Re: [PATCH RFC 01/14] x86/asm: add iosubmit_cmds512() based on movdir64b CPU instruction

[Dave Jiang]

On 11/21/19 3:59 AM, Borislav Petkov wrote:
> On Wed, Nov 20, 2019 at 05:10:41PM -0700, Dave Jiang wrote:
> > I'll add the check on the destination address. The call is modeled after
> > __iowrite64_copy() / __iowrite32_copy() in lib/iomap_copy.c. Looks like
> > those functions do not check for the alignment requirements either.
>
> So just because they don't check, you don't need to check either?

No what I mean was those primitives are missing the checks and we should 
probably address that at some point.

> Can you guarantee that all callers will always do the right thing?
>
> I mean, if you don't care too much, why even write "(must be 512-bit
> aligned)"? Who cares then if the data is aligned or not...


> > > > + * @dst: destination, in MMIO space (must be 512-bit aligned)
> > > > + * @src: source
> > > > + * @count: number of 512 bits quantities to submit
> > >
> > > Where's that check on the data?
> >
> > I don't follow?
>
> What do you do if the caller doesn't submit data in 512 bits quantities?

How would I detect that? Add a size (in bytes) parameter for the total 
source data?