On Thu, Aug 27, 2020 at 03:01:57AM -0700, Richard Henderson wrote: > On 8/27/20 2:11 AM, Edgar E. Iglesias wrote: > > So a first general regression is that opcode 0 no longer > > traps as an illegal instruction (seems to be treated as an > > add with all r0). > > Oops, will fix.
Thanks. Here's another issue, it seems some branches are jumping to the wrong address. This is a disasm from a failing case: 0x00000000ffd033a0: brlid r15, -636 // 0xffffffffffd03124 0x00000000ffd033a4: or r0, r0, r0 0x00000000ffa73124: Address 0xffa73124 is out of bounds. 0x00000000ffa73128: Address 0xffa73128 is out of bounds. This one is from a working one: 0x00000000ffd033a0: brlid r15, -636 // 0xffffffffffd03124 0x00000000ffd033a4: or r0, r0, r0 -------------- 0x00000000ffd03124: imm -40 0x00000000ffd03128: lwi r3, r0, 268 0x00000000ffd0312c: imm -40 0x00000000ffd03130: lwi r4, r0, 256 0x00000000ffd03134: srl r3, r3 0x00000000ffd03138: bsrli r4, r4, 23 0x00000000ffd0313c: andi r3, r3, 1 0x00000000ffd03140: rtsd r15, 8 0x00000000ffd03144: and r3, r4, r3
