Mailinglist Archive: radeonhd (266 mails)
| < Previous | Next > |
Re: [radeonhd] Re: Succeeded in compiling radeonhd 1.3.0 on first try!
- From: Dave Witbrodt <dawitbro@xxxxxxxxxxxxx>
- Date: Sun, 01 Nov 2009 08:47:30 -0500
- Message-id: <4AED9172.6070108@xxxxxxxxxxxxx>
Yang Zhao wrote:
Oh, the wording confused me:
(II) RADEONHD(0): Using only 262144kB of the total 1048576kB.
I would have understood better if the message read "Mapping" instead of "Using".
OK, thanks.
OK, rereading the log file I see the line you refer to.
So, for RV770, is this because the in-kernel DRM does not yet support the experimental 3D acceleration for r700?
I know that 'radeonhd' supports acceleration, and according to the "experimental_3D" wiki,
http://www.x.org/wiki/radeonhd%3Aexperimental_3D
the "R6xx/R7xx mesa code from 6xx-rewrite branch has been merged to master," so if I'm using Mesa 7.6 that piece of the puzzle should be OK as well, right?
To play with 3D on my hardware, the wiki says "still need non-master drm from ~agd5f/drm r6xx-r7xx-3d branch." The instructions explain how to build DRM modules, but also how to rebuild Mesa. To get 3D on RV770, I would have to:
- build DRM from the r6xx-r7xx-3d branch
- skip the wiki instructions about rebuilding Mesa (since I already
have Mesa 7.6)
Does that sound right?
Thanks again,
Dave W.
--
To unsubscribe, e-mail: radeonhd+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: radeonhd+help@xxxxxxxxxxxx
2009/10/31 Dave Witbrodt <dawitbro@xxxxxxxxxxxxx>:
4. Not sure why only 256MB of 1GB of VRAM gets used.
It's only possible to map up to 256MB of VRAM for the CPU to access
directly. There's nothing wrong, per se: the GPU is still able to
access all of it.
Oh, the wording confused me:
(II) RADEONHD(0): Using only 262144kB of the total 1048576kB.
I would have understood better if the message read "Mapping" instead of "Using".
(II) RADEONHD(0): Using 3840x1920 Framebuffer with 3840 pitch
(II) RADEONHD(0): FB: Allocated ScanoutBuffer at offset 0x00008000
(size = 0x01C20000)
(**) RADEONHD(0): Display dimensions: (518, 324) mm
(**) RADEONHD(0): DPI set to (188, 150)
...
6. The DPI 188,150 value is totally bogus. The 'xdpyinfo' utility reports
94 x 94 with X running.
Seems like it's calculating the DPI from total framebuffer size
instead of actual display size. Should be a harmless cosmetic error.
OK, thanks.
7. I see that XVideo actually gets initialized, but XRandR and AIGLX fail
toward the end of the process. Have I done something wrong? Something I can
repair?
The RandR error is bogus: there is a line further up that says to
ignore the randr failure message.
OK, rereading the log file I see the line you refer to.
AIGLX failure is expected without working 3D accel.
So, for RV770, is this because the in-kernel DRM does not yet support the experimental 3D acceleration for r700?
I know that 'radeonhd' supports acceleration, and according to the "experimental_3D" wiki,
http://www.x.org/wiki/radeonhd%3Aexperimental_3D
the "R6xx/R7xx mesa code from 6xx-rewrite branch has been merged to master," so if I'm using Mesa 7.6 that piece of the puzzle should be OK as well, right?
To play with 3D on my hardware, the wiki says "still need non-master drm from ~agd5f/drm r6xx-r7xx-3d branch." The instructions explain how to build DRM modules, but also how to rebuild Mesa. To get 3D on RV770, I would have to:
- build DRM from the r6xx-r7xx-3d branch
- skip the wiki instructions about rebuilding Mesa (since I already
have Mesa 7.6)
Does that sound right?
Thanks again,
Dave W.
--
To unsubscribe, e-mail: radeonhd+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: radeonhd+help@xxxxxxxxxxxx
| < Previous | Next > |