OpenWrt Forum Archive

Topic: Call for testing: Kamikaze 7.09

The content of this topic has been archived on 4 May 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

Hey everybody,

I would like to release Kamikaze 7.09 in less than a week and I could use some more volunteers for testing it.
The codebase mainly consists of 7.07 with backported fixes from trunk.
You can find updated prerelease builds at http://nbd.openwrt.org/kamikaze/7.09/

Changes since 7.07 include:
  - Documentation updates
  - UCI updates - uncommitted changes are now active on config reads
  - PPP fixes
  - Firewall fixes for dynamic interfaces
  - Config enhancements for dnsmasq
  - Timing fixes for BCM947xx (fixes flash access problems on some models)
  - Fix for BCM947xx and Atheros cards on Linux 2.4
  - Prevent interfaces from accidentally being started twice at boot time
  - Fix QoS for dynamically assigned interfaces
  - Fix IMQ errors when ip6tables is installed (Linux 2.4)

I installed pre-7.09 (r8905) it on my Meraki Mini to test Chillispot which appears not perfectly working *ONLY* for Atheros-2.6 target (see https://dev.openwrt.org/ticket/2404).
I didn't find particular issues, except the one I said.

P.S. adm5120 target will be excluded again?

(Last edited by ggp81 on 24 Sep 2007, 11:29)

is broadcom wireless fixed with 2.6. kernel? (as in Asus WL500gP)
I see a fix for 2.4 but none for 2.6
thank you.

nope, no wifi on 2.6 yet

I have installd nbd 7.09 (22 Sept.) on ASUS WL-500g Premium

For my ISP i had to install and configure pptp, but there are problems.

1.) pptp does not connect on reboot or after disconnect
2) after reboot oder disconnect (every 8 hours) i have to

ifdown wan; ifup wan
...

mostly after 3 times ifdown/ifup the connection working

#### WAN configuration
config interface        wan
        option ifname   "eth0.1"
        option proto    pptp
        option ipproto  static
        option server   10.0.0.138
        option ipaddr   10.0.0.140
        option netmask  255.255.255.0
        option username 'XXXXXXXXXX'
        option password 'yyyyyyyyy'

see also:
http://forum.openwrt.org/viewtopic.php?id=12448

ps: before kamikaze i used whiterussion 0.9 for months without problems

Maybe it would be a good occasion to organize a Release-Testing-Team - now with the 7.09 release being on the "door-step"?

Would be good to have a list of things which need testing and a group of people are ready to twist and bend the pre-release versions

:-)

I'll give 7.09 a whirl on my (wireless client-mode) wl-500gp tonight.

Rev 9009 is running just fine in wireless client mode on my Asus WL-500gP (Broadcom BCM947xx/953xx [2.4]). Customizations:

+ p910n
+ ntpclient
+ webif (rev. 3598)
- iptables

I was wondering if it would be possible to add usb boot option - I have opened a ticket https://dev.openwrt.org/ticket/2409 (with patch). I have tested this on 7.06 7.07 7.09 (2.6 only) on WL500gp

Also wondering if there has been any thing done to fix the segfaults that happen on this router ? (https://dev.openwrt.org/ticket/2035)

(Last edited by alexsamad on 25 Sep 2007, 05:45)

I forgot to say that on Meraki Mini also WPA2 in AP mode doesn't work, and it didn't even in 7.07
I opened a ticket (https://dev.openwrt.org/ticket/2271) dealing with this defect over 1 month ago, and I completely forgot it.

What about MadWifi? Is it still the same version as in 7.07?
And have the problems regarding smbmount etc. been addressed?

On a WRT54GS (v4), both 7.07 and 7.09 cannot configure wireless correctly. Downgrading to 7.06 fixes this. Most likely it is an issue with the iwconfig (it chokes on "iwconfig enc 1234567890") -- has the version included changed?

only using iwconfig or also with /etc/config/wireless?

hi all,
i have wireless running fine on my wrt54g (3.1)
this is a working WEP config (7.06 to 7.09) in /etc/config/wireless

config wifi-device  wl0
    option type     broadcom
    option channel  5
# disable radio to prevent an open ap after reflashing:
    option disabled 0

config wifi-iface
    option device   wl0
    option network    lan
    option mode     ap
    option ssid     mysid
    option hidden   0
    option encryption 'wep'
    option key '1'
    option key1 'XXXXXXXXXX'

(Last edited by av1shay on 28 Sep 2007, 10:03)

Here is my working /etc/config/wireless for 128Bit WEP encryption with Atheros WiFi in the WL500g Premium. On the Notebook I am using the Network-Manager from Feisty Fawn.

Some notes to the WEP key format. The format for the WEP key in /etc/config/wireless is HEX.
64Bit: Length: exact 5 characters, allowed are letters (upper and lower case) and numbers

echo -n 'awerf' | hexdump -e '5/1 "%02x" "\n"' | cut -d ':' -f 1-5
6177657266

128Bit: Length: exact 13 characters, allowed are letters (upper and lower case) and numbers

echo -n 'xdhdkkewioddd' | hexdump -e '13/1 "%02x" "\n"' | cut -d ':' -f 1-13
786468646b6b6577696f646464

/etc/config/wireless

config wifi-device  wifi0
        option type     atheros
        option channel  5
        option disabled '0'

        # REMOVE THIS LINE TO ENABLE WIFI:

config wifi-iface
        option device   wifi0
        option network  lan
        option mode     ap
        option ssid     OpenWrt
        option encryption       'wep'
        option key1     '786468646b6b6577696f646464'

(Last edited by forum2006 on 28 Sep 2007, 10:17)

nbd wrote:

only using iwconfig or also with /etc/config/wireless?

iwconfig chokes on "iwconfig enc 1234567890".

When using the same key in /etc/config/wireless, iwconfig will report "Key too long" (or something similar) for the encryption setting.

I did originally try hex keys when first configuring it with wpa/psk.

nbd wrote:

I would like to release Kamikaze 7.09 in less than a week and I could use some more volunteers for testing it.
The codebase mainly consists of 7.07 with backported fixes from trunk.

Several persons have found out that buildroot-ng ignores instructions to compile the driver for the IDE device present on the board of the WL-700gE when using a 2.6 kernel. It is not a recent regression, I have been able to reproduce it with 7.07. A ticket has been created.

i use 709 for a week now on au1000 and wl500gp with atheros.
first time no tx timeout in adhoc and very fast bandwidth in bridging. even 5ghz turbo mode is working in wds / client bridging.
max throughput 6mbyte in turbo a!
good job

Everytime I build an image, I end up with files in /tmp (note that  TMP != /tmp)

Whilst looking at include/host.mk

I saw this

if find -L /tmp -maxdepth 0 >/dev/null 2>/dev/null; then \

wondered why /tmp is hard coded

I usually end up with /tmp/ipkg & /tmp/sed.script

Apart from that ( and the random segfaults which seem to be fixed now), every thing else seems to be okay

Still issues which prevent it from working so well on Soekris boards.
- There is that problem with grub's configuration. This has been discussed before.
   In menu.lst  the"serial" line should have the word "console" removed, or the system will not boot without a serial console attached.
   It might also be good to add "dumb" because it is painfully slow watching that grub screen over such a slow serial connection!

- There is the problem where users don't know what size to set CONFIG_TARGET_ROOTFS_FSPART to (56 on 64MB models and 117 on 128MB     models).

- I had problems with madwifi rate adaption algorithm Minstrel in the past, so I have replaced it with sample. I think that this should be the default. But that's just my opinion.

- also in madwifi there is the problem of running out of TX buffers, so changing it from 200 to 400 is supposed to fix the issue. I can't say for sure that the problem is gone for good but it might be a good idea to set it to 400 in any case.

I am writing this without my notes in front of me. If you need details/patches/etc. ask and I will provide them.

xaviour wrote:

A ticket has been created.

It seams that the root cause has been found. It is in "package/kernel/modules/block.mk" which is a file that you have been modifying nbd so you might know enough of the file syntax to solve the problem.

The discussion might have continued from here.