[Bug 196563] New: Garbled graphics after resume on an R100-based card

[bugzilla-daemon at bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=196563

            Bug ID: 196563
           Summary: Garbled graphics after resume on an R100-based card
           Product: Drivers
           Version: 2.5
    Kernel Version: 4.11.0-2-686-pae (Debian sid)
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: Video(DRI - non Intel)
          Assignee: drivers_video-dri at kernel-bugs.osdl.org
          Reporter: felix.von.s at posteo.de
        Regression: No

When I suspend and then resume my laptop with R100-based graphics, the display
becomes garbled. A single scanline from the video memory is stretched over the
entire height of the screen; this is usually the scanline containing the caret.
Under X11, one can observe the mouse pointer sprite being corrupted as well.
When switching VTs, flashes of the correct, 'intended' screen contents can be
sometimes seen, but the screen instantly reverts to this single stretched
scanline just described.

The issue can be readily worked around via the usual 'vbetool vbestate' hack;
although even then the display contents can appear incorrect after resume, as
if other programs' framebuffers were pasted onto the screen. Moreover, if the
restore part doesn't run soon enough after resuming, the restored screen
becomes glitched with randomly lit pixels. The glitching is more severe the
more time passes between resuming and restoring VBE state; it is as if the
video RAM wasn't being refreshed correctly until the VBE state is restored.
Thus, best results are achieved by storing the VBE state on a tmpfs. But in any
case, it's nothing a temporary VT switch can't fix.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.