Ticket #322 (Fixed)Thu Oct 18 12:26:47 UTC 2012
objasm 4.01 incorrectly assembles some VMOV instructions
Reported by: | Jeffrey Lee (213) | Severity: | Normal |
Part: | RISC OS: C/C++ toolchain | Release: | |
Milestone: | Status | Fixed |
Details by Jeffrey Lee (213):
Copied from the bugs forum now that the bug tracker is back up:
The ARM -> VFP/NEON version of the doubleword VMOV instruction (section A8.6.332 in the ARMv7 ARM) is being assembled incorrectly. It looks like the Rt field is being corrupted; e.g. “VMOV d26,r3,r4” assembles to 0xec44bb3a (VMOV d26,r11,r4), and “VMOV d27,r5,r6” assembles to 0xec46fb3b (VMOV d27,pc,r6).
The VFP/NEON -> ARM version of the instruction seems to work fine.
Changelog:
Modified by Jeffrey Lee (213) Sun, November 02 2014 - 13:52:32 GMT
- Status changed from Open to Fixed
Fixed with objasm 4.02 (DDE 25)