any troubles with my memory chip not being listed under the options for wrt54g jtag flash tool? I don't have the JTAG cable assembled at this moment and I found the solder points to be kind of tiny. I'm concerned my sloppy soldering will end up joining adjacent holes and I will be there solder sucking multiple times.
Is there a simple thing I can short to get the router back into TFTP mode? All I did was do factory reset. The 7130 was working prior to that.
Anything else I can try before I go out and buy the $5 of stuff for the JTAG cable?
Thanks in advance. I found in general factory defaults is just not a good thing to do unless you are using Linksys routers. It seems to have potential for all sorts of trouble with other OEM units.
I also tried using DHCP on ethernet to try and get an IP and tried to associate using WLAN. I think by resetting to defaults from 7230-4 firmware instead of 7130 firmware I ended up disabling many of the ports. It probably thinks the ethernet port is WAN and WLAN is just horked.
I think my only hope is to clear (and/or replace) nvram, cfe, or both.
Maybe if someone can give me some pins to short on the flash memory chip, that could work also.
Next time there is only 1 (one) nvram value you shoud change (et0macaddr).
Lesson learned....
So does it matter that my unit is using this ROM
Fujitsu 29DL323BE-90PFTN 4MB ROM
and the post you mad suggested /fc:29
MX29LV320B 2Mx16 BotB (4MB) Flash Chip
Is that an equivalent memory write architecture?
Actually all the NVRAM changes I made by hand worked fine. The lesson I learned is factory default is something to use warily and only as a last resort especially if you are on a non-linksys router.
For past systems I've used factory default is kind of like your savior in case stuff goes wrong. With these OEM boards and various vlan/switching implementation, access point design variations, cross swapping firmware, you are much better off immediately creating NVRAM backups on a clean system and restoring, than you are using factory default.