You are not logged in.

#1 2022-01-12 05:03:51

lucem
Member
Registered: 2022-01-12
Posts: 5

[SOLVED] No internet access but netctl-auto is running

I'm running into a problem that I'm not quite sure how to debug so any help would be appreciated.

I have been using wifi-menu to first generate the network profiles and then using the netctl service to connect on start up by starting and enabling the profile. I recently tried using netctl-auto instead and it's been super buggy. It has no problems connecting to the profiles that have been generated already and I can open any website with no problems, but sometimes I will just lose internet access (i.e. when I try to open a website in firefox or ping google.com it fails). I thought this might be related to the resuming from suspend problem described on the wiki, and I have enabled netctl-auto-resume as well, but it doesn't seem to work.

When this happens, I check systemctl, it still shows netctl-auto as active and running and there wasn't anything in journalctl that seems to be related to the network either. ip link also shows my network interface as still up. This doesn’t happen on any of my other decides or on the Windows boot, so I definitely it’s something related to arch. This issue can be fixed if I do sudo systemctl restart netctl-auto@interface.service. However, this is quite annoying and I was wondering if there are any ways to debug and fix this. Thanks.

Last edited by lucem (2022-01-23 02:23:31)

Offline

#2 2022-01-12 07:31:56

Strike0
Member
From: Germany
Registered: 2011-09-05
Posts: 1,333

Re: [SOLVED] No internet access but netctl-auto is running

Perhaps the netctl resume works, but obtaining an IP via the dhcp service times out while the link is brought back up. Which dhcp service do you use, and what's its journalctl status when this happens?
If it is dhcpcd, tries this manually when the problem occurs: https://wiki.archlinux.org/title/Dhcpcd … g_IP_first

Offline

#3 2022-01-12 07:59:13

seth
Member
Registered: 2012-09-03
Posts: 26,570

Re: [SOLVED] No internet access but netctl-auto is running

I thought this might be related to the resuming from suspend … on the Windows boot

3rd link below. Mandatory.

Also that the NIC is up doesn't mean all that much - you can still have lost the lease.
Please post a complete system journal for an affected boot (eg. "sudo journalctl -b -1" for the previous one, please don't copy out of the pager but redirect the output into a file or eg. ix.io, see the 1st link below)

Offline

#4 2022-01-13 01:38:51

lucem
Member
Registered: 2022-01-12
Posts: 5

Re: [SOLVED] No internet access but netctl-auto is running

Hi thanks for replying! I just ran into the issue again. For context, this happened when I'm just using firefox, terminal, and vscode. I tried the things that you suggested me to do:

seth wrote:

3rd link below. Mandatory.

Yes I have disabled the windows fast boot when I first installed arch.

seth wrote:

Please post a complete system journal for an affected boot (eg. "sudo journalctl -b -1" for the previous one, please don't copy out of the pager but redirect the output into a file or eg. ix.io, see the 1st link below)

here is the link to my complete journal with the command you told me to use: http://ix.io/3M2a

Strike0 wrote:

If it is dhcpcd, tries this manually when the problem occurs: https://wiki.archlinux.org/title/Dhcpcd … g_IP_first

I am indeed using dhcpcd and here is the output of ` systemctl status dhcpcd` when this happens: http://ix.io/3M2c

Then, I followed the instructions in the wiki link and here are the output: http://ix.io/3M2e

After I did `sudo dhcpcd`, my internet started working again (accessing webpages/ping google.com). Do you have any ideas why this happened? Perhaps I didn't start/enable the dhcpcd service correctly? Sorry I'm new to this and I don't exactly understand how this works but I thought that I was already using wpa_supplicant and it conflicts with dhcpcd? Or are the two separate things? Thanks again for all the help!

Offline

#5 2022-01-13 07:42:14

seth
Member
Registered: 2012-09-03
Posts: 26,570

Re: [SOLVED] No internet access but netctl-auto is running

Yes I have disabled the windows fast boot when I first installed arch.

Make sure it still is because windows updates frequently re-enable it. Not sure whether it's relevant here atm.


8 Seconds into the boot, netctl connects to the AP and runs dhcpcd to sucessfully obtain a lease (you're *NOT* running the dhcpcd service and you're not supposed to do)

Jan 12 11:53:33 apolloxvii audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=systemd-rfkill comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: authenticate with a8:97:cd:c8:64:24
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: send auth to a8:97:cd:c8:64:24 (try 1/3)
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: authenticated
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: associate with a8:97:cd:c8:64:24 (try 1/3)
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: RX AssocResp from a8:97:cd:c8:64:24 (capab=0x511 status=0 aid=91)
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: associated
Jan 12 11:53:34 apolloxvii kernel: wlp59s0: Limiting TX power to 30 (30 - 0) dBm as advertised by a8:97:cd:c8:64:24
Jan 12 11:53:34 apolloxvii kernel: IPv6: ADDRCONF(NETDEV_CHANGE): wlp59s0: link becomes ready
Jan 12 11:53:34 apolloxvii dhcpcd[1074]: dhcpcd-9.4.1 starting
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: DUID 00:04:4c:4c:45:44:00:47:53:10:80:39:b1:c0:4f:35:51:32
Jan 12 11:53:34 apolloxvii kernel: 8021q: 802.1Q VLAN Support v1.8
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: wlp59s0: connected to Access Point: Frontier0176_5G
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: wlp59s0: IAID d0:b6:b8:45
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: wlp59s0: soliciting a DHCP lease
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: wlp59s0: offered 192.168.254.30 from 192.168.254.254
Jan 12 11:53:34 apolloxvii dhcpcd[1077]: wlp59s0: probing address 192.168.254.30/24
Jan 12 11:53:40 apolloxvii dhcpcd[1077]: wlp59s0: leased 192.168.254.30 for 14400 seconds
Jan 12 11:53:40 apolloxvii dhcpcd[1077]: wlp59s0: adding route to 192.168.254.0/24
Jan 12 11:53:40 apolloxvii dhcpcd[1077]: wlp59s0: adding default route via 192.168.254.254
After I did `sudo dhcpcd`, my internet started working again

On the context of the provided journal (that does not record this) that would be mere coincidence. If there's a running dhcpcd server (is) further instances of dhcpcd act as IPC clients (they talk to the running server and "sudo dhcpcd" doesn't tell it anything)
Also dhcpcd still needs something™ to provide a carrier which is either a dhcpcd hook (not in the journal) or established by netctl/wpa_supplicant/iwd

Is the posted journal spposed to cover an affected boot? Did it happen ~19:58?

Jan 12 19:58:32 apolloxvii kernel: ath10k_pci 0000:3b:00.0: timed out waiting peer stats info
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: wmi command 90113 timeout, restarting hardware
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request stats (-11)
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request peer stats info: -108
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to read hi_board_data address: -16
Jan 12 19:58:40 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to receive initialized event from target: 00000000
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to receive initialized event from target: 00000000
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to wait for target init: -110
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request stats (-108)
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: device successfully recovered

Offline

#6 2022-01-13 15:20:40

Strike0
Member
From: Germany
Registered: 2011-09-05
Posts: 1,333

Re: [SOLVED] No internet access but netctl-auto is running

lucem wrote:

Perhaps I didn't start/enable the dhcpcd service correctly? Sorry I'm new to this and I don't exactly understand how this works but I thought that I was already using wpa_supplicant and it conflicts with dhcpcd? Or are the two separate things? Thanks again for all the help!

wpa_supplicant is the tool which connects your wireless adapter to the router, but does nothing concerning IP addresses or routing. dhcpcd is the tool which obtains the IP address after the link is ready. netctl manages both. My mention of dhcpcd service was a shot in the dark triggered by you mentioning resume-from-suspend problems and work arounds.

Please clarify Seth's last question regarding the journal (19:58) first, so we know it covers the error bugging you. Also state again when the problem happens (at boot/resume/runtime?).

Offline

#7 2022-01-13 19:29:01

lucem
Member
Registered: 2022-01-12
Posts: 5

Re: [SOLVED] No internet access but netctl-auto is running

seth wrote:

On the context of the provided journal (that does not record this) that would be mere coincidence. If there's a running dhcpcd server (is) further instances of dhcpcd act as IPC clients (they talk to the running server and "sudo dhcpcd" doesn't tell it anything)
Also dhcpcd still needs something™ to provide a carrier which is either a dhcpcd hook (not in the journal) or established by netctl/wpa_supplicant/iwd

I had recorded the journal first before running "sudo dhcpcd", so the journal doesn't cover the command, but the next time this happens I can run "sudo dhcpcd" first before recording the journal and see if it works. How do I provide this carrier? I looked in "/etc/netctl/hook" and "/etc/wpa_supplicant" and there are both empty. I did also find "/etc/dhcpcd.conf" file, do I need to make some changes there?

seth wrote:

Is the posted journal spposed to cover an affected boot? Did it happen ~19:58?

Jan 12 19:58:32 apolloxvii kernel: ath10k_pci 0000:3b:00.0: timed out waiting peer stats info
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: wmi command 90113 timeout, restarting hardware
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request stats (-11)
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request peer stats info: -108
Jan 12 19:58:37 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to read hi_board_data address: -16
Jan 12 19:58:40 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to receive initialized event from target: 00000000
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to receive initialized event from target: 00000000
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: failed to wait for target init: -110
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: could not request stats (-108)
Jan 12 19:58:43 apolloxvii kernel: ath10k_pci 0000:3b:00.0: device successfully recovered

Yes exactly, it happened around 19:58.

Strike0 wrote:

Also state again when the problem happens (at boot/resume/runtime?).

So the problem happened during runtime. I was just browsing on firefox and when I clicked on a new link, and it started buffering longer than normal, I checked if the problem happened again with "ping google.com"

Thank you guys for your help!

Offline

#8 2022-01-13 20:45:10

seth
Member
Registered: 2012-09-03
Posts: 26,570

Re: [SOLVED] No internet access but netctl-auto is running

netctl provides the carrier by running wpa_supplicant (i don't think there's iwd support atm) out of it's own config - the dhcpcd hook would be https://wiki.archlinux.org/title/Dhcpcd … supplicant but you don't need that!
Your network configuration is ok and not the problem here.

Yes exactly, it happened around 19:58.

What means the device dropped out (HW or driver issue) and was then recovered 10 seconds later.

There're quite some firmware crashes reported for the chip (incl. on this forum) but the consequences seem more severe (HW can't even be reset)
How frequent is this for you?

There's a ridiculous amount of

Jan 12 15:18:08 apolloxvii kernel: logitech-djreceiver 0003:046D:C537.0007: Unexpected input report number 128

Can you try w/o the logitech dongle? (Just a wild shot)

Offline

#9 2022-01-14 05:57:11

lucem
Member
Registered: 2022-01-12
Posts: 5

Re: [SOLVED] No internet access but netctl-auto is running

seth wrote:

What means the device dropped out (HW or driver issue) and was then recovered 10 seconds later.

There're quite some firmware crashes reported for the chip (incl. on this forum) but the consequences seem more severe (HW can't even be reset)
How frequent is this for you?

It varies. On some days, it would literally happen every 5-10 minutes or so, but in the past 2-3 days, it's only happened once a day. However, I don't think the internet recovered after 10 seconds even after the device had recovered but the internet was definitely out for a few minutes until I did "sudo dhcpcd" like I mentioned. In every other case, the internet has always been out until I either reboot or restart netctl with "sudo systemctl restart netctl-auto..."

seth wrote:

There's a ridiculous amount of

Jan 12 15:18:08 apolloxvii kernel: logitech-djreceiver 0003:046D:C537.0007: Unexpected input report number 128

Can you try w/o the logitech dongle? (Just a wild shot)

Haha yeah I didn't realize my logitech receiver cause so many problems until I took a look at the logs. I can try not using my mouse for a while and see if it changes anything, although I don't think it's related because I have been using this mouse for a year or so and I haven't had any networks problems until now. Do you know how to fix this error? Do I just need to download the proper logitech driver? I never thought of doing that since it just worked.

Offline

#10 2022-01-14 08:17:27

Strike0
Member
From: Germany
Registered: 2011-09-05
Posts: 1,333

Re: [SOLVED] No internet access but netctl-auto is running

Well, you could use a simple cable mouse for a while to check if it changes anything regarding the wifi driver crashing. Seth might have mentioned it, because your log shows the kernel is flooded with those djreceiver errors just before the 12:58 crash of the wifi module happened. Trying to fix the mouse itself better be a separate issue/thread to investigate.

I had a quick search regarding your used wifi-card chipset and indeed a number of users have problems. There are also work-arounds for notebooks in the wiki with respect to it. One links to: https://bbs.archlinux.org/viewtopic.php?id=254362
Have a read at that. The affected users reverted to downgrading the firmware module of the wifi card as a work around which appears to work for them. Downgrading is not ideal, but it appears it is a regression of the kernel module. Probably your problem started when the regular linux-firmware package was upgraded. Given all the tries in the linked bbs thread to fix the problem, I'd say this route to downgrade the ath10k module is what you should try too. If downgrading helps, the next step would be to follow-up the kernel bug reports.

Offline

#11 2022-01-14 09:00:11

seth
Member
Registered: 2012-09-03
Posts: 26,570

Re: [SOLVED] No internet access but netctl-auto is running

About the mouse see https://github.com/libratbag/libratbag/issues/1124
The last comment suggests to blacklist hid-logitech-dj module but also to setup the mouse on windows.
Whether it can be possibly relevant to the wifi situation depends on whether they somewhere end up on the same bus.

The revivied NIC might still require to reconnect to the AP and certainly restaring the netctl service but *maybe* even just nudging dhcpcd might get you there - post a journal covering that so we can see what happens in the aftermath.

you wrote:

Jan 12 11:53:28 apolloxvii kernel: ath10k_pci 0000:3b:00.0: firmware ver WLAN.RM.4.4.1-00157-QCARMSWPZ-1 api 6 features wowlan,ignore-otp,mfp crc32 90eebefb

the other thread wrote:

apr 06 11:49:38 stryke kernel: ath10k_pci 0000:3b:00.0: firmware ver WLAN.RM.4.4.1-00140-QCARMSWPZ-1 api 6 features wowlan,ignore-otp,mfp crc32 29eb8ca1

Though you're currently loading /usr/lib/firmware/ath10k/QCA6174/hw3.0/firmware-6.bin as well, it's a different version and nb. that you don't have the "unsupported HTC service id: 1536" and the other thread doesn't have the wmi timeout and most threads about the chip involve an outright firmware crash, so it's possible that the firmware version from the other thread is more stable.

Offline

#12 2022-01-23 02:23:05

lucem
Member
Registered: 2022-01-12
Posts: 5

Re: [SOLVED] No internet access but netctl-auto is running

The problem hasn't reoccurred since the last time where the internet started working after doing "sudo dhcpcd" as described above, so I'll just mark this resolved for now. Thanks for your help!

Offline

Board footer

Powered by FluxBB