Re: [PATCH v5 3/3] x86/tdx: Add Quote generation support

[Sathyanarayanan Kuppuswamy]

On 5/5/22 3:15 PM, Kai Huang wrote:
> On Thu, 2022-05-05 at 13:53 -0700, Sathyanarayanan Kuppuswamy wrote:
> > Hi Kai,
> >
> > On 5/4/22 4:28 PM, Kai Huang wrote:
> > > On Wed, 2022-05-04 at 15:49 -0700, Sathyanarayanan Kuppuswamy wrote:
> > > > --- a/arch/x86/coco/tdx/tdx.c
> > > > +++ b/arch/x86/coco/tdx/tdx.c
> > > > @@ -15,6 +15,7 @@
> > > >     #include <asm/idtentry.h>
> > > >     #include <asm/irq_regs.h>
> > > >     #include <asm/desc.h>
> > > > +#include <asm/io.h>
> > > >
> > > >     /* TDX module Call Leaf IDs */
> > > >     #define TDX_GET_INFO                   1
> > > > @@ -680,8 +681,15 @@ static bool try_accept_one(phys_addr_t *start,
> > > > unsigned long len,
> > > >      */
> > > >     static bool tdx_enc_status_changed(unsigned long vaddr, int numpages,
> > > > bool enc)
> > > >     {
> > > > -       phys_addr_t start = __pa(vaddr);
> > > > -       phys_addr_t end   = __pa(vaddr + numpages * PAGE_SIZE);
> > > > +       phys_addr_t start;
> > > > +       phys_addr_t end;
> > > > +
> > > > +       if (is_vmalloc_addr((void *)vaddr))
> > > > +               start =  page_to_phys(vmalloc_to_page((void*)vaddr));
> > > > +       else
> > > > +               start = __pa(vaddr);
> > > > +
> > > > +       end = start + numpages * PAGE_SIZE;
> > > >
> > > >            if (!enc) {
> > > >                    /* Set the shared (decrypted) bits: */
> > >
> > > Looks set_memory_decrypted() only works for direct-mapping, so you should not
> > > use this.  Instead, you can pass shared bit in 'prot' argument (using
> > > pgprot_decrypted()) when you call vmap(), and explicitly call MapGPA().
> >
> > Is it because of the above change, or you see other direct-mapping
> > dependencies in set_memory_*() functions?
> set_memory_xx()  is supposedly only for direct-mapping.  Please use my
> suggestion above.

I did not find any other direct-mapping dependency in set_memory_*()
functions other than what I have fixed. If I missed anything, please
let me know.

Also, even if set_memory_*() functions does not support vmalloc'ed
memory,  IMO, it is better to add this support to it.

I want to avoid custom solution if it is possible to use generic
function.

>

--
Sathyanarayanan Kuppuswamy
Linux Kernel Developer