Loading radeon module with KMS enabled leads to a black screen

delerious010 · 2011-03-02 03:13:06

I have a HD4350 and have tried both the HDMI > HDMI and DVI > HDMI connectors to plug into my TV.

With either setup, or even with both plugged in, enabling KMS with either early or late start leads to a black screen. Specifically, by the way the TV fails to detect a resolution even when switching inputs, I've got to assume that the video card stops producing output over either port. Popping into the first terminal and running a reboot command works fine, so the system isn't completely locked down, it's really just video related.

Disabling KMS enables the screen to function properly in either or all setups. However, due to the new ati-dri and libgl requiring KMS or segfaulting, this does'nt seem to be a sustainable setup .. yay.

I've tried with the current packages as well as a variety of setups with the experimental packages in the other forum thread, however they all produce the same black screen when loading in the module.

What steps could I take to get additional output in order to catch a specific error message ? Or how could I try to prevent this ? I'm pretty much out of ideas at this point !

All help or comments much appreciated,
- Jon.

delerious010 · 2011-03-03 00:41:24

Did more poking around...
Managed to get the output of dmesg, it displays that KMS loads and that the 3 outputs from the card are displayed ( VGA, DVI and HDMI ). Sadly, much to my very great regret, it does not display any error messages.

delerious010 · 2011-03-03 00:43:29

May as well add that specifying a video="DVI-I:1920x1080-24@60" does not work nor does it work more with HDMI-A.

delerious010 · 2011-03-03 01:07:08

Working with https://bugs.freedesktop.org/show_bug.cgi?id=27731, attempted drm-radeon-testing kernel with radeon.modeset=1 and radeon.audio=0, still a black screen.

slacknatcher · 2011-03-03 06:23:04

I had also experienced this black screen on my main monitor (which goes through a VGA port) on my radeon 3300 HD

I couldn't solve the problem so i took a drastic measure and downloaded "new" drivers from the Radeon-repo, hey work great but at least for me they don't provide KMS, although they provide Gallium-3D for specific cards .

On the link above around page 76? are discussing about black screen which could be related to us, i hope the transition from xf86-video-ati from 6.14.0-1 -> 6.14.0-2 solves the problem.

Could you also post a "dmesg | tail -n 20" please?

slacknatcher · 2011-03-03 06:47:55

i was in archlinux website and saw on recently updated package the "ati-dri" package , so i uninstalled the ones from radeon-repo and tried extra's drivers to see if this solved the problem, they worked !, no more black screen at least for me.

try updating now.

delerious010 · 2011-03-03 13:18:30

Slack :
Sadly, I've tried both the latest ati-dri and and xf86-video-ati packages currently in the repos and both lead to a black screen when loading X without KMS. This is the same Segmentation Fault as reported in another forum post. To get around this, I've had to downgrade the following packages :
/var/cache/pacman/pkg/ati-dri-7.10.0.git20110206-2-x86_64.pkg.tar.tar
/var/cache/pacman/pkg/xf86-video-ati-6.13.2-2-x86_64.pkg.tar.xz

I'll give the Gallium drivers a try later on tonight, this time remembering to snapshot my root just in case -_-

As for the dmesg, what info you looking for in there, the tail's got nothing of relevance in it. I've pasted the output of the following if you find this helpful ( dmesg.log which is a copy from when I booted up with KMS enabled ) :

cat dmesg.log | egrep -i 'ati|radeon|xf86|kms|dri|glx|drm|hdmi|dvi'

Command line: BOOT_IMAGE=/vmlinuz26 root=/dev/disk/by-label/system rootflags=rw,subvol=root rootfstype=btrfs video=DVI-I:1920x1080-24@60 radeon.modeset=1 ro
Using ACPI (MADT) for SMP configuration information
Allocating PCI resources starting at dff00000 (gap: dff00000:10100000)
Kernel command line: BOOT_IMAGE=/vmlinuz26 root=/dev/disk/by-label/system rootflags=rw,subvol=root rootfstype=btrfs video=DVI-I:1920x1080-24@60 radeon.modeset=1 ro
Preemptable hierarchical RCU implementation.
Fast TSC calibration using PIT
Calibrating delay loop (skipped), value calculated using timer frequency.. 5226.62 BogoMIPS (lpj=8706596)
Performance Events: AMD PMU driver.
PCI: Using configuration type 1 for base access
ACPI Exception: AE_NOT_FOUND, Evaluating _PRW (20101013/scan-723)
ACPI Exception: AE_NOT_FOUND, Evaluating _PRW (20101013/scan-723)
ACPI Exception: AE_NOT_FOUND, Evaluating _PRW (20101013/scan-723)
ACPI Exception: AE_NOT_FOUND, Evaluating _PRW (20101013/scan-723)
Block layer SCSI generic (bsg) driver version 0.4 loaded (major 253)
Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
PM: Hibernation image not present or could not be loaded.
[drm] Initialized drm 1.1.0 20060810
[drm] radeon kernel modesetting enabled.
radeon 0000:07:00.0: PCI INT A -> Link[APC6] -> GSI 16 (level, low) -> IRQ 16
radeon 0000:07:00.0: setting latency timer to 64
[drm] initializing kernel modesetting (RV710 0x1002:0x954F).
[drm] register mmio base: 0xFDCE0000
[drm] register mmio size: 65536
radeon 0000:07:00.0: VRAM: 512M 0x0000000000000000 - 0x000000001FFFFFFF (512M used)
radeon 0000:07:00.0: GTT: 512M 0x0000000020000000 - 0x000000003FFFFFFF
[drm] Detected VRAM RAM=512M, BAR=256M
[drm] RAM width 64bits DDR
[drm] radeon: 512M of VRAM memory ready
[drm] radeon: 512M of GTT memory ready.
radeon 0000:07:00.0: irq 40 for MSI/MSI-X
radeon 0000:07:00.0: radeon: using MSI.
[drm] radeon: irq initialized.
[drm] GART: num cpu pages 131072, num gpu pages 131072
[drm] Loading RV710 Microcode
radeon 0000:07:00.0: WB enabled
[drm] ring test succeeded in 1 usecs
[drm] radeon: ib pool ready.
[drm] ib test succeeded in 0 usecs
[drm] Enabling audio support
failed to evaluate ATIF got AE_BAD_PARAMETER
[drm] Radeon Display Connectors
[drm] Connector 0:
[drm]   VGA
[drm]   DDC: 0x7e40 0x7e40 0x7e44 0x7e44 0x7e48 0x7e48 0x7e4c 0x7e4c
[drm]   Encoders:
[drm]     CRT2: INTERNAL_KLDSCP_DAC2
[drm] Connector 1:
[drm]   HDMI-A
[drm]   HPD1
[drm]   DDC: 0x7e50 0x7e50 0x7e54 0x7e54 0x7e58 0x7e58 0x7e5c 0x7e5c
[drm]   Encoders:
[drm]     DFP1: INTERNAL_UNIPHY
[drm] Connector 2:
[drm]   DVI-I
[drm]   HPD4
[drm]   DDC: 0x7f10 0x7f10 0x7f14 0x7f14 0x7f18 0x7f18 0x7f1c 0x7f1c
[drm]   Encoders:
[drm]     CRT1: INTERNAL_KLDSCP_DAC1
[drm]     DFP2: INTERNAL_UNIPHY2
[drm] Internal thermal controller without fan control
[drm] radeon: power management initialized
[drm] fb mappable at 0xE0142000
[drm] vram apper at 0xE0000000
[drm] size 8294400
[drm] fb depth is 24
[drm]    pitch is 7680
fb0: radeondrmfb frame buffer device
drm: registered panic notifier
[drm] Initialized radeon 2.7.0 20080528 for 0000:07:00.0 on minor 0

Mr.Elendig · 2011-03-03 16:23:00

https://wiki.archlinux.org/index.php/In … modules.22 (yes I know it says intel, but it is the same for any kms implimentation.)

And you can't use the new ati drivers without kms.

Last edited by Mr.Elendig (2011-03-03 16:23:18)

delerious010 · 2011-03-03 18:03:27

My Elendig,

That I can't use the new ATI drivers without KMS is precisely why I've posted here.
KMS refuses to output via either DVI-to-HDMI or HDMI-to-HDMI no matter what I do, so I'm fairly SOL right now.

As I've mentioned previously, I've tried so force output over either the DVI or HDMI ports with the video= statement on by kernel line to no avail. I've also tried adding either just "radeon" or "intel_agp radeon" to my MODULES line in mkinitcpio.conf to test various early start methods and neither have resolved this issue yet. As soon as the radeon module gets loaded and KMS/the framebuffer initialize, my screen goes blank.

Looking this over again, though, I see that the 1 thing I was doing differently from your example, and from the ATI wiki doc is that after the refresh rate ( 60 ), I have not been appending either "d" or "e". Any idea what those letters refer to and where I can find additional documentation on this ?

slacknatcher · 2011-03-03 18:38:32

meanwhile could you try the radeon-repo packages? as a for the moment measure

Mr.Elendig · 2011-03-03 19:03:32

d = disable the output.

Disable the outputs you don't use, and explicitly spesify the ones you do use. It should work just fine.

delerious010 · 2011-03-04 00:38:16

First off guys, thanks for providing feedback, it's really quite appreciated !

Alright, so today I've tried both the standard arch kernel 2.6.37 as well as that from the radeon repo. Each using their own radeon driver and firmwares. And each providing the same dull black screen whenever the radeon kernel with kms enabled loads.

The last attempt was with the following kernel parameters .. it's got about everything, including a kitchen sink thrown in for good measure :

BOOT_IMAGE=/vmlinuz26-drm-radeon-testing root=/dev/disk/by-label/system rootflags=rw,subvol=__active rootfstype=btrfs radeon.modeset=1 radeon-audio=0 video=HDMI-A:e video=DVI-I:d video=HDMI-0:e video=DVI-0:d video=VGA:d video=VGA-0:d video=HPD1:e

dmsg, kms/drm bits :

[drm] Initialized drm 1.1.0 20060810
[drm] radeon kernel modesetting enabled.
ACPI: PCI Interrupt Link [APC6] enabled at IRQ 16
radeon 0000:07:00.0: PCI INT A -> Link[APC6] -> GSI 16 (level, low) -> IRQ 16
radeon 0000:07:00.0: setting latency timer to 64
[drm] initializing kernel modesetting (RV710 0x1002:0x954F).
[drm] register mmio base: 0xFDCE0000
[drm] register mmio size: 65536
ATOM BIOS: 
radeon 0000:07:00.0: VRAM: 512M 0x0000000000000000 - 0x000000001FFFFFFF (512M used)
radeon 0000:07:00.0: GTT: 512M 0x0000000020000000 - 0x000000003FFFFFFF
[drm] Detected VRAM RAM=512M, BAR=256M
[drm] RAM width 64bits DDR
[TTM] Zone  kernel: Available graphics memory: 2027264 kiB.
[TTM] Initializing pool allocator.
[drm] radeon: 512M of VRAM memory ready
[drm] radeon: 512M of GTT memory ready.
[drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[drm] Driver supports precise vblank timestamp query.
radeon 0000:07:00.0: irq 46 for MSI/MSI-X
radeon 0000:07:00.0: radeon: using MSI.
[drm] radeon: irq initialized.
[drm] GART: num cpu pages 131072, num gpu pages 131072
[drm] Loading RV710 Microcode
radeon 0000:07:00.0: WB enabled
[drm] ring test succeeded in 1 usecs
[drm] radeon: ib pool ready.
[drm] ib test succeeded in 0 usecs
[drm] Enabling audio support
failed to evaluate ATIF got AE_BAD_PARAMETER
[drm] Radeon Display Connectors
[drm] Connector 0:
[drm]   VGA
[drm]   DDC: 0x7e40 0x7e40 0x7e44 0x7e44 0x7e48 0x7e48 0x7e4c 0x7e4c
[drm]   Encoders:
[drm]     CRT2: INTERNAL_KLDSCP_DAC2
[drm] Connector 1:
[drm]   HDMI-A
[drm]   HPD1
[drm]   DDC: 0x7e50 0x7e50 0x7e54 0x7e54 0x7e58 0x7e58 0x7e5c 0x7e5c
[drm]   Encoders:
[drm]     DFP1: INTERNAL_UNIPHY
[drm] Connector 2:
[drm]   DVI-I
[drm]   HPD4
[drm]   DDC: 0x7f10 0x7f10 0x7f14 0x7f14 0x7f18 0x7f18 0x7f1c 0x7f1c
[drm]   Encoders:
[drm]     CRT1: INTERNAL_KLDSCP_DAC1
[drm]     DFP2: INTERNAL_UNIPHY2
[drm] Internal thermal controller without fan control
[drm] radeon: power management initialized
[drm] fb mappable at 0xE0142000
[drm] vram apper at 0xE0000000
[drm] size 8294400
[drm] fb depth is 24
[drm]    pitch is 7680
Console: switching to colour frame buffer device 240x67
fb0: radeondrmfb frame buffer device
drm: registered panic notifier
[drm] Initialized radeon 2.8.0 20080528 for 0000:07:00.0 on minor 0

delerious010 · 2011-03-04 00:48:19

Oh Crap ...
Crap
Crap
Crap
Crap
Crap

So it appears that I did indeed find my issue a few days ago. However, a stupid freaking typo prevented me from implementing it ... instead of adding radeon.audio=0 I added radeon-audio=0 to my kernel boot line.

So yet, HDMI audio is causing my woes. The patch that was added to 2.6.38, and therefore is available in the drm-radeon-testing kernel provides support for disabling audio, and therefore enabling video.

Yay.

delerious010 wrote:

Working with https://bugs.freedesktop.org/show_bug.cgi?id=27731, attempted drm-radeon-testing kernel with radeon.modeset=1 and radeon.audio=0, still a black screen.

delerious010 · 2011-03-04 01:05:20

Ok so recap.

Ati Radeon HD4350, connecting either via HDMI or DVI>HDMI to my TV would output nothing but a black screen if KMS was enabled and as soon as the radeon driver would load in CLI or X.

The fix to this is found in the 2.6.38 kernel and activated simply by appending radeon.audio=0 to kernel boot line in grub. Side note, as was mentioned elsewhere, UDEV takes tons of time to load if you don't explicitly define radeon.modeset=1 on the command line.

With this, I've now managed to use the latest package versions, compiz, etc...

slacknatcher · 2011-03-04 15:47:22

good at least now you have a functional output, hope 2.6.38 solve this

Last edited by slacknatcher (2011-03-04 15:52:39)

Arch Linux

#1 2011-03-02 03:13:06

Loading radeon module with KMS enabled leads to a black screen

#2 2011-03-03 00:41:24

Re: Loading radeon module with KMS enabled leads to a black screen

#3 2011-03-03 00:43:29

Re: Loading radeon module with KMS enabled leads to a black screen

#4 2011-03-03 01:07:08

Re: Loading radeon module with KMS enabled leads to a black screen

#5 2011-03-03 06:23:04

Re: Loading radeon module with KMS enabled leads to a black screen

#6 2011-03-03 06:47:55

Re: Loading radeon module with KMS enabled leads to a black screen

#7 2011-03-03 13:18:30

Re: Loading radeon module with KMS enabled leads to a black screen

#8 2011-03-03 16:23:00

Re: Loading radeon module with KMS enabled leads to a black screen

#9 2011-03-03 18:03:27

Re: Loading radeon module with KMS enabled leads to a black screen

#10 2011-03-03 18:38:32

Re: Loading radeon module with KMS enabled leads to a black screen

#11 2011-03-03 19:03:32

Re: Loading radeon module with KMS enabled leads to a black screen

#12 2011-03-04 00:38:16

Re: Loading radeon module with KMS enabled leads to a black screen

#13 2011-03-04 00:48:19

Re: Loading radeon module with KMS enabled leads to a black screen

#14 2011-03-04 01:05:20

Re: Loading radeon module with KMS enabled leads to a black screen

#15 2011-03-04 15:47:22

Re: Loading radeon module with KMS enabled leads to a black screen

Board footer