[PATCHv2] arm64: Make sure permission updates happen for pmd/pud

[PATCHv2] arm64: Make sure permission updates happen for pmd/pud

[Laura Abbott]

Commit 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
disallowed block mappings for ioremap since that code does not honor
break-before-make. The same APIs are also used for permission updating
though and the extra checks prevent the permission updates from happening,
even though this should be permitted. This results in read-only permissions
not being fully applied. Visibly, this can occasionaly be seen as a failure
on the built in rodata test when the test data ends up in a section or
as an odd RW gap on the page table dump. Fix this by using
pgattr_change_is_safe instead of p*d_present for determining if the
change is permitted.

Reported-by: Peter Robinson <pbrobinson@gmail.com>
Fixes: 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
Signed-off-by: Laura Abbott <labbott@redhat.com>
---
v2: Switch to using pgattr_change_is_safe per suggestion of Will
---
 arch/arm64/mm/mmu.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
index 2dbb2c9f1ec1..493ff75670ff 100644
--- a/arch/arm64/mm/mmu.c
+++ b/arch/arm64/mm/mmu.c
@@ -933,13 +933,15 @@ int pud_set_huge(pud_t *pudp, phys_addr_t phys, pgprot_t prot)
 {
 	pgprot_t sect_prot = __pgprot(PUD_TYPE_SECT |
 					pgprot_val(mk_sect_prot(prot)));
+	pud_t new_pud = pfn_pud(__phys_to_pfn(phys), sect_prot);
 
-	/* ioremap_page_range doesn't honour BBM */
-	if (pud_present(READ_ONCE(*pudp)))
+	/* Only allow permission changes for now */
+	if (!pgattr_change_is_safe(READ_ONCE(pud_val(*pudp)),
+				   pud_val(new_pud)))
 		return 0;
 
 	BUG_ON(phys & ~PUD_MASK);
-	set_pud(pudp, pfn_pud(__phys_to_pfn(phys), sect_prot));
+	set_pud(pudp, new_pud);
 	return 1;
 }
 
@@ -947,13 +949,15 @@ int pmd_set_huge(pmd_t *pmdp, phys_addr_t phys, pgprot_t prot)
 {
 	pgprot_t sect_prot = __pgprot(PMD_TYPE_SECT |
 					pgprot_val(mk_sect_prot(prot)));
+	pmd_t new_pmd = pfn_pmd(__phys_to_pfn(phys), sect_prot);
 
-	/* ioremap_page_range doesn't honour BBM */
-	if (pmd_present(READ_ONCE(*pmdp)))
+	/* Only allow permission changes for now */
+	if (!pgattr_change_is_safe(READ_ONCE(pmd_val(*pmdp)),
+				   pmd_val(new_pmd)))
 		return 0;
 
 	BUG_ON(phys & ~PMD_MASK);
-	set_pmd(pmdp, pfn_pmd(__phys_to_pfn(phys), sect_prot));
+	set_pmd(pmdp, new_pmd);
 	return 1;
 }
 
-- 
2.17.0

[PATCHv2] arm64: Make sure permission updates happen for pmd/pud

[Kees Cook]

On Wed, May 23, 2018 at 11:43 AM, Laura Abbott <labbott@redhat.com> wrote:
> Commit 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> disallowed block mappings for ioremap since that code does not honor
> break-before-make. The same APIs are also used for permission updating
> though and the extra checks prevent the permission updates from happening,
> even though this should be permitted. This results in read-only permissions
> not being fully applied. Visibly, this can occasionaly be seen as a failure
> on the built in rodata test when the test data ends up in a section or
> as an odd RW gap on the page table dump. Fix this by using
> pgattr_change_is_safe instead of p*d_present for determining if the
> change is permitted.
>
> Reported-by: Peter Robinson <pbrobinson@gmail.com>
> Fixes: 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> Signed-off-by: Laura Abbott <labbott@redhat.com>

Reviewed-by: Kees Cook <keescook@chromium.org>

Thanks for fixing this!

-Kees

> ---
> v2: Switch to using pgattr_change_is_safe per suggestion of Will
> ---
>  arch/arm64/mm/mmu.c | 16 ++++++++++------
>  1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 2dbb2c9f1ec1..493ff75670ff 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -933,13 +933,15 @@ int pud_set_huge(pud_t *pudp, phys_addr_t phys, pgprot_t prot)
>  {
>         pgprot_t sect_prot = __pgprot(PUD_TYPE_SECT |
>                                         pgprot_val(mk_sect_prot(prot)));
> +       pud_t new_pud = pfn_pud(__phys_to_pfn(phys), sect_prot);
>
> -       /* ioremap_page_range doesn't honour BBM */
> -       if (pud_present(READ_ONCE(*pudp)))
> +       /* Only allow permission changes for now */
> +       if (!pgattr_change_is_safe(READ_ONCE(pud_val(*pudp)),
> +                                  pud_val(new_pud)))
>                 return 0;
>
>         BUG_ON(phys & ~PUD_MASK);
> -       set_pud(pudp, pfn_pud(__phys_to_pfn(phys), sect_prot));
> +       set_pud(pudp, new_pud);
>         return 1;
>  }
>
> @@ -947,13 +949,15 @@ int pmd_set_huge(pmd_t *pmdp, phys_addr_t phys, pgprot_t prot)
>  {
>         pgprot_t sect_prot = __pgprot(PMD_TYPE_SECT |
>                                         pgprot_val(mk_sect_prot(prot)));
> +       pmd_t new_pmd = pfn_pmd(__phys_to_pfn(phys), sect_prot);
>
> -       /* ioremap_page_range doesn't honour BBM */
> -       if (pmd_present(READ_ONCE(*pmdp)))
> +       /* Only allow permission changes for now */
> +       if (!pgattr_change_is_safe(READ_ONCE(pmd_val(*pmdp)),
> +                                  pmd_val(new_pmd)))
>                 return 0;
>
>         BUG_ON(phys & ~PMD_MASK);
> -       set_pmd(pmdp, pfn_pmd(__phys_to_pfn(phys), sect_prot));
> +       set_pmd(pmdp, new_pmd);
>         return 1;
>  }
>
> --
> 2.17.0
>



-- 
Kees Cook
Pixel Security

[PATCHv2] arm64: Make sure permission updates happen for pmd/pud

[Peter Robinson]

On Wed, May 23, 2018 at 7:43 PM, Laura Abbott <labbott@redhat.com> wrote:
> Commit 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> disallowed block mappings for ioremap since that code does not honor
> break-before-make. The same APIs are also used for permission updating
> though and the extra checks prevent the permission updates from happening,
> even though this should be permitted. This results in read-only permissions
> not being fully applied. Visibly, this can occasionaly be seen as a failure
> on the built in rodata test when the test data ends up in a section or
> as an odd RW gap on the page table dump. Fix this by using
> pgattr_change_is_safe instead of p*d_present for determining if the
> change is permitted.
>
> Reported-by: Peter Robinson <pbrobinson@gmail.com>
> Fixes: 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> Signed-off-by: Laura Abbott <labbott@redhat.com>

Tested-by: Peter Robinson <pbrobinson@gmail.com>

Tested on Macbin, mustang, pine64, RPi3+ and db410c and fixes the issue I saw.

> ---
> v2: Switch to using pgattr_change_is_safe per suggestion of Will
> ---
>  arch/arm64/mm/mmu.c | 16 ++++++++++------
>  1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 2dbb2c9f1ec1..493ff75670ff 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -933,13 +933,15 @@ int pud_set_huge(pud_t *pudp, phys_addr_t phys, pgprot_t prot)
>  {
>         pgprot_t sect_prot = __pgprot(PUD_TYPE_SECT |
>                                         pgprot_val(mk_sect_prot(prot)));
> +       pud_t new_pud = pfn_pud(__phys_to_pfn(phys), sect_prot);
>
> -       /* ioremap_page_range doesn't honour BBM */
> -       if (pud_present(READ_ONCE(*pudp)))
> +       /* Only allow permission changes for now */
> +       if (!pgattr_change_is_safe(READ_ONCE(pud_val(*pudp)),
> +                                  pud_val(new_pud)))
>                 return 0;
>
>         BUG_ON(phys & ~PUD_MASK);
> -       set_pud(pudp, pfn_pud(__phys_to_pfn(phys), sect_prot));
> +       set_pud(pudp, new_pud);
>         return 1;
>  }
>
> @@ -947,13 +949,15 @@ int pmd_set_huge(pmd_t *pmdp, phys_addr_t phys, pgprot_t prot)
>  {
>         pgprot_t sect_prot = __pgprot(PMD_TYPE_SECT |
>                                         pgprot_val(mk_sect_prot(prot)));
> +       pmd_t new_pmd = pfn_pmd(__phys_to_pfn(phys), sect_prot);
>
> -       /* ioremap_page_range doesn't honour BBM */
> -       if (pmd_present(READ_ONCE(*pmdp)))
> +       /* Only allow permission changes for now */
> +       if (!pgattr_change_is_safe(READ_ONCE(pmd_val(*pmdp)),
> +                                  pmd_val(new_pmd)))
>                 return 0;
>
>         BUG_ON(phys & ~PMD_MASK);
> -       set_pmd(pmdp, pfn_pmd(__phys_to_pfn(phys), sect_prot));
> +       set_pmd(pmdp, new_pmd);
>         return 1;
>  }
>
> --
> 2.17.0
>

[PATCHv2] arm64: Make sure permission updates happen for pmd/pud

[Will Deacon]

Hi Laura,

On Wed, May 23, 2018 at 11:43:46AM -0700, Laura Abbott wrote:
> Commit 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> disallowed block mappings for ioremap since that code does not honor
> break-before-make. The same APIs are also used for permission updating
> though and the extra checks prevent the permission updates from happening,
> even though this should be permitted. This results in read-only permissions
> not being fully applied. Visibly, this can occasionaly be seen as a failure
> on the built in rodata test when the test data ends up in a section or
> as an odd RW gap on the page table dump. Fix this by using
> pgattr_change_is_safe instead of p*d_present for determining if the
> change is permitted.
> 
> Reported-by: Peter Robinson <pbrobinson@gmail.com>
> Fixes: 15122ee2c515 ("arm64: Enforce BBM for huge IO/VMAP mappings")
> Signed-off-by: Laura Abbott <labbott@redhat.com>
> ---
> v2: Switch to using pgattr_change_is_safe per suggestion of Will
> ---

Thanks for re-spinning so quickly. I'll queue as a fix with the relevant
tags.

Will