This is MMIO-specific tunes on the size. Let's skip it for non-MMIO
translations.

Signed-off-by: Peter Xu <peterx@redhat.com>
---
 exec.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/exec.c b/exec.c
index 0adae94..32e5394 100644
--- a/exec.c
+++ b/exec.c
@@ -455,7 +455,7 @@ address_space_translate_internal(AddressSpaceDispatch *d, hwaddr addr, hwaddr *x
      * everything works fine.  If the incoming length is large, however,
      * the caller really has to do the clamping through memory_access_size.
      */
-    if (memory_region_is_ram(mr)) {
+    if (is_mmio && memory_region_is_ram(mr)) {
         diff = int128_sub(section->size, int128_make64(addr));
         *plen = int128_get64(int128_min(diff, int128_make64(*plen)));
     }
-- 
2.7.4

On 15/05/2017 10:50, Peter Xu wrote:
> This is MMIO-specific tunes on the size. Let's skip it for non-MMIO
> translations.

Can you explain this better?  This is not specific to MMIO, in fact it's
for RAM...

Paolo

> Signed-off-by: Peter Xu <peterx@redhat.com>
> ---
>  exec.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/exec.c b/exec.c
> index 0adae94..32e5394 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -455,7 +455,7 @@ address_space_translate_internal(AddressSpaceDispatch *d, hwaddr addr, hwaddr *x
>       * everything works fine.  If the incoming length is large, however,
>       * the caller really has to do the clamping through memory_access_size.
>       */
> -    if (memory_region_is_ram(mr)) {
> +    if (is_mmio && memory_region_is_ram(mr)) {
>          diff = int128_sub(section->size, int128_make64(addr));
>          *plen = int128_get64(int128_min(diff, int128_make64(*plen)));

On Mon, May 15, 2017 at 11:04:52AM +0200, Paolo Bonzini wrote:
> 
> 
> On 15/05/2017 10:50, Peter Xu wrote:
> > This is MMIO-specific tunes on the size. Let's skip it for non-MMIO
> > translations.
> 
> Can you explain this better?  This is not specific to MMIO, in fact it's
> for RAM...

Yes, I misread the codes and comments, sorry...

Please ignore this patch.

-- 
Peter Xu