On Mon, Jun 01, 2020 at 03:57:55PM +0100, Andrew Cooper wrote: > c/s 9267a439c "x86/ucode: Document the behaviour of the microcode_ops hooks" > identified several poor behaviours of the start_update()/end_update_percpu() > hooks. > > AMD have subsequently confirmed that OSVW don't, and are not expected to, > change across a microcode load, rendering all of this complexity unecessary.
Is there a reference to some AMD PM or similar document that can be added here for completeness? > Instead of fixing up the logic to not leave the OSVW state reset in a number > of corner cases, delete the logic entirely. This in turn allows for the > removal of the poorly-named 'start_update' parameter to > microcode_update_one(). > > Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com> Reviewed-by: Roger Pau Monné <roger....@citrix.com> Thanks, Roger.
