Printable View
It doesn't make any difference...Quote:
Originally Posted by lly
one USB-stick and one hhd on a 2.0 hub work fine, if I add a second hdd, there's trouble... I'm starting to thing the hub is broken...
cat /proc/bus/usb/devices only shows usb-hubs when the problem is there...
Also, the usb init bug was in 002-bcm5354.patch. So i don't think my WL500gP V1 was affected.
fix is on the way, take a patience, pleaseQuote:
Originally Posted by DrChair
any reason why the post-boot script would not be loading on r599 ?
my usb-stick is loading as /tmp/mnt/disc0_1 , where I have used the macsat guide to configuring as /opt
if i reboot, login , umount disc0_1, then run the script, all is fine.
login as: qaz
qaz@192.168.112.51's password:
[qaz@TV_WL500 root]$ umount /tmp/mnt/disc0_1
[qaz@TV_WL500 root]$ /usr/local/sbin/post-boot
[qaz@TV_WL500 root]$ cat /tmp/var/log/post-boot
2009-09-23 17:39:52 /usr/local/sbin/post-boot started
2009-09-23 17:39:52 /dev/discs found, USB key present
2009-09-23 17:39:52 USB key partition mounted as /opt
2009-09-23 17:39:52 starting rc.unslung
2009-09-23 17:39:52 /usr/local/sbin/post-boot finished
thanks
Update.....
post-boot IS working... (not really a surprise I guess)
I trimmed it back to just logging an slowly added actions to the script.
In order to get the Usb mount command to work, I have needed to place a Sleep 10 at the top of the script, before the mount command. My USB key must be a slow one. note that post-boot starts before the NTP service collects data.
1970-01-01 10:00:06 /usr/local/sbin/post-boot started
2009-09-23 22:05:55 slept10
2009-09-23 22:05:55 /dev/discs found, USB key present
2009-09-23 22:05:55 USB Key partions mounted as /opt
2009-09-23 22:05:55 starting rc.Unslung
2009-09-23 22:05:55 finished
hope my frustration helps someone else.
two question for r473
1) I've put the BT dl machine into DMZ, however, I still get a lot of following message.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:35 kernel: ip_conntrack: table full, dropping packet.
Sep 28 09:30:36 kernel: ip_conntrack: table full, dropping packet.
If that's in DMZ, it should not be via NAT and thus should not have table full problem?
2) I installed samba3, everytime I restarted, I find that there are 2 instances of smbd run, which takes up 20m of memory. Is it due to 2 USB interface of WL500gP and thus need to start the daemon twice?
Reply is highly appreciated!
increase the max connections in the webinterface at the firewall pages, might work:)
Arrg got a Problem:
bought a USB - HDD Case (Wintec EX-MOB-85) for my WL500G Premium V1
Now, everytime this Device is connected it seems that the DNS resolving in my network is destroyed.
The Asus Router is working in Access Point Mode, as another Router is Gateway, with NAT, DHCP etc..
Strange that internal Adresses are reachable.
No madder if I start Router first or HDD (as its desciped by Wengi) - Problem is there.
If I disconnect the HDD or unpower it, the Routers do their Job naizly - as there never was a problem...
Firmware is: WL500gp-1.9.2.7-d-r655.trx
any suggestions?
here more:
PHP Code:Oct 5 18:42:51 kernel: hub.c: new USB device 01:03.2-1, assigned address 3
Oct 5 18:42:52 kernel: Vendor: Maxtor 4 Model: D060H3 Rev: DAH0
Oct 5 18:42:52 kernel: Type: Direct-Access ANSI SCSI revision: 02
Oct 5 18:42:52 kernel: Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
Oct 5 18:42:52 kernel: SCSI device sda: 120069936 512-byte hdwr sectors (61476 MB)
Oct 5 18:42:52 kernel: /dev/scsi/host0/bus0/target0/lun0:
Oct 5 18:42:52 kernel: USB device connected to the SCSI subsystem
this seems normal...PHP Code:[admin@WL-002215329E61 root]$ ps
PID USER VSZ STAT COMMAND
1 admin 1312 S /sbin/init
2 admin 0 SW [keventd]
3 admin 0 SWN [ksoftirqd_CPU0]
4 admin 0 SW [kswapd]
5 admin 0 SW [bdflush]
6 admin 0 SW [kupdated]
7 admin 0 SW [mtdblockd]
60 admin 1092 S dropbear
64 admin 1516 S syslogd -m 0 -O /tmp/syslog.log -S -l 7 -b 2
65 admin 1120 S httpd
66 admin 1516 S klogd
67 admin 1072 S nas /tmp/nas.lan.conf /tmp/nas.lan.pid lan
69 admin 0 SW [khubd]
81 admin 1320 S watchdog
98 admin 0 SW [usb-storage-0]
99 admin 0 SW [scsi_eh_0]
118 admin 1152 S dropbear
119 admin 1520 S -sh
120 admin 1516 R ps
I didn't created pre/post Scripts allready, because I first want to solve this problem
anyway thx to lly and Miron and all the others around here :cool:
As I can understand, DNS resolver runs on gateway router, right? And clients connected to AP? Thus, check for problem by issuingQuote:
Originally Posted by !gm
on gateway, AP, client step-by-step. In case of nslookup fails on first, you should try to find problem there, not on WL500G. Otherwise, recheck what DNS servers you fill in on clients.Code:nslookup <some_host>
on the Gateway Router nslookup times out
----------------
the WL500GP does nothing, when nslookup is typed in (had to cancel it with Ctrl C)
a client behind the Asus finds the Gateway Router, but then gets a timeout, same for a client directly behind the Gateway Router
so the Problem is at the Gateway Router. But how can this be affected by an HDD that I plug into USB on the Asus? via ethernet - but why are local adresses reachable :confused: (yes i did flushdns on the clients)
lly, Can you add AES-loop support into the kernel?
It would be nice for all people who want encrypt his harddisk.
http://loop-aes.sourceforge.net/loop...-v3.2h.tar.bz2
First of all, I really worry about performance of this solution on our slow CPU.Quote:
Originally Posted by Lesiuk
Second, this loop driver are incompatible with standart one. i.e. we have to replace standart loop driver with this one. Until I know something about it's stability & absence of artifacts, I don't want include it to mainstream. Have you additional info & web-links?
Moreover, it assume patching of util-linux package. Firmware uses util-linux from busybox, so patches provided in distributive has to be ported to apply on busybox. Not so easy task.
http://linio.terramail.pl/loopaes.htmlQuote:
Originally Posted by lly
http://translate.google.pl/translate...istory_state0=
That is howto in polish language about aes-loop, that guy tested it on P133.
So i think it should work nice on that router.
If you make special version of firmware with that enabled i can test it.
I'm not sure that such functionality will needed by much people.Quote:
Originally Posted by Lesiuk
kernel-2.4.37.diff can be applied to our kernel clearly, and kernel builds successfully after that. Kernel size increases +15K compressed (+60K uncompressed). So, you feel free to rebuild FW with this driver as you wish, according with readme.
procedure is simple
- remove loop.h & loop.c from kernel as described in readme
- patch -p1 -i kernel-2.4.37.diff at src/linux/linux
- make oldconfig at src/linux/linux and set new parameters
- make && make install at src/gateway
Sorry, we have much urgent tasks (related to RT-N series) and can't spend a lot of time inserting this package into FW. As I wrote in previous letter, most complex task is util-linux patching.
There is AES-LOOP package in openwrt repo.
I think they are using busybox util-linux.
Can we just copy their code?
http://loblog.wordpress.com/2008/05/...aze-kernel-24/Quote:
I’m running OpenWrt “Kamkaze” on my Asus WL-500 gP (266 MHz MIPS architecture with 32 MiB RAM). Since there is no WiFi support in kernel 2.6 yet, I’m stuck to kernel 2.4. If we want to encrypt a filesystem on this setup, we need a fast and sleek encryption method running on kernel 2.4 .
UPDATE:
Ok, now i see.
They are using util-linux-ng.
No.Quote:
Originally Posted by Lesiuk
If it is not a top secret, for which specific task you plan to use aes-loop? Or it is just for fun?
Resonable compromiss (in case you convince me in real necessity of this module) I see in that we set loop device as module (currently it in kernel)
CONFIG_BLK_DEV_LOOP=m
and add aes-loop kernel patches to sources tree, leave they turned OFF. So, this allow to everyone to compile module without patching and use it on our public builds.
I second that... if you look at a modern day machine you can do a benchmark, which gives you a decrypt speed of about 250mb/s this is straight in the ram which can go over 1500mb/s.Quote:
Originally Posted by lly
and that with a quad core..
so how would a slow 3,5mb/s usb harddrive with even a slower cpu perform?