Re: [PATCH v7 5/6] KVM: arm64: Initialize the ptdump parser with stage-2 attributes

[Sebastian Ene <sebastianene@google.com>]

On Mon, Jul 01, 2024 at 09:42:43AM +0100, Vincent Donnefort wrote:
> O Fri, Jun 21, 2024 at 12:32:29PM +0000, 'Sebastian Ene' via kernel-team wrote:
> > Define a set of attributes used by the ptdump parser to display the
> > properties of a guest memory region covered by a pagetable descriptor.
> > Build a description of the pagetable levels and initialize the parser
> > with this configuration.
> > 
> > Signed-off-by: Sebastian Ene <sebastianene@google.com>
> > ---
> >  arch/arm64/kvm/ptdump.c | 143 ++++++++++++++++++++++++++++++++++++++--
> >  1 file changed, 137 insertions(+), 6 deletions(-)
> > 
> > diff --git a/arch/arm64/kvm/ptdump.c b/arch/arm64/kvm/ptdump.c
> > index 36dc7662729f..cc1d4fdddc6e 100644
> > --- a/arch/arm64/kvm/ptdump.c
> > +++ b/arch/arm64/kvm/ptdump.c
> > @@ -14,6 +14,61 @@
> >  #include <kvm_ptdump.h>
> >  
> >  
> > +#define MARKERS_LEN		(2)
> > +#define KVM_PGTABLE_MAX_LEVELS	(KVM_PGTABLE_LAST_LEVEL + 1)
> > +
> > +struct kvm_ptdump_guest_state {
> > +	struct kvm		*kvm;
> > +	struct pg_state		parser_state;
> > +	struct addr_marker	ipa_marker[MARKERS_LEN];
> > +	struct pg_level		level[KVM_PGTABLE_MAX_LEVELS];
> > +	struct ptdump_range	range[MARKERS_LEN];
> > +};
> > +
> > +static const struct prot_bits stage2_pte_bits[] = {
> > +	{
> > +		.mask	= PTE_VALID,
> > +		.val	= PTE_VALID,
> > +		.set	= " ",
> > +		.clear	= "F",
> > +	}, {
> > +		.mask	= KVM_PTE_LEAF_ATTR_HI_S2_XN | PTE_VALID,
> > +		.val	= KVM_PTE_LEAF_ATTR_HI_S2_XN | PTE_VALID,
> > +		.set	= "XN",
> > +		.clear	= "  ",
> > +	}, {
> > +		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R | PTE_VALID,
> > +		.val	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R | PTE_VALID,
> > +		.set	= "R",
> > +		.clear	= " ",
> > +	}, {
> > +		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W | PTE_VALID,
> > +		.val	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W | PTE_VALID,
> > +		.set	= "W",
> > +		.clear	= " ",
> > +	}, {
> > +		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_AF | PTE_VALID,
> > +		.val	= KVM_PTE_LEAF_ATTR_LO_S2_AF | PTE_VALID,
> > +		.set	= "AF",
> > +		.clear	= "  ",
> > +	}, {
> > +		.mask	= PTE_NG,
> > +		.val	= PTE_NG,
> > +		.set	= "FnXS",
> > +		.clear	= "  ",
> > +	}, {
> > +		.mask	= PTE_CONT | PTE_VALID,
> > +		.val	= PTE_CONT | PTE_VALID,
> > +		.set	= "CON",
> > +		.clear	= "   ",
> > +	}, {
> > +		.mask	= PTE_TABLE_BIT,
> >
> > +		.val	= PTE_TABLE_BIT,
> > +		.set	= "   ",
> > +		.clear	= "BLK",
> > +	},

Hello Vincent,

> 
> When doing a kvm_pgtable_stage2_set_owner(), the walker will init a leaf which
> has both the table-bit and the valid-bit unset. I believe this would lead to
> spurious BLK annotations here.
> 
> The following should fix this problem:
> 
>   .mask		= PTE_TABLE_BIT | PTE_VALID,
>   .val		= PTE_VALID,
>   .set		= "BLK",
>   .clear	= "   ",
> 

Let me try this, thanks for the suggestion !

> > +};
> > +
> >  static int kvm_ptdump_visitor(const struct kvm_pgtable_visit_ctx *ctx,
> >  			      enum kvm_pgtable_walk_flags visit)
> >  {
> > @@ -40,15 +95,79 @@ static int kvm_ptdump_show_common(struct seq_file *m,
> >  	return kvm_pgtable_walk(pgtable, 0, BIT(pgtable->ia_bits), &walker);
> >  }
> >
> 
> [...]

Seb