Bug 15292

Summary: 0x9501:0x1002:0x2542:HIS Radeon HD 3870 with 512MB
Product: xorg Reporter: Jimmy Jazz <Jimmy.Jazz>
Component: Driver/radeonhdAssignee: Luc Verhaegen <lverhaegen>
Status: RESOLVED NOTABUG QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: 7.3 (2007.09)   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
xorg log file none

Description Jimmy Jazz 2008-03-31 13:05:40 UTC
Created attachment 15586 [details]
xorg log file

Hello,

The video card has been tested at resolutions 1680x1050 and 1280x1024 on a flat panel with the following options activated:

(**) RADEONHD(0): Option "shadowfb" "on"
(**) RADEONHD(0): Option "forcereduced" "on"
(**) RADEONHD(0): Option "HPD" "auto"

and works quite well. 
The mainboard uses an AMD Phenom(tm) 9500 Quad-Core Processor with 4Go of RAM and an AMD 770 controller... i.e. full AMD :)

But the radeonhd driver is unable to use all the 512Mo of video memory and has a really initial small framebuffer of 16kb. Moreover the driver complains about an unavailable VRAM scratch space.

lspci returns,
01:00.0 VGA compatible controller: ATI Technologies Inc Unknown device 9501

And here is an excerpt of the xorg log output,

(--) Chipset RV670 found
(**) RADEONHD(0): Depth 24, (**) framebuffer bpp 32
(**) RADEONHD(0): Option "shadowfb" "on"
(**) RADEONHD(0): Option "forcereduced" "on"
(**) RADEONHD(0): Option "HPD" "auto"
(II) RADEONHD(0): Unknown card detected: 0x9501:0x1002:0x2542.
        If - and only if - your card does not work or does not work optimally
        please contact radeonhd@opensuse.org to help rectify this.
        Use the subject: 0x9501:0x1002:0x2542: <name of board>.
(--) RADEONHD(0): Detected an RV670 on an unidentified card
(II) RADEONHD(0): Mapped IO at 0x7ff13c568000 (size 0x00010000)
(II) RADEONHD(0): Getting BIOS copy from legacy VBIOS location
(II) RADEONHD(0): ATOM BIOS Rom:
        SubsystemVendorID: 0x1002 SubsystemID: 0x2542
        IOBaseAddress: 0xde00
        Filename: S3B33904.102
        BIOS Bootup Message: ^M
113-B33904-102 RV670 GDDR4_16Mx32 256bit 512MB 777e/1126m                   ^M

(II) RADEONHD(0): Framebuffer space used by Firmware (kb): 16
(II) RADEONHD(0): Start of VRAM area used by Firmware: 0x1fffc000
(II) RADEONHD(0): The detected amount of videoram exceeds the PCI BAR aperture.
(II) RADEONHD(0): Using only 262144kB of the total 524288kB.
(--) RADEONHD(0): VideoRAM: 262144 kByte
(II) RADEONHD(0): Framebuffer space used by Firmware (kb): 16
(II) RADEONHD(0): Start of VRAM area used by Firmware: 0x1fffc000
(II) RADEONHD(0): AtomBIOS requests 16kB of VRAM scratch space
(II) RADEONHD(0): AtomBIOS VRAM scratch base: 0x1fffc000
(WW) RADEONHD(0): rhdAtomAllocateFbScratch: FW FB scratch area 536854528 (size: 16384) extends beyond available framebuffer size 268435456
(II) RADEONHD(0): Cannot get VRAM scratch space. Allocating in main memory instead
(II) RADEONHD(0): Default Engine Clock: 777000
(II) RADEONHD(0): Default Memory Clock: 1126000
(II) RADEONHD(0): Maximum Pixel ClockPLL Frequency Output: 1200000
(II) RADEONHD(0): Minimum Pixel ClockPLL Frequency Output: 0
(II) RADEONHD(0): Maximum Pixel ClockPLL Frequency Input: 13500
(II) RADEONHD(0): Minimum Pixel ClockPLL Frequency Input: 1000
(II) RADEONHD(0): Maximum Pixel Clock: 400000
(II) RADEONHD(0): Reference Clock: 27000
(II) Loading sub module "i2c"
(II) LoadModule: "i2c"(II) Module "i2c" already built-in
(II) RADEONHD(0): Reference Clock: 27000
(II) RADEONHD(0): I2C bus "RHD I2C line 0" initialized.
(II) RADEONHD(0): I2C bus "RHD I2C line 1" initialized.
(II) RADEONHD(0): I2C bus "RHD I2C line 2" initialized.
(II) RADEONHD(0): I2C bus "RHD I2C line 3" initialized.
(II) Loading sub module "ddc"
(II) LoadModule: "ddc"(II) Module "ddc" already built-in
(II) RADEONHD(0): Detected VGA mode.
(II) RADEONHD(0): Minimum Pixel ClockPLL Frequency Output: 0
(II) RADEONHD(0): Maximum Pixel ClockPLL Frequency Output: 1200000
(II) RADEONHD(0): Maximum Pixel Clock: 400000
(II) RADEONHD(0): Reference Clock: 27000
...
(II) RADEONHD(0): Mapped IO at 0x7f243e137000 (size 0x00010000)
(==) RADEONHD(0): Write-combining range (0xd0000000,0x10000000)
(II) RADEONHD(0): Mapped FB at 0x7f24298ef000 (size 0x10000000)
(EE) RADEONHD(0): RHDVGASave: VGA FB Offset (0xE0000000) is out of range of the Cards Internal FB Address (0xD0000000)
(II) RADEONHD(0): Using 31743 scanlines of offscreen memory
(**) RADEONHD(0): Option "BackingStore" "on"
(**) RADEONHD(0): Backing store enabled
(**) RADEONHD(0): Option "SilkenMouse" "on"
(**) RADEONHD(0): Silken mouse enabled
(II) RADEONHD(0): RandR 1.2 enabled, ignore the following RandR disabled message.
(EE) RADEONHD(0): TMDSBVoltageControl: unhandled chipset: 0x9501.
(II) RADEONHD(0): LVTMA_MACRO_CONTROL: 0x0533040B
(II) RADEONHD(0): LVTMA_TRANSMITTER_ADJUST: 0x02A01000
(II) RADEONHD(0): LVTMA_PREEMPHASIS_CONTROL: 0x41002045
(II) RADEONHD(0): Using HW cursor
(**) Option "dpms"
(**) RADEONHD(0): DPMS enabled
Comment 1 Egbert Eich 2008-03-31 14:49:18 UTC
None of this is a problem:
We can only map half of the VRAM ie 256 MB as you can see from the log:
(II) RADEONHD(0): Using only 262144kB of the total 524288kB.
(--) RADEONHD(0): VideoRAM: 262144 kByte
The firmware itself reserves 16kB:
(II) RADEONHD(0): Framebuffer space used by Firmware (kb): 16
The unavailable VRAM scratch space is not issue ATM. It is more a notice for debugging purposes.
Comment 2 Jimmy Jazz 2008-04-01 04:47:40 UTC
(In reply to comment #1)
> None of this is a problem:
> We can only map half of the VRAM ie 256 MB as you can see from the log:
> (II) RADEONHD(0): Using only 262144kB of the total 524288kB.
> (--) RADEONHD(0): VideoRAM: 262144 kByte
> The firmware itself reserves 16kB:
> (II) RADEONHD(0): Framebuffer space used by Firmware (kb): 16
> The unavailable VRAM scratch space is not issue ATM. It is more a notice for
> debugging purposes.
> 

Certainly a silly question, but does it mean radeonhd driver won't be able to support more than 256MB of vram or is it just a restriction of the 2D mode of the videocard because of the bus or something ? 
Otherwise does it apply to the 3D mode  too ?

If only half of the 512MB of vram is mappable I will instead consider buying  a 256MB only radeonhd videocard  in the near future with the benefit to consume less power :)

I really appreciate your work. The driver is rock stable. It is fast enough for a commun use with the shadow framebuffer mode activated and the videocard is still silent. I'm really eager to test compiz-fusion with the upcoming accelerated driver as soon as it will be available for testing.
Comment 3 Matthias Hopf 2008-04-01 05:58:54 UTC
Only 256MB of the Videoram will be accessible by the processor directly, the rest is only accessible by the GPU itself. 2D acceleration relies on software fallbacks, so the framebuffer and pixmaps have to be CPU addressable. As e.g. textures are typically uploaded by DMA, the whole memory can be used there. That said, we don't have 3D yet :-]

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.