8.02 "Another Cr*ppy Release Edition" ACRE

[bridgman]

Sid, it's not a "hardware problem with the overlay" but starting with R5xx most of the video processing hardware was removed from the overlay block and the work is now being done with shaders using TexturedVideo.

My understanding is that TvTime only supports a limited set of output formats and those don't include any of the formats we support as inputs to the driver, so either TvTime will have to add more output formats or we will need to add shader code to support more input formats.

We are continuing to wonk on video support in the driver so even if TvTime doesn't change I expect this will get addressed.

[ivanovic]

My understanding is that TvTime only supports a limited set of output formats and those don't include any of the formats we support as inputs to the driver, so either TvTime will have to add more output formats or we will need to add shader code to support more input formats.

You adding it is currently the only option because TVTime seems to not be actively developed anymore.

We are continuing to wonk on video support in the driver so even if TvTime doesn't change I expect this will get addressed.

Currently the closed driver supports YV12 and I420. The bad thing is that TVTime needs YUY2 and xdtv needs YUYV. So if you additionally implemented those two modes, most overlay users would probably be really happy (got the supported ones from my rv670 based card by running xvinfo while fglrx is loaded and textured video active).

[Silverlight]

Is it just me, or does Firefox work with composite in this driver? (no more lagging when turning the cube)

[n3Cre0]

Firefox (and thunderbird) scrolling is very fast now, same as when using metacity. Really this driverrelease must be the best one over for me. I got no issues

[sid350]

on x200m and CeleronM430 scrolling is still slow, but faster then before. On X1600XT and AMDx2 4200+ scrolling is good, but without compiz it is a bit faster. Also diagonal tearing is noticeable when you are scrolling very fast.

[enzobelmont]

on x200m and CeleronM430 scrolling is still slow, but faster then before. On X1600XT and AMDx2 4200+ scrolling is good, but without compiz it is a bit faster. Also diagonal tearing is noticeable when you are scrolling very fast.

same here!

i hope _ATI_ solve this soon.

BTW; any news of somebody Taking advantage of the recently released specs???

[1stFlight]

Has anyone else experience the fglrx driver crashing with MythTV?

output from /var/log/messages:

Mar  2 20:57:35 localhost kernel: [  619.420000] [fglrx] PCIe has already been initialized. Reinitializing ...
Mar  2 20:57:35 localhost kernel: [  619.430000] [fglrx] Reserve Block - 0 offset =  0X7ffb000 length = 0X5000
Mar  2 20:57:35 localhost kernel: [  619.430000] [fglrx] Reserve Block - 1 offset =  0X0 length = 0X1000000
Mar  2 20:57:35 localhost kernel: [  619.430000] [fglrx] Reserve Block - 2 offset =  0Xffff000 length = 0X1000
Mar  2 20:57:35 localhost kernel: [  619.430000] [fglrx] Reserve Block - 3 offset =  0Xffbf000 length = 0X40000
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] interrupt source 20008000 successfully enabled
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] enable ID = 0x00000008
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] Receive enable interrupt message with irqEnableMask: 20008000
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] interrupt source 10000000 successfully enabled
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] enable ID = 0x00000009
Mar  2 20:57:35 localhost kernel: [  619.460000] [fglrx] Receive enable interrupt message with irqEnableMask: 10000000
Mar  2 20:57:54 localhost kernel: [  638.650000] UDF-fs: No VRS found
Mar  2 20:57:57 localhost kernel: [  641.060000] UDF-fs: No VRS found
Mar  2 20:59:14 localhost kernel: [  718.450000] [fglrx] PCIe has already been initialized. Reinitializing ...
Mar  2 20:59:14 localhost kernel: [  718.460000] [fglrx] Reserve Block - 0 offset =  0X7ffb000 length = 0X5000
Mar  2 20:59:14 localhost kernel: [  718.460000] [fglrx] Reserve Block - 1 offset =  0X0 length = 0X1000000
Mar  2 20:59:14 localhost kernel: [  718.460000] [fglrx] Reserve Block - 2 offset =  0Xffff000 length = 0X1000
Mar  2 20:59:14 localhost kernel: [  718.460000] [fglrx] Reserve Block - 3 offset =  0Xffbf000 length = 0X40000
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] interrupt source 20008000 successfully enabled
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] enable ID = 0x00000008
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] Receive enable interrupt message with irqEnableMask: 20008000
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] interrupt source 10000000 successfully enabled
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] enable ID = 0x00000009
Mar  2 20:59:14 localhost kernel: [  718.500000] [fglrx] Receive enable interrupt message with irqEnableMask: 10000000

with Ubuntu 7.10 ATI X1950XT

[mal13]

I uninstalled 8.02 driver and went to use mesa....
Ati still suffers logout/switchuser/shutdown problem on kubuntu 7.10. All those work w/ mesa so i rather have usable system for 2 users than use ati (for now).
When someone confirms that logout/switchuser.. problem is fixed i'll go back to fglrx.

[pchev]

With 8.02 and Kubuntu 7.10 I still have to disable atieventsd as describe at the end of this page http://wiki.cchtml.com/index.php/Troubleshooting
After that, logout, switch console, all work fine.

[mal13]

Well i changed allready to Radeonhd. Maybe i'll come back to fglrx later... or i'll wait til hardyheron gets out.
Anyway.. about atieventsd, i found this aka patch:
http://ati.cchtml.com/show_bug.cgi?id=992 Comment #21.