Re: [XFree86] XVideo/XvImage/XvShmPutImage on ATI Radeon with 4.3.99.4twice as slow as with 4.3.0

2003-05-27 Thread Thomas Winischhofer
Hans Korneder wrote: This _might_ be related to the O_SYNC bug discussed in the Athlon related mystery thread on the devel list, please read the archive at http://www.mail-archive.com/devel%40xfree86.org/ (Don't let yourself be missguided by the thread name) Thanks for the hint. Removing O_SYNC

Re: [XFree86] XVideo/XvImage/XvShmPutImage on ATI Radeon with 4.3.99.4twice as slow as with 4.3.0

2003-05-27 Thread Marc Aurele La France
On Tue, 27 May 2003, Thomas Winischhofer wrote: Hans Korneder wrote: This _might_ be related to the O_SYNC bug discussed in the Athlon related mystery thread on the devel list, please read the archive at http://www.mail-archive.com/devel%40xfree86.org/ (Don't let yourself be missguided by

Re: [XFree86] XVideo/XvImage/XvShmPutImage on ATI Radeon with 4.3.99.4twice as slow as with 4.3.0

2003-05-27 Thread Thomas Winischhofer
Marc Aurele La France wrote: Right. I slowed things down a bit by forcing the painting of the colour key through the entire wrapped screen, rather than calling an XAA entry directly. This makes this part of Xv support independent of XAA (i.e. Accel/NoAccel), banking, shadowing, etc. Mind you,

Re: [XFree86] XVideo/XvImage/XvShmPutImage on ATI Radeon with 4.3.99.4twice as slow as with 4.3.0

2003-05-27 Thread Marc Aurele La France
On Wed, 28 May 2003, Thomas Winischhofer wrote: Marc Aurele La France wrote: Right. I slowed things down a bit by forcing the painting of the colour key through the entire wrapped screen, rather than calling an XAA entry directly. This makes this part of Xv support independent of XAA