Now that runqemu was refactor to use python it would be nice to have a module that can be imported in other parts of the project. This would be beneficial specially for selftest and runtime testing. The current implementation has to run qemu using a shell, and because of that there are checks to be sure that qemu is killed after the parent process is killed or ends unexpectedly.
Mariano, would you add these 2 things to the bug description: 1. source of the requirement (which effort or module requires it?) 2. benefits that are sought with this enhancement
(In reply to comment #1) > Mariano, would you add these 2 things to the bug description: > > 1. source of the requirement (which effort or module requires it?) The main consumer that I see is the QemuRunner class in oeqa library, currently runqemu has to be called using a shell because of the command line parameters used in the script. > 2. benefits that are sought with this enhancement In QemuRunner class we have some hacks to kill the shell that called runqemu, and the qemu-system process in case QemuRunner's process ends unexpectedly. I've been working lately with QemuRunner and it seems qemu-system is not always killed when a bitbake worker (using testimage) is killed; I must admit that it happens sporadically, about 10% of the time. So, the benefit here is to have call directly runqemu from python without using a shell, with this we would have better control of the qemu process avoiding hacks.
I don't think I'll have time to work on this in 2.3 time frame. moving to 2.4, increasing estimate.
Still a good idea. Any volunteers?
We've had many changes in runqemu since this was written. Open a new defect if there are similar ERs you would like to see. -- YP bug review.