rfkill soft blocking after hardware switch unblock - all of sudden

There are similar threads, but i would not like bumping them - this didn't happen to my setup just a while back.

problem:
wifi does not work

symptoms:
when trying to start the network in /etc/rc.d the script rc.inet1 complains the RFKILL is blocking it

cause:
I accidentally flipped the RFKILL switch and suddenly all hell broke loose. Despite UN-fliping it, and rfkill ceasing the hardblock, it stubbornly raises the softblock that remains after.

testing:
One can follow the rfkill status with:

# watch -n .5 rfkill list

and operate the rfkill hardware toggle.

solution:
file:/etc/modprobe.d/rfkill.conf

options rfkill master_switch_mode=1
options rfkill default_state=0

the behavior can be further changed, but i was contended with just "restore" so hence the "1"

FWIW

I wonder why this thread keeps popping out as not read to me?

Originally Posted by SCerovec I wonder why this thread keeps popping out as not read to me?

Until it is replied to that appears to be how things work

I'm still getting an soft rfkill on occasion.

Can't see in the logs what's causing it.

So far I found out i can detect an event passing over udev whan I enable "debug" in /etc/udev/udev.conf.

It was not TLP's fault.

Stay tuned...

Can you run these commands and provide the output?

rfkill list all

rfkill unblock all

Some weeks ago I installed a non Slackware distro on a Panasonic Toughbook CF-52. The BIOS ACPI table does not support Linux and there never was a BIOS update. Uninteresting to your problem description except for my cure. I had to perform a specific dance to enable the wireless. Something like this:

rmmod athk52
rmmod athk
rfkill unblock all
Flip the wireless hardware switch off.
Wait about 10 seconds.
Flip the wireless hardware switch on.
modprobe athk52

I don't know if a similar dance for your lappie will help.

Originally Posted by jostber Can you run these commands and provide the output? rfkill list all rfkill unblock all

# rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: no
5: hci0: Bluetooth
	Soft blocked: no
	Hard blocked: no
# rfkill unblock all
bash-4.3# rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: no
5: hci0: Bluetooth
	Soft blocked: no
	Hard blocked: no

the observed behavior is correct so far
my problem is that *something* injects the soft block for no obvious reason
and my grievance is that neither wpa_supplicant nor rc.inet1 handle the undue softblock

Insult to injury, NetworkManager handles it in a breeze, but alas, i resorted to wpa_supplicant as the only one being able to connect to an weird issue on my personal wlan.

So here I am seeking the culprit.

I had some tiny progress lately, too early to report, stay tuned...

@jostber

bash-4.3# rfkill list all <flipped the rfkill switch to OFF>
0: phy0: Wireless LAN
	Soft blocked: yes
	Hard blocked: yes
5: (null): Bluetooth
	Soft blocked: yes
	Hard blocked: no
bash-4.3# rfkill list all <still the rfkill switch is OFF>
0: phy0: Wireless LAN
	Soft blocked: yes
	Hard blocked: yes
bash-4.3# rfkill unblock all <still the rfkill switch is OFF>
bash-4.3# rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: yes
bash-4.3# rfkill list all <flipped the rfkill switch to ON>
0: phy0: Wireless LAN
	Soft blocked: yes < gremlins! >
	Hard blocked: no
6: hci0: Bluetooth
	Soft blocked: yes < gremlins! >
	Hard blocked: no
bash-4.3# rfkill unblock all
bash-4.3# rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no <still the rfkill switch is ON>
	Hard blocked: no
6: hci0: Bluetooth
	Soft blocked: no
	Hard blocked: no

and occasional "injection" of soft block during normal operation

okay pulled out the heavy tools:
fil: /etc/udev/rules.d/50-rfkill.rules

#####
#
# preventing soft block
#
####

KERNEL=="rfkill", RUN+="/usr/local/sbin/rfkillhandler"

the handler:
file: /usr/local/sbin/rfkillhandler

#!/bin/bash

TIME=$(/bin/date +%H:%M:%S)
STATUS="skipped"
LOGFILE=/var/log/rfkillhandler
touch $LOGFILE
PARENT=$(ps $PPID)
declare -a IDX
declare -a SOFT
declare -a HARD
IFS=$'\t\n'
S=$(rfkill list all)
IFS=$' \t\n'
c=-1

LOGGER() {
    echo $TIME "[rfkillhandler]: action:["$ACTION"<"$PARENT"] - " $STATUS >> $LOGFILE
}
if [ $ACTION != 'change' ]; then
    STATUS="ignored"
    LOGGER
    exit 0
fi

# to skip splitting over space
IFS=$'\t\n'
for a in $S
do
  if [ "${a:1:1}" == ":" ]; then
    c=$(( $c + 1 ))
    #echo "index:$c"
    #the index
    IDX[$c]=${a:0:1}
  else
    # the states
    #echo "testing [${a:0:12}]"
    if [ "${a:0:12}" == "Soft blocked" ]; then
      x=${a:13}
      #echo "softblock:$x"
      SOFT[c]=$x
    fi
    if [ "${a:0:12}" == "Hard blocked" ]; then
      x=${a:13}
      #echo "hardblock:$x"
      HARD[c]=$x
    fi
  fi
done
# restoring
IFS=$' \t\n'

STATUS="parsed [$r]"

for (( r=0 ; r <= $c ; r++))
do
  #echo "hardblock: [${HARD[$r]}]"
  if [ "${HARD[$r]}" == " no" ]; then
    if [ "${SOFT[r]}" == " yes" ]; then
      rfkill unblock ${IDX[r]}
      STATUS="unblocked:["${IDX[r]}"]"
    else
    STATUS="skipped:["${IDX[r]}"]"
    fi
  else
  echo "hardblocked : yes"
  STATUS="hardblocked:["${IDX[r]}"]"
  fi
LOGGER
done

STATUS="checked [$r]"

LOGGER

exit 0

it remains to be seen if it will remedy the soft lock (works so far)

It is interesting to see the different approaches to solving the softblock issue. SCerovec's udev / handler solution was way more elaborate than what I did: I modified rc.inet1, adding the /usr/sbin/rfkill unblock all, and have never had a pesky soft block since--would be nice to figure out what the cause of these soft blocks are though...

#! /bin/sh
# /etc/rc.d/rc.inet1
# This script is used to bring up the various network interfaces.
#
# @(#)/etc/rc.d/rc.inet1 10.2  Sun Jul 24 12:45:56 PDT 2005  (pjv)

##############################
# UNBLOCK NETWORK INTERFACES #
##############################
/usr/sbin/rfkill unblock all

############################
# READ NETWORK CONFIG FILE #
############################

# Get the configuration information from /etc/rc.d/rc.inet1.conf:
. /etc/rc.d/rc.inet1.conf

I must admit i still have an icon bound on the launcher that issues the rfkill unblock all, and it happens that i still have to resort to it - less often than before...

that rfkill belongs there, wish PV would put it there one day

FWIW:
It kicks regularly in about 7-8 minutes after switching to battery operation.

Despite all the measures taken.