Hi, regarding your tip about writing it in pure assembler - I don't really see that solving the problem because I will eventually end up with a register as an operand anyway. Maybe you actually meant something similar to this...
If I put the variable in g1 just before issuing the sethi instruction (with a constant), then maybe I can read the g1 register when the breakpoint triggers (as a consequnce when the sethi instruction is executed by the simulator). So basically, its a mov instruction (that moves the variable value to the g1 register) preceeding the sethi instruction. /M -- This message posted from opensolaris.org
