Created attachment 29585 [details] [review]
Xorg.0.log

Created attachment 29586 [details]
Xorg.conf

Created attachment 29587 [details]
command xrandr

can you post xrandr --verbose?

I suspect its an xrandr bug but need to see the verbose.

Created attachment 29588 [details]
xrandr--verbose

could you attach two logs from the following scenarios

start X, 
xrandr --output VGA-0 --off

start X,
xrandr --output VGA-1 --off

just want to make sure I can see the sequence of events in the log.

Created attachment 29589 [details]
xrandr VGA-0--off--verbose

Created attachment 29590 [details]
xrandr VGA-0 --auto--verbose

for start X, 
xrandr --output VGA-0 --off

start X,
xrandr --output VGA-1 --off

there is no log info

(In reply to comment #0)
> When connected two monitors(monitor A and monitor B)
> monitor A --- VGA-0
> monitor B --- VGA-1
> type xrandr --output VGA-0 --off
> monitor A is off and monitor B is on. However, when type
> xrandr --output VGA-1 --off
> monitor A is still off and monitor B is on

Sorry a typo

> When connected two monitors(monitor A and monitor B)
> monitor A --- VGA-0
> monitor B --- VGA-1
> type xrandr --output VGA-0 --off
> monitor A is off and monitor B is on. However, when type
> xrandr --output VGA-0 --auto
> monitor A is still off and monitor B is on

(In reply to comment #4)
> can you post xrandr --verbose?
> I suspect its an xrandr bug but need to see the verbose.

Thanks airlied
 
I tried to investigate bug 23970, and found radeon driver will step into radeon_crtc_set_origin instead of radeon_crtc_mode_fixup, mode_prepare, mode_commit.
 
However, in radeon_crtc_set_origin, radeon driver fails to set DAC_CNTL2 to enable TVDAC, and also CRTC2_GEN_CNTL and so on.
 
Then I try to add code in radeon_crtc_set_origin and force driver to execute mode fixup, mode_prepare and mode_commit (both crtc and output), then radeon driver can run xrandr command right.
 
But I think it is not set_origin's responsibility to enable CRTC_CNTL and DAC_CNTL, I just wanna why radeon driver introduced radeon_crtc_set_origin and its role in xrandr panning?

Is this still an issue with KMS or a newer version of the driver?

Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.