Re: [Qemu-devel] [PATCH 4/5] hw/arm/pxa2xx: Convert pxa2xx-ssp to VMState

[Peter Crosthwaite]

On Fri, Jun 5, 2015 at 4:32 PM, Peter Maydell <address@hidden> wrote:
> On 6 June 2015 at 00:18, Peter Crosthwaite <address@hidden> wrote:
>> On Thu, May 28, 2015 at 5:09 AM, Peter Maydell <address@hidden> wrote:
>>> The pxa2xx-ssp device is already a QOM device but is still
>>> using the old-style register_savevm(); convert to VMState.
>>>
>>> Signed-off-by: Peter Maydell <address@hidden>
>>> ---
>>>  hw/arm/pxa2xx.c | 89 
>>> +++++++++++++++++++++------------------------------------
>>>  1 file changed, 33 insertions(+), 56 deletions(-)
>>>
>>> diff --git a/hw/arm/pxa2xx.c b/hw/arm/pxa2xx.c
>>> index 770902f..09401f9 100644
>>> --- a/hw/arm/pxa2xx.c
>>> +++ b/hw/arm/pxa2xx.c
>>> @@ -457,7 +457,7 @@ typedef struct {
>>>
>>>      MemoryRegion iomem;
>>>      qemu_irq irq;
>>> -    int enable;
>>> +    uint32_t enable;
>>>      SSIBus *bus;
>>>
>>>      uint32_t sscr[2];
>>> @@ -470,10 +470,39 @@ typedef struct {
>>>      uint8_t ssacd;
>>>
>>>      uint32_t rx_fifo[16];
>>> -    int rx_level;
>>> -    int rx_start;
>>> +    uint32_t rx_level;
>>> +    uint32_t rx_start;
>>>  } PXA2xxSSPState;
>>>
>>> +static bool pxa2xx_ssp_vmstate_validate(void *opaque, int version_id)
>>> +{
>>> +    PXA2xxSSPState *s = opaque;
>>> +
>>> +    return s->rx_start < sizeof(s->rx_fifo);
>>
>> Does this need to be ARRAY_SIZE to account for unit32_t indexing?
>
> Yes, good catch.
>

Ok that's all I found. So otherwise,

Reviewed-by: Peter Crosthwaite <address@hidden>

I assume this is intended to break backwards compat on the VMSD with
the conversion from put_byte loop to VMSTATE_UINT32_ARRAY? I'm not
sure what our policy is on these old ARM machines and preserving the
backwards compat.

Regards,
Peter

> -- PMM
>