Bug 15352 - RandR: multi-referenced monitor sections induce strange behavior
Summary: RandR: multi-referenced monitor sections induce strange behavior
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: git
Hardware: All All
: medium major
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
: 15142 (view as bug list)
Depends on:
Blocks:
 
Reported: 2008-04-04 12:29 UTC by Matthias Hopf
Modified: 2018-06-12 18:44 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments

Description Matthias Hopf 2008-04-04 12:29:59 UTC
I just stumbled over some strange behavior, and it definitely changed
past Xserver 1.4.0 (don't know the exact version yet, bisecting this
would be probably worth it).

A snippet of the config I'm testing is


Section "Monitor"
  Identifier "M1"
EndSection
Section "Monitor"
  Identifier "M2"
  Option     "Right-Of" "M1"
EndSection

Section "Device"
  Identifier "D"
  Driver "radeonhd"
  Option "monitor-DVI-I_1/digital" "M1"
  Option "monitor-DVI-I_2/digital" "M2"
EndSection

Section "Screen"
  [...]
  Identifier "S"
  Device "D"
  Monitor "M1"
EndSection

[...]


So basically a static multi-monitor setup for RandR 1.2, with one of the
monitors additionally referenced in the Screen section. This monitor
reference shouldn't be used at all, because the driver sets
use_screen_monitor to FALSE for all outputs.

This used to work - however, beginning with 1.4.90 the multi-monitor
setup fails if and only if I use the two digital DVI outputs on this
card, with DVI and VGA it works fine, with DVI/d and DVI/a it also works
fine. If I remove the reference to the M1 monitor in the Screen section
it works fine. If I add another monitor and reference it in the Screen
section instead it works fine. It's probably difficult to reproduce this
with another driver, but this shouldn't be a driver issue.


Does anybody have a clue WTF is going on here, is this considered
standard behavior or rather a bug? If it's standard, why is its behavior
depending on the outputs (and not on the monitors)?


- Mailed on xorg-devel:
Date: Thu, 13 Mar 2008 18:55:44 +0100
From: Matthias Hopf <mhopf@suse.de>
To: xorg@lists.freedesktop.org
Subject: RandR: multi-referenced monitor sections & problems with them
Comment 1 Matthias Hopf 2008-04-04 12:32:08 UTC
This bug is tracked for openSUSE in

  https://bugzilla.novell.com/show_bug.cgi?id=370656


Comment 2 Matthias Hopf 2008-04-04 12:32:18 UTC
Stefan Dirsch:

Nearly the same problem with nv driver on G80. As long as 

  Monitor "M1"

in Section Screen is not removed/commented out, the monitors cannot be
positioned side by side with the error message similar to this:

(EE) Monitor(1) cannot be positioned right of Monitor[0] without a mode
     specified.

Specifying modes explicitetly for both monitor doesn't help at all. Only
the workaround mentioned above does.
Comment 3 Matthias Hopf 2008-04-04 12:34:02 UTC
*** Bug 15142 has been marked as a duplicate of this bug. ***
Comment 4 Matthias Hopf 2009-12-01 05:46:00 UTC
Apparently, the Screen section monitor is assigned to the first output. In this
case this happened to be TV_7PIN_DIN.

So adding
Option "monitor-TV_7PIN_DIN" "M1"
instead had the same effect.


This is actually a driver issue, fixed in radeonhd in the meantime.

Fixing the base issue is probably not. Monitors are not to be used by multiple instances, because the mode pool between then would be shared. I.e. if one is disconnected, the other one won't see any modes, which is the base issue of this bug.
Comment 5 Adam Jackson 2018-06-12 18:44:19 UTC
Mass closure: This bug has been untouched for more than six years, and is not obviously still valid. Please file a new report if you continue to experience issues with a current server.


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.