[PATCH v2 3/5] d3d9: Handle multisample depth resolve in d3d9_device_SetRenderState().

[Zebediah Figura]

On 12/12/19 7:44 AM, Henri Verbeet wrote:
> On Wed, 11 Dec 2019 at 07:16, Zebediah Figura <z.figura12 at gmail.com> wrote:
>> @@ -289,8 +289,18 @@ static HRESULT WINAPI d3d9_CheckDeviceFormat(IDirect3D9Ex *iface, UINT adapter,
>>      }
>>
>>      wined3d_mutex_lock();
>> -    hr = wined3d_check_device_format(d3d9->wined3d, adapter, device_type, wined3dformat_from_d3dformat(adapter_format),
>> -            usage, bind_flags, wined3d_rtype, wined3dformat_from_d3dformat(format));
>> +    if ((enum wined3d_format_id)format == WINED3DFMT_RESZ)
>> +    {
>> +        DWORD levels;
>> +        hr = wined3d_check_device_multisample_type(d3d9->wined3d, adapter, device_type,
>> +                WINED3DFMT_D24_UNORM_S8_UINT, FALSE, WINED3D_MULTISAMPLE_NONE, &levels);
> Did you intend to use WINED3D_MULTISAMPLE_NON_MASKABLE here?

Yes, on further examination I did mean to do that. Thanks.

> 
> It seems a little superfluous to export
> wined3d_device_resolve_depth_buffer(); the implementation on top of
> wined3d_device_resolve_sub_resource() wasn't that complicated.
> 

Okay, I'll revert back to that implementation. I misread your earlier
response to imply that duplicating between d3d8 and d3d9 would be
dispreferred.