[Bug 39911] New: dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

Bug ID: 39911
           Summary: dxd9 PresentationInterval not propagated on device
                    Reset
           Product: Wine
           Version: 1.9.0
          Hardware: x86
                OS: Mac OS X
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: directx-d3d
          Assignee: [e-mail]
          Reporter: [e-mail]

Background: The steam game XIIZeal relies on vsync to run the game engine
timing; under wine (on OSX) fullscreen runs at an unlimited speed, while on
windows it runs at the refresh rate of your display.

Looking into this this is the sequence of events as best I can tell:

It creates a D3D9 Device with PresentationInterval set to
D3DPRESENT_INTERVAL_IMMEDIATE
Wine creates the device, swap chain, OpenGL context and eventually calls
wglSwapInterval(0).

Then the game does a device Reset, in that reset the PresentationInterval is
set to D3DPRESENT_INTERVAL_ONE. Wine does not ever call wglSwapInterval(1)
after this reset, thus the framerate is never limited.

If I hack up wined3d_device_reset() to call wglSwapInterval() based on the
PresentationInterval the game behaves as expected. However I don't know if this
is the best way to approach this issue.

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #1 from Zakk  ---
Created attachment 54103
  --> https://bugs.winehq.org/attachment.cgi?id=54103
Change swapinterval in device_reset

This is the patch I'm using to work around this issue. It works but I'm not
sure if there's a better place to put this logic.

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #2 from Henri Verbeet  ---
Is wined3d_device_reset() getting called from d3d9ex by any chance?
context_create() (called from create_primary_opengl_context()) should take care
of setting up the swap interval, but if wined3d_device_reset() gets called with
"reset_state" FALSE that never happens. We'll probably want to introduce a
function context_set_swap_interval() and call that from both context_create()
and wined3d_device_reset().

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #3 from Zakk  ---
Yes, now that I look deeper the game is creating the device with
Direct3DCreate9Ex, so that explains the lack of swap interval change on reset

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #4 from Henri Verbeet  ---
Created attachment 54110
  --> https://bugs.winehq.org/attachment.cgi?id=54110
patch

Does this patch help?

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #5 from Zakk  ---
Yep, that fixes the issue

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

Henri Verbeet  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |6b8fb1c22ea26f9fafda7613385
                   |                            |0af10ac3c7256
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #6 from Henri Verbeet  ---
Should be fixed by commit 6b8fb1c22ea26f9fafda76133850af10ac3c7256.

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #7 from Alexandre Julliard  ---
Closing bugs fixed in 1.9.8.

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

Michael Stefaniuc  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
   Target Milestone|---                         |1.8.x

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #8 from Michael Stefaniuc  ---
Henri, is it worth backporting this patch to Stable? It depends on

commit e3e5dcd2cd548bed34cc193383da96c1898fee2c
Author: Henri Verbeet 
Date:   Wed Mar 2 19:24:13 2016 +0100

wined3d: Use a single allocation for texture sub-resource objects.

which has a huge dependency list.

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

--- Comment #9 from Henri Verbeet  ---
(In reply to Michael Stefaniuc from comment #8)
> Henri, is it worth backporting this patch to Stable? It depends on
>
> commit e3e5dcd2cd548bed34cc193383da96c1898fee2c
> Author: Henri Verbeet 
> Date:   Wed Mar 2 19:24:13 2016 +0100
> wined3d: Use a single allocation for texture sub-resource objects.
>
> which has a huge dependency list.

It shouldn't be hard to backport. At first sight, it's probably just the second
argument to the context_acquire() call in swapchain_update_swap_interval(). You
should be able to just replace that with something like
"surface_from_resource(swapchain->front_buffer->sub_resources[0])".

[Bug 39911] dxd9 PresentationInterval not propagated on device Reset

https://bugs.winehq.org/show_bug.cgi?id=39911

Michael Stefaniuc  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|1.8.x                       |---

--- Comment #10 from Michael Stefaniuc  ---
Removing 1.8.x milestone from bugs included in 1.8.4.