Opened 15 years ago

Closed 3 years ago

#4523 closed bug (invalid)

black screen on ATI Mobility Radeon X700

Reported by: makitka Owned by: euan
Priority: normal Milestone: R1
Component: Drivers/Graphics/radeon Version: R1/Development
Keywords: Cc: siarzhuk
Blocked By: #7662 Blocking:
Platform: All

Description

sorry for english. i am from russia

after boot on my laptop there is nothing else but only a black screen. seems than system works - because same problem was in Zeta with Radeon driver - and there was a black screen and a sound after 10 seconds (mediaserver start). videocard has two outputs (notebook Acer Acpire 1690), maybe image translates to a wrong output? in VESA works fine, but there is not 1280x800 support in VESA, so..

Attachments (3)

listdev (3.2 KB ) - added by x-ist 12 years ago.
SYSLOG00.TXT (126.7 KB ) - added by x-ist 12 years ago.
SYSLOG01.TXT (144.4 KB ) - added by x-ist 12 years ago.
more verbose readeon driver logging

Download all attachments as: .zip

Change History (25)

comment:1 by michael.weirauch, 15 years ago

This sounds very much like the problem I had with the Mobility Radeon X600 on my old (and now dead) Samsung X20. Can you check if the screen is really black or if just the backlight is not enabled.

For me, the backlight was not enabled. But when looking closely, one could see the windows/desktop. You can try using the vesa driver/accelerant (safe mode options in boot loader) for the time being and check if that works for you. It did for me.

comment:2 by peterm1561, 15 years ago

The same problem here. I own a very old (ex-zeta) Compaq Armada with ATI Rage mobility P/M AGP 2x (device 4c4d) This worked very well in Zeta with the ati mobility driver from Bebits. On haiku it worked right away. Wonderful...until the laptop came in power save mode.... black screen. It seems that the screen is not initialized properly.

My other laptop (HP NC8000) runs haiku nicely but in this case the screen is black right from boot. If you look carefully you can see that the screen is working OK but that it a problem with the back light.

device Display controller (VGA compatible controller, VGA controller) [3|0|0]

vendor 1002: ATI Technologies Inc device 4e50: RV350 [Mobility Radeon 9600 M10]

comment:3 by siarzhuk, 15 years ago

Cc: imker@… added

The message to euan: Can this problem have the same origins as #1924 observed by me? Note that it still observed on my notebook. Do you have new "fresh" ideas about this behavior? Thanks.

comment:4 by euan, 15 years ago

yes I would think that it is the same issue. There was changes done in the xorg drivers regarding the "bios_scratch" registers after the documentation was released. Apparently some of the register manipulations were incorrect.

its a doubly hard problem for me to help with as 1 i can't boot haiku on real hardware, and 2 I don't have an X800 / X600 laptop. :(

Previous attempts at solving the problem seemed to lead to ACPI / bios problem. I was never sure if that was an accurate diagnosis though.

If anyone wants to try, put some tracing in where the bios scratch regs are read (before they are touched), then compare to the now fully understood lists of regs:

http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/tree/src/radeon_reg.h

Look for RADEON_BIOS_0_SCRATCH

If I get some time this week I'll also have a look.

comment:5 by peterm1561, 15 years ago

I have to correct my earlier comment. It seems to me that my compaq armada with ati rage mobility has another problem which is not based in the ati-driver. I checked it again it seems more acpi related. The laptop is not going into sleep mode, but in coma-mode and refuses to wake up.

I re-checked my hp nc8000 with RV350 [Mobility Radeon 9600 M10] and in this case you can see the screen (a bit) but the backlight does not work. I'm not a developer but let me know if I can do anything to help solving this problem.

comment:6 by peterm1561, 15 years ago

I installed one of the latest builds and the problem is solved. All works well. Thanks!

comment:7 by euan, 15 years ago

That's interesting because I don't think there has been any changes made to the driver. Perhaps the Backlight is now being initalized by ACPI? Could try booting with ACPI disabled from the boot menu to see if that reverts the problem?

comment:8 by peterm1561, 15 years ago

This problem is back in ALPHA2, but not on day 1 but after a few days of happy working with R1A2.

in reply to:  8 comment:9 by siarzhuk, 15 years ago

Replying to euan:

That's interesting because I don't think there has been any changes made to the driver. Perhaps the Backlight is now being initalized by ACPI? Could try booting with ACPI disabled from the boot menu to see if that reverts the problem?

Regarding the last comment from peterm1561: Registers are initiaized by random data? By the way - activating ACPI prevent me to work with my X700xl (described in #1924) at all. In case disabled ACPI it was possible to switch disaply in 1024x768 mode by using Fn-SwScreen shortcut. In case ACPI is ON - it is not possible anymore.

comment:10 by scottmc, 13 years ago

Blocking: 7662 added

comment:11 by taos, 13 years ago

Problem still occurs with alpha3 (as well as with post-alpha3 nightlies, e.g. hrev42460). Observed when booting Acer TravelMate 8101 WLMi without fail-safe video mode enabled.

listdev output shows same graphics hardware as in the original bug report (but a little different from the hardware described in related ticket #1924):

device Display controller (VGA compatible controller, VGA controller) [3|0|0]
  vendor 1002: ATI Technologies Inc
  device 5653: Radeon Mobility X700 (PCIE)

comment:12 by diver, 13 years ago

Cc: siarzhuk added; imker@… removed
Version: R1/alpha1R1/Development

comment:13 by x-ist, 12 years ago

Is still with us in hrevr1alpha4-44610. This is the very same laptop: Acer Aspire 1690 with a Radeon Mobility X700. Now adding some logs.

by x-ist, 12 years ago

Attachment: listdev added

by x-ist, 12 years ago

Attachment: SYSLOG00.TXT added

comment:14 by x-ist, 12 years ago

The recognized ASIC is incorrect (see SYSLOG00.TXT):

Radeon - probeDevice: found Mobility Radeon X700 (M26); ASIC: r420

According to http://dri.freedesktop.org/wiki/ATIRadeon it should be rv410.

comment:15 by x-ist, 12 years ago

Interestingly the connected VGA monitor works fine! Just the built-in display is black :/

in reply to:  15 ; comment:16 by scottmc, 12 years ago

Replying to x-ist:

Interestingly the connected VGA monitor works fine! Just the built-in display is black :/

Could it be that it just isn't turning the backlight on?

in reply to:  16 comment:17 by x-ist, 12 years ago

Replying to scottmc:

Could it be that it just isn't turning the backlight on?

In such a case I should be able to see at least something, right? I tried different light conditions in the room and used a torch light but didn't see anything.

comment:18 by x-ist, 12 years ago

I compiled the radeon driver with ENABLE_LOGGING enabled in /haiku/headers/private/graphics/common/log_col.h to see more syslog output.

Don't know whether I interprete the log correctly, but I see the following oddities:

According to

KERN: Radeon - Radeon_GetBIOSDFPInfo: Panel Info from ATOMBIOS:
KERN: XRes: 1280, YRes: 800, DotClock: 71250

the correct resolution is read from ATOMBIOS.

Following that, display detection is preforrmed and two monitors are detected, although there is just the built-in display panel only. No external monitors attached.

KERN: Radeon - Radeon_DetectDisplays: Searching port 0
KERN: Radeon - Radeon_DetectCRTInt: Found colour CRT connected to CRT-DAC
KERN: Radeon - Radeon_DetectDisplays: Detected monitors: 0x2
KERN: Radeon - Radeon_ConnectorInfo: Port 0:- 
KERN: Monitor:    CRT 
KERN: Conn Type:  VGA
KERN: DDC Port:   None
KERN: TMDS Type:  None
KERN: DAC Type:   Primary
KERN: Radeon - Radeon_ConnectorInfo: Port 1:- 
KERN: Monitor:    None
KERN: Conn Type:  LVDS
KERN: DDC Port:   None
KERN: TMDS Type:  None
KERN: DAC Type:   Unknown

Port 1 should be the native panel right?

Afterwards the display mode is set

KERN: Radeon - SET_DISPLAY_MODE: width=1600, height=900

This resolution is incorrect and probably not supported by the panel (1280x800 would be just right).

Then, a second time displays are detected (why?) and the following bad line follows:

KERN: Radeon - Radeon_SetupDefaultMonitorRouting:
   display_devices=2, whished_num_heads=1, use_laptop_panel=0

So the laptop panel is not going to be used :(

by x-ist, 12 years ago

Attachment: SYSLOG01.TXT added

more verbose readeon driver logging

comment:19 by x-ist, 12 years ago

This problem is most probably related to hardware. The VGA port is the primary one as can be seen in the syslog excerpt above, and Haiku uses it, failing to detect (or use) the built-in panel. That' why the external monitor works fine, but the panel does not. The Linux folks had that problem as well: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/22985

I'm trying to get this straight in the accelerant, and have the backlight switching on already, but the display still remains black. I wished someone with more expertise could have a look on that. Btw. does anyone know what happened to Thomas Kurschel, the author of that driver?

comment:20 by SoonDead, 6 years ago

Can confirm that this still happens, screen is black after boot, not just the backlight but no image, if I plug something into the VGA port it is also black.

Fail safe video driver works, but the resolution is very low this way.

I understand that writing drivers for everything is an uphill battle, but the x700 driver is not working properly, the broadcom4300 wifi driver detects no wireless networks, and was not able to produce sound with the IXP SB400 AC'97 Audio Controller, so I'm still very far from setting up my own haiku based media center.

comment:21 by waddlesplash, 6 years ago

Blocked By: 7662 added
Blocking: 7662 removed

comment:22 by euan, 3 years ago

Resolution: invalid
Status: newclosed

Closing these tickets due to the passing of time. Please consider opening new tickets with current syslogs if still using these graphics accelerators and seeing these issues.

Note: See TracTickets for help on using tickets.