2007/9/26, Luc Verhaegen :

On Fri, Sep 21, 2007 at 04:38:10PM +0200, Ludovic Rousseau wrote:

...

Hello,

On my ATI Radeon HD 2600 PRO (in an iMac 24" aluminium)

$ ./rhd_conntest 01:00.0 Checking connectors on 0x9583, 0x106B, 0x0083 (@01:00:00): Load Detection: RHD_OUTPUT_NONE HotPlug: RHD_HPD_NONE DDC: RHD_DDC_1

The card is connected to the internal pannel.

I'm sorry Ludovic,

But all r600 based laptop cards are currently beyond our capabilities.

Argh!

We still need further information on the LVTMA device on r600 and we're hoping to see this soon.

Only then can we consider working on these devices.

I'm sorry for this, but we are moving as fast as we can :)

OK. I will continue reading the mailing list so I know when I can help test alpha/beta versions for my card. I will then use Linux inside vmware on this machine for now. Thanks -- Dr. Ludovic Rousseau -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

2007/9/26, Luc Verhaegen :

On Fri, Sep 21, 2007 at 04:38:10PM +0200, Ludovic Rousseau wrote:

...

Hello,

On my ATI Radeon HD 2600 PRO (in an iMac 24" aluminium)

$ ./rhd_conntest 01:00.0 Checking connectors on 0x9583, 0x106B, 0x0083 (@01:00:00): Load Detection: RHD_OUTPUT_NONE HotPlug: RHD_HPD_NONE DDC: RHD_DDC_1

The card is connected to the internal pannel.

I'm sorry Ludovic,

But all r600 based laptop cards are currently beyond our capabilities.

We still need further information on the LVTMA device on r600 and we're hoping to see this soon.

Only then can we consider working on these devices.

I'm sorry for this, but we are moving as fast as we can :)

Using git version 737d940b83001dbba858d9025a25024bda439726 as of today I still have the same errors: (EE) RADEONHD(0): RHDLVTMAInit: any other device than an R5xx is still unsupported. (EE) RADEONHD(0): Failed to detect a connected monitor (EE) Screen(s) found, but none have a usable configuration. Complete log attached Do you need more info from ATI about LVTMA device on r600? Is ATI blocking these information? Or are you "just" short in time to implement the needed parts? Thanks for your work! -- Dr. Ludovic Rousseau

On Nov 25, 07 22:43:03 +0100, Ludovic Rousseau wrote:

2007/9/26, Luc Verhaegen :

...

But all r600 based laptop cards are currently beyond our capabilities.

We still need further information on the LVTMA device on r600 and we're hoping to see this soon.

Only then can we consider working on these devices.

With git version 45b217d263e232f1f57778e340371c95105c87e8 (as of today) my card is now working. Yeah! thanks.

The flat panel is used at its native resolution of 1920x1200 but the DPI resolution is not correct. The font size in Gnome is _very_ big even with a font of 10 point.

Should be fixed in master. CU Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

On Nov 28, 07 16:41:50 +0100, Ludovic Rousseau wrote:

$ xrandr Screen 0: minimum 320 x 200, current 1920 x 1200, maximum 1920 x 1920 DVI-I_1/digital disconnected DVI-I_1/analog disconnected PANEL connected 1920x1200+0+0 262mm x 324mm 1920x1200 60.2*+ TV_7PIN_DIN disconnected

The screen width physical size is still 262mm instead of something like 524mm.

I assume this is the monitor lying to you. Is the 324mm ok? In that case, you'll have to specify the dimension of the screen manually in xorg.conf

(II) RADEONHD(0): Max H-Image Size [cm]: horiz.: 52 vert.: 32

Hm, that looks ok... I'm not doing anything with the EDID data, this is base RandR's behavior. Still:

(II) RADEONHD(0): Supported additional Video Mode: (II) RADEONHD(0): clock: 156.0 MHz Image Size: 262 x 324 mm

Ah, there it gets it from. So the monitor is lying in its additional video mode image size, and that one has precedence over the general info block... Until we have a *real* monitor-specific quirk database, I'm sorry but you have to configure that manually. I wouldn't dare to design a logic that sometimes takes the one value, sometimes the other. This is prone to fail. Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

On Dec 01, 07 13:09:48 +0100, Ludovic Rousseau wrote:

2007/11/28, Matthias Hopf :

...

...

(II) RADEONHD(0): Supported additional Video Mode: (II) RADEONHD(0): clock: 156.0 MHz Image Size: 262 x 324 mm

Ah, there it gets it from. So the monitor is lying in its additional video mode image size, and that one has precedence over the general info block...

I have to add here: I have *no* idea why this image size has precedence over the main image size. I guess some other monitor got it wrong the other way, and this kludge was introduced without asking anybody else. This is done in the general (read: non-driver-specific) RandR code. But that is just wild guessing. Maybe something to ask on xorg-devel.

I tried to set the physical display in my xorg.conf (attached). (II) RADEONHD(0): Setting screen physical size to 262 x 324 (...)

Why is radeonhd not using my settings?

You're now stumbling over a bug that is related to Xserver 1.4 (assuming you run that one), which has been fixed in git. AFAIR. Nothing radeonhd related.

Is the driver using a different monitor from the one I configured (called panel24)?

Nope. RandR is just ignoring the displaysize of a configured monitor (and only that). Sorry, I know this is painful. You're not the first to stumble over that. Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

On Jan 25, 08 09:38:24 +0100, Ludovic Rousseau wrote:

...

...

Is the driver using a different monitor from the one I configured (called panel24)?

Nope. RandR is just ignoring the displaysize of a configured monitor (and only that).

Sorry, I know this is painful. You're not the first to stumble over that.

Just to report that I found the solution to my problem of wrong physical dimensions. The solution is described at [1].

My xorg.conf is now: Section "Device" Identifier "generic video card" Driver "radeonhd" EndSection

Section "Monitor" Identifier "PANEL" DisplaySize 524 324 EndSection

This only works with recent Xservers. The DisplaySize was ignored for quite some time.

Auto detection of everything is fine unless your hardware is lying and you need to correct it :-)

Indeed, X.org is getting somewhere :) Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org