On 09/05/2018 02:13 PM, Alexander Duyck wrote: > Instead of keeping the value in CONFIG_DEBUG_VM I am adding a new CONFIG > value called CONFIG_DEBUG_VM_PAGE_INIT_POISON that will control the page > poisoning independent of the CONFIG_DEBUG_VM option. I guess this is a reasonable compromise. If folks see odd 'struct page' corruption, they'll have to know to go turn this on and reboot, though.
