On Fri, Jun 13, 2003 at 12:00:12AM +0200, Michel D?nzer wrote: >It has no choice > but to take the data as it gets it from the client?
Yes, but initially, the interesting piece of a header is the first char, and the kernel checks the first char (and then correctly skips the length of the union when it "knows" that the rest of the header doesn't apply to this particular type) hence it doesn't matter diddly squat what the value of the other bytes are. They are used on different headers, yes - but if there are packets output that aren't filling in all the data, passing zero instead of whatever the value should be is no more likely to work, is it? AFAICT that's how it works at least from memory of May 2002. Here's the cvs diff, as I say, it wasn't added to fix anything, just to avoid confusion when looking at the output of verbose - although on first glance at the verbose output I similary thought it was a bug. (As an aside, the patch itself fixes one bug emitted in a block but adds another that breaks the "#define RADEON_OLD_PACKETS 0" path) http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/dri/xc/xc/lib/GL/mesa/src/drv/radeon/radeon_ioctl.c.diff?r1=1.11.2.21&r2=1.11.2.22 -- Michael. ------------------------------------------------------- This SF.NET email is sponsored by: eBay Great deals on office technology -- on eBay now! Click here: http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5 _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
