[PATCH] x86/uaccess: Avoid barrier_nospec() in copy_from_user()

[David Laight]

From: Linus Torvalds
> Sent: 14 October 2024 12:21
> 
> On Mon, 14 Oct 2024 at 02:59, David Laight <David.Laight at aculab.com> wrote:
> >
> > Isn't LAM just plain stupid unless the hardware validates the bits
> > against the TLB?
> 
> What? No. You can't do that. At some point, the TLB isn't filled, and
> then you have to do the access with lots of linear address bits
> masked.

And the 'high' bits get reloaded from the PTE.
(Although I suspect there isn't space for them.)

Reloading the high bits from the VA that caused the fault would
be 'interesting' - certainly better than just ignoring them.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)