On Sat, Jan 28, 2012 at 12:23, Blue Swirl <blauwir...@gmail.com> wrote:
> On Fri, Jan 27, 2012 at 21:27, Stefan Berger <stef...@linux.vnet.ibm.com>
> wrote:
>> On 01/27/2012 04:10 PM, Stefan Berger wrote:
>>>
>>> After bisecting the following commit seems to be the culprit for the
>>> suspend/resume problems that I am seeing with the current tip
>>> (73093354418602a2ff5e43cb91a21b17fbf047d8).
>>>
>>> commit 59abb06198ee9471e29c970f294eae80c0b39be1
>>> Author: Blue Swirl <blauwir...@gmail.com>
>>> Date: Sun Jan 22 11:00:44 2012 +0000
>>>
>>> Once I revert this patch on the tip everything works fine again...
>>>
>>> Stefan
>>>
>>>
>> And this patch here gets it to work:
>>
>> diff --git a/exec-obsolete.h b/exec-obsolete.h
>> index 03cf35e..a673386 100644
>> --- a/exec-obsolete.h
>> +++ b/exec-obsolete.h
>> @@ -101,7 +101,7 @@ static inline void
>> cpu_physical_memory_mask_dirty_range(ram_
>> end = start + length;
>> mask = ~dirty_flags;
>> p = ram_list.phys_dirty + (start >> TARGET_PAGE_BITS);
>> - for (addr = start; addr <= end; addr += TARGET_PAGE_SIZE) {
>> + for (addr = start; addr < end; addr += TARGET_PAGE_SIZE) {
>> *p++ &= mask;
>> }
>> }
>
> Nice. Please add a description (something like "Fix suspend/resume
> broken by off-by-one error in 7309...") and Signed-off-by: line.
On second thought, the code is also broken for case start =
(ram_addr_t)-TARGET_PAGE_SIZE, length = TARGET_PAGE_SIZE. I'll make
another patch.
