Xen devel - Jul 2013 - [PATCH] x86: fix cache flushing condition in map_pages_to_xen()

This fixes yet another shortcoming of the function (exposed by 8bfaa2c2
["x86: add locking to map_pages_to_xen()"]''s adjustment to
msix_put_fixmap()): It must not flush caches when transitioning to a
non-present mapping. Doing so causes the CLFLUSH to fault, if used in
favor of WBINVD.

To help code readability, factor out the whole flush flags updating
in map_pages_to_xen() into a helper macro.

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -5430,6 +5430,15 @@ l1_pgentry_t *virt_to_xen_l1e(unsigned l
                          flush_area_local((const void *)v, f) : \
                          flush_area_all((const void *)v, f))
 
+#define flush_flags(oldf) do {                                  \
+    unsigned int o_ = (oldf);                                   \
+    if ( (o_) & _PAGE_GLOBAL )                                  \
+        flush_flags |= FLUSH_TLB_GLOBAL;                        \
+    if ( (flags & _PAGE_PRESENT) &&                             \
+         (((o_) ^ flags) & PAGE_CACHE_ATTRS) )                  \
+        flush_flags |= FLUSH_CACHE;                             \
+} while (0)
+
 int map_pages_to_xen(
     unsigned long virt,
     unsigned long mfn,
@@ -5465,11 +5474,7 @@ int map_pages_to_xen(
 
                 if ( l3e_get_flags(ol3e) & _PAGE_PSE )
                 {
-                    if ( l3e_get_flags(ol3e) & _PAGE_GLOBAL )
-                        flush_flags |= FLUSH_TLB_GLOBAL;
-                    if ( (lNf_to_l1f(l3e_get_flags(ol3e)) ^ flags) &
-                         PAGE_CACHE_ATTRS )
-                        flush_flags |= FLUSH_CACHE;
+                    flush_flags(lNf_to_l1f(l3e_get_flags(ol3e)));
                     flush_area(virt, flush_flags);
                 }
                 else
@@ -5481,27 +5486,14 @@ int map_pages_to_xen(
                         if ( !(l2e_get_flags(ol2e) & _PAGE_PRESENT) )
                             continue;
                         if ( l2e_get_flags(ol2e) & _PAGE_PSE )
-                        {
-                            if ( l2e_get_flags(ol2e) & _PAGE_GLOBAL )
-                                flush_flags |= FLUSH_TLB_GLOBAL;
-                            if ( (lNf_to_l1f(l2e_get_flags(ol2e)) ^ flags)
&
-                                 PAGE_CACHE_ATTRS )
-                                flush_flags |= FLUSH_CACHE;
-                        }
+                            flush_flags(lNf_to_l1f(l2e_get_flags(ol2e)));
                         else
                         {
                             unsigned int j;
 
                             pl1e = l2e_to_l1e(ol2e);
                             for ( j = 0; j < L1_PAGETABLE_ENTRIES; j++ )
-                            {
-                                ol1e = pl1e[j];
-                                if ( l1e_get_flags(ol1e) & _PAGE_GLOBAL )
-                                    flush_flags |= FLUSH_TLB_GLOBAL;
-                                if ( (l1e_get_flags(ol1e) ^ flags) &
-                                     PAGE_CACHE_ATTRS )
-                                    flush_flags |= FLUSH_CACHE;
-                            }
+                                flush_flags(l1e_get_flags(pl1e[j]));
                         }
                     }
                     flush_area(virt, flush_flags);
@@ -5595,24 +5587,14 @@ int map_pages_to_xen(
 
                 if ( l2e_get_flags(ol2e) & _PAGE_PSE )
                 {
-                    if ( l2e_get_flags(ol2e) & _PAGE_GLOBAL )
-                        flush_flags |= FLUSH_TLB_GLOBAL;
-                    if ( (lNf_to_l1f(l2e_get_flags(ol2e)) ^ flags) &
-                         PAGE_CACHE_ATTRS )
-                        flush_flags |= FLUSH_CACHE;
+                    flush_flags(lNf_to_l1f(l2e_get_flags(ol2e)));
                     flush_area(virt, flush_flags);
                 }
                 else
                 {
                     pl1e = l2e_to_l1e(ol2e);
                     for ( i = 0; i < L1_PAGETABLE_ENTRIES; i++ )
-                    {
-                        if ( l1e_get_flags(pl1e[i]) & _PAGE_GLOBAL )
-                            flush_flags |= FLUSH_TLB_GLOBAL;
-                        if ( (l1e_get_flags(pl1e[i]) ^ flags) &
-                             PAGE_CACHE_ATTRS )
-                            flush_flags |= FLUSH_CACHE;
-                    }
+                        flush_flags(l1e_get_flags(pl1e[i]));
                     flush_area(virt, flush_flags);
                     free_xen_pagetable(pl1e);
                 }
@@ -5687,10 +5669,8 @@ int map_pages_to_xen(
             if ( (l1e_get_flags(ol1e) & _PAGE_PRESENT) )
             {
                 unsigned int flush_flags = FLUSH_TLB | FLUSH_ORDER(0);
-                if ( l1e_get_flags(ol1e) & _PAGE_GLOBAL )
-                    flush_flags |= FLUSH_TLB_GLOBAL;
-                if ( (l1e_get_flags(ol1e) ^ flags) & PAGE_CACHE_ATTRS )
-                    flush_flags |= FLUSH_CACHE;
+
+                flush_flags(l1e_get_flags(ol1e));
                 flush_area(virt, flush_flags);
             }
 
@@ -5769,6 +5749,8 @@ int map_pages_to_xen(
     return 0;
 }
 
+#undef flush_flags
+
 void destroy_xen_mappings(unsigned long s, unsigned long e)
 {
     bool_t locking = system_state > SYS_STATE_boot;
@@ -5908,6 +5890,8 @@ void destroy_xen_mappings(unsigned long 
     flush_area(NULL, FLUSH_TLB_GLOBAL);
 }
 
+#undef flush_area
+
 void __set_fixmap(
     enum fixed_addresses idx, unsigned long mfn, unsigned long flags)
 {




_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

On 17/07/13 16:09, Jan Beulich wrote:
> This fixes yet another shortcoming of the function (exposed by 8bfaa2c2
> ["x86: add locking to map_pages_to_xen()"]''s adjustment
to
> msix_put_fixmap()): It must not flush caches when transitioning to a
> non-present mapping. Doing so causes the CLFLUSH to fault, if used in
> favor of WBINVD.
>
> To help code readability, factor out the whole flush flags updating
> in map_pages_to_xen() into a helper macro.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>
> --- a/xen/arch/x86/mm.c
> +++ b/xen/arch/x86/mm.c
> @@ -5430,6 +5430,15 @@ l1_pgentry_t *virt_to_xen_l1e(unsigned l
>                           flush_area_local((const void *)v, f) : \
>                           flush_area_all((const void *)v, f))
>  
> +#define flush_flags(oldf) do {                                  \
> +    unsigned int o_ = (oldf);                                   \
> +    if ( (o_) & _PAGE_GLOBAL )                                  \
> +        flush_flags |= FLUSH_TLB_GLOBAL;                        \
> +    if ( (flags & _PAGE_PRESENT) &&                           
\
> +         (((o_) ^ flags) & PAGE_CACHE_ATTRS) )                  \
> +        flush_flags |= FLUSH_CACHE;                             \
> +} while (0)
> +
I have to admit to being surprised that the compiler is even happy with
a macro aliasing a variable, but please can it be renamed to something
else (perhaps "set_flush_flags" ?) for the sanity of other humans
trying
to read the code.

Furthermore, are we not introducing consistency errors?

Previously, we occasionally decided to flush specific cache lines, and
are now conditionally not flushing the cache lines depending on the
mappings.

Should the fix not be "If we need to flush parts of the cache and dont
have mappings to what we want to flush, use wbinvd()" ?

~Andrew
>  int map_pages_to_xen(
>      unsigned long virt,
>      unsigned long mfn,
> @@ -5465,11 +5474,7 @@ int map_pages_to_xen(
>  
>                  if ( l3e_get_flags(ol3e) & _PAGE_PSE )
>                  {
> -                    if ( l3e_get_flags(ol3e) & _PAGE_GLOBAL )
> -                        flush_flags |= FLUSH_TLB_GLOBAL;
> -                    if ( (lNf_to_l1f(l3e_get_flags(ol3e)) ^ flags) &
> -                         PAGE_CACHE_ATTRS )
> -                        flush_flags |= FLUSH_CACHE;
> +                    flush_flags(lNf_to_l1f(l3e_get_flags(ol3e)));
>                      flush_area(virt, flush_flags);
>                  }
>                  else
> @@ -5481,27 +5486,14 @@ int map_pages_to_xen(
>                          if ( !(l2e_get_flags(ol2e) & _PAGE_PRESENT) )
>                              continue;
>                          if ( l2e_get_flags(ol2e) & _PAGE_PSE )
> -                        {
> -                            if ( l2e_get_flags(ol2e) & _PAGE_GLOBAL )
> -                                flush_flags |= FLUSH_TLB_GLOBAL;
> -                            if ( (lNf_to_l1f(l2e_get_flags(ol2e)) ^ flags)
&
> -                                 PAGE_CACHE_ATTRS )
> -                                flush_flags |= FLUSH_CACHE;
> -                        }
> +                            flush_flags(lNf_to_l1f(l2e_get_flags(ol2e)));
>                          else
>                          {
>                              unsigned int j;
>  
>                              pl1e = l2e_to_l1e(ol2e);
>                              for ( j = 0; j < L1_PAGETABLE_ENTRIES; j++
)
> -                            {
> -                                ol1e = pl1e[j];
> -                                if ( l1e_get_flags(ol1e) &
_PAGE_GLOBAL )
> -                                    flush_flags |= FLUSH_TLB_GLOBAL;
> -                                if ( (l1e_get_flags(ol1e) ^ flags) &
> -                                     PAGE_CACHE_ATTRS )
> -                                    flush_flags |= FLUSH_CACHE;
> -                            }
> +                                flush_flags(l1e_get_flags(pl1e[j]));
>                          }
>                      }
>                      flush_area(virt, flush_flags);
> @@ -5595,24 +5587,14 @@ int map_pages_to_xen(
>  
>                  if ( l2e_get_flags(ol2e) & _PAGE_PSE )
>                  {
> -                    if ( l2e_get_flags(ol2e) & _PAGE_GLOBAL )
> -                        flush_flags |= FLUSH_TLB_GLOBAL;
> -                    if ( (lNf_to_l1f(l2e_get_flags(ol2e)) ^ flags) &
> -                         PAGE_CACHE_ATTRS )
> -                        flush_flags |= FLUSH_CACHE;
> +                    flush_flags(lNf_to_l1f(l2e_get_flags(ol2e)));
>                      flush_area(virt, flush_flags);
>                  }
>                  else
>                  {
>                      pl1e = l2e_to_l1e(ol2e);
>                      for ( i = 0; i < L1_PAGETABLE_ENTRIES; i++ )
> -                    {
> -                        if ( l1e_get_flags(pl1e[i]) & _PAGE_GLOBAL )
> -                            flush_flags |= FLUSH_TLB_GLOBAL;
> -                        if ( (l1e_get_flags(pl1e[i]) ^ flags) &
> -                             PAGE_CACHE_ATTRS )
> -                            flush_flags |= FLUSH_CACHE;
> -                    }
> +                        flush_flags(l1e_get_flags(pl1e[i]));
>                      flush_area(virt, flush_flags);
>                      free_xen_pagetable(pl1e);
>                  }
> @@ -5687,10 +5669,8 @@ int map_pages_to_xen(
>              if ( (l1e_get_flags(ol1e) & _PAGE_PRESENT) )
>              {
>                  unsigned int flush_flags = FLUSH_TLB | FLUSH_ORDER(0);
> -                if ( l1e_get_flags(ol1e) & _PAGE_GLOBAL )
> -                    flush_flags |= FLUSH_TLB_GLOBAL;
> -                if ( (l1e_get_flags(ol1e) ^ flags) & PAGE_CACHE_ATTRS
)
> -                    flush_flags |= FLUSH_CACHE;
> +
> +                flush_flags(l1e_get_flags(ol1e));
>                  flush_area(virt, flush_flags);
>              }
>  
> @@ -5769,6 +5749,8 @@ int map_pages_to_xen(
>      return 0;
>  }
>  
> +#undef flush_flags
> +
>  void destroy_xen_mappings(unsigned long s, unsigned long e)
>  {
>      bool_t locking = system_state > SYS_STATE_boot;
> @@ -5908,6 +5890,8 @@ void destroy_xen_mappings(unsigned long 
>      flush_area(NULL, FLUSH_TLB_GLOBAL);
>  }
>  
> +#undef flush_area
> +
>  void __set_fixmap(
>      enum fixed_addresses idx, unsigned long mfn, unsigned long flags)
>  {
>
>
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

>>> On 17.07.13 at 17:40, Andrew Cooper
<andrew.cooper3@citrix.com> wrote:
> On 17/07/13 16:09, Jan Beulich wrote:
>> This fixes yet another shortcoming of the function (exposed by 8bfaa2c2
>> ["x86: add locking to map_pages_to_xen()"]''s
adjustment to
>> msix_put_fixmap()): It must not flush caches when transitioning to a
>> non-present mapping. Doing so causes the CLFLUSH to fault, if used in
>> favor of WBINVD.
>>
>> To help code readability, factor out the whole flush flags updating
>> in map_pages_to_xen() into a helper macro.
>>
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>>
>> --- a/xen/arch/x86/mm.c
>> +++ b/xen/arch/x86/mm.c
>> @@ -5430,6 +5430,15 @@ l1_pgentry_t *virt_to_xen_l1e(unsigned l
>>                           flush_area_local((const void *)v, f) : \
>>                           flush_area_all((const void *)v, f))
>>  
>> +#define flush_flags(oldf) do {                                  \
>> +    unsigned int o_ = (oldf);                                   \
>> +    if ( (o_) & _PAGE_GLOBAL )                                  \
>> +        flush_flags |= FLUSH_TLB_GLOBAL;                        \
>> +    if ( (flags & _PAGE_PRESENT) &&                       
\
>> +         (((o_) ^ flags) & PAGE_CACHE_ATTRS) )                  \
>> +        flush_flags |= FLUSH_CACHE;                             \
>> +} while (0)
>> +
> 
> I have to admit to being surprised that the compiler is even happy with
> a macro aliasing a variable, but please can it be renamed to something
> else (perhaps "set_flush_flags" ?) for the sanity of other humans
trying
> to read the code.
Ugly. I actually picked the same name intentionally.
> Furthermore, are we not introducing consistency errors?
> 
> Previously, we occasionally decided to flush specific cache lines, and
> are now conditionally not flushing the cache lines depending on the
> mappings.
> 
> Should the fix not be "If we need to flush parts of the cache and dont
> have mappings to what we want to flush, use wbinvd()" ?
No. The cache flushing (a half hearted attempt only anyway) is
necessary only when changing cachability of a mapping (and in
fact the condition could be further relaxed to cover only the case
where a translation for the same physical address changes
cachability, but that would require not only comparing the
single MFNs from the old and new PTE, but the sets of all the
ones affected by the current function invocation). It could also
be relaxed to e.g. not get done when transitioning from UC or
to WB. But that''s all pretty pointless.

As to the "half hearted" above: What we do currently isn''t
really
correct anyway: When transitioning a translation to a different
caching mode, we''d have to tear down the translation, flush TLB,
flush cache, and _only then_ establish the new translation. But
what we do right now seems to be good enough in practice.

Jan

On 17/07/2013 17:01, "Jan Beulich" <JBeulich@suse.com> wrote:
>>>> On 17.07.13 at 17:40, Andrew Cooper
<andrew.cooper3@citrix.com> wrote:
>> On 17/07/13 16:09, Jan Beulich wrote:
>>> This fixes yet another shortcoming of the function (exposed by
8bfaa2c2
>>> ["x86: add locking to map_pages_to_xen()"]''s
adjustment to
>>> msix_put_fixmap()): It must not flush caches when transitioning to
a
>>> non-present mapping. Doing so causes the CLFLUSH to fault, if used
in
>>> favor of WBINVD.
>>> 
>>> To help code readability, factor out the whole flush flags updating
>>> in map_pages_to_xen() into a helper macro.
>>> 
>>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
>>> 
>>> --- a/xen/arch/x86/mm.c
>>> +++ b/xen/arch/x86/mm.c
>>> @@ -5430,6 +5430,15 @@ l1_pgentry_t *virt_to_xen_l1e(unsigned l
>>>                           flush_area_local((const void *)v, f) : \
>>>                           flush_area_all((const void *)v, f))
>>>  
>>> +#define flush_flags(oldf) do {                                  \
>>> +    unsigned int o_ = (oldf);                                   \
>>> +    if ( (o_) & _PAGE_GLOBAL )                                
\
>>> +        flush_flags |= FLUSH_TLB_GLOBAL;                        \
>>> +    if ( (flags & _PAGE_PRESENT) &&                   
\
>>> +         (((o_) ^ flags) & PAGE_CACHE_ATTRS) )                
\
>>> +        flush_flags |= FLUSH_CACHE;                             \
>>> +} while (0)
>>> +
>> 
>> I have to admit to being surprised that the compiler is even happy with
>> a macro aliasing a variable, but please can it be renamed to something
>> else (perhaps "set_flush_flags" ?) for the sanity of other
humans trying
>> to read the code.
> 
> Ugly. I actually picked the same name intentionally.
I''m not too strongly opinionated on this one, but it did make me look
twice.
I think it would be cleaner something like:
#define flush_flags(oldf) ({ unsigned int f_ = 0; ...; _f; })
...
flush_flags |= flush_flags(...);
The ''name collision'' I''m fine with, whereas going
directly at a caller''s
variable within a macro is rather grubby behaviour. ;)

Either way though I find the patch acceptable. And I will Ack it as already
posted.

 -- Keir
>> Furthermore, are we not introducing consistency errors?
>> 
>> Previously, we occasionally decided to flush specific cache lines, and
>> are now conditionally not flushing the cache lines depending on the
>> mappings.
>> 
>> Should the fix not be "If we need to flush parts of the cache and
dont
>> have mappings to what we want to flush, use wbinvd()" ?
> 
> No. The cache flushing (a half hearted attempt only anyway) is
> necessary only when changing cachability of a mapping (and in
> fact the condition could be further relaxed to cover only the case
> where a translation for the same physical address changes
> cachability, but that would require not only comparing the
> single MFNs from the old and new PTE, but the sets of all the
> ones affected by the current function invocation). It could also
> be relaxed to e.g. not get done when transitioning from UC or
> to WB. But that''s all pretty pointless.
> 
> As to the "half hearted" above: What we do currently
isn''t really
> correct anyway: When transitioning a translation to a different
> caching mode, we''d have to tear down the translation, flush TLB,
> flush cache, and _only then_ establish the new translation. But
> what we do right now seems to be good enough in practice.
> 
> Jan
>

On 17/07/2013 16:09, "Jan Beulich" <JBeulich@suse.com> wrote:
> This fixes yet another shortcoming of the function (exposed by 8bfaa2c2
> ["x86: add locking to map_pages_to_xen()"]''s adjustment
to
> msix_put_fixmap()): It must not flush caches when transitioning to a
> non-present mapping. Doing so causes the CLFLUSH to fault, if used in
> favor of WBINVD.
> 
> To help code readability, factor out the whole flush flags updating
> in map_pages_to_xen() into a helper macro.
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>
I''d prefer the macro modified as I just posted, but either way:
Acked-by: Keir Fraser <keir@xen.org>

Wednesday, July 17, 2013, 6:31:41 PM, you wrote:
> On 17/07/2013 16:09, "Jan Beulich" <JBeulich@suse.com>
wrote:
>> This fixes yet another shortcoming of the function (exposed by 8bfaa2c2
>> ["x86: add locking to map_pages_to_xen()"]''s
adjustment to
>> msix_put_fixmap()): It must not flush caches when transitioning to a
>> non-present mapping. Doing so causes the CLFLUSH to fault, if used in
>> favor of WBINVD.
>> 
>> To help code readability, factor out the whole flush flags updating
>> in map_pages_to_xen() into a helper macro.
>> 
>> Signed-off-by: Jan Beulich <jbeulich@suse.com>
> I''d prefer the macro modified as I just posted, but either way:
> Acked-by: Keir Fraser <keir@xen.org>
Fixes my issue (reported in thread
http://lists.xenproject.org/archives/html/xen-devel/2013-07/msg01609.html)

Tested-by: Sander Eikelenboom <linux@eikelenboom.it>

>>>> Keir Fraser <keir.xen@gmail.com> 07/17/13 6:31 PM
>>>
>On 17/07/2013 17:01, "Jan Beulich" <JBeulich@suse.com>
wrote:
>>>>> On 17.07.13 at 17:40, Andrew Cooper
<andrew.cooper3@citrix.com> wrote:
>>>> On 17/07/13 16:09, Jan Beulich wrote:
>>>> +#define flush_flags(oldf) do {                                
\
>>>> +    unsigned int o_ = (oldf);                                 
\
>>>> +    if ( (o_) & _PAGE_GLOBAL )                            
\
>>> +        flush_flags |= FLUSH_TLB_GLOBAL;                        \
>>>> +    if ( (flags & _PAGE_PRESENT) &&               
\
>>>> +         (((o_) ^ flags) & PAGE_CACHE_ATTRS) )            
\
>>>> +        flush_flags |= FLUSH_CACHE;                           
\
>>>> +} while (0)
>>>> +
>>> 
>>> I have to admit to being surprised that the compiler is even happy
with
>>> a macro aliasing a variable, but please can it be renamed to
something
>>> else (perhaps "set_flush_flags" ?) for the sanity of
other humans trying
>>> to read the code.
>> 
>> Ugly. I actually picked the same name intentionally.
>
>I''m not too strongly opinionated on this one, but it did make me
look twice.
>I think it would be cleaner something like:
>#define flush_flags(oldf) ({ unsigned int f_ = 0; ...; _f; })
>...
>flush_flags |= flush_flags(...);
>The ''name collision'' I''m fine with, whereas going
directly at a caller''s
>variable within a macro is rather grubby behaviour. ;)
But that would still leave the macro access "flags" directly. And
while I realize
that this is (slightly) odd behavior for a macro, that''s precisely the
reason why
I #undef it right after the function. Perhaps this could be made even more clear
by moving #define and #undef inside the function...

Jan

On 18/07/2013 06:49, "Jan Beulich" <jbeulich@suse.com> wrote:
>>> Ugly. I actually picked the same name intentionally.
>> 
>> I''m not too strongly opinionated on this one, but it did make
me look twice.
>> I think it would be cleaner something like:
>> #define flush_flags(oldf) ({ unsigned int f_ = 0; ...; _f; })
>> ...
>> flush_flags |= flush_flags(...);
>> The ''name collision'' I''m fine with, whereas
going directly at a caller''s
>> variable within a macro is rather grubby behaviour. ;)
> 
> But that would still leave the macro access "flags" directly.
Hmmm true. Don''t really want to expand the verbiage at each macro call
site
any further, so I guess your existing approach is also my preferred one.
> And while I 
> realize
> that this is (slightly) odd behavior for a macro, that''s precisely
the reason
> why
> I #undef it right after the function. Perhaps this could be made even more
> clear
> by moving #define and #undef inside the function...
Yes, that might be slightly preferable.

 -- Keir