Yashi, recompiling into the kernel isn't necessary, just wondering where the missing enhanced usb module is that handles high speed usb 2.0 which would be:

ehci_2.6.15.3-brcm-1_mipsel.ipk

which doesn't seem to exist in your package list, only ohci and usb-core.

I'd like to setup networking using vlan1 as the wan IP instead of using br0 that handles WAN and wireless since I want to break up wireless, wan, and lan on 3 seperate subnets.

vconfig add eth0 0   creates my desired lan IP on vlan0

however

vconfig add eth0 1   although it creates vlan1, is not transparent with eth0.  Doing a udhcpc -i vlan1 will NOT bring up the vlan1 interface on the wan side.

udhcp will only work on the eth0 interface, which is NOT what i want obviously.

How would i best set vlan1 as the wan interface device?

The only way i'm able to do it is create a bridge from vlan1 to eth0 but that's not what i want either.

Ok,
with latest build ( OpenWrt Kamikaze 3250 compiled Sat Feb 18 16:38:35 EST 2006),
and  tftp flashing, serial console does work again....
But, as it is told,
config does not work at all ^^'

@flyashi
Hey, don't be that sad,
you're doing a great job and we're really happy that your doing it for us!
So, if you got something we could do to help you,
just tell us ^^

CFE version 1.0.34 for BCM95365R (32bit,SP,LE)
Build Date: Tue Feb 24 03:21:41 CST 2004 (root@jackylinux)
Copyright (C) 2000,2001,2002 Broadcom Corporation.

Add MAC client version(DNI).
Initializing Arena.
Initializing Devices.
et0: Broadcom BCM47xx 10/100 Mbps Ethernet Controller
CPU type 0x29007: 200MHz
Total memory: 0x2000000 bytes (32MB)

Total memory used by CFE:  0x81BB1280 - 0x82000000 (4517248)
Initialized Data:          0x81BB1280 - 0x81BB3E90 (11280)
BSS Area:                  0x81BB3E90 - 0x81BB45D0 (1856)
Local Heap:                0x81BB45D0 - 0x81FB45D0 (4194304)
Stack Area:                0x81FB45D0 - 0x81FB65D0 (8192)
Text (code) segment:       0x81FB65E0 - 0x81FFFFB0 (301520)
Boot area (physical):      0x01B70000 - 0x01BB0000
Relocation Factor:         I:E23B65E0 - D:01BB0280

configure vlans
*****************************************************************
*********************** VLAN Driver initial  ********************
*****************************************************************
Process LAN port(2-5) vlan Architecture...
SUCCESS: trying to create VLAN 0 for switch
SUCCESS: trying to add LAN port

Process WAN port(2-5) vlan Architecture...
SUCCESS: trying to create VLAN 0 for switch
SUCCESS: trying to add WAN port
SUCCESS: enable ports  success
configure vlans...done
Device eth0:  hwaddr 00-0F-B5-0D-B4-85, ipaddr 192.168.1.1, mask 255.255.255.0
        gateway not set, nameserver not set
Loader:elf Filesys:raw Dev:flash0.os File: Options:(null)
***************************
****  MAC Client V1.0  ****
***************************
et0macaddr value :flag =0 value=00-0f-b5-0d-b4-85
et1macaddr value :flag =0 value=00-0f-b5-0d-b4-86
MAC exist at least one
system ethernet mac exist and not default....
Skip mac client process.....
Loading: 0x80001000/3732 Entry at 0x80001000
Closing network.
et0: link down
Starting program at 0x80001000
Linux version 2.6.15.3 (root@HOME) (gcc version 3.4.4) #1 Tue Feb 7 23:48:24 ES6
CPU revision is: 00029007
early_nvram_init: WGT634U NVRAM found.
Determined physical RAM map:
memory: 02000000 @ 00000000 (usable)
Built 1 zonelists
Kernel command line: root=/dev/mtdblock2 rootfstype=squashfs,jffs2 init=/etc/pr0
Primary instruction cache 8kB, physically tagged, 2-way, linesize 16 bytes.
Primary data cache 4kB, 2-way, linesize 16 bytes.
Synthesized TLB refill handler (19 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (30 instructions).
PID hash table entries: 256 (order: 8, 4096 bytes)
Using 100.000 MHz high precision timer.
Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
Memory: 29904k/32768k available (1992k kernel code, 2848k reserved, 247k data, )
Mount-cache hash table entries: 512
Checking for 'wait' instruction...  unavailable.
NET: Registered protocol family 16
PCI: fixing up bridge
PCI: Fixing up device 0000:01:00.0
Squashfs 2.2 (released 2005/07/03) (C) 2002-2005 Phillip Lougher
devfs: 2004-01-31 Richard Gooch (rgooch@atnf.csiro.au)
devfs: boot_options: 0x1
JFFS2 version 2.2. (NAND) (C) 2001-2003 Red Hat, Inc.
Initializing Cryptographic API
io scheduler noop registered
io scheduler anticipatory registered
Serial: 8250/16550 driver $Revision: 1.90 $ 2 ports, IRQ sharing disabled
serial8250: ttyS0 at MMIO 0x0 (irq = 3) is a 16550A
serial8250: ttyS1 at MMIO 0x0 (irq = 3) is a 16550A
b44.c:v0.97 (Nov 30, 2005)
PCI: Enabling device 0000:00:01.0 (0000 -> 0002)
eth0: Broadcom 47xx 10/100BaseT Ethernet 00:0f:b5:0d:b4:85
Physically mapped flash: Found 1 x16 devices at 0x0 in 16-bit bank
Intel/Sharp Extended Query Table at 0x0031
Using buffer write method
cfi_cmdset_0001: Erase suspend on write enabled
Flash device: 0x800000 at 0x1c000000
bootloader size flag: 3
Creating 5 MTD partitions on "Physically mapped flash":
0x00000000-0x00060000 : "cfe"
0x000a0000-0x007e0000 : "linux"
0x00148000-0x00241000 : "rootfs"
mtd: partition "rootfs" doesn't start on an erase block boundary -- force read-y
0x00060000-0x00080000 : "nvram"
0x00260000-0x007e0000 : "OpenWrt"
NET: Registered protocol family 2
IP route cache hash table entries: 512 (order: -1, 2048 bytes)
TCP established hash table entries: 2048 (order: 1, 8192 bytes)
TCP bind hash table entries: 2048 (order: 1, 8192 bytes)
TCP: Hash tables configured (established 2048 bind 2048)
TCP reno registered
ip_conntrack version 2.4 (256 buckets, 2048 max) - 244 bytes per conntrack
ip_tables: (C) 2000-2002 Netfilter core team
TCP vegas registered
NET: Registered protocol family 1
NET: Registered protocol family 17
802.1Q VLAN Support v1.8 Ben Greear <greearb@candelatech.com>
All bugs added by David S. Miller <davem@redhat.com>
VFS: Mounted root (squashfs filesystem) readonly.
Mounted devfs on /dev
Freeing unused kernel memory: 120k freed
Algorithmics/MIPS FPU Emulator v1.5
insmod: diag.ko: no module by that name found
/sbin/mount_root: /sbin/mount_root: 36: mtd: not found
/sbin/mount_root: /sbin/mount_root: 36: mtd: not found
Partition marked as clean
/sbin/mount_root: /sbin/mount_root: 36: mtd: not found
jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x002eeed8: 0x3032 id
jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x002eeedc: 0x3032 id
umount: Couldn't umount /rom/dev: No such file or directory
init started:  BusyBox v1.1.0 (2006.02.01-03:48+0000) multi-call binary

Please press Enter to activate this console. b44: eth0: Link is up at 100 Mbps,.
b44: eth0: Flow control is off for TX and off for RX.
wlan: 0.8.4.2 (svn 1451)
ath_hal: module license 'Proprietary' taints kernel.
ath_hal: 0.9.16.16 (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413)
ath_rate_sample: 1.2 (svn 1451)
ath_pci: 0.9.4.5 (svn 1451)
PCI: Enabling device 0000:01:01.0 (0000 -> 0002)
PCI: Fixing up device 0000:01:01.0
wifi0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
wifi0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 3s
wifi0: turboG rates: 6Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps
wifi0: H/W encryption support: WEP AES AES_CCM TKIP
wifi0: mac 5.9 phy 4.3 radio 4.6
wifi0: Use hw queue 1 for WME_AC_BE traffic
wifi0: Use hw queue 0 for WME_AC_BK traffic
wifi0: Use hw queue 2 for WME_AC_VI traffic
wifi0: Use hw queue 3 for WME_AC_VO traffic
wifi0: Use hw queue 8 for CAB traffic
wifi0: Use hw queue 9 for beacons
couldn't load module 'wlan_scan_ap' (-89)
unable to load wlan_scan_ap
wifi0: Atheros 5212: mem=0x40000000, irq=2
wlan: mac acl policy registered
device eth0 entered promiscuous mode

BusyBox v1.1.0 (2006.02.01-03:48+0000) Built-in shell (ash)
Enter 'help' for a list of built-in commands.

  _______                     ________        __
|       |.-----.-----.-----.|  |  |  |.----.|  |_
|   -   ||  _  |  -__|     ||  |  |  ||   _||   _|
|_______||   __|_____|__|__||________||__|  |____|
          |__| W I R E L E S S   F R E E D O M
KAMIKAZE (bleeding edge, r3250) -------------------
  * 10 oz Vodka       Shake well with ice and strain
  * 10 oz Triple sec  mixture into 10 shot glasses.
  * 10 oz lime juice  Salute!
---------------------------------------------------
root@OpenWrt:/#

kmod-usb-ehci isn't in make menuconfig; I used kmod-usb2 and it worked... don't know if it was actually usb2 or not.

Glad to see that the OS boots at least. My /etc/config/network modifications seemed to apply after an ifup... i love the name "ifup" too, it's like a tacit admission, "yea, I f***'ed up"

- me

P.S. I'm still looking for the complete list; here's what I can remember:

To /etc/firewall.user, add:

iptables -A FORWARD -i br0 -o $WAN -j ACCEPT

To /etc/config/network, add "ath0" to br0.

To /etc/modules.d/20-madwifi, to the line:
ath_pci
add:
ath_pci autocreate=ap

To /etc/init.d/S20madwifi, add:

wlanconfig ath0 create wlandev wifi0 wlanmode ap
iwconfig ath0 essid "w00t it works"
ifup ath0

Now run firewall.user and ifup br0. Some of these may not be necessary on newer Kamikaze's...this was on a 3178 iirc.

Good luck to all,

- Yasha

thanks Yasha!  You're the man for providing the only source in town for these builds!  Its very much appreciated.

It looks like maybe your cron script missed rebuilding the kmod-switch package?  The version
in your repo is for 2.6.15.1.

That package would seem to have the key network modules for the roboswitch that the kernel complains  about not having that i posted earlier.

In case you don't see the source, it's available on the package link on at:

https://svn.openwrt.org/openwrt/trunk/o … ge/switch/

Thanx again!

(Last edited by Ydef on 19 Feb 2006, 03:27)

Just a simple and stupid question:
How to make the flash "writeble" for doing these changes, described by yashi?
(and how to boot from usb stick as root?)

^^
Thanks,
DeusExMachina

Images are back!

I did a make clean; make; hopefully that cleared up any mismatched package problems.

If you used SquashFS, see http://wiki.openwrt.org/OpenWrtDocs/Usi … f1b8be545d for editing files. To boot from USB disk, see http://wiki.openwrt.org/UsbStorageHowto … a50d94382a .

Let me know if this fixed anything. If not I'll just give up and shoot myself now.

- Yasha

P.S. Just kidding. About the make clean; make ... I actually deleted the entire trunk/openwrt directory. I just hope it was worth it. Lemme know, thanks.

Excellent!  Thanks for recompiling the kmod-switch module for the present kernel.  Now I'm finally able at long last to bring up the wan side dhcp IP on vlan1, meaning that as long as you include this package in your default image then networking should work out of the box and mirror the type of default networking found on the wrt units.

Another suggestion would be to include all usb packages, especially the usb2 package you pointed out to me earlier on the default image, since after all this is a usb storage router so it would make sense to have the usb port work by default.

Haven't installed your latest image yet, wanna get everything down to a save script like you do so when I reflash I need only run that script to have all my customizations restored. 

Two other modules that the kernel seems to look for but can't find on bootstrap are diag.ko and sg.ko

sg.ko should come with one of the usb packages i would think since it's the scsi generic driver.

Did you forget to compile the scsi generic as a module in your kernel config?

Actually, was also wondering if you could make a copy of your kernel config available for download, as that would save me some effort when I go to compile custom modules for this unit.

Also, any idea about diag.ko as another compile time option that you might have missed to compile as a module?

Thanks again for your all help!

(Last edited by Ydef on 20 Feb 2006, 19:39)

What do you mean, "kernel config"? my .config from menuconfig, or the kernel settings on the router? Well, I can't send you the latter; I can send you the former if you'd like, but it's pretty much default + everything else as modules.

If you really want the USB drivers compiled in, sure; the thing is, only the WGT an asus have USB; on all other routers it would just be wasted space. I'll try to make a script that compiles a separate, USB-included version, if you REALLY need it in the default image.

I don't know why kmod-switch wasn't included before; hopefully it's included now; and since it works, networking should *just work* after a reflash, hopefully, possibly, maybe, and if the gods are on our side, and a southernly wind comes after a full moon on the 3rd Friday of next month.

I recall that during my boot, I also had problems with diag.ko; don't know why... it seemed to work fine though, so... why fix it if it ain't broke?

I don't see any kmod's or anything else for that matter relating to SCSI... what do you want me to select?

So, I'll make a .config with USB and copy that in and make a separate usb-enabled images, in addition to the regular images.

Anything else?

- Yasha

Yes, your .config from menuconfig is what I meant, sorry for not being more specific.

Ah, no all I meant is that the usb modules should be included in the default image.  Maybe you made this additional already, as I haven't checked out your most recent images yet so sorry if that  request was redundant. 

As long as switch_robo.ko and switch_core.ko from the kmod-switch package are in /lib/modules/2.6.15.3 on the default image then networking will work out of the box ... which is what it looks like you've done so no need to be *that* pessimistic.

No need really, I can just sometimes overly anal ironing all errors out of my bootlog ... a bad habit that i should really get over.  For the time being i'll just comment out the insmod diag.ko line in /etc/preinit.

Yes, one request specific to the wgt634u that would i think be thoroughly appreciated by the community , i was wondering if you would mind making an image for this router that boots from the usb drive. 

The kernel cmdline args should look something like:

console=ttyS1,115200 root=/dev/scsi/host0/bus0/target0/lun0/part1 rootdelay=20 rw

I think the missing sg.ko module is the culprit behind why a hard link from that hellish /dev path is not made to /dev/sda1.  The kernel config option should be:

CONFIG_CHR_DEV_SG=m

Actually, now that I think about it, since it would be the root filesystem here, i think it would be more prudent to compile all of these options into the image:

CONFIG_BLK_DEV_SD=y
CONFIG_SD_EXTRA_DEVS=40CONFIG_USB=y
CONFIG_USB_DEVICEFS=y
CONFIG_USB_EHCI_HCD=y
CONFIG_USB_OHCI=y
CONFIG_USB_STORAGE=y
CONFIG_USB_SERIAL=m
CONFIG_USB_SERIAL_GENERIC=y
CONFIG_CHR_DEV_SG=y

So if you include that in your config compile time options you could replace that above line with a root=/dev/sda1 if you could possibly make this additonal image available? 

(Last edited by Ydef on 20 Feb 2006, 22:05)

YES O.o
An Image with an EASY and FAST option to set on directly "usb boot" would be fantastic!
Or just one that does it all the time, doesn't matter, but that would be great, cause no one would set webserver and so on directly on a FLASH (changing data saving on flash memory is ILL, cause they don't are done for beeing some kind of webserver hdd for mysql dbs! XD)
So, that would be awesome I absolutly agree with Ydef!

DeusExMachina

I'm reading the USB Storage Howto and I've a few issues with the usb boot script.

First, I think the WGT uses OHCI, not UHCI. I think I tried UHCI and it didn't work; OHCI did. Can someone confirm this?

Also, insmod-usb2 would replace usb-ehci.

Also, it talks of nvram variables, unavailbe in Kamikaze.

Also, the scripts for jffs2 and squashfs are different. I'm preferential to squashfs because of it's failsafe. (Not yet available on WGT though, right? Reset button driver doesn't work?) Anyway, I'll make a USB bootable image, sure. But it's gonna take a lot of testing and modding and stuff.

Long story short: Sure, I'll do it this weekend.

- me

P.S. The differences between the jffs2 and squashfs are going to be a challange to integrate compile-time...but.. i'll think of something. Or one of you will and email me.... right? Thanks

FYI, the diag.ko module is supposed to be responsible for the reset button and LEDs, in case you guys didn't know that obvious fact already....

Really Really?! O.o
Um well
THAT
EXPLAINS
A LOT!
^.^
XD

Nope I didn't know, and I think the others neither...
^^'

Thanks!!!

DeusExMachina

(Last edited by DeusExMachina on 21 Feb 2006, 22:41)

Actually, I ditched the ohci module altogether and am currently using my usb drive with just the following modules loaded:

ehci_hcd 25744 0 - Live 0xc020d000
uhci_hcd 27792 0 - Live 0xc0205000
usb_storage 27440 3 - Live 0xc00a0000
sd_mod 12080 4 - Live 0xc0079000
scsi_mod 71360 2 usb_storage,sd_mod, Live 0xc0171000
usbcore 103584 4 ehci_hcd,uhci_hcd,usb_storage, Live 0xc0226000

the ohci module would always bring up the low speed crap, which is not what i wanted.

Right, although everything is supposed to be configurable now by setting the variables in /etc/config/network that were previously set in nvram.  If anyone knows better please correct me on the evolution/update of the /etc/config/network script as the de facto replacement for the nvram in these cfe type routers.

Actually, I don't think that there's been an adaptable solution to using squashfs with usbdrive root due to pivot_root leaving squashfs symlinks empty to bite the dust, so I wouldn't sweat worrying about squashfs at all.

Actually, I finally got off my lazy ass an am recompiling the latest svn for the wgt to boot off the usb drive so if i don't run into any complications i could possibly provide relief duty with an image for you to put up for this functionality.   Thanks again for all your effort in this, especially in consideration that you don't even have this router!

@yashi
Hi Yashi, saw your new 'lil Server.....
AND WAS SHOCKED!
Boy, PLEASE DO NOT PUT YOUR HDD UPSIDE DOWN!
NEVER,
EVER,
if you want to power on the thingy.
'Cause you can install hdds these days in "normal" position and on the side, but NOT flipped upside down, no hdd will make that very long, so please, flipp it XD

Thanks
DeusExMachina

Yea, well, it's too small for OpenWrt anyway. Not really good for anything without a bigger HDD. But, if you INSIST, sure, i'll screw it into one of the drive bays.

- Yasha

P.S. Anyone got a small-capacity (4-40GB) IDE lying around and want it to serve a public service (compiling OpenWrt images?) email me

Okay...so i flashed my WGT634U using the web interface and the jffs2 image. It rebooted and well...I can't access it anymore. I'm told I can fix this only with a serial console. How would I go about making this happen? Help? :-/

Yea, you should never try Kamikaze without a serial console.

http://wiki.openwrt.org/OpenWrtDocs/Har … 3c08d62d2a has the serial pinouts of the WGT; you'll also need either another router with a serial console, a cell phone serial cable (or similar), or you can make your own using a MAX232 series chip (I used a MAX233, supposed to be 5V level but worked with the 3.3 on the WGT).

If you use another router, just connect the ground, RX to TX, and TX to RX.
If you use a MAX232 series chip, just follow the datasheet.
If you use a cell phone serial cable, just follow the pinouts on http://www.pinouts.ru.

Good luck,

- Yasha

What?  Are you saying that you can't mount a HDD upside-down?   I have never heard that, and have in fact run servers for many months with upside-down drives without problems.

If that was the case, there would be a LOT of bad Ipods out there after people set them down face-up? (Drive upside down at this point)

-Phil

Uh, I would have to agree with pEEf,  this sounds like one of the more ridiculous things I've heard in a long time, although please do prove me wrong if i'm out of line since if there's any truth to this, it would be a good thing to know.

(Last edited by Ydef on 24 Feb 2006, 07:09)

Ok well maybe you missunderstood me, 'cause I'm not an native English Speaker.... so....
Um well, yes I read it in serval newspapers and computer websites,
and I'm actually think the same way.

So um well, I think the c't mentioned it sometime ago (german computernewspaper for professionals, not such yellow-press-gamer-kiddy-shit) and serval other papers. So um well, and there's just the question: If you can mount it the right way, why flipping it upside down?

As I remeber, the "wearout" (damn, this word doesn't really sound technical ~.~) of the bearings and motor is a lot higher, than in the correct position.

So um well, heres an quote of pcguide.com "Make sure that you don't mount the drive upside-down or backwards. The label of the drive goes up, and the circuit board down. The connectors should face into the middle of the case so the interface cable can reach them."

Maybe thats a little prove, but actually, at the moment I don't find many websites about this problem in english (actually, I never find anything if I REALLY NEED IT.... anyone felling similar? XD^^')

DeusExMachina

What specifically needs to get done to get Kamikaze working with pppoe, dsl and a static ip ? NVRAM settings don't seem to be available in Kamikaze. I am using Kamikaze Revision 3254. The LAN part seems to work just fine. Also, clear documentation needs to be put together in the form of a FAQ, once this is up and working. (I'll volunteer to do this, once its up and running.)

Hm.  Very Interesting.  Well thanks for posting that tidbit.  I've never really paid any attention to whether i've had a drive upside down or rightside up before.  Maybe that's the reason behind some unfortunately short lived drives that i've had in the past.

Just set the nvram variables in /etc/config/network.  You shouldn't have a problem bringing up the wan on vlan1.

Did you install the pppoe packages and insmod the modules?