[Xen-devel,v4,10/16] xen/mm: Switch map_pages_to_xen to use MFN typesafe

Message ID 20180221140259.29360-11-julien.grall@arm.com
State New
Headers show
Series
  • xen: Convert page_to_mfn and mfn_to_page to use typesafe MFN
Related show

Commit Message

Julien Grall Feb. 21, 2018, 2:02 p.m.
The current prototype is slightly confusing because it takes a virtual
address and a physical frame (not address!). Switching to MFN will improve
safety and reduce the chance to mistakenly invert the 2 parameters.

Signed-off-by: Julien Grall <julien.grall@arm.com>

---

Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Julien Grall <julien.grall@arm.com>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: George Dunlap <George.Dunlap@eu.citrix.com>
Cc: Ian Jackson <ian.jackson@eu.citrix.com>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Tim Deegan <tim@xen.org>
Cc: Wei Liu <wei.liu2@citrix.com>
Cc: Gang Wei <gang.wei@intel.com>
Cc: Shane Wang <shane.wang@intel.com>
Cc: Kevin Tian <kevin.tian@intel.com>

    Changes in v4:
        - Patch added
---
 xen/arch/arm/mm.c                  |  4 ++--
 xen/arch/x86/mm.c                  | 48 +++++++++++++++++++-------------------
 xen/arch/x86/setup.c               | 14 +++++------
 xen/arch/x86/smpboot.c             |  2 +-
 xen/arch/x86/tboot.c               | 11 +++++----
 xen/arch/x86/x86_64/mm.c           | 27 ++++++++++++---------
 xen/arch/x86/x86_64/mmconfig_64.c  |  4 ++--
 xen/common/efi/boot.c              |  2 +-
 xen/common/vmap.c                  | 10 +++++---
 xen/drivers/acpi/apei/erst.c       |  2 +-
 xen/drivers/acpi/apei/hest.c       |  2 +-
 xen/drivers/passthrough/vtd/dmar.c |  2 +-
 xen/include/asm-arm/mm.h           |  2 +-
 xen/include/xen/mm.h               |  2 +-
 14 files changed, 71 insertions(+), 61 deletions(-)

Comments

Tian, Kevin Feb. 23, 2018, 4:59 a.m. | #1
> From: Julien Grall [mailto:julien.grall@arm.com]
> Sent: Wednesday, February 21, 2018 10:03 PM
> 
> The current prototype is slightly confusing because it takes a virtual
> address and a physical frame (not address!). Switching to MFN will improve
> safety and reduce the chance to mistakenly invert the 2 parameters.
> 
> Signed-off-by: Julien Grall <julien.grall@arm.com>

Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Wei Liu Feb. 23, 2018, 5:21 p.m. | #2
On Wed, Feb 21, 2018 at 02:02:53PM +0000, Julien Grall wrote:
> The current prototype is slightly confusing because it takes a virtual
> address and a physical frame (not address!). Switching to MFN will improve
> safety and reduce the chance to mistakenly invert the 2 parameters.
> 
> Signed-off-by: Julien Grall <julien.grall@arm.com>

Reviewed-by: Wei Liu <wei.liu2@citrix.com>
Jan Beulich March 2, 2018, 3:06 p.m. | #3
>>> On 21.02.18 at 15:02, <julien.grall@arm.com> wrote:
> --- a/xen/arch/x86/x86_64/mm.c
> +++ b/xen/arch/x86/x86_64/mm.c
> @@ -40,6 +40,10 @@ asm(".file \"" __FILE__ "\"");
>  #include <asm/mem_sharing.h>
>  #include <public/memory.h>
>  
> +/* Override macros from asm/page.h to make them work with mfn_t */
> +#undef page_to_mfn
> +#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))

I can't spot where this is needed in this file.

> @@ -234,7 +238,7 @@ void vunmap(const void *va)
>  #ifndef _PAGE_NONE
>      destroy_xen_mappings(addr, addr + PAGE_SIZE * pages);
>  #else /* Avoid tearing down intermediate page tables. */
> -    map_pages_to_xen(addr, 0, pages, _PAGE_NONE);
> +    map_pages_to_xen(addr, _mfn(0), pages, _PAGE_NONE);

INVALID_MFN?

Also please again see about using PFN_DOWN() for some of the
sizes passed to the function as you touch those places anyway.

Jan
Jan Beulich March 2, 2018, 3:08 p.m. | #4
>>> On 02.03.18 at 16:06, <JBeulich@suse.com> wrote:
>>>> On 21.02.18 at 15:02, <julien.grall@arm.com> wrote:
>> --- a/xen/arch/x86/x86_64/mm.c
>> +++ b/xen/arch/x86/x86_64/mm.c
>> @@ -40,6 +40,10 @@ asm(".file \"" __FILE__ "\"");
>>  #include <asm/mem_sharing.h>
>>  #include <public/memory.h>
>>  
>> +/* Override macros from asm/page.h to make them work with mfn_t */
>> +#undef page_to_mfn
>> +#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
> 
> I can't spot where this is needed in this file.

Oh, I'm sorry, I should have looked at the file as well, not just
at the hunks in the patch.

Jan
Julien Grall March 5, 2018, 2:07 p.m. | #5
Hi,

On 02/03/18 15:06, Jan Beulich wrote:
>>>> On 21.02.18 at 15:02, <julien.grall@arm.com> wrote:
>> --- a/xen/arch/x86/x86_64/mm.c
>> +++ b/xen/arch/x86/x86_64/mm.c
>> @@ -40,6 +40,10 @@ asm(".file \"" __FILE__ "\"");
>>   #include <asm/mem_sharing.h>
>>   #include <public/memory.h>
>>   
>> +/* Override macros from asm/page.h to make them work with mfn_t */
>> +#undef page_to_mfn
>> +#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
> 
> I can't spot where this is needed in this file.
> 
>> @@ -234,7 +238,7 @@ void vunmap(const void *va)
>>   #ifndef _PAGE_NONE
>>       destroy_xen_mappings(addr, addr + PAGE_SIZE * pages);
>>   #else /* Avoid tearing down intermediate page tables. */
>> -    map_pages_to_xen(addr, 0, pages, _PAGE_NONE);
>> +    map_pages_to_xen(addr, _mfn(0), pages, _PAGE_NONE);
> 
> INVALID_MFN?

I can but then we end up to the same clumsiness as you mention in #8. So 
what's your preference?

> 
> Also please again see about using PFN_DOWN() for some of the
> sizes passed to the function as you touch those places anyway.

Will do.

> 
> Jan
>
Jan Beulich March 5, 2018, 2:39 p.m. | #6
>>> On 05.03.18 at 15:07, <julien.grall@arm.com> wrote:
> On 02/03/18 15:06, Jan Beulich wrote:
>>>>> On 21.02.18 at 15:02, <julien.grall@arm.com> wrote:
>>> --- a/xen/arch/x86/x86_64/mm.c
>>> +++ b/xen/arch/x86/x86_64/mm.c
>>> @@ -40,6 +40,10 @@ asm(".file \"" __FILE__ "\"");
>>>   #include <asm/mem_sharing.h>
>>>   #include <public/memory.h>
>>>   
>>> +/* Override macros from asm/page.h to make them work with mfn_t */
>>> +#undef page_to_mfn
>>> +#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
>> 
>> I can't spot where this is needed in this file.
>> 
>>> @@ -234,7 +238,7 @@ void vunmap(const void *va)
>>>   #ifndef _PAGE_NONE
>>>       destroy_xen_mappings(addr, addr + PAGE_SIZE * pages);
>>>   #else /* Avoid tearing down intermediate page tables. */
>>> -    map_pages_to_xen(addr, 0, pages, _PAGE_NONE);
>>> +    map_pages_to_xen(addr, _mfn(0), pages, _PAGE_NONE);
>> 
>> INVALID_MFN?
> 
> I can but then we end up to the same clumsiness as you mention in #8. So 
> what's your preference?

My preference is to skip the increments for INVALID_MFN in patch 8.
Since you don't want to do that, I guess we'll have to live with zero
being used everywhere.

Jan
Julien Grall March 5, 2018, 2:44 p.m. | #7
Hi,

On 05/03/18 14:39, Jan Beulich wrote:
>>>> On 05.03.18 at 15:07, <julien.grall@arm.com> wrote:
>> On 02/03/18 15:06, Jan Beulich wrote:
>>>>>> On 21.02.18 at 15:02, <julien.grall@arm.com> wrote:
>>>> --- a/xen/arch/x86/x86_64/mm.c
>>>> +++ b/xen/arch/x86/x86_64/mm.c
>>>> @@ -40,6 +40,10 @@ asm(".file \"" __FILE__ "\"");
>>>>    #include <asm/mem_sharing.h>
>>>>    #include <public/memory.h>
>>>>    
>>>> +/* Override macros from asm/page.h to make them work with mfn_t */
>>>> +#undef page_to_mfn
>>>> +#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
>>>
>>> I can't spot where this is needed in this file.
>>>
>>>> @@ -234,7 +238,7 @@ void vunmap(const void *va)
>>>>    #ifndef _PAGE_NONE
>>>>        destroy_xen_mappings(addr, addr + PAGE_SIZE * pages);
>>>>    #else /* Avoid tearing down intermediate page tables. */
>>>> -    map_pages_to_xen(addr, 0, pages, _PAGE_NONE);
>>>> +    map_pages_to_xen(addr, _mfn(0), pages, _PAGE_NONE);
>>>
>>> INVALID_MFN?
>>
>> I can but then we end up to the same clumsiness as you mention in #8. So
>> what's your preference?
> 
> My preference is to skip the increments for INVALID_MFN in patch 8.
> Since you don't want to do that, I guess we'll have to live with zero
> being used everywhere.

It is not that I don't want. I don't have any x86 setup and not that 
confident to provide a change bigger than switch from plain unsigned int 
to typesafe and light clean-up.

I can provide a patch if you really want. But likely it is going to be 
quicker for you to write it.

Cheers,

Patch

diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c
index 97dcdd5d50..f17907ace8 100644
--- a/xen/arch/arm/mm.c
+++ b/xen/arch/arm/mm.c
@@ -1065,11 +1065,11 @@  out:
 }
 
 int map_pages_to_xen(unsigned long virt,
-                     unsigned long mfn,
+                     mfn_t mfn,
                      unsigned long nr_mfns,
                      unsigned int flags)
 {
-    return create_xen_entries(INSERT, virt, _mfn(mfn), nr_mfns, flags);
+    return create_xen_entries(INSERT, virt, mfn, nr_mfns, flags);
 }
 
 int populate_pt_range(unsigned long virt, unsigned long nr_mfns)
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index ecbfc95ae2..d33b6bfa9d 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -213,7 +213,7 @@  static void __init init_frametable_chunk(void *start, void *end)
         while ( step && s + (step << PAGE_SHIFT) > e + (4 << PAGE_SHIFT) )
             step >>= PAGETABLE_ORDER;
         mfn = alloc_boot_pages(step, step);
-        map_pages_to_xen(s, mfn_x(mfn), step, PAGE_HYPERVISOR);
+        map_pages_to_xen(s, mfn, step, PAGE_HYPERVISOR);
     }
 
     memset(start, 0, end - start);
@@ -793,12 +793,12 @@  static int update_xen_mappings(unsigned long mfn, unsigned int cacheattr)
         XEN_VIRT_START + ((mfn - PFN_DOWN(xen_phys_start)) << PAGE_SHIFT);
 
     if ( unlikely(alias) && cacheattr )
-        err = map_pages_to_xen(xen_va, mfn, 1, 0);
+        err = map_pages_to_xen(xen_va, _mfn(mfn), 1, 0);
     if ( !err )
-        err = map_pages_to_xen((unsigned long)mfn_to_virt(mfn), mfn, 1,
+        err = map_pages_to_xen((unsigned long)mfn_to_virt(mfn), _mfn(mfn), 1,
                      PAGE_HYPERVISOR | cacheattr_to_pte_flags(cacheattr));
     if ( unlikely(alias) && !cacheattr && !err )
-        err = map_pages_to_xen(xen_va, mfn, 1, PAGE_HYPERVISOR);
+        err = map_pages_to_xen(xen_va, _mfn(mfn), 1, PAGE_HYPERVISOR);
     return err;
 }
 
@@ -4655,7 +4655,7 @@  l1_pgentry_t *virt_to_xen_l1e(unsigned long v)
 
 int map_pages_to_xen(
     unsigned long virt,
-    unsigned long mfn,
+    mfn_t mfn,
     unsigned long nr_mfns,
     unsigned int flags)
 {
@@ -4687,13 +4687,13 @@  int map_pages_to_xen(
         ol3e = *pl3e;
 
         if ( cpu_has_page1gb &&
-             !(((virt >> PAGE_SHIFT) | mfn) &
+             !(((virt >> PAGE_SHIFT) | mfn_x(mfn)) &
                ((1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1)) &&
              nr_mfns >= (1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) &&
              !(flags & (_PAGE_PAT | MAP_SMALL_PAGES)) )
         {
             /* 1GB-page mapping. */
-            l3e_write_atomic(pl3e, l3e_from_pfn(mfn, l1f_to_lNf(flags)));
+            l3e_write_atomic(pl3e, l3e_from_mfn(mfn, l1f_to_lNf(flags)));
 
             if ( (l3e_get_flags(ol3e) & _PAGE_PRESENT) )
             {
@@ -4737,7 +4737,7 @@  int map_pages_to_xen(
             }
 
             virt    += 1UL << L3_PAGETABLE_SHIFT;
-            mfn     += 1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT);
+            mfn     = mfn_add(mfn, 1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT));
             nr_mfns -= 1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT);
             continue;
         }
@@ -4752,17 +4752,17 @@  int map_pages_to_xen(
             if ( ((l3e_get_pfn(ol3e) & ~(L2_PAGETABLE_ENTRIES *
                                          L1_PAGETABLE_ENTRIES - 1)) +
                   (l2_table_offset(virt) << PAGETABLE_ORDER) +
-                  l1_table_offset(virt) == mfn) &&
+                  l1_table_offset(virt) == mfn_x(mfn)) &&
                  ((lNf_to_l1f(l3e_get_flags(ol3e)) ^ flags) &
                   ~(_PAGE_ACCESSED|_PAGE_DIRTY)) == 0 )
             {
                 /* We can skip to end of L3 superpage if we got a match. */
                 i = (1u << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) -
-                    (mfn & ((1 << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1));
+                    (mfn_x(mfn) & ((1 << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1));
                 if ( i > nr_mfns )
                     i = nr_mfns;
                 virt    += i << PAGE_SHIFT;
-                mfn     += i;
+                mfn     = mfn_add(mfn, i);
                 nr_mfns -= i;
                 continue;
             }
@@ -4800,14 +4800,14 @@  int map_pages_to_xen(
         if ( !pl2e )
             return -ENOMEM;
 
-        if ( ((((virt >> PAGE_SHIFT) | mfn) &
+        if ( ((((virt >> PAGE_SHIFT) | mfn_x(mfn)) &
                ((1u << PAGETABLE_ORDER) - 1)) == 0) &&
              (nr_mfns >= (1u << PAGETABLE_ORDER)) &&
              !(flags & (_PAGE_PAT|MAP_SMALL_PAGES)) )
         {
             /* Super-page mapping. */
             ol2e = *pl2e;
-            l2e_write_atomic(pl2e, l2e_from_pfn(mfn, l1f_to_lNf(flags)));
+            l2e_write_atomic(pl2e, l2e_from_mfn(mfn, l1f_to_lNf(flags)));
 
             if ( (l2e_get_flags(ol2e) & _PAGE_PRESENT) )
             {
@@ -4830,7 +4830,7 @@  int map_pages_to_xen(
             }
 
             virt    += 1UL << L2_PAGETABLE_SHIFT;
-            mfn     += 1UL << PAGETABLE_ORDER;
+            mfn     = mfn_add(mfn, 1UL << PAGETABLE_ORDER);
             nr_mfns -= 1UL << PAGETABLE_ORDER;
         }
         else
@@ -4849,17 +4849,17 @@  int map_pages_to_xen(
 
                 /* Skip this PTE if there is no change. */
                 if ( (((l2e_get_pfn(*pl2e) & ~(L1_PAGETABLE_ENTRIES - 1)) +
-                       l1_table_offset(virt)) == mfn) &&
+                       l1_table_offset(virt)) == mfn_x(mfn)) &&
                      (((lNf_to_l1f(l2e_get_flags(*pl2e)) ^ flags) &
                        ~(_PAGE_ACCESSED|_PAGE_DIRTY)) == 0) )
                 {
                     /* We can skip to end of L2 superpage if we got a match. */
                     i = (1u << (L2_PAGETABLE_SHIFT - PAGE_SHIFT)) -
-                        (mfn & ((1u << (L2_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1));
+                        (mfn_x(mfn) & ((1u << (L2_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1));
                     if ( i > nr_mfns )
                         i = nr_mfns;
                     virt    += i << L1_PAGETABLE_SHIFT;
-                    mfn     += i;
+                    mfn     = mfn_add(mfn, i);
                     nr_mfns -= i;
                     goto check_l3;
                 }
@@ -4894,7 +4894,7 @@  int map_pages_to_xen(
 
             pl1e  = l2e_to_l1e(*pl2e) + l1_table_offset(virt);
             ol1e  = *pl1e;
-            l1e_write_atomic(pl1e, l1e_from_pfn(mfn, flags));
+            l1e_write_atomic(pl1e, l1e_from_mfn(mfn, flags));
             if ( (l1e_get_flags(ol1e) & _PAGE_PRESENT) )
             {
                 unsigned int flush_flags = FLUSH_TLB | FLUSH_ORDER(0);
@@ -4904,12 +4904,12 @@  int map_pages_to_xen(
             }
 
             virt    += 1UL << L1_PAGETABLE_SHIFT;
-            mfn     += 1UL;
+            mfn     = mfn_add(mfn, 1UL);
             nr_mfns -= 1UL;
 
             if ( (flags == PAGE_HYPERVISOR) &&
                  ((nr_mfns == 0) ||
-                  ((((virt >> PAGE_SHIFT) | mfn) &
+                  ((((virt >> PAGE_SHIFT) | mfn_x(mfn)) &
                     ((1u << PAGETABLE_ORDER) - 1)) == 0)) )
             {
                 unsigned long base_mfn;
@@ -4962,7 +4962,7 @@  int map_pages_to_xen(
         if ( cpu_has_page1gb &&
              (flags == PAGE_HYPERVISOR) &&
              ((nr_mfns == 0) ||
-              !(((virt >> PAGE_SHIFT) | mfn) &
+              !(((virt >> PAGE_SHIFT) | mfn_x(mfn)) &
                 ((1UL << (L3_PAGETABLE_SHIFT - PAGE_SHIFT)) - 1))) )
         {
             unsigned long base_mfn;
@@ -5014,7 +5014,7 @@  int map_pages_to_xen(
 
 int populate_pt_range(unsigned long virt, unsigned long nr_mfns)
 {
-    return map_pages_to_xen(virt, mfn_x(INVALID_MFN), nr_mfns, MAP_SMALL_PAGES);
+    return map_pages_to_xen(virt, INVALID_MFN, nr_mfns, MAP_SMALL_PAGES);
 }
 
 /*
@@ -5275,7 +5275,7 @@  void __set_fixmap(
     enum fixed_addresses idx, unsigned long mfn, unsigned long flags)
 {
     BUG_ON(idx >= __end_of_fixed_addresses);
-    map_pages_to_xen(__fix_to_virt(idx), mfn, 1, flags);
+    map_pages_to_xen(__fix_to_virt(idx), _mfn(mfn), 1, flags);
 }
 
 void *__init arch_vmap_virt_end(void)
@@ -5546,7 +5546,7 @@  static void __memguard_change_range(void *p, unsigned long l, int guard)
     if ( guard )
         flags &= ~_PAGE_PRESENT;
 
-    map_pages_to_xen(_p, mfn_x(virt_to_mfn(p)), _l >> PAGE_SHIFT, flags);
+    map_pages_to_xen(_p, virt_to_mfn(p), _l >> PAGE_SHIFT, flags);
 }
 
 void memguard_guard_range(void *p, unsigned long l)
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index ac530ece2c..fabd689f6e 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -354,7 +354,7 @@  void *__init bootstrap_map(const module_t *mod)
     if ( end - start > BOOTSTRAP_MAP_LIMIT - map_cur )
         return NULL;
 
-    map_pages_to_xen(map_cur, start >> PAGE_SHIFT,
+    map_pages_to_xen(map_cur, maddr_to_mfn(start),
                      (end - start) >> PAGE_SHIFT, PAGE_HYPERVISOR);
     map_cur += end - start;
     return ret;
@@ -977,7 +977,7 @@  void __init noreturn __start_xen(unsigned long mbi_p)
         {
             end = min(e, limit);
             set_pdx_range(s >> PAGE_SHIFT, end >> PAGE_SHIFT);
-            map_pages_to_xen((unsigned long)__va(s), s >> PAGE_SHIFT,
+            map_pages_to_xen((unsigned long)__va(s), maddr_to_mfn(s),
                              (end - s) >> PAGE_SHIFT, PAGE_HYPERVISOR);
         }
 
@@ -1292,7 +1292,7 @@  void __init noreturn __start_xen(unsigned long mbi_p)
 
             if ( map_e < end )
             {
-                map_pages_to_xen((unsigned long)__va(map_e), PFN_DOWN(map_e),
+                map_pages_to_xen((unsigned long)__va(map_e), maddr_to_mfn(map_e),
                                  PFN_DOWN(end - map_e), PAGE_HYPERVISOR);
                 init_boot_pages(map_e, end);
                 map_e = end;
@@ -1302,12 +1302,12 @@  void __init noreturn __start_xen(unsigned long mbi_p)
         {
             /* This range must not be passed to the boot allocator and
              * must also not be mapped with _PAGE_GLOBAL. */
-            map_pages_to_xen((unsigned long)__va(map_e), PFN_DOWN(map_e),
+            map_pages_to_xen((unsigned long)__va(map_e), maddr_to_mfn(map_e),
                              PFN_DOWN(e - map_e), __PAGE_HYPERVISOR_RW);
         }
         if ( s < map_s )
         {
-            map_pages_to_xen((unsigned long)__va(s), s >> PAGE_SHIFT,
+            map_pages_to_xen((unsigned long)__va(s), maddr_to_mfn(s),
                              (map_s - s) >> PAGE_SHIFT, PAGE_HYPERVISOR);
             init_boot_pages(s, map_s);
         }
@@ -1318,7 +1318,7 @@  void __init noreturn __start_xen(unsigned long mbi_p)
         set_pdx_range(mod[i].mod_start,
                       mod[i].mod_start + PFN_UP(mod[i].mod_end));
         map_pages_to_xen((unsigned long)mfn_to_virt(mod[i].mod_start),
-                         mod[i].mod_start,
+                         _mfn(mod[i].mod_start),
                          PFN_UP(mod[i].mod_end), PAGE_HYPERVISOR);
     }
 
@@ -1331,7 +1331,7 @@  void __init noreturn __start_xen(unsigned long mbi_p)
 
         if ( e > s ) 
             map_pages_to_xen((unsigned long)__va(kexec_crash_area.start),
-                             s, e - s, PAGE_HYPERVISOR);
+                             _mfn(s), e - s, PAGE_HYPERVISOR);
     }
 #endif
 
diff --git a/xen/arch/x86/smpboot.c b/xen/arch/x86/smpboot.c
index db264bd7bf..52ad43e0ed 100644
--- a/xen/arch/x86/smpboot.c
+++ b/xen/arch/x86/smpboot.c
@@ -626,7 +626,7 @@  unsigned long alloc_stub_page(unsigned int cpu, unsigned long *mfn)
     BUILD_BUG_ON(NR_CPUS * PAGE_SIZE > (1u << L3_PAGETABLE_SHIFT));
 
     stub_va = XEN_VIRT_END - (cpu + 1) * PAGE_SIZE;
-    if ( map_pages_to_xen(stub_va, mfn_x(page_to_mfn(pg)), 1,
+    if ( map_pages_to_xen(stub_va, page_to_mfn(pg), 1,
                           PAGE_HYPERVISOR_RX | MAP_SMALL_PAGES) )
     {
         if ( !*mfn )
diff --git a/xen/arch/x86/tboot.c b/xen/arch/x86/tboot.c
index d36bf33407..71e757c553 100644
--- a/xen/arch/x86/tboot.c
+++ b/xen/arch/x86/tboot.c
@@ -336,22 +336,23 @@  static void tboot_gen_frametable_integrity(const uint8_t key[TB_KEY_SIZE],
 
 void tboot_shutdown(uint32_t shutdown_type)
 {
-    uint32_t map_base, map_size;
+    mfn_t map_base;
+    uint32_t map_size;
     int err;
 
     g_tboot_shared->shutdown_type = shutdown_type;
 
     /* Create identity map for tboot shutdown code. */
     /* do before S3 integrity because mapping tboot may change xenheap */
-    map_base = PFN_DOWN(g_tboot_shared->tboot_base);
+    map_base = maddr_to_mfn(g_tboot_shared->tboot_base);
     map_size = PFN_UP(g_tboot_shared->tboot_size);
 
-    err = map_pages_to_xen(map_base << PAGE_SHIFT, map_base, map_size,
+    err = map_pages_to_xen(mfn_to_maddr(map_base), map_base, map_size,
                            __PAGE_HYPERVISOR);
     if ( err != 0 )
     {
-        printk("error (%#x) mapping tboot pages (mfns) @ %#x, %#x\n", err,
-               map_base, map_size);
+        printk("error (%#x) mapping tboot pages (mfns) @ %"PRI_mfn", %#x\n",
+               err, mfn_x(map_base), map_size);
         return;
     }
 
diff --git a/xen/arch/x86/x86_64/mm.c b/xen/arch/x86/x86_64/mm.c
index 9b37da6698..cfbc8ecf16 100644
--- a/xen/arch/x86/x86_64/mm.c
+++ b/xen/arch/x86/x86_64/mm.c
@@ -40,6 +40,10 @@  asm(".file \"" __FILE__ "\"");
 #include <asm/mem_sharing.h>
 #include <public/memory.h>
 
+/* Override macros from asm/page.h to make them work with mfn_t */
+#undef page_to_mfn
+#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
+
 unsigned int __read_mostly m2p_compat_vstart = __HYPERVISOR_COMPAT_VIRT_START;
 
 l2_pgentry_t *compat_idle_pg_table_l2;
@@ -111,14 +115,14 @@  static int hotadd_mem_valid(unsigned long pfn, struct mem_hotadd_info *info)
     return (pfn < info->epfn && pfn >= info->spfn);
 }
 
-static unsigned long alloc_hotadd_mfn(struct mem_hotadd_info *info)
+static mfn_t alloc_hotadd_mfn(struct mem_hotadd_info *info)
 {
-    unsigned mfn;
+    mfn_t mfn;
 
     ASSERT((info->cur + ( 1UL << PAGETABLE_ORDER) < info->epfn) &&
             info->cur >= info->spfn);
 
-    mfn = info->cur;
+    mfn = _mfn(info->cur);
     info->cur += (1UL << PAGETABLE_ORDER);
     return mfn;
 }
@@ -317,7 +321,8 @@  static void destroy_m2p_mapping(struct mem_hotadd_info *info)
  */
 static int setup_compat_m2p_table(struct mem_hotadd_info *info)
 {
-    unsigned long i, va, smap, emap, rwva, epfn = info->epfn, mfn;
+    unsigned long i, va, smap, emap, rwva, epfn = info->epfn;
+    mfn_t mfn;
     unsigned int n;
     l3_pgentry_t *l3_ro_mpt = NULL;
     l2_pgentry_t *l2_ro_mpt = NULL;
@@ -378,7 +383,7 @@  static int setup_compat_m2p_table(struct mem_hotadd_info *info)
         memset((void *)rwva, 0xFF, 1UL << L2_PAGETABLE_SHIFT);
         /* NB. Cannot be GLOBAL as the ptes get copied into per-VM space. */
         l2e_write(&l2_ro_mpt[l2_table_offset(va)],
-                  l2e_from_pfn(mfn, _PAGE_PSE|_PAGE_PRESENT));
+                  l2e_from_mfn(mfn, _PAGE_PSE|_PAGE_PRESENT));
     }
 #undef CNT
 #undef MFN
@@ -438,7 +443,7 @@  static int setup_m2p_table(struct mem_hotadd_info *info)
                 break;
         if ( n < CNT )
         {
-            unsigned long mfn = alloc_hotadd_mfn(info);
+            mfn_t mfn = alloc_hotadd_mfn(info);
 
             ret = map_pages_to_xen(
                         RDWR_MPT_VIRT_START + i * sizeof(unsigned long),
@@ -473,7 +478,7 @@  static int setup_m2p_table(struct mem_hotadd_info *info)
             }
 
             /* NB. Cannot be GLOBAL: guest user mode should not see it. */
-            l2e_write(l2_ro_mpt, l2e_from_pfn(mfn,
+            l2e_write(l2_ro_mpt, l2e_from_mfn(mfn,
                    /*_PAGE_GLOBAL|*/_PAGE_PSE|_PAGE_USER|_PAGE_PRESENT));
         }
         if ( !((unsigned long)l2_ro_mpt & ~PAGE_MASK) )
@@ -692,7 +697,7 @@  void __init zap_low_mappings(void)
     flush_local(FLUSH_TLB_GLOBAL);
 
     /* Replace with mapping of the boot trampoline only. */
-    map_pages_to_xen(trampoline_phys, trampoline_phys >> PAGE_SHIFT,
+    map_pages_to_xen(trampoline_phys, maddr_to_mfn(trampoline_phys),
                      PFN_UP(trampoline_end - trampoline_start),
                      __PAGE_HYPERVISOR);
 }
@@ -769,7 +774,7 @@  static int setup_frametable_chunk(void *start, void *end,
 {
     unsigned long s = (unsigned long)start;
     unsigned long e = (unsigned long)end;
-    unsigned long mfn;
+    mfn_t mfn;
     int err;
 
     ASSERT(!(s & ((1 << L2_PAGETABLE_SHIFT) - 1)));
@@ -1368,7 +1373,7 @@  int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
     i = virt_to_mfn(HYPERVISOR_VIRT_END - 1) + 1;
     if ( spfn < i )
     {
-        ret = map_pages_to_xen((unsigned long)mfn_to_virt(spfn), spfn,
+        ret = map_pages_to_xen((unsigned long)mfn_to_virt(spfn), _mfn(spfn),
                                min(epfn, i) - spfn, PAGE_HYPERVISOR);
         if ( ret )
             goto destroy_directmap;
@@ -1377,7 +1382,7 @@  int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)
     {
         if ( i < spfn )
             i = spfn;
-        ret = map_pages_to_xen((unsigned long)mfn_to_virt(i), i,
+        ret = map_pages_to_xen((unsigned long)mfn_to_virt(i), _mfn(i),
                                epfn - i, __PAGE_HYPERVISOR_RW);
         if ( ret )
             goto destroy_directmap;
diff --git a/xen/arch/x86/x86_64/mmconfig_64.c b/xen/arch/x86/x86_64/mmconfig_64.c
index 958b6cf2f4..c68004fc4d 100644
--- a/xen/arch/x86/x86_64/mmconfig_64.c
+++ b/xen/arch/x86/x86_64/mmconfig_64.c
@@ -125,8 +125,8 @@  static void __iomem *mcfg_ioremap(const struct acpi_mcfg_allocation *cfg,
         return NULL;
 
     if (map_pages_to_xen(virt,
-                         (cfg->address >> PAGE_SHIFT) +
-                         (cfg->start_bus_number << (20 - PAGE_SHIFT)),
+                         mfn_add(maddr_to_mfn(cfg->address),
+                                 (cfg->start_bus_number << (20 - PAGE_SHIFT))),
                          size >> PAGE_SHIFT, prot))
         return NULL;
 
diff --git a/xen/common/efi/boot.c b/xen/common/efi/boot.c
index 469bf980cc..64d12685d3 100644
--- a/xen/common/efi/boot.c
+++ b/xen/common/efi/boot.c
@@ -1464,7 +1464,7 @@  void __init efi_init_memory(void)
             if ( (unsigned long)mfn_to_virt(emfn - 1) >= HYPERVISOR_VIRT_END )
                 prot &= ~_PAGE_GLOBAL;
             if ( map_pages_to_xen((unsigned long)mfn_to_virt(smfn),
-                                  smfn, emfn - smfn, prot) == 0 )
+                                  _mfn(smfn), emfn - smfn, prot) == 0 )
                 desc->VirtualStart =
                     (unsigned long)maddr_to_virt(desc->PhysicalStart);
             else
diff --git a/xen/common/vmap.c b/xen/common/vmap.c
index 11785ffb0a..1f50c91789 100644
--- a/xen/common/vmap.c
+++ b/xen/common/vmap.c
@@ -9,6 +9,10 @@ 
 #include <xen/vmap.h>
 #include <asm/page.h>
 
+/* Override macros from asm/page.h to make them work with mfn_t */
+#undef page_to_mfn
+#define page_to_mfn(pg) _mfn(__page_to_mfn(pg))
+
 static DEFINE_SPINLOCK(vm_lock);
 static void *__read_mostly vm_base[VMAP_REGION_NR];
 #define vm_bitmap(x) ((unsigned long *)vm_base[x])
@@ -208,7 +212,7 @@  void *__vmap(const mfn_t *mfn, unsigned int granularity,
 
     for ( ; va && nr--; ++mfn, cur += PAGE_SIZE * granularity )
     {
-        if ( map_pages_to_xen(cur, mfn_x(*mfn), granularity, flags) )
+        if ( map_pages_to_xen(cur, *mfn, granularity, flags) )
         {
             vunmap(va);
             va = NULL;
@@ -234,7 +238,7 @@  void vunmap(const void *va)
 #ifndef _PAGE_NONE
     destroy_xen_mappings(addr, addr + PAGE_SIZE * pages);
 #else /* Avoid tearing down intermediate page tables. */
-    map_pages_to_xen(addr, 0, pages, _PAGE_NONE);
+    map_pages_to_xen(addr, _mfn(0), pages, _PAGE_NONE);
 #endif
     vm_free(va);
 }
@@ -258,7 +262,7 @@  static void *vmalloc_type(size_t size, enum vmap_region type)
         pg = alloc_domheap_page(NULL, 0);
         if ( pg == NULL )
             goto error;
-        mfn[i] = _mfn(page_to_mfn(pg));
+        mfn[i] = page_to_mfn(pg);
     }
 
     va = __vmap(mfn, 1, pages, 1, PAGE_HYPERVISOR, type);
diff --git a/xen/drivers/acpi/apei/erst.c b/xen/drivers/acpi/apei/erst.c
index 14acf5d773..7fc4de5de9 100644
--- a/xen/drivers/acpi/apei/erst.c
+++ b/xen/drivers/acpi/apei/erst.c
@@ -799,7 +799,7 @@  int __init erst_init(void)
 		printk(KERN_WARNING "Failed to get ERST table: %s\n", msg);
 		return -EINVAL;
 	}
-	map_pages_to_xen((unsigned long)__va(erst_addr), PFN_DOWN(erst_addr),
+	map_pages_to_xen((unsigned long)__va(erst_addr), maddr_to_mfn(erst_addr),
 			 PFN_UP(erst_addr + erst_len) - PFN_DOWN(erst_addr),
 			 PAGE_HYPERVISOR);
 	erst_tab = __va(erst_addr);
diff --git a/xen/drivers/acpi/apei/hest.c b/xen/drivers/acpi/apei/hest.c
index f74e7c2a06..70734ab0e2 100644
--- a/xen/drivers/acpi/apei/hest.c
+++ b/xen/drivers/acpi/apei/hest.c
@@ -184,7 +184,7 @@  void __init acpi_hest_init(void)
 		       acpi_format_exception(status));
 		goto err;
 	}
-	map_pages_to_xen((unsigned long)__va(hest_addr), PFN_DOWN(hest_addr),
+	map_pages_to_xen((unsigned long)__va(hest_addr), maddr_to_mfn(hest_addr),
 			 PFN_UP(hest_addr + hest_len) - PFN_DOWN(hest_addr),
 			 PAGE_HYPERVISOR);
 	hest_tab = __va(hest_addr);
diff --git a/xen/drivers/passthrough/vtd/dmar.c b/xen/drivers/passthrough/vtd/dmar.c
index d713a8ca5d..46decd4eb1 100644
--- a/xen/drivers/passthrough/vtd/dmar.c
+++ b/xen/drivers/passthrough/vtd/dmar.c
@@ -1008,7 +1008,7 @@  int __init acpi_dmar_init(void)
     if ( ACPI_SUCCESS(acpi_get_table_phys(ACPI_SIG_DMAR, 0,
                                           &dmar_addr, &dmar_len)) )
     {
-        map_pages_to_xen((unsigned long)__va(dmar_addr), PFN_DOWN(dmar_addr),
+        map_pages_to_xen((unsigned long)__va(dmar_addr), maddr_to_mfn(dmar_addr),
                          PFN_UP(dmar_addr + dmar_len) - PFN_DOWN(dmar_addr),
                          PAGE_HYPERVISOR);
         dmar_table = __va(dmar_addr);
diff --git a/xen/include/asm-arm/mm.h b/xen/include/asm-arm/mm.h
index 7678e29c15..023e2eb213 100644
--- a/xen/include/asm-arm/mm.h
+++ b/xen/include/asm-arm/mm.h
@@ -138,7 +138,7 @@  extern vaddr_t xenheap_virt_start;
 #endif
 
 #ifdef CONFIG_ARM_32
-#define is_xen_heap_page(page) is_xen_heap_mfn(page_to_mfn(page))
+#define is_xen_heap_page(page) is_xen_heap_mfn(__page_to_mfn(page))
 #define is_xen_heap_mfn(mfn) ({                                 \
     unsigned long mfn_ = (mfn);                                 \
     (mfn_ >= mfn_x(xenheap_mfn_start) &&                        \
diff --git a/xen/include/xen/mm.h b/xen/include/xen/mm.h
index f2c6738ad2..caad06e753 100644
--- a/xen/include/xen/mm.h
+++ b/xen/include/xen/mm.h
@@ -165,7 +165,7 @@  bool scrub_free_pages(void);
 /* Map machine page range in Xen virtual address space. */
 int map_pages_to_xen(
     unsigned long virt,
-    unsigned long mfn,
+    mfn_t mfn,
     unsigned long nr_mfns,
     unsigned int flags);
 /* Alter the permissions of a range of Xen virtual address space. */