So I have finally been forced to use floats in an embedded system!!! Yes it does have an FPU and should even support double point precision so what's the problem ? Well now I have to watch my code even more due to the many cavets of using floats and doubles in C. Just have a look at these few questions and decide for yourself.

https://frama-c.com/2011/11/08/Floating-point-quiz.html

I have never questioned daemonising a Linux program that is a service and has no console UI.

I have just done that as I usually do however now the question has been asked ...

Full story »

Usually to SSH into a server we use the following command ...

ssh [hostname]

And then we log into the server.

However when the server isn't running on the default TCP port (22) or when our local username isn't the same as our account on the remote server we have to do something like this ...

ssh [username]@[hostname] -p [port]

If we regularly log-in to this server from the same machine there is a way to just use the 1st way even though the username and port is different.

Full story »

Linux filesystems are usually configured by the /etc/fstab file.

Instead of mounting by block device name (eg. /dev/sda1) and having to remember which partition is which, we can also mount by Label name.

This is easy on a SATA hard-disk or even a USB memory-stick. But how does this work for a Memory Technology Device, like a NOR flash for example?

Full story »

I have saved many hours of my own time and my colleagues time by having a tunnel to a host on a network we normally cannot connect remotely to.
No need for a VPN here, and as explained by this post if done right it is just as secure as a secure VPN.

https://hobo.house/2016/06/20/fun-and-profit-with-reverse-ssh-tunnels-and-autossh/