Michael Ellerman [off-list ref] writes:

Ram Pai [off-list ref] writes:

quoted

Handle Data and  Instruction exceptions caused by memory
protection-key.

The CPU will detect the key fault if the HPTE is already
programmed with the key.

However if the HPTE is not  hashed, a key fault will not
be detected by the  hardware. The   software will detect
pkey violation in such a case.

That seems like the wrong trade off to me.

It means every fault has to go through arch_vma_access_permitted(),
which is at least a function call in the best case, even when pkeys are
not in use, and/or the range in question is not protected by a key.

We don't really need to call arch_vma_access_permitted() in
arch/powerpc/ do_page_fault(). Core kernel does that in
handle_mm_fault(). So if the first fault is a bad access handle_mm_fault
handle this. If it is a valid access we insert the right hash page table
entry and then we do a wrong access, we detect that a key fault in the
low level hash fault handler. IIUC, the call the
arch_vma_access_permitted() from arch/powerpc/ can go away?

-aneesh