On 4/22/22 02:32, Brendan Shanks wrote:
On Apr 21, 2022, at 3:30 PM, Mohamad Al-Jaf mohamadaljaf@gmail.com wrote:
On Thu, Apr 21, 2022 at 1:10 PM Nikolay Sivov nsivov@codeweavers.com wrote:
Like I said for v1, if this only helps with some error message I don't think it's worth having, without a plan on how to implement it.
As of now, this along with msvproc is all that's needed. In MS Edge at least, it spawns a GPU process with video playback so it looks like it's using hardware acceleration somehow. It doesn't call any functions in msmpeg2vdec so I don't know what you mean by implementing it. Well, the way Media Foundation is implemented in Wine seems different from Windows so it might not even need to be implemented in a traditional sense.
Interesting, at least in the latest Chromium it looks like this is loaded explicitly before the sandbox is initialized (I assume because video decoding will need it but the sandbox would prevent it from being loaded), this failure only results in an error message. (https://source.chromium.org/chromium/chromium/src/+/main:media/gpu/windows/d...)
However, at some point it also checks whether the DLL is loaded (with GetModuleHandle), if this fails it assumes that it’s on Windows N (which lacks media foundation) and doesn’t try to use video decode acceleration. (https://source.chromium.org/chromium/chromium/src/+/main:media/gpu/windows/d...)
So a stub DLL really does enable acceleration.
Looks like it will fail anyway later, on ID3D11VideoDevice, or decoder creation.
Brendan