Hi Tom,
On 16 Dec 2019, at 23:52, Tom Rymes trymes@rymes.com wrote:
Folks,
I have been having this issue for quite some time, and though it isn't a major issue, it is keeping me from using some older hardware that is, due to the Intel vulnerabilities, perversely more secure (it's 32-bit).
No, it is not.
Loads of hardening features are limited in their effectiveness on 32 bit:
https://blog.ipfire.org/post/32-bit-is-dead-long-live-32-bit
The long story short is that I have devices with wireless cards that use the b43 driver, and they refuse to come up because they see the rfkill switch as being on. These devices worked fine before Core 105 was released. Nothing we do seems to resolve this, and IPFire doesn't include the rfkill utility (it does include the rfkill kernel module).
How difficult would it be to include the rfkill utility? That would allow me to use it to try and override the switch.
https://mirrors.edge.kernel.org/pub/software/network/rfkill/
Here's an old forum thread we posted when this became an issue back with core 105: https://forum.ipfire.org/viewtopic.php?t=17398
Tom
PS: After I wrote this, I downloaded an rfkill RPM for a different architecture and extracted the binaries. It runs and it allowed me to unblock the device and get it working. I do think it would be a helpful addition to the distribution.
I can package it for you. It is not using any other libraries, so it should be easy.
But I am an an airport now, so cannot do it today :)
-Michael