Marc Zyngier <[email protected]> writes:
On Mon, 09 Feb 2026 22:14:04 +0000, Colton Lewis <[email protected]> wrote:
We may want a partitioned PMU but not have FEAT_FGT to untrap the specific registers that would normally be untrapped. Add a handler for those registers in the fast path so we can still get a performance boost from partitioning.
The idea is to handle traps for all the PMU registers quickly by writing directly to the hardware when possible instead of hooking into the emulated vPMU as the standard handlers in sys_regs.c do.
This seems extremely premature. My assumption is that PMU traps are rare, and that doing a full exit should be acceptable. Until you demonstrate the contrary, I don't want this sort of massive bloat in the most performance-critical path.
After some consideration I agree. I will try a full exit and see if that is sufficient.
