On 2021-11-04 4:48 p.m., Helge Deller wrote:
I believe the alternative code patching isn't critical. It just patches in NOPs, so even if the newly patched NOPs aren't visible yet (when the TLB handler is executed), it uses the old instructions which shouldn't harm anything. They were executed before the whole time.
But we flush the data cache when we hit a page boundary. This makes these NOPs visible. Maybe we should do both flushes after all words have been written. Dave -- John David Anglin dave.anglin@xxxxxxxx
