On Wed, Mar 09, 2005 at 04:17:25PM -0500, Hughes, Gregory Mark wrote: > Jim, > > I've been working with PA-RISC processors from the days of the 710. I > have never seen any processor related difference in floating point > results using PA1.1 instructions. There is the possibility of a > hardware fault in your particular system
Greg, You are probably more aware which CPUs implement which subset of floating point instructions. parisc-linux kernel might still have bugs in the FP exception handling code for older CPUs. I expect most testing was on PA7100LC (712/80) or PA7300LC (B180) CPUs. ie later processors than what is in the 710 (PCX-S, IIRC) box. Jim, It would be worth checking if you are in fact excercising that code and how. You have time to rebuild your kernel with #define FPUDEBUG 1 in arch/parisc/math-emu/drivers.c? No promises that anyone will fix the problem, but without this info no one can. thanks, grant -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
