Bug 109207 - `override_redirect` windows are not visible on amdgpu
Summary: `override_redirect` windows are not visible on amdgpu
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/AMDgpu (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2019-01-01 22:52 UTC by Levi Webb
Modified: 2019-03-03 23:35 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments

Description Levi Webb 2019-01-01 22:52:15 UTC
`InputOutput` windows with the `override_redirect` flag set in the `XSetWindowAttributes` structure, alongside the `CWOverrideRedirect` mask set when calling `XCreateWindow` should display a window after mapping (parent to root window).

This is not the case on amdgpu, the window receives a single `VisibilityFullyObscured` event and is never seen. The `map_state` attribute is always `IsViewable`, raising the window has no effect, focusing the window grabs input focus but does not show the window.

All other drivers behave accordingly here by providing a visible window. This flag is essential for allowing certain window behavior independent of an active window manager, including an application I maintain.

Affected system info:

$ pacman -Ss xorg-server
extra/xorg-server 1.20.3-1 (xorg) [installed]
    Xorg X server
...
$ pacman -Ss amdgpu
extra/xf86-video-amdgpu 18.1.0-1 (xorg-drivers) [installed]
    X.org amdgpu video driver
$ glxinfo | grep "renderer string"
OpenGL renderer string: Radeon RX 560 Series (POLARIS11, DRM 3.27.0, 4.19.2-arch1-1-ARCH, LLVM 7.0.0)

Tested with nouveau, nvidia, intel, llvmpipe to assert that the behavior is consistent. Feel free to ask for additional details.
Comment 1 Michel Dänzer 2019-01-03 16:50:54 UTC
Please attach the Xorg log files from amdgpu and another driver (ideally modesetting).

Also, if you could provide a specific test case reproducing the problem, that would be helpful.

Is this a regression from older versions of xf86-video-amdgpu? Does it still happen with current upstream xf86-video-amdgpu Git master?
Comment 2 Michel Dänzer 2019-03-02 10:46:10 UTC
I'm afraid there can be no progress on this without the information I asked for.
Comment 3 Levi Webb 2019-03-03 23:35:03 UTC
For some obscure reason, I can no longer reproduce the bug on a newer kernel. I can only assume some other unrelated fix to the driver resolved this issue as well.

Apologies for the late follow up.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.