aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorCraig Topper <craig.topper@gmail.com>2011-11-15 06:39:01 +0000
committerCraig Topper <craig.topper@gmail.com>2011-11-15 06:39:01 +0000
commit44ec9fddc22abeabaf2565691ed02b1f47f532af (patch)
tree330a8f90816e68aad5c23ffd813a122efbb406ff
parentf178418729d91e2f15a8eb27c0e4787b8ec7a3ce (diff)
downloadexternal_llvm-44ec9fddc22abeabaf2565691ed02b1f47f532af.zip
external_llvm-44ec9fddc22abeabaf2565691ed02b1f47f532af.tar.gz
external_llvm-44ec9fddc22abeabaf2565691ed02b1f47f532af.tar.bz2
Fix PR11370 for real. Prevents converting 256-bit FP instruction to AVX2 256-bit integer instructions when AVX2 isn't enabled.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@144629 91177308-0d34-0410-b5e6-96231b3b80d8
-rw-r--r--lib/Target/X86/X86InstrInfo.cpp26
1 files changed, 17 insertions, 9 deletions
diff --git a/lib/Target/X86/X86InstrInfo.cpp b/lib/Target/X86/X86InstrInfo.cpp
index 9f7b21f..359a9df 100644
--- a/lib/Target/X86/X86InstrInfo.cpp
+++ b/lib/Target/X86/X86InstrInfo.cpp
@@ -3573,14 +3573,17 @@ static const unsigned ReplaceableInstrsAVX2[][3] = {
// FIXME: Some shuffle and unpack instructions have equivalents in different
// domains, but they require a bit more work than just switching opcodes.
-static const unsigned *lookup(unsigned opcode, unsigned domain, bool hasAVX2) {
+static const unsigned *lookup(unsigned opcode, unsigned domain) {
for (unsigned i = 0, e = array_lengthof(ReplaceableInstrs); i != e; ++i)
if (ReplaceableInstrs[i][domain-1] == opcode)
return ReplaceableInstrs[i];
- if (domain != 3 || hasAVX2) // only use PackedInt domain if AVX2 is enabled
- for (unsigned i = 0, e = array_lengthof(ReplaceableInstrsAVX2); i != e; ++i)
- if (ReplaceableInstrsAVX2[i][domain-1] == opcode)
- return ReplaceableInstrsAVX2[i];
+ return 0;
+}
+
+static const unsigned *lookupAVX2(unsigned opcode, unsigned domain) {
+ for (unsigned i = 0, e = array_lengthof(ReplaceableInstrsAVX2); i != e; ++i)
+ if (ReplaceableInstrsAVX2[i][domain-1] == opcode)
+ return ReplaceableInstrsAVX2[i];
return 0;
}
@@ -3588,16 +3591,21 @@ std::pair<uint16_t, uint16_t>
X86InstrInfo::getExecutionDomain(const MachineInstr *MI) const {
uint16_t domain = (MI->getDesc().TSFlags >> X86II::SSEDomainShift) & 3;
bool hasAVX2 = TM.getSubtarget<X86Subtarget>().hasAVX2();
- return std::make_pair(domain,
- domain && lookup(MI->getOpcode(), domain, hasAVX2) ? 0xe : 0);
+ uint16_t validDomains = 0;
+ if (domain && lookup(MI->getOpcode(), domain))
+ validDomains = 0xe;
+ else if (domain && lookupAVX2(MI->getOpcode(), domain))
+ validDomains = hasAVX2 ? 0xe : 0x6;
+ return std::make_pair(domain, validDomains);
}
void X86InstrInfo::setExecutionDomain(MachineInstr *MI, unsigned Domain) const {
assert(Domain>0 && Domain<4 && "Invalid execution domain");
uint16_t dom = (MI->getDesc().TSFlags >> X86II::SSEDomainShift) & 3;
assert(dom && "Not an SSE instruction");
- bool hasAVX2 = TM.getSubtarget<X86Subtarget>().hasAVX2();
- const unsigned *table = lookup(MI->getOpcode(), dom, hasAVX2);
+ const unsigned *table = lookup(MI->getOpcode(), dom);
+ if (!table) // try the other table
+ table = lookupAVX2(MI->getOpcode(), dom);
assert(table && "Cannot change domain");
MI->setDesc(get(table[Domain-1]));
}