On Sat, 2004-11-06 at 12:25 +0100, Nicolai Haehnle wrote: > > Okay, the new syslog has all the debug info. I notice the following line: > > Nov 7 07:37:58 disoft-dc [drm:radeon_cp_init_ring_buffer] writeback test > failed > > The Radeon DRM source code has a comment indicating that writeback doesn't > work everywhere, but I think it's safe to assume that all R300-based chips > should be capable of writeback.
The comment says 'not everywhere' instead of 'not with all chips' for a reason... > This would indeed point towards a problem in the AGP setup in one way or > another, It's known that a lot of AGP bridges have issues with transfers from the chip to the AGP aperture. It would be better to allocate the page to write back to from PCI memory, but that's only guaranteed to work if dev_priv->fb_location != 0. > and that means that the ring buffer won't work properly. Without a working > ring buffer setup, it's only natural for a lockup to occur. Actually, the RPTR only matters for detecting that the ring buffer is filling up, which may never be the case as long as you're using indirect buffers. (E.g., RPTR writeback was completely broken for the longest time in the r128 DRM, nobody seemed to notice...) > Perhaps we should fail completely when the writeback test failed on > R300-based hardware. I don't think so. It may be a good idea to read the CP_RB_RPTR register directly in GET_RING_HEAD() if the writeback test failed though. -- Earthling Michel DÃnzer | Debian (powerpc), X and DRI developer Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer ------------------------------------------------------- This SF.Net email is sponsored by: Sybase ASE Linux Express Edition - download now for FREE LinuxWorld Reader's Choice Award Winner for best database on Linux. http://ads.osdn.com/?ad_idU88&alloc_id065&op=click -- _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
