Forum › Forums › New users › New Users and General Questions › WLAN/Bluetooth Card Continual Disconnects Block Function.
- This topic has 32 replies, 6 voices, and was last updated Jan 10-12:27 am by stevesr0.
-
AuthorPosts
-
December 13, 2021 at 12:53 pm #72848
Anonymous
::It’s all too confusing for me, lots of prose, no command is in code-tags….
Please in code-tags:
inxi -SNxx
rfkill listIs the package isenkram-cli installed and running?
https://www.debian.org/releases/bullseye/amd64/ch06s04#completing-installed-systemDecember 13, 2021 at 6:47 pm #72871Memberstevesr0
::Hi Moddit,
I appreciate the comment about grey hair – that would be a decided improvement in my physical state <g>.
I do have a usb wireless adapter, which works OK. I have been working at dissecting this as a learning experience.The device is replaceable; the only time I have disassembled this system (to replace a fan), I found it to be a very tedious operation. So I have avoided doing that so far in favor of trying to solve it as a software (including firmware issue).
My only justification for posting this is (a) personal benefit of learning things from responses I get (vide infra – I hadn’t heard of the isenkram-cli package until male mentioned it in his comment) and (b) in case others have a similar problem.
Final comment: There is a howto debug and report problems with the iwlwifi firmware published by the developers, which might be of interest to others with possible iwlwifi problems. It is located at https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/debugging. (N.B. The site wireless.wiki.kernel.org covers ALL of linux wireless, not just iwlwifi.)
stevesr0
- This reply was modified 1 year, 4 months ago by stevesr0.
- This reply was modified 1 year, 4 months ago by stevesr0.
December 13, 2021 at 6:55 pm #72873Memberstevesr0
::Hi male,
Thanks for comments.
I installed isenkram-cli (thanks for the suggestion). It said no firmware was missing.
stevesr0
System: Host: stevesFujitsu Kernel: 5.15.0-2-amd64 x86_64 bits: 64 compiler: gcc v: 11.2.0 Desktop: Openbox 3.6.1 dm: startx Distro: antiX-19.1-sid_x64-net Marielle Franco 18 December 2019 base: Debian GNU/Linux bookworm/sid Network: Device-1: Marvell 88E8055 PCI-E Gigabit Ethernet vendor: Fujitsu Limited. driver: sky2 v: 1.30 port: 2000 bus-ID: 04:00.0 chip-ID: 11ab:4363 Device-2: Intel PRO/Wireless 4965 AG or AGN [Kedron] Network driver: iwl4965 v: in-tree: bus-ID: 0c:00.0 chip-ID: 8086:4229 rfkill ID TYPE DEVICE SOFT HARD 0 wlan phy0 unblocked blocked- This reply was modified 1 year, 4 months ago by stevesr0.
- This reply was modified 1 year, 4 months ago by stevesr0.
- This reply was modified 1 year, 4 months ago by stevesr0.
December 13, 2021 at 8:23 pm #72886MemberModdIt
::Hi steves, thanks for humerous answer. Intent too, here it is usualy fix asap so user happy.
Opening the case of your device would allow a substitution so you can see if the card is ok and first if the antenna is still
correctly clipped to the package studs. Then again you have persons of 10 grades higher than me helping so maybe need not.On opening for access. My EEPC is glued back together, the case clips are so flimsy, they break when you look
at them, I had the case pretty warm which usualy helps, no good effect.on my T series one screw opens a flap, access to wifi/BT card and modem Card in a minute or so.
Designed by experienced engineers who listened to service providers, not craphead fresh from university costcutters.December 13, 2021 at 8:59 pm #72887Anonymous
::If rfkill shows “hardblocked” for the Intel_Card, you have no chance.
What is this machine without a hardware switch?
inxi -MClues to a possible switch may give you
sudo dmesg | egrep 'radio|kill|switch'In the BIOS you had also already looked?
December 13, 2021 at 10:51 pm #72889Memberstevesr0
::Hi Moddit and Male,
1. The card connections certainly could be the issue. If I break down and open the box, I will check that <g>. The hard drive and battery are readily available, the wlan/BT card not so much (at least for me). Curiousity will eventually force me to open it, if I don’t find a magic software answer.
2. The machine HAS a hardware switch – otherwise it is unusable <g>.
3. The one software thing that the linux wireless site suggests for this card is to disable concurrent use of the bluetooth and wlan radios. I have tried without success so far. The wireless site (and several others) recommend placing “options bt_coex_active=0” in /etc/modprobe.d/iwlwifi.conf. I created that file and it didn’t do anything obvious. Reading further, I learned that the current state of the parameters for different modules could be found in /sys/module/iwlwifi/parameters/bt_coex_active. However, I don’t have such a file. A file /sys/module/iwlegacy/bt_coex_active exists. This file consists of a single character “Y”.
So, I created a file /etc/modprobe.d/iwlegacy.conf and placed the options bt_coex_active=0 in that file. Rebooting, I saw several items booting which stated ignore this line in this file (libkmod/libkmod-config.c:712 kmod_config_parse: ignoring bad line starting with ‘options’).There is a /sys/module/iwl4965/ directory which doesn’t have a bt_coex_active file. I created an /etc/modprobe.d/iwl4965.conf file but this was also ignored on the next boot.
I accidentally tried to create a file in /sys/module/ as sudo and was denied permission! (I suspect it is the extra attributes that can lock files and aren’t visible with ls -asl, but I have to stop for today.)
So, I would love to know how to make the system block concurrent operation of BT and WLAN on this card…
Thanks for comments.
stevesr0
December 14, 2021 at 8:55 am #72901MemberModdIt
::At the beginning of the thread you posted below switch on, no hard, maybe the switch contacts are corroded so you get malfunction, common issue.
Please do not use wd40, it will ruin the swtch. Contact spray which really works is acidic and needs a follow up with a fitting neutralising
lubricant partner spray. It removes the plating on contacts very quickly so must work fast or you get an non recoverable dead switch.UPDATE – RFKILL list results:
Physical switch OFF:
phy0: Soft Block NO, Hard Block YES
phy1: Soft Block NO, Hard Block NO
(No Bluetooth line)Physical switch ON: # All listed softblock
phy0: Soft Block YES, Hard Block NO
phy1: Soft Block YES, Hard Block NO
Bluetooth: Soft Block YES, Hard Block NOYour comment on cheap internal cards made me take a look in ebay, decent cards are crazy price in comparison to some months ago.
The USB dongles and PCI cards stable, for now. We are getting raging inflation in EU, all essentials, food, energy, clothes
prices up very fast. Look at company profits you see why. Btards…- This reply was modified 1 year, 4 months ago by ModdIt.
December 14, 2021 at 12:58 pm #72920Anonymous
::Since one must come first of all on it to offer helping users a rfkill output WITH locked
hardware switch… That makes me shake my head.lsmod
will, I assume, show you the modules iwlegacy and iwl4965.
Lock both modules with
echo "blacklist iwlegacy" | sudo tee /etc/modprobe.d/blacklist_iwlegacy.conf
and
echo "blacklist iwl4965" | sudo tee /etc/modprobe.d/blacklist_iwl4965.confrestart OR unload the modules
sudo modprobe -rfv iwlegacy iwl4965
===========================================
Load the module
sudo modprobe iwlegacyand test. If you don’t succeed, unload it again and do the same with the other module.
Remove the “good” module from the block list.
===========================================
regarding bluetooth you could try loading the modules with the corresponding options
You can query these options e.g. with
modinfo iwlegacy | grep parmso I can imagine to load the module right away with the right options
sudo modprobe -v iwlegacy bt_coex_active=1is that successful, then make it permanent in /etc/modprobe.d/iwlegacy_options.conf with the entry:
options iwlegacy bt_coex_active=1December 15, 2021 at 1:15 am #72960Memberstevesr0
::Thanks Moddit and male,
I will check out both sets of suggestions and report back after that.
I did break down and open the box. To my (inexperienced) eye, the wlan/BT card connections seem ok.
stevesr0
December 15, 2021 at 9:58 am #72977MemberModdIt
::While open, first check with a multimeter, is the switch ok, Hope you have one, mine has paid for itself multiple times. Any measurable resistance bridge or replace the switch. I also always replace BIOS battery and check for any signs of corrosion around it with 8x magnifier. Anything suspicious clean the area carefully. Danger depens very much on where you live, in a really dry area you are pretty safe, I lived in Phillipines, in the rainy season you can pretty much sit and watch circuit board corrosion progressing.
Anyway, wish you succesful fault finding.December 15, 2021 at 12:57 pm #72986MemberRobin
::in a really dry area you are pretty safe,
Merely a single addition to this sentence: This is true only under the aspect of corrosion. But in a dry area you should be much more carefully in concerns of static electricity discharge when handling your opened device. This is in moist environment much less dangerous to the circuitry (simply because the humidity conducts better than dry air, allowing the static charge to decrease in short period, so the chance to damage your device while coming near or touching a pcb track or a soldering eyelet of one of the high resistance CMOS ICs is decreased noticeably.)
Windows is like a submarine. Open a window and serious problems will start.
December 17, 2021 at 1:40 am #73095Memberstevesr0
::Hi Moddit, male and robin,
Thanks for comments.
I have just tried the software suggestions made by male. The problem I have is that when I turn on the physical switch, the malfunctioning card takes over and I have to shut down the machine via power button. Before turning the switch on, I was able to note that response to removing iwlegacy and iwl4965 and reloading them individually gave different results. I am going to have to repeat these tests and update this response, because if I recheck before sending this from this machine, the system will certainly lock up again. (N.B. While the terminal is frozen, the browser seems to be continuing to work – go figure?!)
So I will post this and update shortly.
SOFTWARE TESTING UPDATE: To confuse things, when I load iwlegacy, iwl4965 stays blocked. But when I load iwl4965, it also loads iwlegacy.
When I load iwlegacy, the system doesn’t lock up when I turn the physical switch on, but does when I load iwl4965.
So, this doesn’t suggest to me that a module conflict is the cause of the problem. (But it was interesting to try these commands and blacklisting the modules DID seem to stop them from autoloading at boot. So, thanks for the suggestion,male.)
I will get back to the hardware testing when I figure out how to safely use my multimeter <g>.
stevesr0
- This reply was modified 1 year, 4 months ago by stevesr0.
- This reply was modified 1 year, 4 months ago by stevesr0.
December 17, 2021 at 10:34 am #73108Anonymous
::Again, I probably only half understand what you’re doing. Why do you insist on a hardware failure?
The presence of the module iwlegacy was 90% guessed by me because I miss the output lsmod.I think we agree that loading both modules on this laptop is wrong. Hence my suggestion to blacklist both and then load and test them individually.
You don’t say a word about whether you had a successful wifi connection with iwlegacy. Why don’t you show rfkill list ?
Are all the “test files” you created earlier cleaned up cleanly?Before you throw the thing away, you might want to test the procedure with a live CD or a fresh install of antiX21 if no wifi connection is established there either.
My 2Cent.December 17, 2021 at 8:32 pm #73161Memberstevesr0
::Hi male,
I am not trying to be frustrating.
In answer to your questions.
(1) Why I insist that the failure is due to hardware? I can’t be sure, but mainly the fact that it doesn’t connect even using the original windows hard drive that the machine arrived with. I haven’t had a successful connection with this wireless/BT combo card since my opening post months ago.
(2) I don’t know how to separate iwlegacy from iwl4965, because when I load iwl4965, it loads iwlegacy. When I load iwlegacy alone, I don’t see messages in dmesg that indicates that the wlan function is started. When I load iwl4965, I (at least sometimes) see messages about the available wireless channels.)
(3) the rfkill command has gotten crazy. Sometimes it lists nothing, even with the wlan/BT hardware switch off. Sometimes it says wlan is hard and soft blocked when I haven’t blocked it via software command. Sometimes it stops working – this usually indicates that the terminal has stopped working and I have to do a hard shutdown (power button).
(4) In addition to the usual modules, I believe this card is supposed to use iwldvm, and that isn’t loaded by default. When I did load that today, an iwlwifi module also appears in lsmod. However, even with iwldvm, iwlwifi, iwlegacy, iwl4965, mac80211 and cfg80211 loaded, it doesn’t make things work.
I am perplexed and feel I need to set aside further evaluation of this problem until (a) I learn of a solution for a card with similar software and problems OR (b) I get to use a multimeter and discover an obvious problem with the card hardware or that the switch isn’t really turning on normally.
Thanks for all suggestions/assistance to date.
stevesr0
January 3, 2022 at 1:39 am #74508Memberstevesr0
::Update.
I have tried a second intel 4965 wlan/BT mini pcie card with identical issues.
Looking at dmesg, it seems that when the card is active, a continual stream of connects and disconnects occur with a usb device. While I haven’t figured out how to check this directly when running Windows, tonight when I booted with the physical switch on, the system stopped functioning after the welcome screen gave way to the desktop. I shut off view the power switch and then booted without a problem with the physical switch off.
I have not managed to shutoff individual modules through any software technique I have tried (blacklisting, removal of bluez).
In linux, the dmesg messages are
usb:1-2 device using uhci_hcd
new usb device found idvendor=0c24 idproduct=000f bcdDevice=19.15
device string Mfr=0 Product=0 Serial#=0
then a device # x disconnectedthen the sequence repeats incrementing the device number each time by 1.
Because this is happening in both the original Windows setup and in linux, I think there is a hardware problem, involving either cable or motherboard. Since the messages involve USB, I am considering getting a wlan only card if one is available. Other than that perhaps trying to replace the 3 cables connected to the card or taping over the pins used by the BT module (to see if this will enable the wlan module to work.)
Appreciate any suggestions from those who have had any similar problems.
(For the record, this is NOT my “work” machine – it is my toy to learn about the problems one can experience with a computer. However, I know others have had similar problems with such cards and so if there is some fix that emerges it might be useful to others.)
Thanks in advance.
stevesr0
-
AuthorPosts
- You must be logged in to reply to this topic.