WiFi speed is pretty low when compared to LAN

Hi there,

I am a Linux newbie. I recently switched to the purple side of Linux :slight_smile: from Windows 10.
Now, I am facing a situation with my WiFi speed.
Speedtest is showing above 100+ Mbps upload/download speed via LAN.
But, in WiFi, it’s a mere 8+ Mbps speed for download, and 20+ Mbps speed for upload.
Following is my system information.

lspci -k|grep Network

01:00.0 Network controller: Intel Corporation Wireless 8265 / 8275 (rev 78)
lspci -k|grep wifi

Kernel driver in use: iwlwifi
Kernel modules: iwlwifi

sudo dmesg | grep iwl
[sudo] password for bineeth: 

[    9.010730] iwlwifi 0000:01:00.0: enabling device (0000 -> 0002)
[    9.186151] iwlwifi 0000:01:00.0: loaded firmware version 36.ca7b901d.0 8265-36.ucode op_mode iwlmvm
[    9.596829] iwlwifi 0000:01:00.0: Detected Intel(R) Dual Band Wireless AC 8265, REV=0x230
[    9.657391] iwlwifi 0000:01:00.0: base HW address: 40:ec:99:c5:a0:ee
[    9.732525] ieee80211 phy0: Selected rate control algorithm 'iwl-mvm-rs'

I’ve also tried connecting my system to the Mobile network. And there, I am getting a reasonable download speed of 14+ Mbps via WiFi.

I think, this has to do something with the iwlwifi driver.

Based on my network card configuration, I believe my system is supposed to use firmware iwlwifi-3168-22.ucode.

I’ve tried removing the higher versions. And I could see the system is picking iwlwifi-3168-22.ucode firmware in the boot log (Note: In dmesg output, firmware version 36.x is loaded. That’s because I re-installed the OS today, I tried switching firmware before that).

But, still no luck. Can someone help me to solve this, please?

Others have reported the same issue with this chip. Here is all the info on iwlwifi. There is some good info there if you read through it. It has all the firmware links also.

https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi#wi-fibluetooth_coexistence

Example: options iwlwifi bt_coex_active=0 swcrypto=1 11n_disable=8

I would suggest you try some of the suggestions on the Arch wiki page also. You can also set it to run one band or the other. Sometimes that is a problem. There is also bluetooth co-exist which can have an effect on it. You’ll have to try some things and experiment.

https://wiki.archlinux.org/title/Network_configuration/Wireless#iwlwifi

Edit: Another issue that can affect the Wifi from Windows is the fast startup feature on Windows 10. It should be turned off.

1 Like

Thanks! @ricklinux for the reply. I’ve tried all these. And fast startup feature is not enabled in my machine. But, nothing has worked for me so far :frowning: .

It’s not the fast boot in the Bios. It’s the fast start up Windows just to make sure. But i can tell you that that chip is known to have this problem. You can try different firmware. I gave you the links. There isn’t much else i can do to help.

Okay. I understand @ricklinux. Thank you very much for the help :slight_smile: .

1 Like

This is an 802.11 AC dual band 2.4 Ghz and 5.0 Ghz. Did you try disabling the N in your router?

Edit: Also set the power save option off

Create or edit the following

/etc/modprobe.d/iwlwifi.conf

add

options iwlwifi 11n_disable=1 swcrypto=1 bt_coex_active=0 power_save=0

Edit: Did you try it disabling the 2.4 Ghz also so it only uses 5.0 Ghz band.

1 Like

Also, keep in mind that wifi is slower than Ethernet generally.

1 Like

Yes it is… i only get half the speed on Wifi and this one is a fast AC chip.

Hi @ricklinux, thanks again for the suggestion.
I tried this today. And the following was the result.

Band 5.0 GHz, Mode: 802.11g:

bineeth@Workstation-10SRCT01WW ~]$ speedtest
Retrieving speedtest.net configuration...
Testing from BSNL (117.216.24.161)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Keralavision Broadband (Trivandrum) [4.66 km]: 31.795 ms
Testing download speed................................................................................
Download: 8.88 Mbit/s
Testing upload speed......................................................................................................
Upload: 6.26 Mbit/s

Band 5.0 GHz, Mode: 802.11 b/g/n mix:

[bineeth@Workstation-10SRCT01WW ~]$ speedtest
Retrieving speedtest.net configuration...
Testing from BSNL (117.216.24.161)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Jtel (Nagercoil) [61.17 km]: 64.074 ms
Testing download speed................................................................................
Download: 13.28 Mbit/s
Testing upload speed......................................................................................................
Upload: 3.34 Mbit/s

And via LAN:

[bineeth@Workstation-10SRCT01WW ~]$ speedtest
Retrieving speedtest.net configuration...
Testing from BSNL (117.216.24.161)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Keralavision Broadband (Trivandrum) [4.66 km]: 28.979 ms
Testing download speed................................................................................
Download: 122.96 Mbit/s
Testing upload speed......................................................................................................
Upload: 102.18 Mbit/s

I remember that I’ve got around 85.0+ Mbps download and 35.0+ upload speed via WiFi, during the first days of my installation.
And I think, this problem happened after I’ve done an update.
I don’t remember exactly when that happened.
But, I think it should be after 14th August (I used to update my OS only on Saturday).

I’ve never faced this kind of issue in Windows.
I am not complaining. I know that this is not the fault of Endeavour (or Linux).
I assume there should be some extra tweaks in Intel’s Windows driver, that’s not present in Linux counterpart.

It’s because of the Driver issues, right? Or, is that a limitation of the Linux Kernel? Sorry if this is a dumb question; as I said I am new to Linux.

Look in /var/log/pacman.log to find out which packages changed when.

1 Like

As @jonathon say’s look at the pacman log and on or around the date in question and see what packages were updated. You can always use downgrade to downgrade a suspected update to a previous version. It’s not a limitation of Linux but it could be the module (driver) or firmware or another package that affects the WiFi in some way. i.e. Bluetooth

You can only check and see what you think it might be based on that date and try downgrading it.

Edit: You are not the first one with that Intel chip to have the same issue. It’s well documented on Arch and other sites.

Welcome! :smile:

One thing worth trying is to check if some system configuration file updates should be merged. That can be done e.g. with eos-pacdiff.

If you find any config updates related to wifi stuff with this tool, please check the changes and if they seem reasonable, merge the changes.

The new config file names have a .pacnew suffix.

Thanks, @ricklinux @jonathon and @manuel, for the suggestions and help :smiley:

I’ll check what you’ve suggested.

And @ricklinux yes, I am aware that I am not the only one facing this issue :slight_smile:

I’ve decided to not bother you guys again with this unsolvable problem.

But, please let me ask a final question.

Is this something we can solve by fixing some logic issue in the iwlwifi source code?

I don’t think so. The code is probably too complicated anyway for people who are not already familiar with it.

But have you tried the LTS (or other) kernel? Don’t know if it helps, but worth a shot.

Yes, I tried the LTS kernel, @manuel. But, I ended up with no luck.

Did you try disabling n? There are all kinds of options.

/etc/modprobe.d/iwlwifi.conf
options iwlwifi 11n_disable=1

Or

options iwlwifi 11n_disable=1 led_mode=1 swcrypto=1

and this one

options iwlwifi 11n_disable=8

Edit: It is also advisable to try disabling the n in your router settings.

Thanks @ricklinux :slight_smile: , yes I did try them. I even tried disabling the Bluetooth, libvirt, and Nftables daemon. But, no luck so far.

iwlwifi-issue

Thanks! @manuel for the suggestion :slight_smile:.

Following is the output of eos-pacdiff.

[

bineeth@Workstation-10SRCT01WW ~]$ eos-pacdiff
Starting pacdiff & diffuse as root ...
Root Password: ==> pacsave file found for /etc/iptables/iptables.rules
:: (V)iew, (S)kip, (R)emove pacsave, (O)verwrite with pacsave, (Q)uit: [v/s/r/o/q] s
==> pacsave file found for /etc/iptables/ip6tables.rules
:: (V)iew, (S)kip, (R)emove pacsave, (O)verwrite with pacsave, (Q)uit: [v/s/r/o/q] s
==> pacnew file found for /etc/pacman.d/mirrorlist
:: (V)iew, (S)kip, (R)emove pacnew, (O)verwrite with pacnew, (Q)uit: [v/s/r/o/q] s

I think I can safely ignore iptables related changes, as I am using Nftables now.
I can’t find anything related to iwlwifi.

Thanks! @jonathon :slight_smile: . Following are the entries from pacman.log for 14th August.

[2021-08-14T10:49:12+0530] [PACMAN] Running 'pacman -Syu'
[2021-08-14T10:49:12+0530] [PACMAN] synchronizing package lists
[2021-08-14T10:49:15+0530] [PACMAN] starting full system upgrade
[2021-08-14T10:49:45+0530] [ALPM] running '60-mkinitcpio-remove.hook'...
[2021-08-14T10:49:45+0530] [ALPM] transaction started
[2021-08-14T10:49:45+0530] [ALPM] upgraded libphonenumber (8.12.25-1 -> 8.12.29-1)
[2021-08-14T10:49:46+0530] [ALPM] upgraded evolution-data-server (3.40.3-1 -> 3.40.4-1)
[2021-08-14T10:49:46+0530] [ALPM] upgraded gnome-maps (40.3-1 -> 40.4-1)
[2021-08-14T10:49:47+0530] [ALPM] upgraded imlib2 (1.7.2-1 -> 1.7.3-1)
[2021-08-14T10:49:47+0530] [ALPM] upgraded libstemmer (0+713-2 -> 2.1.0-1)
[2021-08-14T10:50:02+0530] [ALPM] upgraded linux (5.13.9.arch1-1 -> 5.13.10.arch1-1)
[2021-08-14T10:50:07+0530] [ALPM] upgraded linux-headers (5.13.9.arch1-1 -> 5.13.10.arch1-1)
[2021-08-14T10:50:07+0530] [ALPM] upgraded r8168 (8.049.02-10 -> 8.049.02-11)
[2021-08-14T10:50:07+0530] [ALPM] upgraded sdl2 (2.0.16-1 -> 2.0.16-2)
[2021-08-14T10:50:08+0530] [ALPM] transaction completed
[2021-08-14T10:50:08+0530] [ALPM] running '30-systemd-update.hook'...
[2021-08-14T10:50:08+0530] [ALPM] running '60-depmod.hook'...
[2021-08-14T10:50:09+0530] [ALPM] running '90-mkinitcpio-install.hook'...
[2021-08-14T10:50:09+0530] [ALPM-SCRIPTLET] ==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
[2021-08-14T10:50:09+0530] [ALPM-SCRIPTLET]   -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
[2021-08-14T10:50:09+0530] [ALPM-SCRIPTLET] ==> Starting build: 5.13.10-arch1-1
[2021-08-14T10:50:09+0530] [ALPM-SCRIPTLET]   -> Running build hook: [base]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [udev]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [autodetect]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [modconf]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [block]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: xhci_pci
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [keyboard]
[2021-08-14T10:50:10+0530] [ALPM-SCRIPTLET]   -> Running build hook: [keymap]
[2021-08-14T10:50:11+0530] [ALPM-SCRIPTLET]   -> Running build hook: [filesystems]
[2021-08-14T10:50:11+0530] [ALPM-SCRIPTLET]   -> Running build hook: [fsck]
[2021-08-14T10:50:11+0530] [ALPM-SCRIPTLET] ==> Generating module dependencies
[2021-08-14T10:50:11+0530] [ALPM-SCRIPTLET] ==> Creating xz-compressed initcpio image: /boot/initramfs-linux.img
[2021-08-14T10:50:16+0530] [ALPM-SCRIPTLET] ==> Image generation successful
[2021-08-14T10:50:16+0530] [ALPM-SCRIPTLET] ==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
[2021-08-14T10:50:16+0530] [ALPM-SCRIPTLET]   -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
[2021-08-14T10:50:16+0530] [ALPM-SCRIPTLET] ==> Starting build: 5.13.10-arch1-1
[2021-08-14T10:50:16+0530] [ALPM-SCRIPTLET]   -> Running build hook: [base]
[2021-08-14T10:50:17+0530] [ALPM-SCRIPTLET]   -> Running build hook: [udev]
[2021-08-14T10:50:17+0530] [ALPM-SCRIPTLET]   -> Running build hook: [modconf]
[2021-08-14T10:50:17+0530] [ALPM-SCRIPTLET]   -> Running build hook: [block]
[2021-08-14T10:50:17+0530] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: aic94xx
[2021-08-14T10:50:17+0530] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: wd719x
[2021-08-14T10:50:18+0530] [ALPM-SCRIPTLET] ==> WARNING: Possibly missing firmware for module: xhci_pci
[2021-08-14T10:50:19+0530] [ALPM-SCRIPTLET]   -> Running build hook: [keyboard]
[2021-08-14T10:50:19+0530] [ALPM-SCRIPTLET]   -> Running build hook: [keymap]
[2021-08-14T10:50:19+0530] [ALPM-SCRIPTLET]   -> Running build hook: [filesystems]
[2021-08-14T10:50:20+0530] [ALPM-SCRIPTLET]   -> Running build hook: [fsck]
[2021-08-14T10:50:21+0530] [ALPM-SCRIPTLET] ==> Generating module dependencies
[2021-08-14T10:50:21+0530] [ALPM-SCRIPTLET] ==> Creating xz-compressed initcpio image: /boot/initramfs-linux-fallback.img
[2021-08-14T10:50:34+0530] [ALPM-SCRIPTLET] ==> Image generation successful

But, my untrained eyes cannot find anything suspicious.