Wifi is hard blocked after suspend

archdude · 2018-04-29 14:40:23

I've been through numerous threads on arch forums where users reported the wifi being disabled after the laptop wakes from suspend. I have an HP laptop and I have faced this issue with all Linux variants I have tried.

Problem: Wifi is hard blocked after suspend

sudo rfkill list

Before suspend:

1: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: no

After suspend:

1: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: yes

The problem is that there is no hardware button provided on machine, this happens when Linux suspends after inactivity or on lid being closed. This is definitely an issue with Linux because windows OS shows no such issues.

Helpful info:

Output of

lspci | grep Network

is:

08:00.0 Network controller: Intel Corporation Wireless 3160 (rev 83)

Output of

lsmod | grep iwlwifi

is:

iwlwifi               323584  1 iwlmvm
cfg80211              749568  3 iwlmvm,iwlwifi,mac80211

Output of

dmesg

logs around the suspend event:

[  692.484835] PM: suspend entry (deep)
[  692.484837] PM: Syncing filesystems ... done.
[  692.657161] Freezing user space processes ... (elapsed 0.101 seconds) done.
[  692.758584] OOM killer disabled.
[  692.758586] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[  692.760098] Suspending console(s) (use no_console_suspend to debug)
[  692.796779] sd 0:0:0:0: [sda] Synchronizing SCSI cache
[  692.812892] sd 0:0:0:0: [sda] Stopping disk
[  693.467686] ACPI: EC: interrupt blocked
[  693.500929] ACPI: Preparing to enter system sleep state S3
[  693.508213] ACPI: EC: event blocked
[  693.508215] ACPI: EC: EC stopped
[  693.508216] PM: Saving platform NVS memory
[  693.508254] Disabling non-boot CPUs ...
[  693.549132] smpboot: CPU 1 is now offline
[  693.614901] smpboot: CPU 2 is now offline
[  693.658161] smpboot: CPU 3 is now offline
[  693.661608] ACPI: Low-level resume complete
[  693.661679] ACPI: EC: EC started
[  693.661680] PM: Restoring platform NVS memory
[  693.664342] Enabling non-boot CPUs ...
[  693.664378] x86: Booting SMP configuration:
[  693.664379] smpboot: Booting Node 0 Processor 1 APIC 0x2
[  693.666285]  cache: parent cpu1 should not be sleeping
[  693.666439] CPU1 is up
[  693.666460] smpboot: Booting Node 0 Processor 2 APIC 0x1
[  693.667147]  cache: parent cpu2 should not be sleeping
[  693.667321] CPU2 is up
[  693.667347] smpboot: Booting Node 0 Processor 3 APIC 0x3
[  693.667865]  cache: parent cpu3 should not be sleeping
[  693.668249] CPU3 is up
[  693.679350] ACPI: Waking up from system sleep state S3
[  693.694815] ACPI: EC: interrupt unblocked
[  693.792731] pciehp 0000:00:1c.1:pcie004: Slot(1): Link Down
[  693.793635] iwlwifi 0000:08:00.0: reporting RF_KILL (radio disabled)
[  693.793690] iwlwifi 0000:08:00.0: RF_KILL bit toggled to disable radio.
[  693.794124] usb usb2: root hub lost power or was reset
[  693.794127] usb usb3: root hub lost power or was reset
[  693.799450] sd 0:0:0:0: [sda] Starting disk
[  693.808076] r8169 0000:09:00.0 eno1: link down
[  693.842403] ACPI: EC: event unblocked
[  694.035873] usb 1-1: reset high-speed USB device number 2 using ehci-pci
[  694.109996] ata2: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[  694.110027] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[  694.121426] ata2.00: configured for UDMA/100
[  694.132584] usb 2-5: reset full-speed USB device number 3 using xhci_hcd
[  694.195324] ata1.00: configured for UDMA/133
[  694.395952] usb 2-3: reset high-speed USB device number 2 using xhci_hcd
[  694.727370] psmouse serio1: synaptics: queried max coordinates: x [..5610], y [..4700]
[  694.764778] psmouse serio1: synaptics: queried min coordinates: x [1362..], y [1238..]
[  696.678334] OOM killer enabled.
[  696.678336] Restarting tasks ... done.
[  696.712511] usb 2-7: USB disconnect, device number 4
[  696.746843] PM: suspend exit

How can I find the exact cause ?

seth · 2018-04-29 15:02:54

The problem is that there is no hardware button provided on machine

???? Sure?
What exact model of notebook?

You can try to cheat on the BIOS by issueing the wrong acpi_osi, but a hard block is a hard block - you cannot undo it by software. At "best" it's an issue w/ the firmware for the chip.
Another thing you could try is just rescanning the PCI bus, in case it's rather a problem there and the interface isn't really blocked but the situation gets misinterpreted:

echo 1 > /sys/bus/pci/rescan

archdude · 2018-04-29 17:24:44

seth wrote:

The problem is that there is no hardware button provided on machine
???? Sure?
What exact model of notebook?
You can try to cheat on the BIOS by issueing the wrong acpi_osi, but a hard block is a hard block - you cannot undo it by software. At "best" it's an issue w/ the firmware for the chip.
Another thing you could try is just rescanning the PCI bus, in case it's rather a problem there and the interface isn't really blocked but the situation gets misinterpreted:
echo 1 > /sys/bus/pci/rescan

Its HP envy 15 and PCI re-scan didn't help. Though, the wireless chip is listed by lspci.

The point is: Auto suspend turns on the hard block, the machine is not touched, no button pressed, leave it alone for 15 minutes and it goes to sleep. Then press any button, it resumes, but this time its in "Airplane Mode" .

I guess the same as you stated, the firmware that suspends wifi, hard blocks it until rebooted. But hey, the Intel Corporation Wireless 3160 chip here is pretty common, and the same firmware works on other machines too.

Any option to completely re-load the wifi chipset firmware module without rebooting ?

seth · 2018-04-29 18:07:00

And Fn+F12 (the airplane) does not toggle wifi for you?
Also, what do you use for power management? (TLP, laptop-mode-tools, ...) and does this also happen if you disable that service, boot to multi-user.target (to rule our nastyness by your desktop environment) and directly run "systemctl suspend"?

About the firmware, check dmesg on whether the firmware isn't reloaded by rescanning the PCI bus anyway.

archdude · 2018-05-01 12:11:18

seth wrote:

And Fn+F12 (the airplane) does not toggle wifi for you?
Also, what do you use for power management? (TLP, laptop-mode-tools, ...) and does this also happen if you disable that service, boot to multi-user.target (to rule our nastyness by your desktop environment) and directly run "systemctl suspend"?
About the firmware, check dmesg on whether the firmware isn't reloaded by rescanning the PCI bus anyway.

Fn+F12 doesn't have any effect on hard block, it only toggles soft block.

I also booted to shell (before xinit, so as to skip desktop environment), and "systemctl suspend" directly from there also causes hard block. So desktop environment is not the culprit.

what do you use for power management?

How can I check that for my machine ? It must be some default module from archlinux.

Last edited by archdude (2018-05-01 12:14:47)

seth · 2018-05-01 12:17:55

There's no "default module" - you're using what you installed and configured (unless you're using antergos, manjaro, blackarch or some other derivate)

Try lying about the OS, eg.

acpi_osi=! acpi_osi="Windows 2009"

artesano · 2018-05-01 15:36:26

do you have a dual boot machine?

https://bbs.archlinux.org/viewtopic.php … 6#p1594426

axfelix · 2018-05-09 06:15:04

This is happening to me with 4.16 on my Dell Venue 11 Pro 7130, never happened before

Arch Linux

#1 2018-04-29 14:40:23

Wifi is hard blocked after suspend

#2 2018-04-29 15:02:54

Re: Wifi is hard blocked after suspend

#3 2018-04-29 17:24:44

Re: Wifi is hard blocked after suspend

#4 2018-04-29 18:07:00

Re: Wifi is hard blocked after suspend

#5 2018-05-01 12:11:18

Re: Wifi is hard blocked after suspend

#6 2018-05-01 12:17:55

Re: Wifi is hard blocked after suspend

#7 2018-05-01 15:36:26

Re: Wifi is hard blocked after suspend

#8 2018-05-09 06:15:04

Re: Wifi is hard blocked after suspend

Board footer