Bug 27470

Created attachment 34678 [details]
expected background image and cursor

Created attachment 34679 [details]
picture of the distortions on the screen

Created attachment 34680 [details]
dmesg output of boot

Does radeon.new_pll=0 help?

(In reply to comment #4)
> Does radeon.new_pll=0 help?

Negative. No change.

Is this still an issue with a newer version of the kernel or driver?
If so, does:
xrandr --output DVI-0 --set coherent 0
help?

(In reply to comment #6)
> Is this still an issue with a newer version of the kernel or driver?
> If so, does:
> xrandr --output DVI-0 --set coherent 0
> help?

Problem still exists with 2.6.37-rc7. In the meantime I got a new display. It's a Dell 3008 WFP with same resolution as the other one. This time I do get an image, but horizontal resolution is exactly half of what it should be. So instead of 2560x1600 I get 1280x1600 scaled up by the display. Resolutions up to 1920x1200 work fine. xrandr --output DVI-0 --set coherent 0 did not change anything.

Please tell me what I can do to solve this or help solving this. This time I actually own the display and I would very much like to use it's full potential with free radeon drivers. I have no problem running bleeding edge stuff and am a programmer myself, so any hint where to get started would be very appreciated.

I have a Dell 3008 WFP as well and it works on all cards I have access to including an rv770. Perhaps there is a problem with one of your DVI ports?  Does the other DVI port work any better?  Please attach your xorg log and dmesg output with the latest kernel and also a copy of your vbios.  To get a copy of your bios:
(as root)
(use lspci to get the bus id)
cd /sys/bus/pci/devices/<pci bus id>
echo 1 > rom
cat rom > /tmp/vbios.rom
echo 0 > rom

Problem persists with the second DVI port and with a different dual link DVI cable.
Now running 2.6.37-rc8.

Created attachment 41717 [details]
dmesg on 2.6.37-rc8 with new display

Created attachment 41718 [details]
Xorg.0.log

Created attachment 41719 [details]
vbios.rom

Just a heads up: problem's still there on 2.6.37 final. 

And for completeness, lspci output for my graphics card:

05:00.0 VGA compatible controller: ATI Technologies Inc RV770 [Radeon HD 4850] (prog-if 00 [VGA controller])
        Subsystem: Giga-byte Technology Device 21c7
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 40
        Region 0: Memory at d0000000 (64-bit, prefetchable) [size=256M]
        Region 2: Memory at fd8e0000 (64-bit, non-prefetchable) [size=64K]
        Region 4: I/O ports at 6c00 [size=256]
        [virtual] Expansion ROM at fd800000 [disabled] [size=128K]
        Capabilities: <access denied>
        Kernel driver in use: radeon

Created attachment 41950 [details] [review]
new pll algo

Does this patch help?

Created attachment 41952 [details]
patched files

Negative. Patch applies to 2.6.37, but the problem still persists:

patching file drivers/gpu/drm/radeon/atombios_crtc.c
Hunk #1 succeeded at 915 (offset -48 lines).
patching file drivers/gpu/drm/radeon/radeon_display.c
Hunk #1 succeeded at 448 (offset -332 lines).
Hunk #2 succeeded at 560 (offset -332 lines).
Hunk #3 succeeded at 619 (offset -332 lines).
Hunk #4 succeeded at 735 (offset -332 lines).
patching file drivers/gpu/drm/radeon/radeon_legacy_crtc.c
patching file drivers/gpu/drm/radeon/radeon_mode.h

Attaching patched files in case you want to check the rather large offsets.

Retested on 2.6.38-rc2. The patch now applies cleanly but still does not help. I also tried the radeonhd and fglrx drivers. All show the same problem. With radeonhd, the highest working resolution is even lower at 1600x1200.

I start suspecting a hardware problem here, since all drivers show such consistent failures.

I bit the bullet and tested on Windows 7 -> same problem there. So I bought a Sapphire Radeon HD5670 and had the tech support confirm that this card does support 2560x1600. But still, using DVI, I get exactly the same problems. Frustration reached a new peak.

But the upside is: this card supports displayport and with UMS I am finally able to use full resolution of my display! But KMS does not find the displayport output and the screen stays blank. And without KMS I get no acceleration at all and no power management.

Created attachment 42960 [details]
dmesg output of a drm-fixes kernel with the following (useless) patch, showing failure to detect DP output

diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index 695de9a..a8c23c6 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -356,7 +356,8 @@ retry:
        atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
        if (args.v1.ucReplyStatus && !args.v1.ucDataOutLen) {
-               if (args.v1.ucReplyStatus == 0x20 && retry_count++ < 10)
+              if (args.v1.ucReplyStatus != 3 && retry_count++ < 2000)

Can you attach a copy of your vbios from the new card?

(as root)
(use lspci to get the bus id)
cd /sys/bus/pci/devices/<pci bus id>
echo 1 > rom
cat rom > /tmp/vbios.rom
echo 0 > rom

Created attachment 42993 [details]
vbios of the 5670

Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.