MachineInstrBundle: Fix reversed isSuperRegisterEq() call
Unfortunately this fix had the effect of exposing the
-verify-machineinstrs FIXME of X86InstrInfo.cpp in two testcases for
which I disabled it for now.
Two testcases also have additional pushq/popq where the corrected code
cannot prove that %rax is dead any longer. Looking at the examples, this
could potentially be fixed by improving computeRegisterLiveness() to check
the live-in lists of the successors blocks when reaching the end of a
block.
This fixes http://llvm.org/PR25951.
llvm-svn: 256799
diff --git a/llvm/lib/CodeGen/MachineInstrBundle.cpp b/llvm/lib/CodeGen/MachineInstrBundle.cpp
index 3eaf4c5..4619daf 100644
--- a/llvm/lib/CodeGen/MachineInstrBundle.cpp
+++ b/llvm/lib/CodeGen/MachineInstrBundle.cpp
@@ -315,7 +315,7 @@
if (!TRI->regsOverlap(MOReg, Reg))
continue;
- bool Covered = TRI->isSuperRegisterEq(MOReg, Reg);
+ bool Covered = TRI->isSuperRegisterEq(Reg, MOReg);
if (MO.readsReg()) {
PRI.Read = true;
if (Covered) {
diff --git a/llvm/lib/Target/X86/X86InstrInfo.cpp b/llvm/lib/Target/X86/X86InstrInfo.cpp
index d4cae7b..bd4d39e 100644
--- a/llvm/lib/Target/X86/X86InstrInfo.cpp
+++ b/llvm/lib/Target/X86/X86InstrInfo.cpp
@@ -4466,6 +4466,10 @@
// (unnecessarily) saving+restoring a dead register. However the
// MachineVerifier expects operands that read from dead registers
// to be marked with the "undef" flag.
+ // An example of this can be found in
+ // test/CodeGen/X86/peephole-na-phys-copy-folding.ll and
+ // test/CodeGen/X86/cmpxchg-clobber-flags.ll when using
+ // -verify-machineinstrs.
BuildMI(MBB, MI, DL, get(Push)).addReg(AX, getKillRegState(true));
}
if (FromEFLAGS) {