BCM4312 and Puppy431 on Lenovo s10-2 [SOLVED]

[jrb]

pup-431-bcm4312v2.iso has been uploaded to http://puppylinux.ca/members/choicepup/ ... adcom4312/.

I burnt to CD, booted Lenovo s10-2 from external CD, used Bootflash install to make a bootable 2GB usb drive. Booted from that and used Puppy universal installer to do a frugal install to /mnt/sda5. All with no problems.

Tried the remaster script, still broken. Oh well, if anyone wants to know how to do a custom remaster for their machine without the remaster script let me know.

BTW: when I did the frugal install it still recommends "nosmp" for the kernel line in menu.1st. Don't do it! See my comments above.

[rjbrewer]

jrb:
re; version number

Certainly not the first time that's happened. Some of Ttuuxxxs'
more creative version numbers really had my head spinning.

Noticed while booting, the line with usb counts all the way to 10.
Doesn't seem to hurt anything, but on a few other linux sites,
( arch linux) they talk about possible effects of removing ssb.

[tempestuous]

ssb.blacklist=yes
...
I tried both ways with stock puppy431. ssb still loaded.

The thoroughness of your testing remains in question.
There's a chance that the kernel boot option did succeed in blocking the loading of the ssb module during initial bootup, but if you didn't have rerwin's modules-preference fix installed, then ssb may have loaded later in the boot sequence.
Also, the Network Wizard may be configured to load the ssb module.

So to be thorough, you need rerwin's fix installed, and you need to disable any Network Wizard config for ssb.

[rjbrewer]

jrb:

Burned v2 iso and did a full and a usb install using universal
installer. Both work fine.

Those Lenovo s10s' are extremely nice little machines;
Santa is being very good to your girls this year.

thanks
rjb

[jrb]

rjb,
I read a blog authored by a couple who bicycled from Prudhome Bay, Alaska to Tierra del Fuego, Arctic Ocean to Antarctic Ocean. Two years and they carried an s10 the whole way and gave it very good reviews.

Knowing my girls, I figured something that tough might hold up. Hoping for the best anyway.

Glad the .iso is working for you, J

tempestuous,
I believe I had rerwin's .pet installed (what did I forget and when did I forget it?) when I ran those tests. I'll try to get back and test again with usb boot but it will have to wait until all my other Christmas preps are done.

[n01fu]

I have tried this workaround on a HP 2140, and although it allows me to see the wireless as eth1, i have to manually go through the rmmod each time.

any suggestions?

[jrb]

I have tried this workaround on a HP 2140, and although it allows me to see the wireless as eth1, i have to manually go through the rmmod each time.

Could you give more detail? Are you booting from pup-431-bcm4312v2?
Is eth1 there before you do the manual rmmod? The more you tell us the better advice we can give.

Cheers, J

[jrb]

I suggest you could go one step further, and incorporate my modified ssb module into your initrd, and then into your ISO. If you do so, you must also replace the modules.alias file which has been accordingly modified to remove the defunct device ID's.

Done!

As you have suggested I placed your ssb.ko.gz and modules.alias in initrd.gz and made a new pup-431bcm4312v3.iso.

I booted the Lenovo s10-2 with the new initrd.gz and it worked quite well. No ssb conflict. Hopefully this will take care of Scottbg1's problem.

Much appreciated, J

This .iso has tempestuous' modified ssb.ko which will not conflict with the wl.ko module needed for broadcom4312 wifi cards. It does not contain the b43 or b43 legacy modules.
http://puppylinux.ca/members/choicepup/ ... 4312v3.iso

[mikeb]

tempestuous wrote:
I suggest you could go one step further, and incorporate my modified ssb module into your initrd, and then into your ISO. If you do so, you must also replace the modules.alias file which has been accordingly modified to remove the defunct device ID's.

Done!

The one that can co exist with wl as threatened?...where is this from just for reference

regards

mike

[tempestuous]

I modified the Puppy-2.6.30.5 kernel source so that the conflicting device ID's were disabled in the ssb module.

If you're interested, the relevant source files are found at /usr/src/linux-2.6.30.5/drivers/ssb/
and the exact source file to modify is b43_pci_bridge.c

[mikeb]

Ah ha...well I'm not the compile king (did make a working wl.ko though )

So basically one to be incorporated into newer puppy releases as from what I understood this approach keeps everything compatible.

Nice work all...lots of puppy users with broadcom devices

Hope holidays went well

mike

[rjbrewer]

Did an upgrade of the v2 full install with v3. Works fine and
long count of usb modules is gone. Very nice.

[jojonouvo]

Hi all, after too many hours with no results I'm asking your help. I have a broadcom pcmcia WPCB-104GB and puppy 431.
My kernel is 2.6.30.5
I can see ssb when I do lsmod but there's no wl.
The card seems to be recognised:

# pccardctl ident
Socket 0:
product info: "Broadcom", "802.11g PCMCIA", "8.0", ""
manfid: 0x02d0, 0x0425
function: 6 (network)
# pccardctl status
Socket 0:
3.3V 16-bit PC Card
Subdevice 0 (function 0) [unbound]

I don't know if that helps but I have this as well:

# lspci
00:00.0 Class 0600: 1002:cab0 (rev 13)
00:01.0 Class 0604: 1002:700f (rev 01)
00:02.0 Class 0c03: 10b9:5237 (rev 03)
00:06.0 Class 0401: 10b9:5451 (rev 02)
00:07.0 Class 0601: 10b9:1533
00:08.0 Class 0703: 10b9:5457
00:0a.0 Class 0607: 1217:6972
00:10.0 Class 0101: 10b9:5229 (rev c4)
00:11.0 Class 0680: 10b9:7101
00:12.0 Class 0200: 100b:0020
01:05.0 Class 0300: 1002:4336
# iwconfig
lo no wireless extensions.

eth0 no wireless extensions.

I tried everything I could read in the forums but never got to see anything but the eth0 in the connection manager.
The problem seems to be solved with jrb's iso but I don't want to reinstall my puppy again, what's left?
Thanx for your help.

[tempestuous]

The problem seems to be solved with jrb's iso

No, not necessarily.

# pccardctl ident
Socket 0:
product info: "Broadcom", "802.11g PCMCIA", "8.0", ""
manfid: 0x02d0, 0x0425

Wow. That device ID is hard to identify. Googling indicates it's a Freebox WPCB-152G, which is possibly a rebranded Belkin F5D7010. The device ID 02d0:0425 is not generally recognised, which might explain the problem. But it should be supported by the standard b43 driver already in Puppy.

1. First check check if the b43 module has automatically loaded

Code: Select all

lsmod

2a. if you don't see b43 listed, try loading it manually

Code: Select all

modprobe b43

2b. or if you see b43 listed, unload/reload it -

Code: Select all

rmmod b43
modprobe b43

3. now check if a second (wifi) interface has been created with -

Code: Select all

ifconfig -a

[jojonouvo]

Thank you tempestuous, it is a freebox card.
b43 module doesn't automatically load, but I can load it:

# modprobe b43
# lsmod
Module Size Used by
b43 127240 0
rfkill 10672 1 b43
mac80211 166056 1 b43
cfg80211 64972 2 b43,mac80211
led_class 4112 1 b43
input_polldev 3764 1 b43
usbhid 26112 0
wl 1275072 0
lib80211 6416 1 wl
...

Then I can see lo:

# ifconfig -a
eth0 Link encap:Ethernet HWaddr 00:C0:9F:13:2A:A5
inet addr:192.168.0.8 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:388 errors:0 dropped:0 overruns:0 frame:0
TX packets:381 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:237637 (232.0 KiB) TX bytes:64148 (62.6 KiB)
Interrupt:11 Base address:0xe000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)

And still nothing else than eth0 in the connection manager. I tried to load b43 from the manager but it didn't find the card. I also rmmod wl which brought no change and tried to rmod ssb but was told that:

# rmmod ssb
ERROR: Module ssb is in use by b43,ohci_hcd

Now here's a tough puppy challenge for me, what's missing?

[jojonouvo]

Any new hint in 2010?

[mikeb]

Erm..I think this discussion drifted onto another thread or the email notification never got sent

mike

[tempestuous]

# lsmod
Module Size Used by
b43 127240 0
...
wl 1275072 0

Both the b43 and wl modules are loaded, and this will obviously create a conflict. It might (?) be possible to unload the wl module, but then it's likely that the b43 needs to be unloaded/reloaded.

You could possibly set PREFLIST in BootManager to use the b43 in preference to wl, but the PREFLIST function only works properly after applying a certain hotfix ... and I'm not sure if jrb's version of Puppy includes this hotfix
... it's getting too confusing.

The best thing to do would be to go back to Puppy 4.3.1 and try to load the b43 module in an environment where the wl module is not installed.

[jojonouvo]

Thanx tempestuous, I'm using the original Puppy 4.3.1, not jrb's version.
I'm able to unload wl and then load b43 but what should I do next?
The wizard still doesn't see eth1..

By the way, tell me if you think I should buy a card that's known to work, it's cheap and maybe we wouldn't need to bother that much on broacom.

[tempestuous]

I'm able to unload wl and then load b43 but what should I do next?

Well let's double-check this: the b43 module can only load successfully once the wl module has been unloaded.
So if they're both loaded, then you must unload both, then reload the b43 module.

Then check if a new (wifi) interface has been created with

Code: Select all

ifconfig -a

If no new interface, then run this diagnostic command

Code: Select all

dmesg

and look for clues about how the b43 module loaded.

if you think I should buy a card that's known to work, it's cheap and maybe we wouldn't need to bother that much on broacom.

Yes, this would be the ideal solution. Broadcom wifi devices have an unhappy history with Linux, and Broadcom has still not fully released its source code (the proprietary driver includes a closed-source binary file).
The situation becomes worse when vendors like Dell and HP mess with the hardware specs.
And in your situation, we're dealing with an "oddball" model.

I'm not up to date with current wifi adaptors, but generally the Ralink-based miniPCI and Cardbus devices are well regarded.