RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-12-06 19:10:00 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

MacroFusion: Fix macro fusion with ExitSU failing in top-down scheduling

Browse Source 
When fusing instructions A and B, we must add all predecessors of B as
predecessors of A to avoid instructions getting scheduling in between.

There is a special case involving ExitSU: Every other node must be
scheduled before it by design and we don't need to make this explicit in
the graph, however when fusing with a different node we need to schedule
every othere node before the fused node too and we need to make this
explicit now: This patch adds a dependency from the fused node to all
roots in the graph.

Differential Revision: https://reviews.llvm.org/D49830

llvm-svn: 338046
This commit is contained in:
Matthias Braun 2018-07-26 17:43:56 +00:00
parent 6ecab04348
commit 2c06ea1b59
2 changed files with 39 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
lib/CodeGen/MacroFusion.cpp
View File

@ -87,7 +87,7 @@ static bool fuseInstructionPair(ScheduleDAGMI &DAG, SUnit &FirstSU,
// Make the FirstSU also dependent on the dependencies of the SecondSU to
// prevent them from being scheduled between the FirstSU and the SecondSU.
if (&FirstSU != &DAG.EntrySU)
if (&FirstSU != &DAG.EntrySU) {
for (const SDep &SI : SecondSU.Preds) {
SUnit *SU = SI.getSUnit();
if (SI.isWeak() || isHazard(SI) || &FirstSU == SU || FirstSU.isSucc(SU))
@ -96,6 +96,16 @@ static bool fuseInstructionPair(ScheduleDAGMI &DAG, SUnit &FirstSU,
FirstSU.print(dbgs(), &DAG); dbgs() << '\n';);
DAG.addEdge(&FirstSU, SDep(SU, SDep::Artificial));
}
// ExitSU comes last by design, which acts like an implicit dependency
// between ExitSU and any bottom root in the graph. We should transfer
// this to FirstSU as well.
if (&SecondSU == &DAG.ExitSU) {
for (SUnit &SU : DAG.SUnits) {
if (SU.Succs.empty())
DAG.addEdge(&FirstSU, SDep(&SU, SDep::Artificial));
}
}
}
++NumFused;
return true;

28
test/CodeGen/AArch64/macro-fusion-last.mir Normal file
View File

@ -0,0 +1,28 @@
# RUN: llc -o - %s -mtriple=aarch64-- -mattr=+arith-bcc-fusion -run-pass postmisched | FileCheck %s --check-prefixes=CHECK,FUSION
# RUN: llc -o - %s -mtriple=aarch64-- -mattr=-arith-bcc-fusion -run-pass postmisched | FileCheck %s --check-prefixes=CHECK,NOFUSION
# Make sure the last instruction is correctly macro-fused when scheduling
# top-down (post-ra).
---
# CHECK-LABEL: name: fuse_last
# CHECK: $x1 = LDRXui $x0, 0
# NOFUSION: $xzr = SUBSXri killed $x2, 0, 0, implicit-def $nzcv
# CHECK: STRXui killed $x0, killed $x1, 0
# FUSION: $xzr = SUBSXri killed $x2, 0, 0, implicit-def $nzcv
# CHECK: Bcc 1, %bb.1, implicit killed $nzcv
name: fuse_last
tracksRegLiveness: true
body: |
bb.0:
liveins: $x0, $x2
$x1 = LDRXui $x0, 0
; There is latency between these two instructions tempting the scheduler to
; move the SUBSXri in between them. However doing so breaks macro fusion.
STRXui $x0, $x1, 0
$xzr = SUBSXri $x2, 0, 0, implicit-def $nzcv
Bcc 1, %bb.1, implicit killed $nzcv
bb.1:
RET_ReallyLR implicit undef $x0
...