Re: [PATCH 1/5] wined3d: Decouple private data from the resource structure.

[Henri Verbeet]

On 24 February 2014 15:37, Stefan Dösinger <stefan(a)codeweavers.com> wrote:
> @@ -1976,6 +1976,19 @@ static inline void context_invalidate_active_texture(struct wined3d_context *con
>          context_invalidate_state(context, STATE_SAMPLER(sampler));
>  }
>
> +struct wined3d_privstore
> +{
> +    struct list content;
> +};
> +
> +struct wined3d_privstore *wined3d_privstore_create(void) DECLSPEC_HIDDEN;
> +void wined3d_privstore_destroy(struct wined3d_privstore *store) DECLSPEC_HIDDEN;
> +HRESULT wined3d_privstore_free_private_data(struct wined3d_privstore *store, const GUID *guid);
> +HRESULT wined3d_privstore_get_private_data(const struct wined3d_privstore *store,
> +        const GUID *guid, void *data, DWORD *data_size, BOOL addref) DECLSPEC_HIDDEN;
> +HRESULT wined3d_privstore_set_private_data(struct wined3d_privstore *store,
> +        const GUID *guid, const void *data, DWORD data_size, DWORD flags) DECLSPEC_HIDDEN;
> +
>  #define WINED3D_RESOURCE_ACCESS_GPU     0x1
>  #define WINED3D_RESOURCE_ACCESS_CPU     0x2

I'm not entirely convinced this is how we want this to work. Calling
HeapAlloc() for a struct list seems less than ideal, and I'd like the
WINED3DSPD_IUNKNOWN handling to be done by the callers.

[PATCH 1/5] wined3d: Decouple private data from the resource structure.

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 2014-02-25 13:27, schrieb Henri Verbeet:
> Calling HeapAlloc() for a struct list seems less than ideal,
> That is easy to fix. It will move the actual structure definition into
> the public header though.
> and I'd like the WINED3DSPD_IUNKNOWN handling to be done by the 
> callers.
> For get_private_data that's fairly easy. I can just return the stored
> flags and the caller decides if it wants to addref anything.

For set_private_data it's tricky. The caller would have to know the
old contents, same for free_private_data and the destruction of the
store. With that amount of logic in the caller we might as well
re-implement the entire thing in each calling library.

But the getter behavior is version specific part, so I think just
moving that out makes sense on its own.
wine-devel(a)winehq.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=+wqG
-----END PGP SIGNATURE-----