maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[CodeGen] fix broken successor probability in MBB dump 

When printing successor probabilities for a MBB, a human readable value is sometimes shown as 200.0%.
The human readable output is based on getProbabilityIterator, which returns 0xFFFFFFFF for getNumerator() and 0x80000000 for getDenominator() for unknown BranchProbability.
By using getSuccProbability as we do for the non-human readable part, we can avoid this problem.

Differential Revision: https://reviews.llvm.org/D52605

llvm-svn: 343297
This commit is contained in:
Hiroshi Inoue 2018-09-28 05:27:32 +00:00
parent 73d18aa028
commit 69bfa40200
2 changed files with 26 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
llvm/lib/CodeGen/MachineBasicBlock.cpp
View File

@ -362,7 +362,7 @@ void MachineBasicBlock::print(raw_ostream &OS, ModuleSlotTracker &MST,
// Print human readable probabilities as comments.
OS << "; ";
for (auto I = succ_begin(), E = succ_end(); I != E; ++I) {
const BranchProbability &BP = *getProbabilityIterator(I);
const BranchProbability &BP = getSuccProbability(I);
if (I != succ_begin())
OS << ", ";
OS << printMBBReference(**I) << '('

25
llvm/test/Other/X86/mbb-dump.ll Normal file
View File

@ -0,0 +1,25 @@
; RUN: llc < %s 2>&1 -print-after=machine-scheduler -mtriple=x86_64-unknown-unknown | FileCheck %s
; expected MBB dump output
; # *** IR Dump After Machine Instruction Scheduler ***:
; # Machine code for function foo: NoPHIs, TracksLiveness
;
; 0B bb.0 (%ir-block.0):
; successors: %bb.1(0x80000000); %bb.1(100.00%)
;
; 16B bb.1.next:
; ; predecessors: %bb.0
; previously, it was broken as
; successors: %bb.1(0x80000000); %bb.1(200.00%)
define void @foo(){
; CHECK: IR Dump After Machine Instruction Scheduler
; CHECK: bb.0
; CHECK: 100.0
; CHECK: bb.1
br label %next
next:
ret void
}