Opened 8 years ago
Closed 8 years ago
#2838 closed merge (fixed)
-fasm codeGen doesn't truncate CMMInt's at the right place
| Reported by: | TristanAllwood | Owned by: | igloo |
|---|---|---|---|
| Priority: | normal | Milestone: | 6.10.2 |
| Component: | Compiler | Version: | 6.10.1 |
| Keywords: | Cc: | ||
| Operating System: | Unknown/Multiple | Architecture: | Unknown/Multiple |
| Type of failure: | Test Case: | ||
| Blocked By: | Blocking: | ||
| Related Tickets: | Differential Rev(s): | ||
| Wiki Page: |
Description
With this sample program:
{-# LANGUAGE MagicHash, UnboxedTuples #-}
module Simpl(complement) where
import GHC.Base
import GHC.Num
complement (I# x#) = I# (word2Int# (int2Word# (4294967295#) `xor#` int2Word# (-1#)))
And compiling with
-O0 -fasm
on 6.10 or a recent head blows up in the assembler due to the constant.
Assembler messages: Error: missing or invalid immediate expression `-4294967296' taken as 0 Error: suffix or operands invalid for `mov'
After some debugging with SPJ, we believe that what is happening is:
CMMIntis storing 4294967295 as an Integer. This magic number is (2^32) - 1.
4294967295 ^ -1 is 00..011..1 ^ 11..111..1 = 11..100..0 which is (-4294967296)
We think internally using Integer arithmatic is fine (in e.g. CmmOpt), however truncation isn't happening correctly when turning the Integer into a real Int; makeImmediate in nativeGen\machineRegs looks like a possible culprit.
Change History (9)
comment:1 follow-up: ↓ 3 Changed 8 years ago by TristanAllwood
- Summary changed from -fast codeGen doesn't truncate CMMInt's at the right place to -fasm codeGen doesn't truncate CMMInt's at the right place
comment:2 Changed 8 years ago by TristanAllwood
given it's wrapped in #ifdef powepc, makeImmediate was probably a red herring.
comment:3 in reply to: ↑ 1 Changed 8 years ago by simonmar
- difficulty set to Unknown
Replying to TristanAllwood:
For reference, the test case does seem to be handled fine with -fnew-codegen in Today's Head.
I suspect this is by accident, as the new codegen is not running CmmOpt.
comment:5 Changed 8 years ago by TristanAllwood
- Operating System changed from Unknown/Multiple to Windows
comment:6 Changed 8 years ago by TristanAllwood
- Operating System changed from Windows to Unknown/Multiple
I can reproduce under both Windows (vista) and linux i686
comment:7 Changed 8 years ago by simonmar
- Owner set to simonmar
comment:8 Changed 8 years ago by simonmar
- Owner changed from simonmar to igloo
- Type changed from bug to merge
Fixed, thanks:
Tue Dec 9 02:55:15 PST 2008 Simon Marlow <marlowsd@gmail.com> * Fix #2838: we should narrow a CmmInt before converting to ImmInteger
comment:9 Changed 8 years ago by igloo
- Resolution set to fixed
- Status changed from new to closed
Merged.
For reference, the test case does seem to be handled fine with -fnew-codegen in Today's Head.