Corruption around mouse cursor when dragged, particularly around Dash, Indicators, window chrome

[Brendan Donegan]

When dragging the cursor around the UI on the Nexus7, the cursor and the areas around the cursor quite often become corrupted. This is particularly easy to reproduce in the Dash. and a few examples are attached as screenshots - but it also happens around the indicators, menus and window chrome.

Steps to reproduce:

1. Open the Dash
2. In turn, hover the cursor over each icon shown
3. Close the Dash and open a terminal window
4. Move the cursor repeatedly into and out of the title bar of the window

Expected Result:

No corruption occurs

Actual Result:

At least half the time, some corruption occurs in the highlight shown around the icon on step 2, and the same on step 4

[Chris Wayne (cwayne)]

I have seen this as well, although only when plugging in a mouse. Marking confirmed.

[Bryce Harrington (bryce)]

I've seen this as well. It's my guess this is due to a bug in the tegra3 driver.

[Sean Feole (sfeole)]

Moving bug back to a New status since it's not currently assigned to anyone.

Per comment #4 if this is a Tegra3 potential bug then it will have to be brought to NVidia's attention.

[Sean Feole (sfeole)]

Marking confirmed and assigning to Alex Chiang

[Sean Feole (sfeole)]

I have not seen this bug in Raring 13.04, if anyone else has let me know.

[Eric Brower (ebrower)]

For somebody experiencing this, can you please try adding/modifying your /etc/X11/xorg.conf as follows, or create a suitable /etc/X11/xorg.conf.d/xyz.conf script:

Section "Device"
    Identifier "Tegra"
    Driver "tegra"
    Option "ARGBHWCursor"
EndSection

If this resolves/works-around the corruption, you'll also want to adapt the following kernel patch to resolve console errors when the cursor moves off-screen:
http://nv-tegra.nvidia.com/gitweb/?p=chromeos/kernel.git;a=commitdiff;h=b50bf5979036750891461a5824bb12273f04fe16

[Andy Whitcroft (apw)]

@Eric -- this kernel patch seems to be from other tegra kernel not the one on which we are based. I can find nothing even remotely similar to even guess as to where a similar change might be made.

@Ogra et al -- if you could test the xorg.conf.d changes though as they may well still hint as to the issue.

[Eric Brower (ebrower)]

@Andy -- taking a look at the Nexus7 kernel, the patch may not be necessary. Let's see if enabling ARGBHWCursor improves the situation, and go from there.