Bug 98651

Summary: HDMI goes black on Radeon HD 6950
Product: Drivers Reporter: Thomas (v10lator)
Component: Video(Other)Assignee: drivers_video-other
Status: RESOLVED DUPLICATE    
Severity: normal CC: adf.lists, kernel
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 4.0.3 Subsystem:
Regression: No Bisected commit-id:

Description Thomas 2015-05-19 09:52:39 UTC
Kernel 4.0.2 boots fine. With 4.0.3 and up the HDMI screen stays black and tells me to check cables.

Bisected:
d5c3b64b9618214a5443c57c64511ab5fa599cec is the first bad commit
commit d5c3b64b9618214a5443c57c64511ab5fa599cec
Author: Alex Deucher <alexander.deucher@amd.com>
Date:   Sun Apr 19 12:01:00 2015 -0400

    drm/radeon: adjust pll when audio is not enabled
    
    commit 7fe04d6fa824ccea704535a597dc417c8687f990 upstream.
    
    Fixes display problems with some monitors when audio
    is not enabled.
    
    Bugs:
    https://bugs.freedesktop.org/show_bug.cgi?id=89505
    https://bugzilla.kernel.org/show_bug.cgi?id=94171
    Plus several reports on IRC.
    
    Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
    Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

:040000 040000 db2832bfa3fb015ebff5cdd46bc7927899395415 87719424febe89c541b5935a945c8a58ced12c7d M drivers
Comment 1 Thomas 2015-05-19 09:56:46 UTC
I just compiled a 4.0.4 Kernel with the bad commit reverted. The screen shows a image but HDMI audio isn't working (pavucontrol says it's unplugged).
Comment 2 Robin Bankhead 2015-05-25 11:31:06 UTC
Not sure whether this is a me-too: I have the same issue since 4.0.3 with an AMD E2-1800 APU with GPU component as follows:

00:01.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Wrestler [Radeon HD 7340]

Notable differences: The behaviour is observed on laptop's primary screen (LVDS1). No access to an HDMI screen to test at the moment, but HDMI audio is enabled (radeon.audio=1).

Also the screen doesn't "stay black" but goes black just when udev loads and radeondrmfb takes over from simplefb (the point at which there is normally a flicker and the console output gets cleared).

Other possibly-germane notes: BIOS is in full UEFI mode, and I have agpgart disabled (despite advice in kernel documentation, GLX seems to have been working fine without it until now; does that doc need updating?).
Comment 3 Andy Furniss 2015-05-25 11:38:19 UTC
(In reply to Thomas from comment #1)
> I just compiled a 4.0.4 Kernel with the bad commit reverted. The screen
> shows a image but HDMI audio isn't working (pavucontrol says it's unplugged).

There is another bug WRT audio that some people hit.

If you are affected you may be able to get audio by doing something to HDMI with xrandr eg. off then auto or change mode then change back or even just --set audio on (even though it's already on - well auto). Doesn't work for everyone, but does for me.
Comment 4 Robin Bankhead 2015-05-25 12:27:44 UTC
Confirming, just backed-out that commit on my laptop and picture on LVDS has returned. The screen-refresh when radeondrmfb initialises seems to take longer than I remember it doing before, though.

I'll test HDMI including audio when I can.

BTW #98751 appears to be a dupe of this (it has more activity, but you were first, assuming that's the determining factor).
Comment 5 Thomas 2015-05-27 15:26:42 UTC

*** This bug has been marked as a duplicate of bug 98751 ***