The trap instruction is emitted as a data-in-text rather than an instruction.
Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
Event Timeline
test/CodeGen/ARM/debugtrap.ll | ||
---|---|---|
13 ↗ | (On Diff #46269) | Not sure why was this valid. "trap" was present only in a comment: '@ trap' |
test/CodeGen/ARM/trap.ll | ||
4 ↗ | (On Diff #46269) | We shouldn't be testing -mtriple=armv7-unknown-nacl in CodeGen, it is a Clang thing. -mattr=+nacl-trap is the right thing to test in the back-end. |
Comment Actions
Hi Alexandros,
The test seems to have been vague on purpose (because it was using .short/.long), but using .inst is definitely the right way.
LGTM. Thanks!