Page 17 of 19

5-3.3.3

Posted: Sun 06 May 2012, 02:09
by sszindian
Have wlan0 working... did NOT need 'the patch' mentioned above. Appears to be holding up fine... posting from it now.

eth0 appears to connect on boot but then the little pemasu-logo pop's up on the desktop and says... 'Releasing eth0' and by gollie... it sure does!!!
cannot get back on with eth0 period. Happens more often than not as I boot more and more.

Noticed.... the regular Network Manager is 'different' than the one in 5-3.3.2??? anyway, somethings a-miss in eth0 for sure!

>>>---Indian------>

Re: 5-3.3.3

Posted: Sun 06 May 2012, 02:41
by James C
sszindian wrote: eth0 appears to connect on boot but then the little pemasu-logo pop's up on the desktop and says... 'Releasing eth0' and by gollie... it sure does!!!
cannot get back on with eth0 period. Happens more often than not as I boot more and more.
Same thing here except it's every reboot.Have tried Frisbee,SNS and Network Wizard..........can get a connection but on reboot it is released and must start over.
Started with new fresh installs several times..... no difference.
Testing is fun..... :lol:

Posted: Sun 06 May 2012, 06:05
by pemasu
Yeah....I think it is SNS and Frisbee dhcpcd conflicting. I have compiled dhcpcd in Upup Precise for SNS and tried to isolate those 2 dhcpcd`s once again.
For me....both works now....if you delete the profile for wireless created by other one.

I normally dont use SNS but now as test purpose....I have used it....with eth0 and wlan0.

Looks like to work for me....

I just would need time to build this test version with latest woof. But I just cant promise anything now...

Posted: Mon 07 May 2012, 06:02
by linux28
:shock: Upup Precise 5.3.3.4?

Posted: Wed 09 May 2012, 15:20
by irishrm
Frugal USB install on acer 5750 i5 4g ram.

I must thank pemasu for including the wl drivers.
Frisbee had no problem connecting me to the internet.

I couldn't make any sense of the PPM so all my apps are downloaded from the puppy lucid 528 repo and are working well.

Precise is a welcome companion to fatdog 64 which is the only other puppy which will give me a wireless connection on this machine.

Comparisons:
Precise is giving me access to all the apps I need.
Fatdog doesn't have 64bit versions of some of these.

Fatdog appears to have better video playback ability.

Thanks again for another great puppy-irishrm.

upup precise

Posted: Wed 09 May 2012, 16:14
by Pelo
Hello Everybody. Could'nt we speak french here ?
I wil try it ASAP (as soon as possible - aviation language -)
I worried about having bought a computer on which i could'nt use puppies.In fact no, all the recent ones are fittable and i can see them on wide screen, and with the sound, naturally : Slacko was the first one, then Racy, then Dpup exprimo, then Hardy Husky, and now Upup. Où vais-je donner de la tête ?

My poor old ACER Aspire 3000 will not use Upup, because Acer is driven by an awful pilot : the B52, no.. B43

Download if very slow : 45 minutes. I stopped it. I shall see to-night.

Au revoir. Pélo

Posted: Tue 15 May 2012, 10:09
by linux28
xim bug:
Lucid Puppy, slacko
Image
Precise Puppy
Image

Posted: Tue 15 May 2012, 15:27
by linux28
xim
bug is the lack of those things?The Moren instructions to those who want to install?For help

xim bug

Posted: Wed 16 May 2012, 06:09
by shinobar
linux28 wrote:xim
bug is the lack of those things?The Moren instructions to those who want to install?For help
Try on a virtual terminal:

Code: Select all

# gtk-query-immodules-2.0 > /usr/lib/gtk-2.0/2.10.0/gtk.immodules

Re: xim bug

Posted: Wed 16 May 2012, 06:15
by linux28
shinobar wrote:
linux28 wrote:xim
bug is the lack of those things?The Moren instructions to those who want to install?For help
Try on a virtual terminal:

Code: Select all

# gtk-query-immodules-2.0 > /usr/lib/gtk-2.0/2.10.0/gtk.immodules
Upup Precise
No, still not work.You know use Upup the Precise, open a text editor, without these input methods:

Posted: Wed 16 May 2012, 06:17
by linux28
Use xim
But does not take

Re: Expanded Broadcom wifi support - continued

Posted: Thu 17 May 2012, 02:06
by rerwin
peebee, zekebaby, pemasu, et al,
After a week essentially away from puppy and then some false starts and dead ends, I now have a more solid Broadcom wifi experimental package. In addition to attempting to resolve the issues I described last time,
http://murga-linux.com/puppy/viewtopic. ... 890#624890,
I think I fixed the dual loading of wl and b43 while retaining support for the b44 ethernet driver (which uses the same ssb-bus arrangement that b43 does).

No matter the order these modules are originally loaded, they get reloaded in the order, wl, bcma, ssb, b43/b44/brcmsmac. (lsmod lists them in reverse order.) I simulated the loading of wl, bcma and ssb on my PC with wifi that uses the b43 driver only. So, I need those of you with other Broadcom wifi setups to try out the package. It should work as well as the version in lupuplus but also supporting the the newer brcmsmac-driver devices. Even with the drivers being loaded in the right sequence, that might not translate into the correct selection of drivers, although I expect that they should "just work". Your tests will verify that.

However, if there are multiple Broadcom wifi devices installed but using different drivers, if some use the wl driver and others b43, the latter might not be detected. Blacklisting wl would avoid that conflict. There should be no problem with only one such device installed.

Note that during installation of the package, a popup appears that is intended for peebee's proprietary-driver packages, which are not included in this experimental package. The drivers already in precise pup are sufficient, for now.
Richard

Re: Expanded Broadcom wifi support - continued

Posted: Thu 17 May 2012, 08:02
by peebee
rerwin wrote:peebee, zekebaby, pemasu, et al,
Richard
Hi Richard

Welcome back - hope it was a break for pleasure....

Pristine frugal install of Precise 5.3.3.3 - then added your experiment 3 pet and rebooted creating a savefile.

lsmod showed that both b43 and wl were loaded after the reboot but dmesg and netwiz showed that wl was actually being used - created a wifi connection successfully using Frisbee.

Rebooted - this time only wl was present and wifi was persistent over reboot.

Blacklisted wl and rebooted - b43 was then being used as expected - again Frisbee maintained the wifi connection without any intervention.

I'll PM you 2 pmodemdiag files after the 1st and 2nd boot.

Cheers
peebee

Posted: Thu 17 May 2012, 13:59
by rerwin
peebee,
Thanks for running the experiment. Your diag files reveal 2 interesting things. The first shows that my logic for determining whether b43 got loaded via ssb or directly from a modalias (therefore, whether to unload it) is incorrect. Not only is the test backward, it still would not detect the difference. I need another way to decide, which I think I know.

The second file shows that ssb is being blacklisted, which accounts for neither ssb nor b43 being loaded. My current logic does not create the blacklist-ssb.conf file, so it must be created elsewhere -- in one of the network tools or a previous broadcom-wifi implementation? A pfile search for it finds no clues in my precise-pup PC. Does anyone know how blacklist-ssb.conf is created?

EDIT: I see now that blacklist-ssb.conf is also in the first diag file. My logic does not consider blacklisting when it reloads ssb, which would explain why it is ignored in the first file, but not why it is respected in the second. (Maybe the blacklist conf file gets added after my logic runs the first time.) My struggle has been to understand all of the factors surrounding respect for blacklisting and module "install" commands. That is the tricky part of this experiment.
Richard

Posted: Thu 17 May 2012, 17:12
by jim3630
Richard, et al.,

don't know if this is useful to your 3rd experiment. when the new broadcom drivers first came out pemasu included brcmsmac with it's firmware and wl and firmware in one of his dpups. brcmsmac showed on wan0 and wl on eth0. i could connect using either by blacklisting the other. blacklisting neither brcmsmac was auto-detected on wan0 and worked without issue. i do not have b43 usage with this BCM4313 802.11b/g/n Wireless LAN Controller [14e4:4727] (rev 01).
will do 3rd if will help.

Posted: Thu 17 May 2012, 23:14
by don570
I tested audacity 2.0 with Precise

http://murga-linux.com/puppy/viewtopic.php?t=78257

____________________________________________________

Posted: Fri 18 May 2012, 03:07
by rerwin
jim3630 wrote:i could connect using either by blacklisting the other. blacklisting neither brcmsmac was auto-detected on wan0 and worked without issue. i do not have b43 usage with this BCM4313 802.11b/g/n Wireless LAN Controller [14e4:4727] (rev 01).
will do 3rd if will help.
Thank you for your offer to help. Yes, your testing of your BCM4313 would be most helpful. Since it does not use b43, my flawed b43 logic should not affect you. Please run without- and with wl blacklisted and get a pmodemdiag file for each case. You should not have to blacklist brcmsmac nor bcma to use wl.
Richard

Posted: Fri 18 May 2012, 09:52
by linux28
xim bug: :shock: :shock:
Still can not solve ah, help

Posted: Fri 18 May 2012, 14:18
by jim3630
here is experiment 3 results

pfix=ram installed experiment 3 pet and rebooted creating a savefile.

wifi connected with wl on wan0 via frisbee without issue.

Blacklisted wl and rebooted this time wifi auto-connected with brcmsmac on wan0 via fribee without issue.

Richard pm'd you 2 pmodemdiag files hope this helps.

Posted: Sat 19 May 2012, 15:51
by rerwin
jim3630,
Thanks for the diag files. They confirm that the package works as intended. The effective piece is the preference entry for wl over bcma.

To complete the verification, when you get the time, could you run with the package but use the SNS and Network wizard, in both cases rebooting a couple of times to ensure that network settings are retained as expected? No diag files are needed unless you find a problem. Thanks.
Richard