>I think this is a bug in your device tree because the KASSERT was >intentional: > > - In the ACPI case, we probe for CPU PMU support before calling > armv8_pmu_init. > - In the FDT case, the PMU attaches to a node described in the device > tree. > >So if you hit this KASSERT, AFAICT it means your device tree is describing >a device that is not there. Unless I'm missing something here.
I tried to fix it to work properly as a kernel module for debugging and improving tprof itself, but the current implement is difficult due to the acpi/fdt pmu interrupt and tprof, so I gave up :-P I'll revert it. thanks! -- ryo shimizu