Re: [PATCH 05/11] x86/fpu: set PKRU state for kernel threads

From: Dave Hansen
Date: Fri Oct 19 2018 - 14:26:56 EST

On 10/19/2018 10:37 AM, Andy Lutomirski wrote:
>> I think it's much more straightforward to just not enforce pkeys.
>> Having this "phantom" value could cause a very odd, nearly
>> undebuggable I/O failure.
> But now we have the reverse. The IO can work if itâs truly async but,
> if the kernel decides to synchronously complete IO (with GUP or
> copy_to_user), itâll fail, right. This isnât exactly friendly
> either.

Yeah, but a synchronous I/O failure is really straightforward to debug
because you get an immediate error message about it. This is certainly
not the weirdest behavior or asymmetry that we would see from
synchronous vs. async I/O.