The Experimental patch in comment #9, although it seems to prevent the hang under some workloads, puts the code in the wrong function. We really think it should be at the end of __vgic_v3_save_state()
We will have a revised patch after some further testing.
The Experimental patch in comment #9, although it seems to prevent the hang under some workloads, puts the code in the wrong function. We really think it should be at the end of __vgic_ v3_save_ state()
We will have a revised patch after some further testing.