https://bugs.llvm.org/show_bug.cgi?id=39272
Bug ID: 39272
Summary: wasm32: Opcodes for some SIMD float operations
disagree with current spec
Product: libraries
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: enhancement
Priority: P
Component: Backend: WebAssembly
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected], [email protected]
In testing the wasm32 SIMD support of LLVM, I ran across a few opcodes that
looked like they weren't currently matching the spec online. It looks like the
issue is that after the SIMD leading opcode (0xfd) it's expected that the SIMD
operation is encoded as a leb128 integer, but currently LLVM is just emitting
an extra byte.
For example f64x2.abs -
https://github.com/llvm-mirror/llvm/blob/d069d45aa888c5b8a44521d4cade32767c09b35e/test/MC/WebAssembly/simd-encodings.s#L382-L383
- is encoded as 0xfd,0x80, but I think that should be 0xfd,0x80,0x01 with
leb128 encoding.
I haven't currently got a super minimal test case for this, but I can
definitely make one if necessary!
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
llvm-bugs mailing list
[email protected]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
