[ATI][KMS][R280] Text console hidden and horizontal black line at bottom of screen (on xorg)

[Daniele Napolitano]

Ubuntu version: 9.10 Karmic
Kernel: 2.6.31-10-generic

Video card:
VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01)

I've follow this steps:
1) Go in the VT1
2) sudo service gdm stop
3) sudo rmmod radeon
4) sudo modprobe radeon modeset=1

Then text console disappear (think change screen resolution) but still black (the monitor does not enter standby).

I try to restart gdm (without visual feedback) with:
sudo service gdm start

X start correctly, so login into GNOME session. All ok.

Try to switch VT (return on VT1) but it is still black. Returning on VT7 (managed by X) appears a black section (~150px) on the bottom of the desktop that hide all behind. Curiously, if I go with the mouse on the border bottom of the screen, appear a strange verticale pixel noise (seems corrupted VRAM) the follow horizontal mouse movements.

dmesg log after loading radeon module:
[ 161.066163] [drm] radeon kernel modesetting enabled.
[ 161.078036] [drm] radeon: Initializing kernel modesetting.
[ 161.078076] [drm] register mmio base: 0xDF800000
[ 161.078079] [drm] register mmio size: 65536
[ 161.079211] [drm] GPU reset succeed (RBBM_STATUS=0x00000140)
[ 161.079218] [drm] Clocks initialized !
[ 161.079223] [drm] Generation 2 PCI interface, using max accessible memory
[ 161.079351] [drm] Detected VRAM RAM=128M, BAR=128M
[ 161.079358] [drm] RAM width 128bits DDR
[ 161.079593] agpgart-intel 0000:00:00.0: AGP 2.0 bridge
[ 161.079616] agpgart-intel 0000:00:00.0: putting AGP V2 device into 4x mode
[ 161.079647] radeon 0000:01:00.0: putting AGP V2 device into 4x mode
[ 161.079670] [drm] radeon: VRAM 128M
[ 161.079673] [drm] radeon: VRAM from 0x00000000 to 0x07FFFFFF
[ 161.079675] [drm] radeon: GTT 64M
[ 161.079677] [drm] radeon: GTT from 0xF8000000 to 0xFBFFFFFF
[ 161.240805] [drm] radeon: irq initialized.
[ 161.240886] [TTM] TTM available graphics memory: 501 MiB
[ 161.240889] [TTM] TTM available object memory: 433 MiB
[ 161.240915] [drm] radeon: 128M of VRAM memory ready
[ 161.240918] [drm] radeon: 64M of GTT memory ready.
[ 161.240934] [drm] radeon: cp idle (0x02000000)
[ 161.240938] [drm] Loading R200 Microcode
[ 161.243574] [drm] radeon: ring at 0x00000000F8000000
[ 161.243595] [drm] ring test succeeded in 1 usecs
[ 161.246885] [drm] radeon: ib pool ready.
[ 161.246980] [drm] ib test succeeded in 0 usecs
[ 161.254794] [drm] Default TV standard: PAL
[ 161.254802] [drm] 27.000000000 MHz TV ref clk
[ 161.254804] [drm] No TV DAC info found in BIOS
[ 161.254809] [drm] DFP table revision: 3
[ 161.256433] [drm] Radeon Display Connectors
[ 161.256438] [drm] Connector 0:
[ 161.256440] [drm] VGA
[ 161.256444] [drm] DDC: 0x60 0x60 0x60 0x60 0x60 0x60 0x60 0x60
[ 161.256446] [drm] Encoders:
[ 161.256448] [drm] CRT1: INTERNAL_DAC1
[ 161.256451] [drm] Connector 1:
[ 161.256452] [drm] DVI-I
[ 161.256456] [drm] DDC: 0x64 0x64 0x64 0x64 0x64 0x64 0x64 0x64
[ 161.256458] [drm] Encoders:
[ 161.256460] [drm] CRT2: INTERNAL_DAC2
[ 161.256462] [drm] DFP1: INTERNAL_TMDS1
[ 161.317519] i2c-adapter i2c-1: unable to read EDID block.
[ 161.317529] radeon 0000:01:00.0: DVI-I-1: no EDID data
[ 161.442115] [drm] fb mappable at 0xF0040000
[ 161.442121] [drm] vram apper at 0xF0000000
[ 161.442123] [drm] size 5242880
[ 161.442126] [drm] fb depth is 24
[ 161.442128] [drm] pitch is 5120
[ 161.443800] fb0: radeondrmfb frame buffer device
[ 161.443808] registered panic notifier
[ 161.443811] [drm] radeon: kernel modesetting successfully initialized.
[ 161.443821] [drm] Initialized radeon 2.0.0 20080528 for 0000:01:00.0 on minor 0
[ 187.263699] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
[ 187.313998] [drm] DAC-5: set mode 1280x1024 14
[ 238.424735] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
[ 241.510133] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
[ 247.293125] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id

Everytime that try to switch VT, it repeat last line.

[John Paul Adrian Glaubitz (glaubitz)]

Hi,

I had exactly the same problem on Karmic (Ubuntu 9.10) with a Radeon RV-370 and xf86-video-ati from git.

I asked the radeon-developers in #radeon on FreeNode and they suggested that "fbcon" is missing. It turned out to be right, by loading that module, console switching worked immediately.

Loading the module (as root):

modprobe fbcon

Making this permanent:

echo "fbcon" >> /etc/modules

Adrian

[Daniele Napolitano (dnax88)]

Thanks Adrian! This solution solves all problems reported here perfectly.

But the radeon module should not automatically invoke fbcon?

[John Paul Adrian Glaubitz (glaubitz)]

Hi,

no, I asked Dave Airlie on that (the radeon kms developer) on IRC:

22:47 < cbmuser> hmm, guess radeon should depend on fbcon then, shouldn't it?
22:48 < airlied> no, its a distro thing
22:48 < airlied> it selects fbcon but it can't make you load it
22:49 < cbmuser> ok, I guess I will report that on Ubuntu then !
22:49 < cbmuser> thanks alot
22:49 * airlied wonders how vesafb works for hem
22:50 < airlied> if they don't have fbcon built in

So he says, that radeon uses fbcon once it's loaded. But your distribution (in this case Ubuntu) should take care of it since the module seems to be rather essential (you always need it when using framebuffer-consoles). vesafb also depends on it.

Personally, however, I agree with you and think that radeon should depend on it if otherwise your framebuffer and X will be broken.

Adrian

[Jeremy Foshee (jeremyfoshee)]

Hi Daniele,

This bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? Can you try with the latest development release of Ubuntu? ISO CD images are available from http://cdimage.ubuntu.com/releases/ .

If it remains an issue, could you run the following command from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux 434190

Also, if you could test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Daniele Napolitano (dnax88)]

Yes. This is fixed in Ubuntu 9.10 Karmic.