On Mon, Jan 9, 2012 at 12:11, Avi Kivity <a...@redhat.com> wrote:
> On 01/08/2012 11:10 PM, Blue Swirl wrote:
>> Instead of each target knowing or guessing the guest page size,
>> just pass the desired size of dirtied memory area. This should also
>> improve performance due to memset() optimizations.
>>
>> diff --git a/exec-obsolete.h b/exec-obsolete.h
>> index f8af27e..019c09a 100644
>> --- a/exec-obsolete.h
>> +++ b/exec-obsolete.h
>> @@ -76,6 +76,20 @@ static inline int
>> cpu_physical_memory_set_dirty_flags(ram_addr_t addr,
>> return ram_list.phys_dirty[addr >> TARGET_PAGE_BITS] |= dirty_flags;
>> }
>>
>> +static inline void cpu_physical_memory_set_dirty_range(ram_addr_t start,
>> + int length,
>> + int dirty_flags)
>> +{
>> + int i, len;
>> + uint8_t *p;
>> +
>> + len = length >> TARGET_PAGE_BITS;
>> + p = ram_list.phys_dirty + (start >> TARGET_PAGE_BITS);
>> + for (i = 0; i < len; i++) {
>> + p[i] |= dirty_flags;
>> + }
>> +}
>> +
>
> Breaks for start = 0xfff, length = 2.
Right. I copied this from cpu_physical_memory_mask_dirty_range() which
suffers from the same problem.
