Slacko 5.3.3-RC2 (5.3.2.4[0])

[peebee]

@ peebee... don't feel bad mate! Does this happen in Racy? Hasn't that got a 3X kernel?

Hi Mick

Yep - happens in Racy5.2.2 - was reported to Barry - that's why the firmware got unpacked.....so the only thing needed was the modprobe.

Has been the same for all k3 puppies except for the recent lupu528+k318 from wuxiandianzi (earlier ones had the problem).

Cheers
peebee

[01micko]

That is a known bug. Karl Godt posted the fix on racy thread.

for cpu scaling? I'll keep it in mind

[pemasu]

http://www.murga-linux.com/puppy//viewt ... 286#592286

[jim3630]

Thanks for explanation jim3630
Is that one of don570's "right click" things? Reason I ask is that I can't see how that would work without a wrapper script.

mick maybe it is one of don's things. don't recall where got it and without a wrapper script is why don't work you'd know better than me.

your gui script works thanks.

[mavrothal]

Hi Mick

Pristine manual frugal install of 5.3.1.3 on my HP550 laptop.

Don't know if you were expecting any change in behaviour - but B43 still not detected and loaded automatically - all fine once loaded manually.

Try to see if adding this line in the /etc/rc.d/rc.sysinit immediately after:
"echo "LOAD KERNEL MODULES"
echo -n "Loading kernel modules..." >/dev/console"
Will help auto detecting the driver.

Code: Select all

udevadm trigger --action=add --subsystem-match="net"

If this fails try this patch

Code: Select all

--- a/etc/rc.d/rc.sysinit	2011-09-21 23:53:37.000000000 +0300
+++ b/etc/rc.d/rc.sysinit	2012-01-23 12:04:33.232216145 +0200
@@ -271,6 +271,27 @@
 echo "LOAD KERNEL MODULES"
 echo -n "Loading kernel modules..." >/dev/console
 
+#my intention is for puppy to work with either of these...
+if [ -f /sbin/udevd ];then
+ #100611 UDEV_LOG=2 to prevent non-critical o/p to screen at bootup and shutdown...
+ UDEVVER=`udevd --version`
+ if [ $UDEVVER -gt 150 ];then
+  #100611 v151 is recommended for kernel 2.6.27+.
+  #110502 change 'never' to 'early', fixes device nodes created with correct owner:group...
+  UDEV_LOG=2 /sbin/udevd --daemon --resolve-names=early
+ else
+  UDEV_LOG=2 /sbin/udevd --daemon
+ fi
+ udevadm trigger
+ udevadm settle  #  May take 2 minutes during the "Loading Modules" step !!!  
+else
+ #note, this script is no longer in Puppy.
+ pup_event_backend_d >/tmp/pup_event_backend_errors 2>&1 & #hotplug daemon. my homebrew replacement for udevd.
+fi
+sleep 0.1
+
+
 MODALIASES="`ls /sys/bus/*/devices/*/modalias`"
 
 #101119 new /sbin/pup_event_backend_modprobe, these must be deleted every boot...
@@ -304,22 +325,6 @@
  #...perhaps a udev rule could have done this?
 fi
 
-#my intention is for puppy to work with either of these...
-if [ -f /sbin/udevd ];then
- #100611 UDEV_LOG=2 to prevent non-critical o/p to screen at bootup and shutdown...
- UDEVVER=`udevd --version`
- if [ $UDEVVER -gt 150 ];then
-  #100611 v151 is recommended for kernel 2.6.27+.
-  #110502 change 'never' to 'early', fixes device nodes created with correct owner:group...
-  UDEV_LOG=2 /sbin/udevd --daemon --resolve-names=early
- else
-  UDEV_LOG=2 /sbin/udevd --daemon
- fi
-else
- #note, this script is no longer in Puppy.
- pup_event_backend_d >/tmp/pup_event_backend_errors 2>&1 & #hotplug daemon. my homebrew replacement for udevd.
-fi
-sleep 0.1
 
 #replay uevents from /sys...
 for ONEMODALIAS in $MODALIASES

PS Also could you post the output of the following command, before and after you manually load broadcom driver.

Code: Select all

for x in $(udevadm trigger --dry-run --verbose); do udevadm info --query=all --path=$x; done

[drblock2]

Slacko 5.3.3-alpha3 refuses to boot on my AMD Athlon (2GHz). It just hangs after announcing initrd.

Alpha2 booted, but gave me blank menus.

The md5sums were fine. I suspect that Athlon does not like the 3.x kernel. The latest Racy 5.2.2 with the 3.0.7 kernel also had strange problems which prompted me to give up on it.

Good that 01micko is rolling back to 2.6.39.4 for the MAIN version.

[mavrothal]

Mick,
I noticed you have /etc/rc.d/rc.udev in the build but it is not called from anywere or ever run. Actually it will not run even manually since /usr/{bin,sbin} is missing from its path.

Fixing the path and calling it from rc.sysinit with

Code: Select all

--- a/etc/rc.d/rc.sysinit	2011-09-21 23:53:37.000000000 +0300
+++ b/etc/rc.d/rc.sysinit	2012-01-23 10:00:03.000000000 +0200
@@ -114,6 +114,10 @@
 #if have just done a switch_root, output a 'done' message...
 [ -d /initrd ] && status_func 0 #note, /initrd does not exist when a full-hd installation.
 
+[ ! -d /dev/.udev ] && mkdir /dev/.udev
+/etc/rc.d/rc.udev start
+
 . /etc/rc.d/MODULESCONFIG #modules loading configuration.
 PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/X11R7/bin
 [ $loglevel ] && LOGLEVEL=$loglevel #boot param.

runs fine and does not give me any problems.
Does not solve any either, but that's because I do not have any device issue with 5.3.1.3

I'm not entirely sure if this is the best place to start it but I can not test this without a devise problem

[peebee]

Try to see if adding this line in the /etc/rc.d/rc.sysinit immediately after:
"echo "LOAD KERNEL MODULES"
echo -n "Loading kernel modules..." >/dev/console"
Will help auto detecting the driver.

Code: Select all

udevadm trigger --action=add --subsystem-match="net"

If this fails try this patch

Hi mavrothal

Thanks for your interest...

Applied your first patch to a pristine pfix=ram version of 5.3.13, rebooted - sad to say the b43 wifi was not detected

I'm afraid the 2nd patch is beyond my level of competency to apply ... can you give me some clues on what I should do to apply it - thanks.

Cheers
peebee

[mavrothal]

Applied your first patch to a pristine pfix=ram version of 5.3.13, rebooted - sad to say the b43 wifi was not detected

I'm afraid the 2nd patch is beyond my level of competency to apply ... can you give me some clues on what I should do to apply it - thanks.

Don't wary you did not miss anything...

Turns out that my home iMac has a Broadcom wifi module
So I tried few things for myself. with similar results.

Ended up realizing that the Broadcom firmware is missing from the build (the folder is there but empty) and udev can not handle it.

Even if I manually load the module ifconfig still fails to bring wlan0 up.
Is it suppose to be added from a pet or something? (sorry haven't followed the Broadcom story closely).

Please point to a proper 5.3.1.3 version so I can take another stub at it (assuming I'm allowed access to the machine ).

[don570]

Thanks for explanation jim3630
Is that one of don570's "right click" things? Reason I ask is that I can't see how that would work without a wrapper script.

Sounds like ghasher or gtkhash wasn't installed properly.
Barry's blog mentioned a bug that he has
fixed involving links in the 'SendTo' folder. Ttuuxx's gtkhash
package unfortunately used the 'SendTo' folder for links and this
was causing the strange behaviour.

Check to see if you have put any links in 'SendTo' folder.
Transfer them to 'OpenWith' folder.

I suggest installing a compatible version again and then
installing my right click package again. There's no harm
done installing it a second time.

But if that doesn't work better start with a fresh installation again.

http://www.murga-linux.com/puppy/viewto ... 035#586035

____________________________________________

[peebee]

Ended up realizing that the Broadcom firmware is missing from the build (the folder is there but empty) and udev can not handle it.

Even if I manually load the module ifconfig still fails to bring wlan0 up.
Is it suppose to be added from a pet or something? (sorry haven't followed the Broadcom story closely).

Please point to a proper 5.3.1.3 version so I can take another stub at it (assuming I'm allowed access to the machine ).

Hi mavrothal

The firmware is there - it's now in /lib/firmware/b43 - Barry took the decision to stop putting compressed firmware into /lib/modules/all-firmware and instead ship with all the firmware already uncompressed into /lib/firmware.

If udev is still expecting the compressed firmware then that would need to be changed OR Barry would have to revert to the previous method in woof.....

At least that's how I understand matters - I could easily be wrong

Cheers
peebee

[rodin.s]

Slacko 5.3.3-alpha3 refuses to boot on my AMD Athlon (2GHz). It just hangs after announcing initrd.

I have the same problem. Alfa-2 boots and works OK. (Intel Celeron 900MHz).

[mavrothal]

The firmware is there - it's now in /lib/firmware/b43 - Barry took the decision to stop putting compressed firmware into /lib/modules/all-firmware and instead ship with all the firmware already uncompressed into /lib/firmware.

Hmm..., b43 is the only common firmware between /lib/firmware and /lib/modules/all-firmware. How come?
Is the /lib/firmware version for this kernel? and if it is does the kernel "knows" it? (Mick?)

In any case, this firmware is not working in my iMac.
"ls -l /dev/.udev/firmware-missing/" says

Code: Select all

lrwxrwxrwx 1 root root 68 2012-01-23 19:53 b43-open\x2fucode16_mimo.fw -> /devices/pci0000:00/0000:00:15.0/0000:04:00.0/ssb0:0/firmware/ssb0:0
lrwxrwxrwx 1 root root 68 2012-01-23 19:53 b43\x2fn0initvals16.fw -> /devices/pci0000:00/0000:00:15.0/0000:04:00.0/ssb0:0/firmware/ssb0:0

and indeed initvals16.fw is not present in /lib/firmware/b43. Though ucode16_mimo.fw, is.

peebee, does your wifi works after you modprobe b43?
Is there anything in /dev/.udev/firmware-missing/ after you modprobe b43?
What about the output of the "for x in $(udevadm trigger --dry-run --verbose); do udevadm info --query=all --path=$x; done" command mentioned before?

[peebee]

peebee, does your wifi works after you modprobe b43?
Is there anything in /dev/.udev/firmware-missing/ after you modprobe b43?
What about the output of the "for x in $(udevadm trigger --dry-run --verbose); do udevadm info --query=all --path=$x; done" command mentioned before?

Hi mavrothal

My b43 is version 4311 and uses the "13" series of the firmware.

It works perfectly after the modprobe b43 is done and dmesg shows that the firmware is loaded.

I believe(?) the firmware is not kernel dependent and is the same in all puppies - it is firmware that is loaded into the b43 so no reason why it should be kernel dependent?

I have no /dev/.udev/firmware-missing file or directory

Sorry missed that request - I've tried to run the command but the output is so verbose it overflows the uxrvt buffer - how can I capture the output to a file???

Cheers
peebee

[mavrothal]

I believe(?) the firmware is not kernel dependent and is the same in all puppies - it is firmware that is loaded into the b43 so no reason why it should be kernel dependent?

I believe th slacko driver is using the V4 firmware.
Boy, Broadcom is a pain! As I understand it you must extract the linux firmware from your actual chip or the broadcom proprietary driver!

Sorry missed that request - I've tried to run the command but the output is so verbose it overflows the uxrvt buffer - how can I capture the output to a file???

Code: Select all

for x in $(udevadm trigger --dry-run --verbose); do udevadm info --query=all --path=$x; done > /root/udef_info.before     # or .after

will generate the udef_info.before (or .after) file in home.

[pemasu]

I checked the wcpufreq in LHP 64 which uses 3.0.4 kernel. Tazoc has fixed the script simply with "||" the two location choices of drivers. Here is fixed wcpufreq pet. There is still one bug in script. The script does not rule out non needed drivers where there is choosable dropdown box for performance, powersave...etc.

[smokey01]

Whoops wrong area.

[pemasu]

Smokey01. Thanks for info, but...maybe wrong thread. I copy your post and repost it to the Exprimo thread.

[01micko]

The firmware is there - it's now in /lib/firmware/b43 - Barry took the decision to stop putting compressed firmware into /lib/modules/all-firmware and instead ship with all the firmware already uncompressed into /lib/firmware.

Hmm..., b43 is the only common firmware between /lib/firmware and /lib/modules/all-firmware. How come?
Is the /lib/firmware version for this kernel? and if it is does the kernel "knows" it? (Mick?)

In any case, this firmware is not working in my iMac.
"ls -l /dev/.udev/firmware-missing/" says

Code: Select all

lrwxrwxrwx 1 root root 68 2012-01-23 19:53 b43-open\x2fucode16_mimo.fw -> /devices/pci0000:00/0000:00:15.0/0000:04:00.0/ssb0:0/firmware/ssb0:0
lrwxrwxrwx 1 root root 68 2012-01-23 19:53 b43\x2fn0initvals16.fw -> /devices/pci0000:00/0000:00:15.0/0000:04:00.0/ssb0:0/firmware/ssb0:0

and indeed initvals16.fw is not present in /lib/firmware/b43. Though ucode16_mimo.fw, is.

peebee, does your wifi works after you modprobe b43?
Is there anything in /dev/.udev/firmware-missing/ after you modprobe b43?
What about the output of the "for x in $(udevadm trigger --dry-run --verbose); do udevadm info --query=all --path=$x; done" command mentioned before?

The Broadcom firmware is proprietary and therefore independent of kernel version. I'd say mavrothal that your issue is slightly different to peebee's with the ucode16_mimo.fw not loading. Perhaps too that the firmware is expected to be compressed???

The initvals16.fw looks like some weird mac dependency for b43, and I can't find it! You may be able to find it and extract it on your OSX system. If you get that to load then maybe ucode16_mimo.fw will load with it?

I have attached the compressed firmware which works with wuxiandianzi's lupu variant with k3.1.8 .. includes b43.tar.gz and b43legacy.tar.gz

[mavrothal]

The Broadcom firmware is proprietary and therefore independent of kernel version. I'd say mavrothal that your issue is slightly different to peebee's with the ucode16_mimo.fw not loading. Perhaps too that the firmware is expected to be compressed???

The initvals16.fw looks like some weird mac dependency for b43, and I can't find it! You may be able to find it and extract it on your OSX system. If you get that to load then maybe ucode16_mimo.fw will load with it?

I have attached the compressed firmware which works with wuxiandianzi's lupu variant with k3.1.8 .. includes b43.tar.gz and b43legacy.tar.gz

Did I say this is a pain?...
The pet does not work compressed or extracted.
I compiled b43-fwcutter and extracted the firmware from broadcom-wl-5.100.138 and now is telling me the firmware is too new and ifconfig still gives me "ifconfig: SIOCSIFFLAGS: Operation not supported"
I think I'll stay with puppies running in the VM in the Mac (at least for today )