Thiago Jung Bauermann <bauer...@linux.vnet.ibm.com> writes: > Michael Ellerman <m...@ellerman.id.au> writes: >> Thiago Jung Bauermann <bauer...@linux.vnet.ibm.com> writes: >>> The hypervisor interface to access 24x7 performance counters (which collect >>> performance information from system power on to system power off) has been >>> extended in POWER9 adding new fields to the request and result element >>> structures. >>> >>> Also, results for some domains now return more than one result element and >>> those need to be added to get a total count. >>> >>> The first two patches fix bugs in the existing code. The following 4 >>> patches are code improvements and the last two finally implement support >>> for the changes in POWER9 described above. >>> >>> POWER8 systems only support version 1 of the interface, while POWER9 >>> systems only support version 2. I tested these patches on POWER8 to verify >>> that there are no regressions, and also on POWER9 DD1. >> >> Where is version 2 documented? > > Only in an internal specification.
Presumably a PAPR ACR, what's the title? And is it approved/accepted? >> And what happens when we boot on a POWER9 in POWER8 compatibility mode? > > It will still use version 2 of the API, and still require aggregation of > result elements. Does this mean that hv_24x7_init should check > cur_cpu_spec->oprofile_cpu_type for "ppc64/power9" instead of > cpu_has_feature(CPU_FTR_ARCH_300)? Yes. cheers