From: Henri Verbeet <hverbeet@gmail.com>
Subject: Re: [PATCH v3 7/7] dxgi/tests: Add tests for IDXGIAdapter3::SetVideoMemoryReservation().
Message-Id: <CAOsNvwxdrq+FE9Qy62fWdj0a6sSKE+WaEcG5LF9_4nYq75Y4Aw@mail.gmail.com>
Date: Tue, 26 Nov 2019 23:32:28 +0330
In-Reply-To: <20191125141101.85701-7-cmccarthy@codeweavers.com>
References: <20191125141101.85701-1-cmccarthy@codeweavers.com> <20191125141101.85701-7-cmccarthy@codeweavers.com>

On Mon, 25 Nov 2019 at 17:44, Conor McCarthy <cmccarthy@codeweavers.com> wrote:
> +    hr = IDXGIAdapter3_SetVideoMemoryReservation(adapter3, 0, DXGI_MEMORY_SEGMENT_GROUP_LOCAL, reservation);
> +    ok(hr == S_OK, "Got unexpected hr %#x.\n", hr);
> +    hr = IDXGIAdapter3_SetVideoMemoryReservation(adapter3, 0, DXGI_MEMORY_SEGMENT_GROUP_NON_LOCAL, 0);
> +    ok(hr == S_OK, "Got unexpected hr %#x.\n", hr);
> +    non_local_hr = IDXGIAdapter3_SetVideoMemoryReservation(adapter3, 0, DXGI_MEMORY_SEGMENT_GROUP_NON_LOCAL, reservation);
> +    ok(non_local_hr == S_OK || non_local_hr == E_INVALIDARG, "Got unexpected hr %#x.\n", non_local_hr);
> +
Does the E_INVALIDARG return correspond to some property of the
adapter by any chance? Like, e.g. being an UMA adapter? Also, where
reasonably possible, it makes sense to send patches at the start of
the series with todo_wine. That means that tests don't depend on the
implementation or fix going in first, and it's more clear what a
particular patch fixes in terms of public API.