Bug 26398

Summary: radeon x3450 xrandr dual dvi blank screen
Product: xorg Reporter: pha <bug>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: tim
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
xrandr with both monitors
none
xorg log
none
xorg log from git master
none
lspci of radeon hd4350
none
xrandr with DVI-0 blank
none
xorg.log with radeon 6.12.4
none
Xorg log with 6.12 branch
none
Xorg log with master
none
Xorg log with master and NewPLL=false
none
Registers dump using vesa and radeon drivers none

Description pha 2010-02-02 13:56:25 UTC
Created attachment 33015 [details]
xrandr with both monitors

i am having problems with dual dvi monitors, no matter which one is connected to first port stays blank
it seems xrandr is working, i can expand display to blanked monitor, i can open windows on it, but it won't come out of standby

here is a snippet from Xorg.0.log
(II) RADEON(0): Port0:
  XRANDR name: DVI-1
  Connector: DVI-I
  CRT1: INTERNAL_KLDSCP_DAC1
  DFP1: INTERNAL_UNIPHY
  DDC reg: 0x7e40
(II) RADEON(0): Port1:
  XRANDR name: DVI-0
  Connector: DVI-I
  CRT2: INTERNAL_KLDSCP_DAC2
  DFP2: INTERNAL_KLDSCP_LVTMA
  DDC reg: 0x7e50


Port0 is always blank, Port1 works fine

i use driver 6.12.4. xorg-server 1.6.3, kernel 2.6.32.7, no KMS
my card is
01:00.0 VGA compatible controller: ATI Technologies Inc Mobility Radeon HD 3470
01:00.0 0300: 1002:95c0
Comment 1 pha 2010-02-02 13:56:52 UTC
Created attachment 33016 [details]
xorg log
Comment 2 pha 2010-02-02 14:02:17 UTC
little clarification, it is a pci-express card although lspci says it is mobility radeon
Xorg.0.log states:
(--) RADEON(0): Chipset: "ATI Radeon HD 3470" (ChipID = 0x95c0)
...
...
(II) RADEON(0): ATOM BIOS Rom: 
        SubsystemVendorID: 0x174b SubsystemID: 0xe390
        IOBaseAddress: 0x3000
        Filename: s4e3901sa.01
        BIOS Bootup Message: 
11x-04E39-001SA-010 RV620 GDDR3_16Mx32 64bit 256MB 800e/800m                
Comment 3 Alex Deucher 2010-02-02 14:35:02 UTC
Can you try xf86-video-ati from git master?
Comment 4 pha 2010-02-04 04:56:40 UTC
i've tried git master and it doesn't work, monitor stays suspended

Comment 5 Alex Deucher 2010-02-04 07:29:23 UTC
(In reply to comment #4)
> i've tried git master and it doesn't work, monitor stays suspended
> 

Can you attach your log from git master?
Comment 6 pha 2010-02-04 07:32:28 UTC
Created attachment 33067 [details]
xorg log from git master
Comment 7 pha 2010-02-14 06:29:07 UTC
i've tried both monitors with intel driver and both work fine
amd binary driver also works, but for some reason it thinks i have 6400x1200 virtual desktop instead of 3200x1200
is there something i could add ?
Comment 8 Tim 2010-02-19 06:32:19 UTC
i have the same issue on debian/lenny. i have a radeon hd4350 with two dvi ports. the fglrx driver works well. 

i attached my lcpci, xrandr and Xorg.0.log

Comment 9 Tim 2010-02-19 06:33:38 UTC
Created attachment 33420 [details]
lspci of radeon hd4350
Comment 10 Tim 2010-02-19 06:34:51 UTC
Created attachment 33421 [details]
xrandr with DVI-0 blank
Comment 11 Tim 2010-02-19 06:36:59 UTC
Created attachment 33422 [details]
xorg.log with radeon 6.12.4
Comment 12 Alex Deucher 2010-02-19 08:23:21 UTC
(In reply to comment #8)
> i have the same issue on debian/lenny. i have a radeon hd4350 with two dvi
> ports. the fglrx driver works well. 
> 
> i attached my lcpci, xrandr and Xorg.0.log
> 

Does this work any better with git master or the latest bits from the 6.12-branch?
Comment 13 Loïc Lefort 2010-02-19 11:02:55 UTC
I have the same card, and I'm experiencing a similar problem using 2 dvi monitors: the monitor connected to DVI-1 always stays black.
I just tried both git master and 6.12 branch, neither works.
Comment 14 Loïc Lefort 2010-02-19 11:04:25 UTC
Created attachment 33431 [details]
Xorg log with 6.12 branch
Comment 15 Loïc Lefort 2010-02-19 11:04:43 UTC
Created attachment 33432 [details]
Xorg log with master
Comment 16 Alex Deucher 2010-02-19 11:06:26 UTC
Does:
Option "NewPLL" "FALSE"
help?
Comment 17 Loïc Lefort 2010-02-19 11:12:32 UTC
No, the monitor connected to DVI-1 stays blank
Comment 18 Loïc Lefort 2010-02-19 11:13:41 UTC
Created attachment 33433 [details]
Xorg log with master and NewPLL=false
Comment 19 Tim 2010-02-23 04:15:46 UTC
I follow these steps: http://wiki.x.org/wiki/radeonBuildHowTo but *without* remove the latest code changes. And every thing works fine!

(II) Loading /opt/xorg/lib/xorg/modules/drivers/radeon_drv.so
(II) Module radeon: vendor="X.Org Foundation"
        compiled for 1.7.4, module version = 6.12.99
        Module class: X.Org Video Driver
        ABI class: X.Org Video Driver, version 6.0
Comment 20 Alex Deucher 2010-02-23 07:38:31 UTC
(In reply to comment #19)
> I follow these steps: http://wiki.x.org/wiki/radeonBuildHowTo but *without*
> remove the latest code changes. And every thing works fine!

what are you removing or not removing?  If you are removing the driver before you've tested it, you won't be using it :)
Comment 21 Alex Deucher 2010-02-23 07:39:18 UTC
(In reply to comment #17)
> No, the monitor connected to DVI-1 stays blank
> 

Can you try again with git master?  You might also try reverting 0c3468d812e3790ce03d9e76779ae81e7b7b82d5
Comment 22 Loïc Lefort 2010-02-23 10:24:48 UTC
Still nothing on the second head with current master, and reverting 0c3468d812e3790ce03d9e76779ae81e7b7b82d5 doesn't help.
Is there any other information that I can provide to help debug this problem ?
Comment 23 Alex Deucher 2010-02-23 10:43:18 UTC
(In reply to comment #22)
> Still nothing on the second head with current master, and reverting
> 0c3468d812e3790ce03d9e76779ae81e7b7b82d5 doesn't help.
> Is there any other information that I can provide to help debug this problem ?
> 

Can you boot up with only the problematic monitor attached and then dump the registers using the vesa driver (assuming the head lights up ok), and then using the radeon driver.  use avivotool:
http://cgit.freedesktop.org/~airlied/radeontool
(as root):
./avivotool regs all > vesa.regs
./avivotool regs all > radeon.regs
Comment 24 Loïc Lefort 2010-02-23 12:34:23 UTC
(In reply to comment #23)
> Can you boot up with only the problematic monitor attached and then dump the
> registers using the vesa driver (assuming the head lights up ok), and then
> using the radeon driver.

It's not a problematic monitor, both monitors work fine when plugged on dvi-0 and in dualhead on a x1650. It's always the monitor plugged on dvi-1 that stays blank, even if I swap them.
Anyway, I did some register dumps :
- using vesa then radeon with both monitors plugged
- using vesa then radeon with the monitor on dvi-0 unplugged.
I did reboot between both tests, but not between the vesa/radeon driver switch. Using vesa, both heads display something, but it's all mangled. Using radeon, dvi-0 is fine, and dvi-1 stays blank.
Comment 25 Loïc Lefort 2010-02-23 12:34:56 UTC
Created attachment 33511 [details]
Registers dump using vesa and radeon drivers
Comment 26 pha 2010-03-30 09:02:32 UTC
i've just tested  xf86-video-ati-6.12.192, monitor is still off
Comment 27 Alex Deucher 2010-10-19 19:12:41 UTC
Is this still an issue with kms or a newer version of the driver?
Comment 28 pha 2010-11-22 11:15:19 UTC
hi,
this bug is fixed at least from 6.13.1, i'm using dual dvi right now
xorg-server 1.7.6, kernel 2.6.36

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.