I was so pleased when I finally discovered the QEMU Monitor.
It is an excellent tool that covers a lot of the gaps in the QEMU documentation.
Curious which devices are memory mapped to which regions on your virtual board and the docs aren’t specific? Check the device info in the monitor!
Want to know what interrupt signals a device might generate and can’t find that info in the docs, use the monitor to check!
It’s honestly a life saver.
A cool tip I recently found to make using the Monitor easier when you are running your QEMU machine in nographic mode is that you can have QEMU read and write the monitor to a file on your system. Then you can use a tool like socat to have one terminal running your QEMU machine and another running your monitor! Super convenient.
QEMU monitor is a really unappreciated little gem.
I have a bunch of helper scripts that create / restore machine snapshots - then I run Jupyter server inside the VM and use create / restore as a time travelling debugger
Note that the 'HMP' (i.e. human readable) version of the monitor can change, so using it in non-trivial scripts is probably a bad idea. The QMP (Json mode) has proper deprecation and fixed behaviour.
For HMP we just try and make stuff easy for people driving it themselves for debug etc, so the output can change if it makes it easier to read, add more info or whatever.
[HMP maintainer, although very little changes these days!]
It is an excellent tool that covers a lot of the gaps in the QEMU documentation.
Curious which devices are memory mapped to which regions on your virtual board and the docs aren’t specific? Check the device info in the monitor!
Want to know what interrupt signals a device might generate and can’t find that info in the docs, use the monitor to check!
It’s honestly a life saver.
A cool tip I recently found to make using the Monitor easier when you are running your QEMU machine in nographic mode is that you can have QEMU read and write the monitor to a file on your system. Then you can use a tool like socat to have one terminal running your QEMU machine and another running your monitor! Super convenient.
This stack exchange answer explains the details:
https://unix.stackexchange.com/questions/426652/connect-to-r...
I have a bunch of helper scripts that create / restore machine snapshots - then I run Jupyter server inside the VM and use create / restore as a time travelling debugger
For HMP we just try and make stuff easy for people driving it themselves for debug etc, so the output can change if it makes it easier to read, add more info or whatever. [HMP maintainer, although very little changes these days!]