Bug 18104

please attach your xorg log and config.

Created attachment 19714 [details]
config.

Created attachment 19715 [details]
config.

Comment on attachment 19715 [details]
config.

The connection is doing weird stuff.

Created attachment 19718 [details]
log.

Alex: log and config attached now.

Is there something else I can help with?

This is still happening on latest git (44abaa). I'm attaching a new log.

Created attachment 20068 [details]
Log for git version (44abaa).

Does it work properly if you start X with only LVDS and then attach the DVI monitor afterward?

(In reply to comment #10)
> Does it work properly if you start X with only LVDS and then attach the DVI
> monitor afterward?

Actually yes, it works fine when I do that.

I tried the same now with the VGA connector and it's the same issue: it works if I plug it in after starting X, but not if it's plugged in in the initialization.

I'm changing the summary to reflect that.

I have the exact same issue with the latest RADEON driver, 6.10.0 currently. I've posted a log file for bug #10682:

/attachment.cgi?id=22958

I've been experiencing this issue for more than one year hence I'm sticking to radeonhd for now, which doesn't exhibit the problem. But it averages less than 30% performance of RADEON, making 3D rendering unusable.

I know I can use xrandr and plug my monitor after I start X but then XFCE4, my desktop manager, sees only one screen, allows only for one wallpaper...

Is this bug solved in the upcoming RADEON release?

(In reply to comment #13)
> I have the exact same issue with the latest RADEON driver, 6.10.0 currently.
> I've posted a log file for bug #10682:
> 
> /attachment.cgi?id=22958
> 
> I've been experiencing this issue for more than one year hence I'm sticking to
> radeonhd for now, which doesn't exhibit the problem. But it averages less than
> 30% performance of RADEON, making 3D rendering unusable.
> 
> I know I can use xrandr and plug my monitor after I start X but then XFCE4, my
> desktop manager, sees only one screen, allows only for one wallpaper...
> 
> Is this bug solved in the upcoming RADEON release?

Yeah, it's quite annoying. I have to issue a bunch of xrandr commands each time I boot.

And of course sometimes I forget to unplug the display when I boot and I have to reboot.

I'm using the latest git and it's still not fixed.

Can you try with the radeon driver from git master?  I've made some changes that might help.  Also, can you attach your xorg log from git master and the output of xrandr --verbose?

(In reply to comment #15)
> Can you try with the radeon driver from git master?  I've made some changes
> that might help.  Also, can you attach your xorg log from git master and the
> output of xrandr --verbose?

It still doesn't work (with 31bd6d2). I'm attaching the log and xrandr output.

Created attachment 23034 [details]
Log of 31bd6d2.

Created attachment 23035 [details]
xrandr output of (31bd6d2)

Created attachment 23036 [details]
xrandr out put of (31bd6d2)

I didn't have the display connected in the previous xrandr output... I guess this is what you want.

Created attachment 23045 [details] [review]
Make sure crtc is enabled before attempting to blank/unblank

Does this patch help?

(In reply to comment #20)
> Created an attachment (id=23045) [details]
> Make sure crtc is enabled before attempting to blank/unblank
> 
> Does this patch help?

Nope, same behavior.

Actually I'm not sure this is really 'Unreversible blank screens when external LCD is plugged at init time' I thought the system was still working but there was no display working, but maybe I was wrong.

What I'm experiencing is that the the whole system hangs, not even the keyboard seems to work.

(In reply to comment #21)
> Actually I'm not sure this is really 'Unreversible blank screens when external
> LCD is plugged at init time' I thought the system was still working but there
> was no display working, but maybe I was wrong.
> 
> What I'm experiencing is that the the whole system hangs, not even the keyboard
> seems to work.
> 

This looks like a dup of bug 16781 which, if so, would mean the driver is stuck in a loop.  Can you access the machine remotely?

(In reply to comment #22)

> This looks like a dup of bug 16781 which, if so, would mean the driver is stuck
> in a loop.  Can you access the machine remotely?
> 

As far as I'm concerned, I could logon remotely and reboot the machine. I also see X is using 100% of my CPU meanwhile, giving the impression something is trapped in an infinite loop -- keyboard doesn't respond to the contrary of what I first reported.

(There might be multiple bug reports that dup 16781. I already tried to flag old ones that resemble this one.)

Please try the latest patch on bug 16781.

*** This bug has been marked as a duplicate of bug 16781 ***

Created attachment 23084 [details]
Xorg log file with all monitors turned black, xf86-video-ati from Gentoo overlays

I'm afraid it doesn't work for me though :( . I've just compiled ati-6.10.99 from Gentoo GIT overlay one hour ago and I still get black screens when X starts while my external monitor is plugged in the DVI-0 output. As usual, X uses 95-100% CPU when locked. I'm attaching xorg log.

(In reply to comment #25)
> Created an attachment (id=23084) [details]
> Xorg log file with all monitors turned black, xf86-video-ati from Gentoo
> overlays
> 
> I'm afraid it doesn't work for me though :( . I've just compiled ati-6.10.99
> from Gentoo GIT overlay one hour ago and I still get black screens when X
> starts while my external monitor is plugged in the DVI-0 output. As usual, X
> uses 95-100% CPU when locked. I'm attaching xorg log.
> 

Are you sure the gentoo 6.10.99 overlay has the latest git bits? I released 6.10.99 a week ago and the latest fixes just went into git in the last two days.

Can I tell from the log I compiled the right version?

(In reply to comment #27)
> Can I tell from the log I compiled the right version?
> 

nope, probably not.

My mistake :s ... I have just emerged xf86-video-9999 (hard masked) and now it works; my DVI output now works and I can see both screens as configured in xorg.conf.

Thanks a whole lot for the work! Thanks, thanks, thanks :-) .