[PATCH 1/3] wined3d: Handle discarded buffers in buffer_resource_sub_resource_map().

[Zebediah Figura (she/her)]

On 1/26/22 07:47, Henri Verbeet wrote:
> On Tue, 25 Jan 2022 at 03:21, Zebediah Figura <zfigura at codeweavers.com> wrote:
>> @@ -927,10 +928,13 @@ static HRESULT buffer_resource_sub_resource_map(struct wined3d_resource *resourc
>>
>>       count = ++resource->map_count;
>>
>> -    if (buffer->buffer_object)
>> +    context = context_acquire(device, NULL, 0);
>> +
>> +    wined3d_buffer_get_memory(buffer, context, &addr);
>> +
>> +    if (addr.buffer_object)
>>       {
> I think that potentially changes the location we map in some cases.
> (In particular, when buffer->buffer_object is non-NULL, but
> WINED3D_LOCATION_BUFFER is not current.) I don't know for sure whether
> that would cause issues for applications, but this code is sensitive
> enough that I'd prefer avoiding wined3d_buffer_get_memory() here
> unless we have a good reason not to.

I guess I see the point here—if we are using a BO in the first place, we 
probably want to actually use the BO instead of reusing sysmem.

Note though that this doesn't seem to preclude 
wined3d_buffer_get_memory(); it just means we want to keep the check for 
buffer->buffer_object instead of addr.buffer_object.

> 
> What is the issue exactly? Is it that buffer->buffer_object may not
> have been created yet for discarded resources?

More generally, no location may yet have been prepared. In particular 
after patch 3/3 no location will be prepared on creation.