aboutsummaryrefslogtreecommitdiff
path: root/llvm
diff options
context:
space:
mode:
Diffstat (limited to 'llvm')
-rw-r--r--llvm/include/llvm/CodeGen/SelectionDAG.h8
-rw-r--r--llvm/include/llvm/Support/Jobserver.h162
-rw-r--r--llvm/include/llvm/Support/ThreadPool.h4
-rw-r--r--llvm/include/llvm/Support/Threading.h18
-rw-r--r--llvm/include/llvm/Target/GenericOpcodes.td2
-rw-r--r--llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp39
-rw-r--r--llvm/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp15
-rw-r--r--llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp25
-rw-r--r--llvm/lib/Support/CMakeLists.txt1
-rw-r--r--llvm/lib/Support/Jobserver.cpp259
-rw-r--r--llvm/lib/Support/Parallel.cpp98
-rw-r--r--llvm/lib/Support/ThreadPool.cpp108
-rw-r--r--llvm/lib/Support/Threading.cpp5
-rw-r--r--llvm/lib/Support/Unix/Jobserver.inc195
-rw-r--r--llvm/lib/Support/Windows/Jobserver.inc79
-rw-r--r--llvm/lib/Target/AMDGPU/SIRegisterInfo.td19
-rw-r--r--llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp4
-rw-r--r--llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp178
l---------llvm/test/CodeGen/AMDGPU/.#llvm.amdgcn.smfmac.gfx950.ll1
-rw-r--r--llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll32
-rw-r--r--llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-ignore-copies-crash.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir240
-rw-r--r--llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir16
-rw-r--r--llvm/test/CodeGen/AMDGPU/dst-sel-hazard.mir24
-rw-r--r--llvm/test/CodeGen/AMDGPU/hazards-gfx950.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir12
-rw-r--r--llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll24
-rw-r--r--llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx10.mir30
-rw-r--r--llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir360
-rw-r--r--llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx9.mir180
-rw-r--r--llvm/test/CodeGen/AMDGPU/machine-scheduler-sink-trivial-remats.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/mai-hazards.mir6
-rw-r--r--llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll24
-rw-r--r--llvm/test/CodeGen/AMDGPU/regalloc-failure-overlapping-insert-assert.mir16
-rw-r--r--llvm/test/CodeGen/AMDGPU/rename-independent-subregs.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir12
-rw-r--r--llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir36
-rw-r--r--llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir4
-rw-r--r--llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir16
-rw-r--r--llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll4
-rw-r--r--llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir24
-rw-r--r--llvm/test/TableGen/ResolveSchedClass.td18
-rw-r--r--llvm/test/ThinLTO/X86/memprof-dups.ll138
-rw-r--r--llvm/test/ThinLTO/X86/memprof_imported_internal.ll10
-rw-r--r--llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/amdgpu_isel.ll.expected4
-rw-r--r--llvm/unittests/Support/CMakeLists.txt1
-rw-r--r--llvm/unittests/Support/JobserverTest.cpp442
-rw-r--r--llvm/utils/TableGen/SubtargetEmitter.cpp10
-rw-r--r--llvm/utils/gn/secondary/llvm/lib/Support/BUILD.gn1
-rw-r--r--llvm/utils/gn/secondary/llvm/unittests/Support/BUILD.gn1
53 files changed, 2297 insertions, 636 deletions
diff --git a/llvm/include/llvm/CodeGen/SelectionDAG.h b/llvm/include/llvm/CodeGen/SelectionDAG.h
index d9d6f0b..62c0806 100644
--- a/llvm/include/llvm/CodeGen/SelectionDAG.h
+++ b/llvm/include/llvm/CodeGen/SelectionDAG.h
@@ -1959,6 +1959,10 @@ public:
LLVM_ABI SDValue makeEquivalentMemoryOrdering(LoadSDNode *OldLoad,
SDValue NewMemOp);
+ /// Get all the nodes in their topological order without modifying any states.
+ LLVM_ABI void getTopologicallyOrderedNodes(
+ SmallVectorImpl<const SDNode *> &SortedNodes) const;
+
/// Topological-sort the AllNodes list and a
/// assign a unique node id for each node in the DAG based on their
/// topological order. Returns the number of nodes.
@@ -2009,7 +2013,9 @@ public:
/// function mirrors \c llvm::salvageDebugInfo.
LLVM_ABI void salvageDebugInfo(SDNode &N);
- LLVM_ABI void dump() const;
+ /// Dump the textual format of this DAG. Print nodes in sorted orders if \p
+ /// Sorted is true.
+ LLVM_ABI void dump(bool Sorted = false) const;
/// In most cases this function returns the ABI alignment for a given type,
/// except for illegal vector types where the alignment exceeds that of the
diff --git a/llvm/include/llvm/Support/Jobserver.h b/llvm/include/llvm/Support/Jobserver.h
new file mode 100644
index 0000000..6bee3b5
--- /dev/null
+++ b/llvm/include/llvm/Support/Jobserver.h
@@ -0,0 +1,162 @@
+//===- llvm/Support/Jobserver.h - Jobserver Client --------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines a client for the GNU Make jobserver protocol. This allows
+// LLVM tools to coordinate parallel execution with a parent `make` process.
+//
+// The jobserver protocol is a mechanism for GNU Make to share its pool of
+// available "job slots" with the subprocesses it invokes. This is particularly
+// useful for tools that can perform parallel operations themselves (e.g., a
+// multi-threaded linker or compiler). By participating in this protocol, a
+// tool can ensure the total number of concurrent jobs does not exceed the
+// limit specified by the user (e.g., `make -j8`).
+//
+// How it works:
+//
+// 1. Establishment:
+// A child process discovers the jobserver by inspecting the `MAKEFLAGS`
+// environment variable. If a jobserver is active, this variable will
+// contain a `--jobserver-auth=<value>` argument. The format of `<value>`
+// determines how to communicate with the server.
+//
+// 2. The Implicit Slot:
+// Every command invoked by `make` is granted one "implicit" job slot. This
+// means a tool can always perform at least one unit of work without needing
+// to communicate with the jobserver. This implicit slot should NEVER be
+// released back to the jobserver.
+//
+// 3. Acquiring and Releasing Slots:
+// On POSIX systems, the jobserver is implemented as a pipe. The
+// `--jobserver-auth` value specifies either a path to a named pipe
+// (`fifo:PATH`) or a pair of file descriptors (`R,W`). The pipe is
+// pre-loaded with single-character tokens, one for each available job slot.
+//
+// - To acquire an additional slot, a client reads a single-character token
+// from the pipe.
+// - To release a slot, the client must write the *exact same* character
+// token back to the pipe.
+//
+// It is critical that a client releases all acquired slots before it exits,
+// even in cases of error, to avoid deadlocking the build.
+//
+// Example:
+// A multi-threaded linker invoked by `make -j8` wants to use multiple
+// threads. It first checks for the jobserver. It knows it has one implicit
+// slot, so it can use one thread. It then tries to acquire 7 more slots by
+// reading 7 tokens from the jobserver pipe. If it only receives 3 tokens,
+// it knows it can use a total of 1 (implicit) + 3 (acquired) = 4 threads.
+// Before exiting, it must write the 3 tokens it read back to the pipe.
+//
+// For more context, see:
+// - GNU Make manual on job slots:
+// https://www.gnu.org/software/make/manual/html_node/Job-Slots.html
+// - LLVM RFC discussion on jobserver support:
+// https://discourse.llvm.org/t/rfc-adding-gnu-make-jobserver-
+// support-to-llvm-for-coordinated-parallelism/87034
+// - Ninja’s jobserver support PR:
+// https://github.com/ninja-build/ninja/pull/2506
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_SUPPORT_JOBSERVER_H
+#define LLVM_SUPPORT_JOBSERVER_H
+
+#include "llvm/ADT/StringRef.h"
+#include <memory>
+#include <string>
+
+namespace llvm {
+
+/// A JobSlot represents a single job slot that can be acquired from or released
+/// to a jobserver pool. This class is move-only.
+class JobSlot {
+public:
+ /// Default constructor creates an invalid instance.
+ JobSlot() = default;
+
+ // Move operations are allowed.
+ JobSlot(JobSlot &&Other) noexcept : Value(Other.Value) {
+ Other.Value = kInvalidValue;
+ }
+ JobSlot &operator=(JobSlot &&Other) noexcept {
+ if (this != &Other) {
+ this->Value = Other.Value;
+ Other.Value = kInvalidValue;
+ }
+ return *this;
+ }
+
+ // Copy operations are disallowed.
+ JobSlot(const JobSlot &) = delete;
+ JobSlot &operator=(const JobSlot &) = delete;
+
+ /// Returns true if this instance is valid (either implicit or explicit).
+ bool isValid() const { return Value >= 0; }
+
+ /// Returns true if this instance represents the implicit job slot.
+ bool isImplicit() const { return Value == kImplicitValue; }
+
+ static JobSlot createExplicit(uint8_t V) {
+ return JobSlot(static_cast<int16_t>(V));
+ }
+
+ static JobSlot createImplicit() { return JobSlot(kImplicitValue); }
+
+ uint8_t getExplicitValue() const;
+ bool isExplicit() const { return isValid() && !isImplicit(); }
+
+private:
+ friend class JobserverClient;
+ friend class JobserverClientImpl;
+
+ JobSlot(int16_t V) : Value(V) {}
+
+ /// The jobserver pipe carries explicit tokens (bytes 0–255). We reserve two
+ /// sentinels in Value for special cases:
+ /// kInvalidValue (-1): no slot held
+ /// kImplicitValue (INT16_MAX): implicit slot granted at startup (no pipe
+ /// I/O)
+ ///
+ /// We use int16_t so Value can store 0–255 explicit tokens and
+ /// sentinels without overflow, enforces fixed 16-bit width, and avoids
+ /// unsigned/signed mix-ups.
+ static constexpr int16_t kInvalidValue = -1;
+ static constexpr int16_t kImplicitValue = INT16_MAX;
+ int16_t Value = kInvalidValue;
+};
+
+/// The public interface for a jobserver client.
+/// This client is a lazy-initialized singleton that is created on first use.
+class JobserverClient {
+public:
+ virtual ~JobserverClient();
+
+ /// Tries to acquire a job slot from the pool. On failure (e.g., if the pool
+ /// is empty), this returns an invalid JobSlot instance. The first successful
+ /// call will always return the implicit slot.
+ virtual JobSlot tryAcquire() = 0;
+
+ /// Releases a job slot back to the pool.
+ virtual void release(JobSlot Slot) = 0;
+
+ /// Returns the number of job slots available, as determined on first use.
+ /// This value is cached. Returns 0 if no jobserver is active.
+ virtual unsigned getNumJobs() const = 0;
+
+ /// Returns the singleton instance of the JobserverClient.
+ /// The instance is created on the first call to this function.
+ /// Returns a nullptr if no jobserver is configured or an error occurs.
+ static JobserverClient *getInstance();
+
+ /// Resets the singleton instance. For testing purposes only.
+ static void resetForTesting();
+};
+
+} // end namespace llvm
+
+#endif // LLVM_SUPPORT_JOBSERVER_H
diff --git a/llvm/include/llvm/Support/ThreadPool.h b/llvm/include/llvm/Support/ThreadPool.h
index c26681c..c20efc7 100644
--- a/llvm/include/llvm/Support/ThreadPool.h
+++ b/llvm/include/llvm/Support/ThreadPool.h
@@ -16,6 +16,7 @@
#include "llvm/ADT/DenseMap.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/Support/Compiler.h"
+#include "llvm/Support/Jobserver.h"
#include "llvm/Support/RWMutex.h"
#include "llvm/Support/Threading.h"
#include "llvm/Support/thread.h"
@@ -180,6 +181,7 @@ private:
void grow(int requested);
void processTasks(ThreadPoolTaskGroup *WaitingForGroup);
+ void processTasksWithJobserver();
/// Threads in flight
std::vector<llvm::thread> Threads;
@@ -208,6 +210,8 @@ private:
/// Maximum number of threads to potentially grow this pool to.
const unsigned MaxThreadCount;
+
+ JobserverClient *TheJobserver = nullptr;
};
#endif // LLVM_ENABLE_THREADS
diff --git a/llvm/include/llvm/Support/Threading.h b/llvm/include/llvm/Support/Threading.h
index d3fe0a5..8884680 100644
--- a/llvm/include/llvm/Support/Threading.h
+++ b/llvm/include/llvm/Support/Threading.h
@@ -142,6 +142,11 @@ constexpr bool llvm_is_multithreaded() { return LLVM_ENABLE_THREADS; }
/// the thread shall remain on the actual CPU socket.
LLVM_ABI std::optional<unsigned>
compute_cpu_socket(unsigned ThreadPoolNum) const;
+
+ /// If true, the thread pool will attempt to coordinate with a GNU Make
+ /// jobserver, acquiring a job slot before processing a task. If no
+ /// jobserver is found in the environment, this is ignored.
+ bool UseJobserver = false;
};
/// Build a strategy from a number of threads as a string provided in \p Num.
@@ -210,6 +215,19 @@ constexpr bool llvm_is_multithreaded() { return LLVM_ENABLE_THREADS; }
return S;
}
+ /// Returns a thread strategy that attempts to coordinate with a GNU Make
+ /// jobserver. The number of active threads will be limited by the number of
+ /// available job slots. If no jobserver is detected in the environment, this
+ /// strategy falls back to the default hardware_concurrency() behavior.
+ inline ThreadPoolStrategy jobserver_concurrency() {
+ ThreadPoolStrategy S;
+ S.UseJobserver = true;
+ // We can still request all threads be created, as they will simply
+ // block waiting for a job slot if the jobserver is the limiting factor.
+ S.ThreadsRequested = 0; // 0 means 'use all available'
+ return S;
+ }
+
/// Return the current thread id, as used in various OS system calls.
/// Note that not all platforms guarantee that the value returned will be
/// unique across the entire system, so portable code should not assume
diff --git a/llvm/include/llvm/Target/GenericOpcodes.td b/llvm/include/llvm/Target/GenericOpcodes.td
index faf7788..e3f995d 100644
--- a/llvm/include/llvm/Target/GenericOpcodes.td
+++ b/llvm/include/llvm/Target/GenericOpcodes.td
@@ -126,7 +126,7 @@ def G_FRAME_INDEX : GenericInstruction {
}
def G_GLOBAL_VALUE : GenericInstruction {
- let OutOperandList = (outs type0:$dst);
+ let OutOperandList = (outs ptype0:$dst);
let InOperandList = (ins unknown:$src);
let hasSideEffects = false;
}
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 95f53fe..6ea2e27 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -12698,6 +12698,45 @@ unsigned SelectionDAG::AssignTopologicalOrder() {
return DAGSize;
}
+void SelectionDAG::getTopologicallyOrderedNodes(
+ SmallVectorImpl<const SDNode *> &SortedNodes) const {
+ SortedNodes.clear();
+ // Node -> remaining number of outstanding operands.
+ DenseMap<const SDNode *, unsigned> RemainingOperands;
+
+ // Put nodes without any operands into SortedNodes first.
+ for (const SDNode &N : allnodes()) {
+ checkForCycles(&N, this);
+ unsigned NumOperands = N.getNumOperands();
+ if (NumOperands == 0)
+ SortedNodes.push_back(&N);
+ else
+ // Record their total number of outstanding operands.
+ RemainingOperands[&N] = NumOperands;
+ }
+
+ // A node is pushed into SortedNodes when all of its operands (predecessors in
+ // the graph) are also in SortedNodes.
+ for (unsigned i = 0U; i < SortedNodes.size(); ++i) {
+ const SDNode *N = SortedNodes[i];
+ for (const SDNode *U : N->users()) {
+ unsigned &NumRemOperands = RemainingOperands[U];
+ assert(NumRemOperands && "Invalid number of remaining operands");
+ --NumRemOperands;
+ if (!NumRemOperands)
+ SortedNodes.push_back(U);
+ }
+ }
+
+ assert(SortedNodes.size() == AllNodes.size() && "Node count mismatch");
+ assert(SortedNodes.front()->getOpcode() == ISD::EntryToken &&
+ "First node in topological sort is not the entry token");
+ assert(SortedNodes.front()->getNumOperands() == 0 &&
+ "First node in topological sort has operands");
+ assert(SortedNodes.back()->use_empty() &&
+ "Last node in topologic sort has users");
+}
+
/// AddDbgValue - Add a dbg_value SDNode. If SD is non-null that means the
/// value is produced by SD.
void SelectionDAG::AddDbgValue(SDDbgValue *DB, bool isParameter) {
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp
index 4b2a00c..fcfbfe6 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGDumper.cpp
@@ -1061,13 +1061,24 @@ static void DumpNodes(const SDNode *N, unsigned indent, const SelectionDAG *G) {
N->dump(G);
}
-LLVM_DUMP_METHOD void SelectionDAG::dump() const {
+LLVM_DUMP_METHOD void SelectionDAG::dump(bool Sorted) const {
dbgs() << "SelectionDAG has " << AllNodes.size() << " nodes:\n";
- for (const SDNode &N : allnodes()) {
+ auto dumpEachNode = [this](const SDNode &N) {
if (!N.hasOneUse() && &N != getRoot().getNode() &&
(!shouldPrintInline(N, this) || N.use_empty()))
DumpNodes(&N, 2, this);
+ };
+
+ if (Sorted) {
+ SmallVector<const SDNode *> SortedNodes;
+ SortedNodes.reserve(AllNodes.size());
+ getTopologicallyOrderedNodes(SortedNodes);
+ for (const SDNode *N : SortedNodes)
+ dumpEachNode(*N);
+ } else {
+ for (const SDNode &N : allnodes())
+ dumpEachNode(N);
}
if (getRoot().getNode()) DumpNodes(getRoot().getNode(), 2, this);
diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp
index e61558c..c35f29d 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp
@@ -144,6 +144,11 @@ UseMBPI("use-mbpi",
cl::init(true), cl::Hidden);
#ifndef NDEBUG
+static cl::opt<bool>
+ DumpSortedDAG("dump-sorted-dags", cl::Hidden,
+ cl::desc("Print DAGs with sorted nodes in debug dump"),
+ cl::init(false));
+
static cl::opt<std::string>
FilterDAGBasicBlockName("filter-view-dags", cl::Hidden,
cl::desc("Only display the basic block whose name "
@@ -932,7 +937,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nInitial selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -952,7 +957,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nOptimized lowered selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -974,7 +979,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nType-legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -998,7 +1003,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nOptimized type-legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1016,7 +1021,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nVector-legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1032,7 +1037,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nVector/type-legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1052,7 +1057,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nOptimized vector-legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1072,7 +1077,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nLegalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1092,7 +1097,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nOptimized legalized selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
#if !defined(NDEBUG) && LLVM_ENABLE_ABI_BREAKING_CHECKS
if (TTI->hasBranchDivergence())
@@ -1116,7 +1121,7 @@ void SelectionDAGISel::CodeGenAndEmitDAG() {
ISEL_DUMP(dbgs() << "\nSelected selection DAG: "
<< printMBBReference(*FuncInfo->MBB) << " '" << BlockName
<< "'\n";
- CurDAG->dump());
+ CurDAG->dump(DumpSortedDAG));
if (ViewSchedDAGs && MatchFilterBB)
CurDAG->viewGraph("scheduler input for " + BlockName);
diff --git a/llvm/lib/Support/CMakeLists.txt b/llvm/lib/Support/CMakeLists.txt
index 7da972f..42b21b5 100644
--- a/llvm/lib/Support/CMakeLists.txt
+++ b/llvm/lib/Support/CMakeLists.txt
@@ -207,6 +207,7 @@ add_llvm_component_library(LLVMSupport
InstructionCost.cpp
IntEqClasses.cpp
IntervalMap.cpp
+ Jobserver.cpp
JSON.cpp
KnownBits.cpp
KnownFPClass.cpp
diff --git a/llvm/lib/Support/Jobserver.cpp b/llvm/lib/Support/Jobserver.cpp
new file mode 100644
index 0000000..9f726eb
--- /dev/null
+++ b/llvm/lib/Support/Jobserver.cpp
@@ -0,0 +1,259 @@
+//===- llvm/Support/Jobserver.cpp - Jobserver Client Implementation -------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Support/Jobserver.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/Statistic.h"
+#include "llvm/ADT/StringExtras.h"
+#include "llvm/Config/llvm-config.h"
+#include "llvm/Support/Debug.h"
+#include "llvm/Support/Error.h"
+#include "llvm/Support/raw_ostream.h"
+
+#include <atomic>
+#include <memory>
+#include <mutex>
+#include <new>
+
+#define DEBUG_TYPE "jobserver"
+
+using namespace llvm;
+
+namespace {
+struct FdPair {
+ int Read = -1;
+ int Write = -1;
+ bool isValid() const { return Read >= 0 && Write >= 0; }
+};
+
+struct JobserverConfig {
+ enum Mode {
+ None,
+ PosixFifo,
+ PosixPipe,
+ Win32Semaphore,
+ };
+ Mode TheMode = None;
+ std::string Path;
+ FdPair PipeFDs;
+};
+
+/// A helper function that checks if `Input` starts with `Prefix`.
+/// If it does, it removes the prefix from `Input`, assigns the remainder to
+/// `Value`, and returns true. Otherwise, it returns false.
+bool getPrefixedValue(StringRef Input, StringRef Prefix, StringRef &Value) {
+ if (Input.consume_front(Prefix)) {
+ Value = Input;
+ return true;
+ }
+ return false;
+}
+
+/// A helper function to parse a string in the format "R,W" where R and W are
+/// non-negative integers representing file descriptors. It populates the
+/// `ReadFD` and `WriteFD` output parameters. Returns true on success.
+static std::optional<FdPair> getFileDescriptorPair(StringRef Input) {
+ FdPair FDs;
+ if (Input.consumeInteger(10, FDs.Read))
+ return std::nullopt;
+ if (!Input.consume_front(","))
+ return std::nullopt;
+ if (Input.consumeInteger(10, FDs.Write))
+ return std::nullopt;
+ if (!Input.empty() || !FDs.isValid())
+ return std::nullopt;
+ return FDs;
+}
+
+/// Parses the `MAKEFLAGS` environment variable string to find jobserver
+/// arguments. It splits the string into space-separated arguments and searches
+/// for `--jobserver-auth` or `--jobserver-fds`. Based on the value of these
+/// arguments, it determines the jobserver mode (Pipe, FIFO, or Semaphore) and
+/// connection details (file descriptors or path).
+Expected<JobserverConfig> parseNativeMakeFlags(StringRef MakeFlags) {
+ JobserverConfig Config;
+ if (MakeFlags.empty())
+ return Config;
+
+ // Split the MAKEFLAGS string into arguments.
+ SmallVector<StringRef, 8> Args;
+ SplitString(MakeFlags, Args);
+
+ // If '-n' (dry-run) is present as a legacy flag (not starting with '-'),
+ // disable the jobserver.
+ if (!Args.empty() && !Args[0].starts_with("-") && Args[0].contains('n'))
+ return Config;
+
+ // Iterate through arguments to find jobserver flags.
+ // Note that make may pass multiple --jobserver-auth flags; the last one wins.
+ for (StringRef Arg : Args) {
+ StringRef Value;
+ if (getPrefixedValue(Arg, "--jobserver-auth=", Value)) {
+ // Try to parse as a file descriptor pair first.
+ if (auto FDPair = getFileDescriptorPair(Value)) {
+ Config.TheMode = JobserverConfig::PosixPipe;
+ Config.PipeFDs = *FDPair;
+ } else {
+ StringRef FifoPath;
+ // If not FDs, try to parse as a named pipe (fifo).
+ if (getPrefixedValue(Value, "fifo:", FifoPath)) {
+ Config.TheMode = JobserverConfig::PosixFifo;
+ Config.Path = FifoPath.str();
+ } else {
+ // Otherwise, assume it's a Windows semaphore.
+ Config.TheMode = JobserverConfig::Win32Semaphore;
+ Config.Path = Value.str();
+ }
+ }
+ } else if (getPrefixedValue(Arg, "--jobserver-fds=", Value)) {
+ // This is an alternative, older syntax for the pipe-based server.
+ if (auto FDPair = getFileDescriptorPair(Value)) {
+ Config.TheMode = JobserverConfig::PosixPipe;
+ Config.PipeFDs = *FDPair;
+ } else {
+ return createStringError(inconvertibleErrorCode(),
+ "Invalid file descriptor pair in MAKEFLAGS");
+ }
+ }
+ }
+
+// Perform platform-specific validation.
+#ifdef _WIN32
+ if (Config.TheMode == JobserverConfig::PosixFifo ||
+ Config.TheMode == JobserverConfig::PosixPipe)
+ return createStringError(
+ inconvertibleErrorCode(),
+ "FIFO/Pipe-based jobserver is not supported on Windows");
+#else
+ if (Config.TheMode == JobserverConfig::Win32Semaphore)
+ return createStringError(
+ inconvertibleErrorCode(),
+ "Semaphore-based jobserver is not supported on this platform");
+#endif
+ return Config;
+}
+
+std::once_flag GJobserverOnceFlag;
+JobserverClient *GJobserver = nullptr;
+
+} // namespace
+
+namespace llvm {
+class JobserverClientImpl : public JobserverClient {
+ bool IsInitialized = false;
+ std::atomic<bool> HasImplicitSlot{true};
+ unsigned NumJobs = 0;
+
+public:
+ JobserverClientImpl(const JobserverConfig &Config);
+ ~JobserverClientImpl() override;
+
+ JobSlot tryAcquire() override;
+ void release(JobSlot Slot) override;
+ unsigned getNumJobs() const override { return NumJobs; }
+
+ bool isValid() const { return IsInitialized; }
+
+private:
+#if defined(LLVM_ON_UNIX)
+ int ReadFD = -1;
+ int WriteFD = -1;
+ std::string FifoPath;
+#elif defined(_WIN32)
+ void *Semaphore = nullptr;
+#endif
+};
+} // namespace llvm
+
+// Include the platform-specific parts of the class.
+#if defined(LLVM_ON_UNIX)
+#include "Unix/Jobserver.inc"
+#elif defined(_WIN32)
+#include "Windows/Jobserver.inc"
+#else
+// Dummy implementation for unsupported platforms.
+JobserverClientImpl::JobserverClientImpl(const JobserverConfig &Config) {}
+JobserverClientImpl::~JobserverClientImpl() = default;
+JobSlot JobserverClientImpl::tryAcquire() { return JobSlot(); }
+void JobserverClientImpl::release(JobSlot Slot) {}
+#endif
+
+namespace llvm {
+JobserverClient::~JobserverClient() = default;
+
+uint8_t JobSlot::getExplicitValue() const {
+ assert(isExplicit() && "Cannot get value of implicit or invalid slot");
+ return static_cast<uint8_t>(Value);
+}
+
+/// This is the main entry point for acquiring a jobserver client. It uses a
+/// std::call_once to ensure the singleton `GJobserver` instance is created
+/// safely in a multi-threaded environment. On first call, it reads the
+/// `MAKEFLAGS` environment variable, parses it, and attempts to construct and
+/// initialize a `JobserverClientImpl`. If successful, the global instance is
+/// stored in `GJobserver`. Subsequent calls will return the existing instance.
+JobserverClient *JobserverClient::getInstance() {
+ std::call_once(GJobserverOnceFlag, []() {
+ LLVM_DEBUG(
+ dbgs()
+ << "JobserverClient::getInstance() called for the first time.\n");
+ const char *MakeFlagsEnv = getenv("MAKEFLAGS");
+ if (!MakeFlagsEnv) {
+ errs() << "Warning: failed to create jobserver client due to MAKEFLAGS "
+ "environment variable not found\n";
+ return;
+ }
+
+ LLVM_DEBUG(dbgs() << "Found MAKEFLAGS = \"" << MakeFlagsEnv << "\"\n");
+
+ auto ConfigOrErr = parseNativeMakeFlags(MakeFlagsEnv);
+ if (Error Err = ConfigOrErr.takeError()) {
+ errs() << "Warning: failed to create jobserver client due to invalid "
+ "MAKEFLAGS environment variable: "
+ << toString(std::move(Err)) << "\n";
+ return;
+ }
+
+ JobserverConfig Config = *ConfigOrErr;
+ if (Config.TheMode == JobserverConfig::None) {
+ errs() << "Warning: failed to create jobserver client due to jobserver "
+ "mode missing in MAKEFLAGS environment variable\n";
+ return;
+ }
+
+ if (Config.TheMode == JobserverConfig::PosixPipe) {
+#if defined(LLVM_ON_UNIX)
+ if (!areFdsValid(Config.PipeFDs.Read, Config.PipeFDs.Write)) {
+ errs() << "Warning: failed to create jobserver client due to invalid "
+ "Pipe FDs in MAKEFLAGS environment variable\n";
+ return;
+ }
+#endif
+ }
+
+ auto Client = std::make_unique<JobserverClientImpl>(Config);
+ if (Client->isValid()) {
+ LLVM_DEBUG(dbgs() << "Jobserver client created successfully!\n");
+ GJobserver = Client.release();
+ } else
+ errs() << "Warning: jobserver client initialization failed.\n";
+ });
+ return GJobserver;
+}
+
+/// For testing purposes only. This function resets the singleton instance by
+/// destroying the existing client and re-initializing the `std::once_flag`.
+/// This allows tests to simulate the first-time initialization of the
+/// jobserver client multiple times.
+void JobserverClient::resetForTesting() {
+ delete GJobserver;
+ GJobserver = nullptr;
+ // Re-construct the std::once_flag in place to reset the singleton state.
+ new (&GJobserverOnceFlag) std::once_flag();
+}
+} // namespace llvm
diff --git a/llvm/lib/Support/Parallel.cpp b/llvm/lib/Support/Parallel.cpp
index 3ac6fc7..8e0c724 100644
--- a/llvm/lib/Support/Parallel.cpp
+++ b/llvm/lib/Support/Parallel.cpp
@@ -7,12 +7,17 @@
//===----------------------------------------------------------------------===//
#include "llvm/Support/Parallel.h"
+#include "llvm/ADT/ScopeExit.h"
#include "llvm/Config/llvm-config.h"
+#include "llvm/Support/ExponentialBackoff.h"
+#include "llvm/Support/Jobserver.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/Threading.h"
#include <atomic>
#include <future>
+#include <memory>
+#include <mutex>
#include <thread>
#include <vector>
@@ -49,6 +54,9 @@ public:
class ThreadPoolExecutor : public Executor {
public:
explicit ThreadPoolExecutor(ThreadPoolStrategy S) {
+ if (S.UseJobserver)
+ TheJobserver = JobserverClient::getInstance();
+
ThreadCount = S.compute_thread_count();
// Spawn all but one of the threads in another thread as spawning threads
// can take a while.
@@ -69,6 +77,10 @@ public:
});
}
+ // To make sure the thread pool executor can only be created with a parallel
+ // strategy.
+ ThreadPoolExecutor() = delete;
+
void stop() {
{
std::lock_guard<std::mutex> Lock(Mutex);
@@ -111,15 +123,62 @@ private:
void work(ThreadPoolStrategy S, unsigned ThreadID) {
threadIndex = ThreadID;
S.apply_thread_strategy(ThreadID);
+ // Note on jobserver deadlock avoidance:
+ // GNU Make grants each invoked process one implicit job slot. Our
+ // JobserverClient models this by returning an implicit JobSlot on the
+ // first successful tryAcquire() in a process. This guarantees forward
+ // progress without requiring a dedicated "always-on" thread here.
+
+ static thread_local std::unique_ptr<ExponentialBackoff> Backoff;
+
while (true) {
- std::unique_lock<std::mutex> Lock(Mutex);
- Cond.wait(Lock, [&] { return Stop || !WorkStack.empty(); });
- if (Stop)
- break;
- auto Task = std::move(WorkStack.back());
- WorkStack.pop_back();
- Lock.unlock();
- Task();
+ if (TheJobserver) {
+ // Jobserver-mode scheduling:
+ // - Acquire one job slot (with exponential backoff to avoid busy-wait).
+ // - While holding the slot, drain and run tasks from the local queue.
+ // - Release the slot when the queue is empty or when shutting down.
+ // Rationale: Holding a slot amortizes acquire/release overhead over
+ // multiple tasks and avoids requeue/yield churn, while still enforcing
+ // the jobserver’s global concurrency limit. With K available slots,
+ // up to K workers run tasks in parallel; within each worker tasks run
+ // sequentially until the local queue is empty.
+ ExponentialBackoff Backoff(std::chrono::hours(24));
+ JobSlot Slot;
+ do {
+ if (Stop)
+ return;
+ Slot = TheJobserver->tryAcquire();
+ if (Slot.isValid())
+ break;
+ } while (Backoff.waitForNextAttempt());
+
+ auto SlotReleaser = llvm::make_scope_exit(
+ [&] { TheJobserver->release(std::move(Slot)); });
+
+ while (true) {
+ std::function<void()> Task;
+ {
+ std::unique_lock<std::mutex> Lock(Mutex);
+ Cond.wait(Lock, [&] { return Stop || !WorkStack.empty(); });
+ if (Stop && WorkStack.empty())
+ return;
+ if (WorkStack.empty())
+ break;
+ Task = std::move(WorkStack.back());
+ WorkStack.pop_back();
+ }
+ Task();
+ }
+ } else {
+ std::unique_lock<std::mutex> Lock(Mutex);
+ Cond.wait(Lock, [&] { return Stop || !WorkStack.empty(); });
+ if (Stop)
+ break;
+ auto Task = std::move(WorkStack.back());
+ WorkStack.pop_back();
+ Lock.unlock();
+ Task();
+ }
}
}
@@ -130,9 +189,20 @@ private:
std::promise<void> ThreadsCreated;
std::vector<std::thread> Threads;
unsigned ThreadCount;
+
+ JobserverClient *TheJobserver = nullptr;
};
-Executor *Executor::getDefaultExecutor() {
+// A global raw pointer to the executor. Lifetime is managed by the
+// objects created within createExecutor().
+static Executor *TheExec = nullptr;
+static std::once_flag Flag;
+
+// This function will be called exactly once to create the executor.
+// It contains the necessary platform-specific logic. Since functions
+// called by std::call_once cannot return value, we have to set the
+// executor as a global variable.
+void createExecutor() {
#ifdef _WIN32
// The ManagedStatic enables the ThreadPoolExecutor to be stopped via
// llvm_shutdown() which allows a "clean" fast exit, e.g. via _exit(). This
@@ -156,16 +226,22 @@ Executor *Executor::getDefaultExecutor() {
ThreadPoolExecutor::Deleter>
ManagedExec;
static std::unique_ptr<ThreadPoolExecutor> Exec(&(*ManagedExec));
- return Exec.get();
+ TheExec = Exec.get();
#else
// ManagedStatic is not desired on other platforms. When `Exec` is destroyed
// by llvm_shutdown(), worker threads will clean up and invoke TLS
// destructors. This can lead to race conditions if other threads attempt to
// access TLS objects that have already been destroyed.
static ThreadPoolExecutor Exec(strategy);
- return &Exec;
+ TheExec = &Exec;
#endif
}
+
+Executor *Executor::getDefaultExecutor() {
+ // Use std::call_once to lazily and safely initialize the executor.
+ std::call_once(Flag, createExecutor);
+ return TheExec;
+}
} // namespace
} // namespace detail
diff --git a/llvm/lib/Support/ThreadPool.cpp b/llvm/lib/Support/ThreadPool.cpp
index c304f0f..6960268 100644
--- a/llvm/lib/Support/ThreadPool.cpp
+++ b/llvm/lib/Support/ThreadPool.cpp
@@ -6,6 +6,7 @@
//
//===----------------------------------------------------------------------===//
//
+//
// This file implements a crude C++11 based thread pool.
//
//===----------------------------------------------------------------------===//
@@ -14,6 +15,8 @@
#include "llvm/Config/llvm-config.h"
+#include "llvm/ADT/ScopeExit.h"
+#include "llvm/Support/ExponentialBackoff.h"
#include "llvm/Support/FormatVariadic.h"
#include "llvm/Support/Threading.h"
#include "llvm/Support/raw_ostream.h"
@@ -33,7 +36,10 @@ ThreadPoolInterface::~ThreadPoolInterface() = default;
#if LLVM_ENABLE_THREADS
StdThreadPool::StdThreadPool(ThreadPoolStrategy S)
- : Strategy(S), MaxThreadCount(S.compute_thread_count()) {}
+ : Strategy(S), MaxThreadCount(S.compute_thread_count()) {
+ if (Strategy.UseJobserver)
+ TheJobserver = JobserverClient::getInstance();
+}
void StdThreadPool::grow(int requested) {
llvm::sys::ScopedWriter LockGuard(ThreadsLock);
@@ -45,7 +51,15 @@ void StdThreadPool::grow(int requested) {
Threads.emplace_back([this, ThreadID] {
set_thread_name(formatv("llvm-worker-{0}", ThreadID));
Strategy.apply_thread_strategy(ThreadID);
- processTasks(nullptr);
+ // Note on jobserver deadlock avoidance:
+ // GNU Make grants each invoked process one implicit job slot.
+ // JobserverClient::tryAcquire() returns that implicit slot on the first
+ // successful call in a process, ensuring forward progress without a
+ // dedicated "always-on" thread.
+ if (TheJobserver)
+ processTasksWithJobserver();
+ else
+ processTasks(nullptr);
});
}
}
@@ -133,6 +147,96 @@ void StdThreadPool::processTasks(ThreadPoolTaskGroup *WaitingForGroup) {
}
}
+/// Main loop for worker threads when using a jobserver.
+/// This function uses a two-level queue; it first acquires a job slot from the
+/// external jobserver, then retrieves a task from the internal queue.
+/// This allows the thread pool to cooperate with build systems like `make -j`.
+void StdThreadPool::processTasksWithJobserver() {
+ while (true) {
+ // Acquire a job slot from the external jobserver.
+ // This polls for a slot and yields the thread to avoid a high-CPU wait.
+ JobSlot Slot;
+ // The timeout for the backoff can be very long, as the shutdown
+ // is checked on each iteration. The sleep duration is capped by MaxWait
+ // in ExponentialBackoff, so shutdown latency is not a problem.
+ ExponentialBackoff Backoff(std::chrono::hours(24));
+ bool AcquiredToken = false;
+ do {
+ // Return if the thread pool is shutting down.
+ {
+ std::unique_lock<std::mutex> LockGuard(QueueLock);
+ if (!EnableFlag)
+ return;
+ }
+
+ Slot = TheJobserver->tryAcquire();
+ if (Slot.isValid()) {
+ AcquiredToken = true;
+ break;
+ }
+ } while (Backoff.waitForNextAttempt());
+
+ if (!AcquiredToken) {
+ // This is practically unreachable with a 24h timeout and indicates a
+ // deeper problem if hit.
+ report_fatal_error("Timed out waiting for jobserver token.");
+ }
+
+ // `make_scope_exit` guarantees the job slot is released, even if the
+ // task throws or we exit early. This prevents deadlocking the build.
+ auto SlotReleaser =
+ make_scope_exit([&] { TheJobserver->release(std::move(Slot)); });
+
+ // While we hold a job slot, process tasks from the internal queue.
+ while (true) {
+ std::function<void()> Task;
+ ThreadPoolTaskGroup *GroupOfTask = nullptr;
+
+ {
+ std::unique_lock<std::mutex> LockGuard(QueueLock);
+
+ // Wait until a task is available or the pool is shutting down.
+ QueueCondition.wait(LockGuard,
+ [&] { return !EnableFlag || !Tasks.empty(); });
+
+ // If shutting down and the queue is empty, the thread can terminate.
+ if (!EnableFlag && Tasks.empty())
+ return;
+
+ // If the queue is empty, we're done processing tasks for now.
+ // Break the inner loop to release the job slot.
+ if (Tasks.empty())
+ break;
+
+ // A task is available. Mark it as active before releasing the lock
+ // to prevent race conditions with `wait()`.
+ ++ActiveThreads;
+ Task = std::move(Tasks.front().first);
+ GroupOfTask = Tasks.front().second;
+ if (GroupOfTask != nullptr)
+ ++ActiveGroups[GroupOfTask];
+ Tasks.pop_front();
+ } // The queue lock is released.
+
+ // Run the task. The job slot remains acquired during execution.
+ Task();
+
+ // The task has finished. Update the active count and notify any waiters.
+ {
+ std::lock_guard<std::mutex> LockGuard(QueueLock);
+ --ActiveThreads;
+ if (GroupOfTask != nullptr) {
+ auto A = ActiveGroups.find(GroupOfTask);
+ if (--(A->second) == 0)
+ ActiveGroups.erase(A);
+ }
+ // If all tasks are complete, notify any waiting threads.
+ if (workCompletedUnlocked(nullptr))
+ CompletionCondition.notify_all();
+ }
+ }
+ }
+}
bool StdThreadPool::workCompletedUnlocked(ThreadPoolTaskGroup *Group) const {
if (Group == nullptr)
return !ActiveThreads && Tasks.empty();
diff --git a/llvm/lib/Support/Threading.cpp b/llvm/lib/Support/Threading.cpp
index 693de0e..9da357a 100644
--- a/llvm/lib/Support/Threading.cpp
+++ b/llvm/lib/Support/Threading.cpp
@@ -14,6 +14,7 @@
#include "llvm/Support/Threading.h"
#include "llvm/Config/config.h"
#include "llvm/Config/llvm-config.h"
+#include "llvm/Support/Jobserver.h"
#include <cassert>
#include <optional>
@@ -51,6 +52,10 @@ int llvm::get_physical_cores() { return -1; }
static int computeHostNumHardwareThreads();
unsigned llvm::ThreadPoolStrategy::compute_thread_count() const {
+ if (UseJobserver)
+ if (auto JS = JobserverClient::getInstance())
+ return JS->getNumJobs();
+
int MaxThreadCount =
UseHyperThreads ? computeHostNumHardwareThreads() : get_physical_cores();
if (MaxThreadCount <= 0)
diff --git a/llvm/lib/Support/Unix/Jobserver.inc b/llvm/lib/Support/Unix/Jobserver.inc
new file mode 100644
index 0000000..53bf7f2
--- /dev/null
+++ b/llvm/lib/Support/Unix/Jobserver.inc
@@ -0,0 +1,195 @@
+//===- llvm/Support/Unix/Jobserver.inc - Unix Jobserver Impl ----*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file implements the UNIX-specific parts of the JobserverClient class.
+//
+//===----------------------------------------------------------------------===//
+
+#include <atomic>
+#include <cassert>
+#include <cerrno>
+#include <fcntl.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+namespace {
+/// Returns true if the given file descriptor is a FIFO (named pipe).
+bool isFifo(int FD) {
+ struct stat StatBuf;
+ if (::fstat(FD, &StatBuf) != 0)
+ return false;
+ return S_ISFIFO(StatBuf.st_mode);
+}
+
+/// Returns true if the given file descriptors are valid.
+bool areFdsValid(int ReadFD, int WriteFD) {
+ if (ReadFD == -1 || WriteFD == -1)
+ return false;
+ // Check if the file descriptors are actually valid by checking their flags.
+ return ::fcntl(ReadFD, F_GETFD) != -1 && ::fcntl(WriteFD, F_GETFD) != -1;
+}
+} // namespace
+
+/// The constructor sets up the client based on the provided configuration.
+/// For pipe-based jobservers, it duplicates the inherited file descriptors,
+/// sets them to close-on-exec, and makes the read descriptor non-blocking.
+/// For FIFO-based jobservers, it opens the named pipe. After setup, it drains
+/// all available tokens from the jobserver to determine the total number of
+/// available jobs (`NumJobs`), then immediately releases them back.
+JobserverClientImpl::JobserverClientImpl(const JobserverConfig &Config) {
+ switch (Config.TheMode) {
+ case JobserverConfig::PosixPipe: {
+ // Duplicate the read and write file descriptors.
+ int NewReadFD = ::dup(Config.PipeFDs.Read);
+ if (NewReadFD < 0)
+ return;
+ int NewWriteFD = ::dup(Config.PipeFDs.Write);
+ if (NewWriteFD < 0) {
+ ::close(NewReadFD);
+ return;
+ }
+ // Set the new descriptors to be closed automatically on exec().
+ if (::fcntl(NewReadFD, F_SETFD, FD_CLOEXEC) == -1 ||
+ ::fcntl(NewWriteFD, F_SETFD, FD_CLOEXEC) == -1) {
+ ::close(NewReadFD);
+ ::close(NewWriteFD);
+ return;
+ }
+ // Set the read descriptor to non-blocking.
+ int flags = ::fcntl(NewReadFD, F_GETFL, 0);
+ if (flags == -1 || ::fcntl(NewReadFD, F_SETFL, flags | O_NONBLOCK) == -1) {
+ ::close(NewReadFD);
+ ::close(NewWriteFD);
+ return;
+ }
+ ReadFD = NewReadFD;
+ WriteFD = NewWriteFD;
+ break;
+ }
+ case JobserverConfig::PosixFifo:
+ // Open the FIFO for reading. It must be non-blocking and close-on-exec.
+ ReadFD = ::open(Config.Path.c_str(), O_RDONLY | O_NONBLOCK | O_CLOEXEC);
+ if (ReadFD < 0 || !isFifo(ReadFD)) {
+ if (ReadFD >= 0)
+ ::close(ReadFD);
+ ReadFD = -1;
+ return;
+ }
+ FifoPath = Config.Path;
+ // The write FD is opened on-demand in release().
+ WriteFD = -1;
+ break;
+ default:
+ return;
+ }
+
+ IsInitialized = true;
+ // Determine the total number of jobs by acquiring all available slots and
+ // then immediately releasing them.
+ SmallVector<JobSlot, 8> Slots;
+ while (true) {
+ auto S = tryAcquire();
+ if (!S.isValid())
+ break;
+ Slots.push_back(std::move(S));
+ }
+ NumJobs = Slots.size();
+ assert(NumJobs >= 1 && "Invalid number of jobs");
+ for (auto &S : Slots)
+ release(std::move(S));
+}
+
+/// The destructor closes any open file descriptors.
+JobserverClientImpl::~JobserverClientImpl() {
+ if (ReadFD >= 0)
+ ::close(ReadFD);
+ if (WriteFD >= 0)
+ ::close(WriteFD);
+}
+
+/// Tries to acquire a job slot. The first call to this function will always
+/// successfully acquire the single "implicit" slot that is granted to every
+/// process started by `make`. Subsequent calls attempt to read a one-byte
+/// token from the jobserver's read pipe. A successful read grants one
+/// explicit job slot. The read is non-blocking; if no token is available,
+/// it fails and returns an invalid JobSlot.
+JobSlot JobserverClientImpl::tryAcquire() {
+ if (!IsInitialized)
+ return JobSlot();
+
+ // The first acquisition is always for the implicit slot.
+ if (HasImplicitSlot.exchange(false, std::memory_order_acquire)) {
+ LLVM_DEBUG(dbgs() << "Acquired implicit job slot.\n");
+ return JobSlot::createImplicit();
+ }
+
+ char Token;
+ ssize_t Ret;
+ LLVM_DEBUG(dbgs() << "Attempting to read token from FD " << ReadFD << ".\n");
+ // Loop to retry on EINTR (interrupted system call).
+ do {
+ Ret = ::read(ReadFD, &Token, 1);
+ } while (Ret < 0 && errno == EINTR);
+
+ if (Ret == 1) {
+ LLVM_DEBUG(dbgs() << "Acquired explicit token '" << Token << "'.\n");
+ return JobSlot::createExplicit(static_cast<uint8_t>(Token));
+ }
+
+ LLVM_DEBUG(dbgs() << "Failed to acquire job slot, read returned " << Ret
+ << ".\n");
+ return JobSlot();
+}
+
+/// Releases a job slot back to the pool. If the slot is implicit, it simply
+/// resets a flag. If the slot is explicit, it writes the character token
+/// associated with the slot back into the jobserver's write pipe. For FIFO
+/// jobservers, this may require opening the FIFO for writing if it hasn't
+/// been already.
+void JobserverClientImpl::release(JobSlot Slot) {
+ if (!Slot.isValid())
+ return;
+
+ // Releasing the implicit slot just makes it available for the next acquire.
+ if (Slot.isImplicit()) {
+ LLVM_DEBUG(dbgs() << "Released implicit job slot.\n");
+ [[maybe_unused]] bool was_already_released =
+ HasImplicitSlot.exchange(true, std::memory_order_release);
+ assert(!was_already_released && "Implicit slot released twice");
+ return;
+ }
+
+ uint8_t Token = Slot.getExplicitValue();
+ LLVM_DEBUG(dbgs() << "Releasing explicit token '" << (char)Token << "' to FD "
+ << WriteFD << ".\n");
+
+ // For FIFO-based jobservers, the write FD might not be open yet.
+ // Open it on the first release.
+ if (WriteFD < 0) {
+ LLVM_DEBUG(dbgs() << "WriteFD is invalid, opening FIFO: " << FifoPath
+ << "\n");
+ WriteFD = ::open(FifoPath.c_str(), O_WRONLY | O_CLOEXEC);
+ if (WriteFD < 0) {
+ LLVM_DEBUG(dbgs() << "Failed to open FIFO for writing.\n");
+ return;
+ }
+ LLVM_DEBUG(dbgs() << "Opened FIFO as new WriteFD: " << WriteFD << "\n");
+ }
+
+ ssize_t Written;
+ // Loop to retry on EINTR (interrupted system call).
+ do {
+ Written = ::write(WriteFD, &Token, 1);
+ } while (Written < 0 && errno == EINTR);
+
+ if (Written <= 0) {
+ LLVM_DEBUG(dbgs() << "Failed to write token to pipe, write returned "
+ << Written << "\n");
+ }
+}
diff --git a/llvm/lib/Support/Windows/Jobserver.inc b/llvm/lib/Support/Windows/Jobserver.inc
new file mode 100644
index 0000000..79028ee
--- /dev/null
+++ b/llvm/lib/Support/Windows/Jobserver.inc
@@ -0,0 +1,79 @@
+//==- llvm/Support/Windows/Jobserver.inc - Windows Jobserver Impl -*- C++ -*-=//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file implements the Windows-specific parts of the JobserverClient class.
+// On Windows, the jobserver is implemented using a named semaphore.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Support/Windows/WindowsSupport.h"
+#include <atomic>
+#include <cassert>
+
+namespace llvm {
+/// The constructor for the Windows jobserver client. It attempts to open a
+/// handle to an existing named semaphore, the name of which is provided by
+/// GNU make in the --jobserver-auth argument. If the semaphore is opened
+/// successfully, the client is marked as initialized.
+JobserverClientImpl::JobserverClientImpl(const JobserverConfig &Config) {
+ Semaphore = (void *)::OpenSemaphoreA(SEMAPHORE_MODIFY_STATE | SYNCHRONIZE,
+ FALSE, Config.Path.c_str());
+ if (Semaphore != nullptr)
+ IsInitialized = true;
+}
+
+/// The destructor closes the handle to the semaphore, releasing the resource.
+JobserverClientImpl::~JobserverClientImpl() {
+ if (Semaphore != nullptr)
+ ::CloseHandle((HANDLE)Semaphore);
+}
+
+/// Tries to acquire a job slot. The first call always returns the implicit
+/// slot. Subsequent calls use a non-blocking wait on the semaphore
+/// (`WaitForSingleObject` with a timeout of 0). If the wait succeeds, the
+/// semaphore's count is decremented, and an explicit job slot is acquired.
+/// If the wait times out, it means no slots are available, and an invalid
+/// slot is returned.
+JobSlot JobserverClientImpl::tryAcquire() {
+ if (!IsInitialized)
+ return JobSlot();
+
+ // First, grant the implicit slot.
+ if (HasImplicitSlot.exchange(false, std::memory_order_acquire)) {
+ return JobSlot::createImplicit();
+ }
+
+ // Try to acquire a slot from the semaphore without blocking.
+ if (::WaitForSingleObject((HANDLE)Semaphore, 0) == WAIT_OBJECT_0) {
+ // The explicit token value is arbitrary on Windows, as the semaphore
+ // count is the real resource.
+ return JobSlot::createExplicit(1);
+ }
+
+ return JobSlot(); // Invalid slot
+}
+
+/// Releases a job slot back to the pool. If the slot is implicit, it simply
+/// resets a flag. For an explicit slot, it increments the semaphore's count
+/// by one using `ReleaseSemaphore`, making the slot available to other
+/// processes.
+void JobserverClientImpl::release(JobSlot Slot) {
+ if (!IsInitialized || !Slot.isValid())
+ return;
+
+ if (Slot.isImplicit()) {
+ [[maybe_unused]] bool was_already_released =
+ HasImplicitSlot.exchange(true, std::memory_order_release);
+ assert(!was_already_released && "Implicit slot released twice");
+ return;
+ }
+
+ // Release the slot by incrementing the semaphore count.
+ (void)::ReleaseSemaphore((HANDLE)Semaphore, 1, NULL);
+}
+} // namespace llvm
diff --git a/llvm/lib/Target/AMDGPU/SIRegisterInfo.td b/llvm/lib/Target/AMDGPU/SIRegisterInfo.td
index 5630580..f98e312 100644
--- a/llvm/lib/Target/AMDGPU/SIRegisterInfo.td
+++ b/llvm/lib/Target/AMDGPU/SIRegisterInfo.td
@@ -367,19 +367,6 @@ def SCC_CLASS : SIRegisterClass<"AMDGPU", [i1], 1, (add SCC)> {
let BaseClassOrder = 10000;
}
-def M0_CLASS : SIRegisterClass<"AMDGPU", [i32], 32, (add M0)> {
- let CopyCost = 1;
- let isAllocatable = 0;
- let HasSGPR = 1;
-}
-
-def M0_CLASS_LO16 : SIRegisterClass<"AMDGPU", [i16, f16, bf16], 16, (add M0_LO16)> {
- let CopyCost = 1;
- let Size = 16;
- let isAllocatable = 0;
- let HasSGPR = 1;
-}
-
// TODO: Do we need to set DwarfRegAlias on register tuples?
def SGPR_LO16 : SIRegisterClass<"AMDGPU", [i16, f16, bf16], 16,
@@ -797,7 +784,7 @@ def SReg_LO16 : SIRegisterClass<"AMDGPU", [i16, f16, bf16], 16,
TMA_LO_LO16, TMA_HI_LO16, TBA_LO_LO16, TBA_HI_LO16, SRC_SHARED_BASE_LO_LO16,
SRC_SHARED_LIMIT_LO_LO16, SRC_PRIVATE_BASE_LO_LO16, SRC_PRIVATE_LIMIT_LO_LO16,
SRC_POPS_EXITING_WAVE_ID_LO16, SRC_VCCZ_LO16, SRC_EXECZ_LO16, SRC_SCC_LO16,
- EXEC_LO_LO16, EXEC_HI_LO16, M0_CLASS_LO16, SRC_FLAT_SCRATCH_BASE_LO_LO16,
+ EXEC_LO_LO16, EXEC_HI_LO16, M0_LO16, SRC_FLAT_SCRATCH_BASE_LO_LO16,
SRC_FLAT_SCRATCH_BASE_HI_LO16)> {
let Size = 16;
let isAllocatable = 0;
@@ -805,7 +792,7 @@ def SReg_LO16 : SIRegisterClass<"AMDGPU", [i16, f16, bf16], 16,
}
def SReg_32_XEXEC : SIRegisterClass<"AMDGPU", [i32, f32, i16, f16, bf16, v2i16, v2f16, v2bf16, i1], 32,
- (add SReg_32_XM0_XEXEC, M0_CLASS)> {
+ (add SReg_32_XM0_XEXEC, M0)> {
let AllocationPriority = 0;
}
@@ -830,7 +817,7 @@ def APERTURE_Class : SIRegisterClass<"AMDGPU", Reg64Types.types, 32,
// Register class for all scalar registers (SGPRs + Special Registers)
def SReg_32 : SIRegisterClass<"AMDGPU", [i32, f32, i16, f16, bf16, v2i16, v2f16, v2bf16, i1], 32,
- (add SReg_32_XM0, M0_CLASS)> {
+ (add SReg_32_XM0, M0)> {
let AllocationPriority = 0;
let HasSGPR = 1;
let BaseClassOrder = 32;
diff --git a/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp b/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp
index ebd957c..e8c849e 100644
--- a/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp
+++ b/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp
@@ -195,9 +195,9 @@ class SPIRVLegalizePointerCast : public FunctionPass {
if (DstType->getElementType() != SrcType->getElementType()) {
// Support bitcast between vectors of different sizes only if
// the total bitwidth is the same.
- auto dstBitWidth =
+ [[maybe_unused]] auto dstBitWidth =
DstType->getElementType()->getScalarSizeInBits() * dstNumElements;
- auto srcBitWidth =
+ [[maybe_unused]] auto srcBitWidth =
SrcType->getElementType()->getScalarSizeInBits() * srcNumElements;
assert(dstBitWidth == srcBitWidth &&
"Unsupported bitcast between vectors of different sizes.");
diff --git a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
index ddb95a4..faeab95 100644
--- a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
+++ b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
@@ -29,6 +29,7 @@
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
+#include "llvm/ADT/StringExtras.h"
#include "llvm/Analysis/MemoryProfileInfo.h"
#include "llvm/Analysis/ModuleSummaryAnalysis.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
@@ -40,6 +41,7 @@
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/GraphWriter.h"
#include "llvm/Support/InterleavedRange.h"
+#include "llvm/Support/SHA1.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/IPO.h"
#include "llvm/Transforms/Utils/CallPromotionUtils.h"
@@ -60,6 +62,9 @@ STATISTIC(FunctionClonesThinBackend,
"Number of function clones created during ThinLTO backend");
STATISTIC(FunctionsClonedThinBackend,
"Number of functions that had clones created during ThinLTO backend");
+STATISTIC(
+ FunctionCloneDuplicatesThinBackend,
+ "Number of function clone duplicates detected during ThinLTO backend");
STATISTIC(AllocTypeNotCold, "Number of not cold static allocations (possibly "
"cloned) during whole program analysis");
STATISTIC(AllocTypeCold, "Number of cold static allocations (possibly cloned) "
@@ -5186,19 +5191,127 @@ bool CallsiteContextGraph<DerivedCCG, FuncTy, CallTy>::assignFunctions() {
return Changed;
}
+// Compute a SHA1 hash of the callsite and alloc version information of clone I
+// in the summary, to use in detection of duplicate clones.
+uint64_t ComputeHash(const FunctionSummary *FS, unsigned I) {
+ SHA1 Hasher;
+ // Update hash with any callsites that call non-default (non-zero) callee
+ // versions.
+ for (auto &SN : FS->callsites()) {
+ // In theory all callsites and allocs in this function should have the same
+ // number of clone entries, but handle any discrepancies gracefully below
+ // for NDEBUG builds.
+ assert(
+ SN.Clones.size() > I &&
+ "Callsite summary has fewer entries than other summaries in function");
+ if (SN.Clones.size() <= I || !SN.Clones[I])
+ continue;
+ uint8_t Data[sizeof(SN.Clones[I])];
+ support::endian::write32le(Data, SN.Clones[I]);
+ Hasher.update(Data);
+ }
+ // Update hash with any allocs that have non-default (non-None) hints.
+ for (auto &AN : FS->allocs()) {
+ // In theory all callsites and allocs in this function should have the same
+ // number of clone entries, but handle any discrepancies gracefully below
+ // for NDEBUG builds.
+ assert(AN.Versions.size() > I &&
+ "Alloc summary has fewer entries than other summaries in function");
+ if (AN.Versions.size() <= I ||
+ (AllocationType)AN.Versions[I] == AllocationType::None)
+ continue;
+ Hasher.update(ArrayRef<uint8_t>(&AN.Versions[I], 1));
+ }
+ return support::endian::read64le(Hasher.result().data());
+}
+
static SmallVector<std::unique_ptr<ValueToValueMapTy>, 4> createFunctionClones(
Function &F, unsigned NumClones, Module &M, OptimizationRemarkEmitter &ORE,
std::map<const Function *, SmallPtrSet<const GlobalAlias *, 1>>
- &FuncToAliasMap) {
+ &FuncToAliasMap,
+ FunctionSummary *FS) {
+ auto TakeDeclNameAndReplace = [](GlobalValue *DeclGV, GlobalValue *NewGV) {
+ // We might have created this when adjusting callsite in another
+ // function. It should be a declaration.
+ assert(DeclGV->isDeclaration());
+ NewGV->takeName(DeclGV);
+ DeclGV->replaceAllUsesWith(NewGV);
+ DeclGV->eraseFromParent();
+ };
+
+ // Handle aliases to this function, and create analogous alias clones to the
+ // provided clone of this function.
+ auto CloneFuncAliases = [&](Function *NewF, unsigned I) {
+ if (!FuncToAliasMap.count(&F))
+ return;
+ for (auto *A : FuncToAliasMap[&F]) {
+ std::string AliasName = getMemProfFuncName(A->getName(), I);
+ auto *PrevA = M.getNamedAlias(AliasName);
+ auto *NewA = GlobalAlias::create(A->getValueType(),
+ A->getType()->getPointerAddressSpace(),
+ A->getLinkage(), AliasName, NewF);
+ NewA->copyAttributesFrom(A);
+ if (PrevA)
+ TakeDeclNameAndReplace(PrevA, NewA);
+ }
+ };
+
// The first "clone" is the original copy, we should only call this if we
// needed to create new clones.
assert(NumClones > 1);
SmallVector<std::unique_ptr<ValueToValueMapTy>, 4> VMaps;
VMaps.reserve(NumClones - 1);
FunctionsClonedThinBackend++;
+
+ // Map of hash of callsite/alloc versions to the instantiated function clone
+ // (possibly the original) implementing those calls. Used to avoid
+ // instantiating duplicate function clones.
+ // FIXME: Ideally the thin link would not generate such duplicate clones to
+ // start with, but right now it happens due to phase ordering in the function
+ // assignment and possible new clones that produces. We simply make each
+ // duplicate an alias to the matching instantiated clone recorded in the map
+ // (except for available_externally which are made declarations as they would
+ // be aliases in the prevailing module, and available_externally aliases are
+ // not well supported right now).
+ DenseMap<uint64_t, Function *> HashToFunc;
+
+ // Save the hash of the original function version.
+ HashToFunc[ComputeHash(FS, 0)] = &F;
+
for (unsigned I = 1; I < NumClones; I++) {
VMaps.emplace_back(std::make_unique<ValueToValueMapTy>());
+ std::string Name = getMemProfFuncName(F.getName(), I);
+ auto Hash = ComputeHash(FS, I);
+ // If this clone would duplicate a previously seen clone, don't generate the
+ // duplicate clone body, just make an alias to satisfy any (potentially
+ // cross-module) references.
+ if (HashToFunc.contains(Hash)) {
+ FunctionCloneDuplicatesThinBackend++;
+ auto *Func = HashToFunc[Hash];
+ if (Func->hasAvailableExternallyLinkage()) {
+ // Skip these as EliminateAvailableExternallyPass does not handle
+ // available_externally aliases correctly and we end up with an
+ // available_externally alias to a declaration. Just create a
+ // declaration for now as we know we will have a definition in another
+ // module.
+ auto Decl = M.getOrInsertFunction(Name, Func->getFunctionType());
+ ORE.emit(OptimizationRemark(DEBUG_TYPE, "MemprofClone", &F)
+ << "created clone decl " << ore::NV("Decl", Decl.getCallee()));
+ continue;
+ }
+ auto *PrevF = M.getFunction(Name);
+ auto *Alias = GlobalAlias::create(Name, Func);
+ if (PrevF)
+ TakeDeclNameAndReplace(PrevF, Alias);
+ ORE.emit(OptimizationRemark(DEBUG_TYPE, "MemprofClone", &F)
+ << "created clone alias " << ore::NV("Alias", Alias));
+
+ // Now handle aliases to this function, and clone those as well.
+ CloneFuncAliases(Func, I);
+ continue;
+ }
auto *NewF = CloneFunction(&F, *VMaps.back());
+ HashToFunc[Hash] = NewF;
FunctionClonesThinBackend++;
// Strip memprof and callsite metadata from clone as they are no longer
// needed.
@@ -5208,40 +5321,17 @@ static SmallVector<std::unique_ptr<ValueToValueMapTy>, 4> createFunctionClones(
Inst.setMetadata(LLVMContext::MD_callsite, nullptr);
}
}
- std::string Name = getMemProfFuncName(F.getName(), I);
auto *PrevF = M.getFunction(Name);
- if (PrevF) {
- // We might have created this when adjusting callsite in another
- // function. It should be a declaration.
- assert(PrevF->isDeclaration());
- NewF->takeName(PrevF);
- PrevF->replaceAllUsesWith(NewF);
- PrevF->eraseFromParent();
- } else
+ if (PrevF)
+ TakeDeclNameAndReplace(PrevF, NewF);
+ else
NewF->setName(Name);
updateSubprogramLinkageName(NewF, Name);
ORE.emit(OptimizationRemark(DEBUG_TYPE, "MemprofClone", &F)
<< "created clone " << ore::NV("NewFunction", NewF));
// Now handle aliases to this function, and clone those as well.
- if (!FuncToAliasMap.count(&F))
- continue;
- for (auto *A : FuncToAliasMap[&F]) {
- std::string Name = getMemProfFuncName(A->getName(), I);
- auto *PrevA = M.getNamedAlias(Name);
- auto *NewA = GlobalAlias::create(A->getValueType(),
- A->getType()->getPointerAddressSpace(),
- A->getLinkage(), Name, NewF);
- NewA->copyAttributesFrom(A);
- if (PrevA) {
- // We might have created this when adjusting callsite in another
- // function. It should be a declaration.
- assert(PrevA->isDeclaration());
- NewA->takeName(PrevA);
- PrevA->replaceAllUsesWith(NewA);
- PrevA->eraseFromParent();
- }
- }
+ CloneFuncAliases(NewF, I);
}
return VMaps;
}
@@ -5401,7 +5491,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
SmallVector<std::unique_ptr<ValueToValueMapTy>, 4> VMaps;
bool ClonesCreated = false;
unsigned NumClonesCreated = 0;
- auto CloneFuncIfNeeded = [&](unsigned NumClones) {
+ auto CloneFuncIfNeeded = [&](unsigned NumClones, FunctionSummary *FS) {
// We should at least have version 0 which is the original copy.
assert(NumClones > 0);
// If only one copy needed use original.
@@ -5415,7 +5505,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
assert(NumClonesCreated == NumClones);
return;
}
- VMaps = createFunctionClones(F, NumClones, M, ORE, FuncToAliasMap);
+ VMaps = createFunctionClones(F, NumClones, M, ORE, FuncToAliasMap, FS);
// The first "clone" is the original copy, which doesn't have a VMap.
assert(VMaps.size() == NumClones - 1);
Changed = true;
@@ -5424,9 +5514,9 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
};
auto CloneCallsite = [&](const CallsiteInfo &StackNode, CallBase *CB,
- Function *CalledFunction) {
+ Function *CalledFunction, FunctionSummary *FS) {
// Perform cloning if not yet done.
- CloneFuncIfNeeded(/*NumClones=*/StackNode.Clones.size());
+ CloneFuncIfNeeded(/*NumClones=*/StackNode.Clones.size(), FS);
assert(!isMemProfClone(*CalledFunction));
@@ -5448,6 +5538,10 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
// below.
auto CalleeOrigName = CalledFunction->getName();
for (unsigned J = 0; J < StackNode.Clones.size(); J++) {
+ // If the VMap is empty, this clone was a duplicate of another and was
+ // created as an alias or a declaration.
+ if (J > 0 && VMaps[J - 1]->empty())
+ continue;
// Do nothing if this version calls the original version of its
// callee.
if (!StackNode.Clones[J])
@@ -5591,7 +5685,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
#endif
// Perform cloning if not yet done.
- CloneFuncIfNeeded(/*NumClones=*/AllocNode.Versions.size());
+ CloneFuncIfNeeded(/*NumClones=*/AllocNode.Versions.size(), FS);
OrigAllocsThinBackend++;
AllocVersionsThinBackend += AllocNode.Versions.size();
@@ -5624,6 +5718,10 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
// Update the allocation types per the summary info.
for (unsigned J = 0; J < AllocNode.Versions.size(); J++) {
+ // If the VMap is empty, this clone was a duplicate of another and
+ // was created as an alias or a declaration.
+ if (J > 0 && VMaps[J - 1]->empty())
+ continue;
// Ignore any that didn't get an assigned allocation type.
if (AllocNode.Versions[J] == (uint8_t)AllocationType::None)
continue;
@@ -5670,7 +5768,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
// we don't need to do ICP, but might need to clone this
// function as it is the target of other cloned calls.
if (NumClones)
- CloneFuncIfNeeded(NumClones);
+ CloneFuncIfNeeded(NumClones, FS);
}
else {
@@ -5690,7 +5788,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
}
#endif
- CloneCallsite(StackNode, CB, CalledFunction);
+ CloneCallsite(StackNode, CB, CalledFunction, FS);
}
} else if (CB->isTailCall() && CalledFunction) {
// Locate the synthesized callsite info for the callee VI, if any was
@@ -5700,7 +5798,7 @@ bool MemProfContextDisambiguation::applyImport(Module &M) {
if (CalleeVI && MapTailCallCalleeVIToCallsite.count(CalleeVI)) {
auto Callsite = MapTailCallCalleeVIToCallsite.find(CalleeVI);
assert(Callsite != MapTailCallCalleeVIToCallsite.end());
- CloneCallsite(Callsite->second, CB, CalledFunction);
+ CloneCallsite(Callsite->second, CB, CalledFunction, FS);
}
}
}
@@ -5846,6 +5944,10 @@ void MemProfContextDisambiguation::performICP(
// check.
CallBase *CBClone = CB;
for (unsigned J = 0; J < NumClones; J++) {
+ // If the VMap is empty, this clone was a duplicate of another and was
+ // created as an alias or a declaration.
+ if (J > 0 && VMaps[J - 1]->empty())
+ continue;
// Copy 0 is the original function.
if (J > 0)
CBClone = cast<CallBase>((*VMaps[J - 1])[CB]);
@@ -5891,6 +5993,10 @@ void MemProfContextDisambiguation::performICP(
// TotalCount and the number promoted.
CallBase *CBClone = CB;
for (unsigned J = 0; J < NumClones; J++) {
+ // If the VMap is empty, this clone was a duplicate of another and was
+ // created as an alias or a declaration.
+ if (J > 0 && VMaps[J - 1]->empty())
+ continue;
// Copy 0 is the original function.
if (J > 0)
CBClone = cast<CallBase>((*VMaps[J - 1])[CB]);
diff --git a/llvm/test/CodeGen/AMDGPU/.#llvm.amdgcn.smfmac.gfx950.ll b/llvm/test/CodeGen/AMDGPU/.#llvm.amdgcn.smfmac.gfx950.ll
new file mode 120000
index 0000000..8747bd5
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/.#llvm.amdgcn.smfmac.gfx950.ll
@@ -0,0 +1 @@
+matt@mattbookAMD.56897 \ No newline at end of file
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll b/llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll
index 2cde060..a54dc9d 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/irtranslator-inline-asm.ll
@@ -66,7 +66,7 @@ define amdgpu_kernel void @asm_simple_agpr_clobber() {
define i32 @asm_vgpr_early_clobber() {
; CHECK-LABEL: name: asm_vgpr_early_clobber
; CHECK: bb.1 (%ir-block.0):
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %8, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %9, !1
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 7; v_mov_b32 $1, 7", 1 /* sideeffect attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %8, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %9, !1
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY %9
; CHECK-NEXT: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[COPY]], [[COPY1]]
@@ -94,7 +94,7 @@ entry:
define i32 @test_single_vgpr_output() nounwind {
; CHECK-LABEL: name: test_single_vgpr_output
; CHECK: bb.1.entry:
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %8
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 7", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %8
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: $vgpr0 = COPY [[COPY]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -106,7 +106,7 @@ entry:
define i32 @test_single_sgpr_output_s32() nounwind {
; CHECK-LABEL: name: test_single_sgpr_output_s32
; CHECK: bb.1.entry:
- ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
+ ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2424842 /* regdef:SReg_32 */, def %8
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: $vgpr0 = COPY [[COPY]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -119,7 +119,7 @@ entry:
define float @test_multiple_register_outputs_same() #0 {
; CHECK-LABEL: name: test_multiple_register_outputs_same
; CHECK: bb.1 (%ir-block.0):
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %8, 2031626 /* regdef:VGPR_32 */, def %9
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_mov_b32 $1, 1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %8, 1835018 /* regdef:VGPR_32 */, def %9
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY %9
; CHECK-NEXT: [[FADD:%[0-9]+]]:_(s32) = G_FADD [[COPY]], [[COPY1]]
@@ -136,7 +136,7 @@ define float @test_multiple_register_outputs_same() #0 {
define double @test_multiple_register_outputs_mixed() #0 {
; CHECK-LABEL: name: test_multiple_register_outputs_mixed
; CHECK: bb.1 (%ir-block.0):
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %8, 3735562 /* regdef:VReg_64 */, def %9
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0; v_add_f64 $1, 0, 0", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %8, 3473418 /* regdef:VReg_64 */, def %9
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s64) = COPY %9
; CHECK-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
@@ -171,7 +171,7 @@ define amdgpu_kernel void @test_input_vgpr_imm() {
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p4) = COPY $sgpr8_sgpr9
; CHECK-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[C]](s32)
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 v0, $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 v0, $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY1]]
; CHECK-NEXT: S_ENDPGM 0
call void asm sideeffect "v_mov_b32 v0, $0", "v"(i32 42)
ret void
@@ -185,7 +185,7 @@ define amdgpu_kernel void @test_input_sgpr_imm() {
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p4) = COPY $sgpr8_sgpr9
; CHECK-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 42
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY [[C]](s32)
- ; CHECK-NEXT: INLINEASM &"s_mov_b32 s0, $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"s_mov_b32 s0, $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[COPY1]]
; CHECK-NEXT: S_ENDPGM 0
call void asm sideeffect "s_mov_b32 s0, $0", "s"(i32 42)
ret void
@@ -212,7 +212,7 @@ define float @test_input_vgpr(i32 %src) nounwind {
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
- ; CHECK-NEXT: INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %9, 2031625 /* reguse:VGPR_32 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"v_add_f32 $0, 1.0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %9, 1835017 /* reguse:VGPR_32 */, [[COPY1]]
; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY %9
; CHECK-NEXT: $vgpr0 = COPY [[COPY2]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -227,7 +227,7 @@ define i32 @test_memory_constraint(ptr addrspace(3) %a) nounwind {
; CHECK-NEXT: liveins: $vgpr0
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(p3) = COPY $vgpr0
- ; CHECK-NEXT: INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 2031626 /* regdef:VGPR_32 */, def %9, 262158 /* mem:m */, [[COPY]](p3)
+ ; CHECK-NEXT: INLINEASM &"ds_read_b32 $0, $1", 8 /* mayload attdialect */, 1835018 /* regdef:VGPR_32 */, def %9, 262158 /* mem:m */, [[COPY]](p3)
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY %9
; CHECK-NEXT: $vgpr0 = COPY [[COPY1]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -244,7 +244,7 @@ define i32 @test_vgpr_matching_constraint(i32 %a) nounwind {
; CHECK-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 1
; CHECK-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY]], [[C]]
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[AND]](s32)
- ; CHECK-NEXT: INLINEASM &";", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %11, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+ ; CHECK-NEXT: INLINEASM &";", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %11, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY %11
; CHECK-NEXT: $vgpr0 = COPY [[COPY2]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -256,13 +256,13 @@ define i32 @test_vgpr_matching_constraint(i32 %a) nounwind {
define i32 @test_sgpr_matching_constraint() nounwind {
; CHECK-LABEL: name: test_sgpr_matching_constraint
; CHECK: bb.1.entry:
- ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
+ ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2424842 /* regdef:SReg_32 */, def %8
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
- ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %10
+ ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 8", 0 /* attdialect */, 2424842 /* regdef:SReg_32 */, def %10
; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY %10
; CHECK-NEXT: [[COPY2:%[0-9]+]]:sreg_32 = COPY [[COPY]](s32)
; CHECK-NEXT: [[COPY3:%[0-9]+]]:sreg_32 = COPY [[COPY1]](s32)
- ; CHECK-NEXT: INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %12, 2621449 /* reguse:SReg_32 */, [[COPY2]], 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3)
+ ; CHECK-NEXT: INLINEASM &"s_add_u32 $0, $1, $2", 0 /* attdialect */, 2424842 /* regdef:SReg_32 */, def %12, 2424841 /* reguse:SReg_32 */, [[COPY2]], 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3)
; CHECK-NEXT: [[COPY4:%[0-9]+]]:_(s32) = COPY %12
; CHECK-NEXT: $vgpr0 = COPY [[COPY4]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
@@ -285,7 +285,7 @@ define void @test_many_matching_constraints(i32 %a, i32 %b, i32 %c) nounwind {
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vgpr_32 = COPY [[COPY2]](s32)
; CHECK-NEXT: [[COPY4:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
; CHECK-NEXT: [[COPY5:%[0-9]+]]:vgpr_32 = COPY [[COPY1]](s32)
- ; CHECK-NEXT: INLINEASM &"; ", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %11, 2031626 /* regdef:VGPR_32 */, def %12, 2031626 /* regdef:VGPR_32 */, def %13, 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY4]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY5]](tied-def 5)
+ ; CHECK-NEXT: INLINEASM &"; ", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %11, 1835018 /* regdef:VGPR_32 */, def %12, 1835018 /* regdef:VGPR_32 */, def %13, 2147483657 /* reguse tiedto:$0 */, [[COPY3]](tied-def 3), 2147614729 /* reguse tiedto:$2 */, [[COPY4]](tied-def 7), 2147549193 /* reguse tiedto:$1 */, [[COPY5]](tied-def 5)
; CHECK-NEXT: [[COPY6:%[0-9]+]]:_(s32) = COPY %11
; CHECK-NEXT: [[COPY7:%[0-9]+]]:_(s32) = COPY %12
; CHECK-NEXT: [[COPY8:%[0-9]+]]:_(s32) = COPY %13
@@ -306,10 +306,10 @@ define void @test_many_matching_constraints(i32 %a, i32 %b, i32 %c) nounwind {
define i32 @test_sgpr_to_vgpr_move_matching_constraint() nounwind {
; CHECK-LABEL: name: test_sgpr_to_vgpr_move_matching_constraint
; CHECK: bb.1.entry:
- ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2621450 /* regdef:SReg_32 */, def %8
+ ; CHECK-NEXT: INLINEASM &"s_mov_b32 $0, 7", 0 /* attdialect */, 2424842 /* regdef:SReg_32 */, def %8
; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY %8
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]](s32)
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %10, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %10, 2147483657 /* reguse tiedto:$0 */, [[COPY1]](tied-def 3)
; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY %10
; CHECK-NEXT: $vgpr0 = COPY [[COPY2]](s32)
; CHECK-NEXT: SI_RETURN implicit $vgpr0
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-ignore-copies-crash.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-ignore-copies-crash.mir
index f2d3272..137488f 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-ignore-copies-crash.mir
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/regbankcombiner-ignore-copies-crash.mir
@@ -24,7 +24,7 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr(s32) = COPY [[C]](s32)
; CHECK-NEXT: [[FMUL:%[0-9]+]]:vgpr(s32) = G_FMUL [[COPY]], [[COPY1]]
; CHECK-NEXT: [[C1:%[0-9]+]]:sgpr(s32) = G_FCONSTANT float 1.000000e+00
- ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %5(s32)
+ ; CHECK-NEXT: INLINEASM &"v_mov_b32 $0, 0", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %5(s32)
; CHECK-NEXT: [[COPY2:%[0-9]+]]:vgpr(s32) = COPY [[C1]](s32)
; CHECK-NEXT: [[AMDGPU_FMED3_:%[0-9]+]]:vgpr(s32) = nnan G_AMDGPU_FMED3 [[FMUL]], %5, [[COPY2]]
; CHECK-NEXT: $vgpr0 = COPY [[AMDGPU_FMED3_]](s32)
@@ -33,7 +33,7 @@ body: |
%2:vgpr(s32) = COPY %1(s32)
%3:vgpr(s32) = G_FMUL %0, %2
%4:sgpr(s32) = G_FCONSTANT float 1.000000e+00
- INLINEASM &"v_mov_b32 $0, 0", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %5:vgpr_32
+ INLINEASM &"v_mov_b32 $0, 0", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %5:vgpr_32
%6:vgpr(s32) = COPY %4(s32)
%7:vgpr(s32) = nnan G_AMDGPU_FMED3 %3(s32), %5(s32), %6(s32)
$vgpr0 = COPY %7(s32)
diff --git a/llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir b/llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir
index 035354c..8a39d9c 100644
--- a/llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir
+++ b/llvm/test/CodeGen/AMDGPU/branch-relax-indirect-branch.mir
@@ -68,7 +68,7 @@ body: |
; CHECK-NEXT: successors: %bb.3(0x04000000), %bb.7(0x7c000000)
; CHECK-NEXT: liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2621450 /* regdef:SReg_32 */, def renamable $sgpr4
+ ; CHECK-NEXT: INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2424842 /* regdef:SReg_32 */, def renamable $sgpr4
; CHECK-NEXT: S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
; CHECK-NEXT: S_CBRANCH_SCC0 %bb.3, implicit killed $scc
; CHECK-NEXT: {{ $}}
@@ -149,7 +149,7 @@ body: |
successors: %bb.3(0x04000000), %bb.2(0x7c000000)
liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
- INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2621450 /* regdef:SReg_32 */, def renamable $sgpr4
+ INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2424842 /* regdef:SReg_32 */, def renamable $sgpr4
S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
S_CBRANCH_SCC1 %bb.2, implicit killed $scc
diff --git a/llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir b/llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir
index 584d6a5..a2f0205 100644
--- a/llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir
+++ b/llvm/test/CodeGen/AMDGPU/branch-relax-no-terminators.mir
@@ -69,7 +69,7 @@ body: |
; CHECK-NEXT: successors: %bb.3(0x04000000), %bb.7(0x7c000000)
; CHECK-NEXT: liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2621450 /* regdef:SReg_32 */, def renamable $sgpr4
+ ; CHECK-NEXT: INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2424842 /* regdef:SReg_32 */, def renamable $sgpr4
; CHECK-NEXT: S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
; CHECK-NEXT: S_CBRANCH_SCC0 %bb.3, implicit killed $scc
; CHECK-NEXT: {{ $}}
@@ -151,7 +151,7 @@ body: |
successors: %bb.3(0x04000000), %bb.2(0x7c000000)
liveins: $vcc_hi, $vcc_lo, $sgpr5, $sgpr6, $sgpr7, $sgpr8, $sgpr9, $sgpr10, $sgpr11, $sgpr12, $sgpr13, $sgpr14, $sgpr15, $sgpr16, $sgpr17, $sgpr18, $sgpr19, $sgpr20, $sgpr21, $sgpr22, $sgpr23, $sgpr24, $sgpr25, $sgpr26, $sgpr27, $sgpr28, $sgpr29, $sgpr30, $sgpr31, $sgpr34, $sgpr35, $sgpr36, $sgpr37, $sgpr38, $sgpr39, $sgpr40, $sgpr41, $sgpr42, $sgpr43, $sgpr44, $sgpr45, $sgpr46, $sgpr47, $sgpr48, $sgpr49, $sgpr50, $sgpr51, $sgpr52, $sgpr53, $sgpr54, $sgpr55, $sgpr56, $sgpr57, $sgpr58, $sgpr59, $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr68, $sgpr69, $sgpr70, $sgpr71, $sgpr72, $sgpr73, $sgpr74, $sgpr75, $sgpr76, $sgpr77, $sgpr78, $sgpr79, $sgpr80, $sgpr81, $sgpr82, $sgpr83, $sgpr84, $sgpr85, $sgpr86, $sgpr87, $sgpr88, $sgpr89, $sgpr90, $sgpr91, $sgpr92, $sgpr93, $sgpr94, $sgpr95, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $sgpr100, $sgpr101, $vgpr0, $vgpr1
- INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2621450 /* regdef:SReg_32 */, def renamable $sgpr4
+ INLINEASM &"v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64\0A v_nop_e64", 1 /* sideeffect attdialect */, 2424842 /* regdef:SReg_32 */, def renamable $sgpr4
S_CMP_LG_U32 killed renamable $sgpr4, 0, implicit-def $scc
S_CBRANCH_SCC1 %bb.2, implicit killed $scc
diff --git a/llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir b/llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir
index c475efb..52cdf5a 100644
--- a/llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir
+++ b/llvm/test/CodeGen/AMDGPU/coalesce-copy-to-agpr-to-av-registers.mir
@@ -20,13 +20,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_64 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1:areg_64 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
undef %2.sub0:areg_64 = COPY %0
%2.sub1:areg_64 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -45,13 +45,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
undef %2.sub0:areg_64_align2 = COPY %0
%2.sub1:areg_64_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -72,7 +72,7 @@ body: |
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0:areg_96 = COPY [[COPY]]
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub1:areg_96 = COPY [[COPY1]]
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2:areg_96 = COPY [[COPY2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY3]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
@@ -80,7 +80,7 @@ body: |
undef %3.sub0:areg_96 = COPY %0
%3.sub1:areg_96 = COPY %1
%3.sub2:areg_96 = COPY %2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %3
SI_RETURN
...
@@ -101,7 +101,7 @@ body: |
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub1:areg_96_align2 = COPY [[COPY1]]
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY3]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
@@ -109,7 +109,7 @@ body: |
undef %3.sub0:areg_96_align2 = COPY %0
%3.sub1:areg_96_align2 = COPY %1
%3.sub2:areg_96_align2 = COPY %2
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %3
SI_RETURN
...
@@ -128,13 +128,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0_sub1:areg_128 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub2_sub3:areg_128 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64 = COPY $vgpr0_vgpr1
%1:vreg_64 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128 = COPY %0
%2.sub2_sub3:areg_128 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %2
SI_RETURN
...
@@ -153,13 +153,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub2_sub3:areg_128_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64 = COPY $vgpr0_vgpr1
%1:vreg_64 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128_align2 = COPY %0
%2.sub2_sub3:areg_128_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %2
SI_RETURN
...
@@ -178,13 +178,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr9
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:sgpr_32 = COPY $sgpr8
%1:sgpr_32 = COPY $sgpr9
undef %2.sub0:areg_64_align2 = COPY %0
%2.sub1:areg_64_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -203,13 +203,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_96 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1_sub2:areg_96 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vreg_64 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96 = COPY %0
%2.sub1_sub2:areg_96 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -228,13 +228,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vreg_64 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1_sub2:areg_96_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vreg_64 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96_align2 = COPY %0
%2.sub1_sub2:areg_96_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -253,13 +253,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0_sub1:areg_96 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub2:areg_96 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64 = COPY $vgpr0_vgpr1
%1:vgpr_32 = COPY $vgpr2
undef %2.sub0_sub1:areg_96 = COPY %0
%2.sub2:areg_96 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -278,13 +278,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0_sub1:areg_96_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64 = COPY $vgpr0_vgpr1
%1:vgpr_32 = COPY $vgpr2
undef %2.sub0_sub1:areg_96_align2 = COPY %0
%2.sub2:areg_96_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -302,12 +302,12 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %2.sub0:areg_64 = COPY %0
%2.sub1:areg_64 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -326,13 +326,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
undef %2.sub0:areg_64_align2 = COPY %0
%2.sub1:areg_64_align2 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -350,12 +350,12 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_96 = COPY %0
%1.sub1:areg_96 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %1
SI_RETURN
...
@@ -373,12 +373,12 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_96_align2 = COPY %0
%1.sub1:areg_96_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %1
SI_RETURN
...
@@ -398,14 +398,14 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_128 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_128 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub3:areg_128 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_128 = COPY %0
%1.sub1:areg_128 = COPY %0
%1.sub2:areg_128 = COPY %0
%1.sub3:areg_128 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %1
SI_RETURN
...
@@ -425,14 +425,14 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_128_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_128_align2 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub3:areg_128_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_128_align2 = COPY %0
%1.sub1:areg_128_align2 = COPY %0
%1.sub2:areg_128_align2 = COPY %0
%1.sub3:areg_128_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %1
SI_RETURN
...
@@ -451,15 +451,15 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_64 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1:areg_64 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
%1:vgpr_32 = COPY $vgpr1
undef %2.sub0:areg_64 = COPY %0
%2.sub1:areg_64 = COPY %1
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
- INLINEASM &"; use $0", 0 /* attdialect */, 2031625 /* reguse:VGPR_32 */, killed %0
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 1835017 /* reguse:VGPR_32 */, killed %0
SI_RETURN
...
@@ -477,14 +477,14 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]]
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_64 = COPY %0
%1.sub1:areg_64 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %1
- INLINEASM &"; use $0", 0 /* attdialect */, 2031625 /* reguse:VGPR_32 */, killed %0
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 1835017 /* reguse:VGPR_32 */, killed %0
SI_RETURN
...
@@ -503,16 +503,16 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_64 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 3735561 /* reguse:VReg_64 */, [[COPY]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 3473417 /* reguse:VReg_64 */, [[COPY]]
; CHECK-NEXT: SI_RETURN
%0:vgpr_32 = COPY $vgpr0
undef %1.sub0:areg_64 = COPY %0
%1.sub1:areg_64 = COPY %0
undef %2.sub0:vreg_64 = COPY %0
%2.sub1:vreg_64 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %1
- INLINEASM &"; use $0", 0 /* attdialect */, 3735561 /* reguse:VReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 3473417 /* reguse:VReg_64 */, killed %2
SI_RETURN
...
@@ -533,13 +533,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_64 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
%0.sub1:vreg_64 = COPY $vgpr1
undef %2.sub0:areg_64 = COPY %0.sub0
%2.sub1:areg_64 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -558,13 +558,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_64 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
%0.sub1:vreg_64 = COPY $vgpr1
undef %2.sub0:areg_64_align2 = COPY %0.sub0
%2.sub1:areg_64_align2 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -585,7 +585,7 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96 = COPY [[COPY]].sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 =COPY $vgpr0
%0.sub1:vreg_96 = COPY $vgpr1
@@ -593,7 +593,7 @@ body: |
undef %3.sub0:areg_96 = COPY %0.sub0
%3.sub1:areg_96 = COPY %0.sub1
%3.sub2:areg_96 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %3
SI_RETURN
...
@@ -614,7 +614,7 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96_align2 = COPY [[COPY]].sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 =COPY $vgpr0
%0.sub1:vreg_96 = COPY $vgpr1
@@ -622,7 +622,7 @@ body: |
undef %3.sub0:areg_96_align2 = COPY %0.sub0
%3.sub1:areg_96_align2 = COPY %0.sub1
%3.sub2:areg_96_align2 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %3
SI_RETURN
...
@@ -641,13 +641,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2_sub3:vreg_128 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_128 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2_sub3:areg_128 = COPY [[COPY]].sub2_sub3
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_128 =COPY $vgpr0_vgpr1
%0.sub2_sub3:vreg_128 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128 = COPY %0.sub0_sub1
%2.sub2_sub3:areg_128 = COPY %0.sub2_sub3
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %2
SI_RETURN
...
@@ -668,13 +668,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_128 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2_sub3:areg_128_align2 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_128 =COPY $vgpr0_vgpr1
%0.sub1:vreg_128 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128_align2 = COPY %0.sub0
%2.sub2_sub3:areg_128_align2 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %2
SI_RETURN
...
@@ -693,13 +693,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:sreg_64 = COPY $sgpr9
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:sreg_64 = COPY $sgpr8
%0.sub1:sreg_64 = COPY $sgpr9
undef %2.sub0:areg_64_align2 = COPY %0.sub0
%2.sub1:areg_64_align2 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -718,13 +718,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1_sub2:areg_96 = COPY [[COPY]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 =COPY $vgpr0
%0.sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96 = COPY %0.sub0
%2.sub1_sub2:areg_96 = COPY %0.sub1_sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -743,13 +743,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1_sub2:areg_96_align2 = COPY [[COPY]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 =COPY $vgpr0
%0.sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96_align2 = COPY %0.sub0
%2.sub1_sub2:areg_96_align2 = COPY %0.sub1_sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -768,13 +768,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2:vreg_96 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_96 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_96 = COPY $vgpr0_vgpr1
%0.sub2:vreg_96 = COPY $vgpr2
undef %2.sub0_sub1:areg_96 = COPY %0.sub0_sub1
%2.sub2:areg_96 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -793,13 +793,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2:vreg_96 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_96_align2 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_96 = COPY $vgpr0_vgpr1
%0.sub2:vreg_96 = COPY $vgpr2
undef %2.sub0_sub1:areg_96_align2 = COPY %0.sub0_sub1
%2.sub2:areg_96_align2 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -817,12 +817,12 @@ body: |
; CHECK-NEXT: undef [[COPY:%[0-9]+]].sub0:vreg_64 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
undef %2.sub0:areg_64 = COPY %0.sub0
%2.sub1:areg_64 = COPY %0.sub0
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -841,13 +841,13 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
undef %1.sub0:areg_96 = COPY %0.sub0
%1.sub1:areg_96 = COPY %0.sub0
%1.sub2:areg_96 = COPY %0.sub0
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %1
SI_RETURN
...
@@ -865,12 +865,12 @@ body: |
; CHECK-NEXT: undef [[COPY:%[0-9]+]].sub0:vreg_64 = COPY $vgpr0
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96_align2 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
undef %1.sub0:areg_96_align2 = COPY %0.sub0
%1.sub1:areg_96_align2 = COPY %0.sub0
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %1
SI_RETURN
...
@@ -890,14 +890,14 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_128 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_128 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub3:areg_128 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
undef %1.sub0:areg_128 = COPY %0.sub0
%1.sub1:areg_128 = COPY %0.sub0
%1.sub2:areg_128 = COPY %0.sub0
%1.sub3:areg_128 = COPY %0.sub0
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %1
SI_RETURN
...
@@ -917,14 +917,14 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_128_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_128_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub3:areg_128_align2 = COPY [[COPY]].sub0
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
undef %1.sub0:areg_128_align2 = COPY %0.sub0
%1.sub1:areg_128_align2 = COPY %0.sub0
%1.sub2:areg_128_align2 = COPY %0.sub0
%1.sub3:areg_128_align2 = COPY %0.sub0
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %1
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %1
SI_RETURN
...
@@ -943,13 +943,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_64 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64 = COPY $vgpr0
%0.sub1:vreg_64 = COPY $vgpr1
undef %2.sub0:areg_64 = COPY %0.sub0
%2.sub1:areg_64 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -968,13 +968,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:vreg_64_align2 = COPY $vgpr1
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_64_align2 = COPY $vgpr0
%0.sub1:vreg_64_align2 = COPY $vgpr1
undef %2.sub0:areg_64_align2 = COPY %0.sub0
%2.sub1:areg_64_align2 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -995,7 +995,7 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96 = COPY [[COPY]].sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 = COPY $vgpr0
%0.sub1:vreg_96 = COPY $vgpr1
@@ -1003,7 +1003,7 @@ body: |
undef %3.sub0:areg_96 = COPY %0.sub0
%3.sub1:areg_96 = COPY %0.sub1
%3.sub2:areg_96 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %3
SI_RETURN
...
@@ -1024,7 +1024,7 @@ body: |
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_96_align2 = COPY [[COPY]].sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96_align2 = COPY $vgpr0
%0.sub1:vreg_96_align2 = COPY $vgpr1
@@ -1032,7 +1032,7 @@ body: |
undef %3.sub0:areg_96_align2 = COPY %0.sub0
%3.sub1:areg_96_align2 = COPY %0.sub1
%3.sub2:areg_96_align2 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %3
SI_RETURN
...
@@ -1051,13 +1051,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2_sub3:vreg_128 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_128 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2_sub3:areg_128 = COPY [[COPY]].sub2_sub3
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_128 = COPY $vgpr0_vgpr1
%0.sub2_sub3:vreg_128 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128 = COPY %0.sub0_sub1
%2.sub2_sub3:areg_128 = COPY %0.sub2_sub3
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %2
SI_RETURN
...
@@ -1076,13 +1076,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2_sub3:vreg_128_align2 = COPY $vgpr2_vgpr3
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2_sub3:areg_128_align2 = COPY [[COPY]].sub2_sub3
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_128_align2 = COPY $vgpr0_vgpr1
%0.sub2_sub3:vreg_128_align2 = COPY $vgpr2_vgpr3
undef %2.sub0_sub1:areg_128_align2 = COPY %0.sub0_sub1
%2.sub2_sub3:areg_128_align2 = COPY %0.sub2_sub3
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %2
SI_RETURN
...
@@ -1101,13 +1101,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1:sreg_64 = COPY $sgpr9
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_64_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1:areg_64_align2 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:sreg_64 = COPY $sgpr8
%0.sub1:sreg_64 = COPY $sgpr9
undef %2.sub0:areg_64_align2 = COPY %0.sub0
%2.sub1:areg_64_align2 = COPY %0.sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1126,13 +1126,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1_sub2:areg_96 = COPY [[COPY]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 = COPY $vgpr0
%0.sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96 = COPY %0.sub0
%2.sub1_sub2:areg_96 = COPY %0.sub1_sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -1150,13 +1150,13 @@ body: |
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub0:areg_96 = COPY [[COPY]]
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub1_sub2:areg_96 = COPY [[COPY1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY2]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY2]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96 = COPY $vgpr0
%0.sub1_sub2:vreg_96 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96 = COPY %0.sub2
%2.sub1_sub2:areg_96 = COPY %0.sub0_sub1
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -1176,13 +1176,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub1_sub2:vreg_96_align2 = COPY $vgpr1_vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0:areg_96_align2 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub1_sub2:areg_96_align2 = COPY [[COPY]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0:vreg_96_align2 = COPY $vgpr0
%0.sub1_sub2:vreg_96_align2 = COPY $vgpr1_vgpr2
undef %2.sub0:areg_96_align2 = COPY %0.sub0
%2.sub1_sub2:areg_96_align2 = COPY %0.sub1_sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1201,13 +1201,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2:vreg_96 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_96 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_96 = COPY $vgpr0_vgpr1
%0.sub2:vreg_96 = COPY $vgpr2
undef %2.sub0_sub1:areg_96 = COPY %0.sub0_sub1
%2.sub2:areg_96 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %2
SI_RETURN
...
@@ -1226,13 +1226,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2:vreg_96_align2 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_96_align2 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_96_align2 = COPY $vgpr0_vgpr1
%0.sub2:vreg_96_align2 = COPY $vgpr2
undef %2.sub0_sub1:areg_96_align2 = COPY %0.sub0_sub1
%2.sub2:areg_96_align2 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1251,13 +1251,13 @@ body: |
; CHECK-NEXT: [[COPY:%[0-9]+]].sub2:vreg_96 = COPY $vgpr2
; CHECK-NEXT: undef [[COPY1:%[0-9]+]].sub0_sub1:areg_96_align2 = COPY [[COPY]].sub0_sub1
; CHECK-NEXT: [[COPY1:%[0-9]+]].sub2:areg_96_align2 = COPY [[COPY]].sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
undef %0.sub0_sub1:vreg_96 = COPY $vgpr0_vgpr1
%0.sub2:vreg_96 = COPY $vgpr2
undef %2.sub0_sub1:areg_96_align2 = COPY %0.sub0_sub1
%2.sub2:areg_96_align2 = COPY %0.sub2
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1274,11 +1274,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64 = COPY $vgpr0_vgpr1
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_64 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64 = COPY $vgpr0_vgpr1
%2:areg_64 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4325385 /* reguse:AReg_64 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4063241 /* reguse:AReg_64 */, killed %2
SI_RETURN
...
@@ -1295,11 +1295,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr0_vgpr1
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_64_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr0_vgpr1
%2:areg_64_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1316,11 +1316,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_96 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
%3:areg_96 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 6422537 /* reguse:AReg_96 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6160393 /* reguse:AReg_96 */, %3
SI_RETURN
...
@@ -1337,11 +1337,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_96_align2 = COPY $vgpr0_vgpr1_vgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_96_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_96_align2 = COPY $vgpr0_vgpr1_vgpr2
%3:areg_96_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 6750217 /* reguse:AReg_96_Align2 */, %3
+ INLINEASM &"; use $0", 0 /* attdialect */, 6488073 /* reguse:AReg_96_Align2 */, %3
SI_RETURN
...
@@ -1358,11 +1358,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_128 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_128 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
%2:areg_128 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 8847369 /* reguse:AReg_128 */, killed %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 8585225 /* reguse:AReg_128 */, killed %2
SI_RETURN
...
@@ -1379,11 +1379,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_128_align2 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_128_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_128_align2 = COPY $vgpr0_vgpr1_vgpr2_vgpr3
%2:areg_128_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %2
SI_RETURN
...
@@ -1400,11 +1400,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr8_sgpr9
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_64_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:sreg_64 = COPY $sgpr8_sgpr9
%2:areg_64_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
@@ -1421,11 +1421,11 @@ body: |
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
; CHECK-NEXT: [[COPY1:%[0-9]+]]:areg_96_align2 = COPY [[COPY]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY1]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY1]]
; CHECK-NEXT: SI_RETURN
%0:vreg_96 = COPY $vgpr0_vgpr1_vgpr2
%2:areg_96_align2 = COPY %0
- INLINEASM &"; use $0", 0 /* attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %2
+ INLINEASM &"; use $0", 0 /* attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %2
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir b/llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir
index c07c9ef..126cbc6 100644
--- a/llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir
+++ b/llvm/test/CodeGen/AMDGPU/coalescer-early-clobber-subreg.mir
@@ -20,10 +20,10 @@ body: |
; CHECK-LABEL: name: foo1
; CHECK: liveins: $vgpr0_vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %2.sub0, 2031627 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1
+ ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %2.sub0, 1835019 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1
; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
; CHECK-NEXT: S_ENDPGM 0
- INLINEASM &"", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %0:vgpr_32, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32
+ INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %0:vgpr_32, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32
undef %2.sub0:vreg_64 = COPY killed %0
%2.sub1:vreg_64 = COPY killed %1
FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
@@ -41,10 +41,10 @@ body: |
; CHECK-LABEL: name: foo2
; CHECK: liveins: $vgpr0_vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 2031627 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1, 2031626 /* regdef:VGPR_32 */, def undef %2.sub0
+ ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1, 1835018 /* regdef:VGPR_32 */, def undef %2.sub0
; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
; CHECK-NEXT: S_ENDPGM 0
- INLINEASM &"", 0 /* attdialect */, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32, 2031626 /* regdef:VGPR_32 */, def %0:vgpr_32
+ INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %1:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %0:vgpr_32
undef %2.sub0:vreg_64 = COPY killed %0
%2.sub1:vreg_64 = COPY killed %1
FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
@@ -62,10 +62,10 @@ body: |
; CHECK-LABEL: name: foo3
; CHECK: liveins: $vgpr0_vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %2.sub0, 2031627 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1
+ ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %2.sub0, 1835019 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1
; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
; CHECK-NEXT: S_ENDPGM 0
- INLINEASM &"", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %1:vgpr_32, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32
+ INLINEASM &"", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %1:vgpr_32, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32
undef %2.sub0:vreg_64 = COPY killed %1
%2.sub1:vreg_64 = COPY killed %0
FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
@@ -83,10 +83,10 @@ body: |
; CHECK-LABEL: name: foo4
; CHECK: liveins: $vgpr0_vgpr1
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 2031627 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1, 2031626 /* regdef:VGPR_32 */, def undef %2.sub0
+ ; CHECK-NEXT: INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def undef early-clobber %2.sub1, 1835018 /* regdef:VGPR_32 */, def undef %2.sub0
; CHECK-NEXT: FLAT_STORE_DWORDX2 $vgpr0_vgpr1, %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
; CHECK-NEXT: S_ENDPGM 0
- INLINEASM &"", 0 /* attdialect */, 2031627 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32, 2031626 /* regdef:VGPR_32 */, def %1:vgpr_32
+ INLINEASM &"", 0 /* attdialect */, 1835019 /* regdef-ec:VGPR_32 */, def early-clobber %0:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %1:vgpr_32
undef %2.sub0:vreg_64 = COPY killed %1
%2.sub1:vreg_64 = COPY killed %0
FLAT_STORE_DWORDX2 killed $vgpr0_vgpr1, killed %2, 0, 0, implicit $exec, implicit $flat_scr :: (store (s64))
diff --git a/llvm/test/CodeGen/AMDGPU/dst-sel-hazard.mir b/llvm/test/CodeGen/AMDGPU/dst-sel-hazard.mir
index 1635b09..4dfdb56 100644
--- a/llvm/test/CodeGen/AMDGPU/dst-sel-hazard.mir
+++ b/llvm/test/CodeGen/AMDGPU/dst-sel-hazard.mir
@@ -370,7 +370,7 @@ body: |
; HAZARD-LABEL: name: inline_sdwa_hazard
; HAZARD: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $exec, $mode
; HAZARD-NEXT: {{ $}}
- ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; HAZARD-NEXT: S_NOP 0
; HAZARD-NEXT: renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
; HAZARD-NEXT: S_ENDPGM 0
@@ -378,10 +378,10 @@ body: |
; NOHAZARD-LABEL: name: inline_sdwa_hazard
; NOHAZARD: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $exec, $mode
; NOHAZARD-NEXT: {{ $}}
- ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; NOHAZARD-NEXT: renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
; NOHAZARD-NEXT: S_ENDPGM 0
- INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
S_ENDPGM 0
...
@@ -397,17 +397,17 @@ body: |
; HAZARD-NEXT: {{ $}}
; HAZARD-NEXT: renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
; HAZARD-NEXT: S_NOP 0
- ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; HAZARD-NEXT: S_ENDPGM 0
;
; NOHAZARD-LABEL: name: sdwa_inline_hazard
; NOHAZARD: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $exec, $mode
; NOHAZARD-NEXT: {{ $}}
; NOHAZARD-NEXT: renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
- ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; NOHAZARD-NEXT: S_ENDPGM 0
renamable $vgpr0 = V_ADD_U16_sdwa 0, $vgpr1, 0, $vgpr2, 0, 1, 0, 3, 3, implicit $exec, implicit killed $vgpr0(tied-def 0)
- INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
S_ENDPGM 0
...
@@ -421,19 +421,19 @@ body: |
; HAZARD-LABEL: name: inline_inline_hazard
; HAZARD: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $exec, $mode
; HAZARD-NEXT: {{ $}}
- ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; HAZARD-NEXT: S_NOP 0
- ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; HAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; HAZARD-NEXT: S_ENDPGM 0
;
; NOHAZARD-LABEL: name: inline_inline_hazard
; NOHAZARD: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $exec, $mode
; NOHAZARD-NEXT: {{ $}}
- ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
- ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
+ ; NOHAZARD-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
; NOHAZARD-NEXT: S_ENDPGM 0
- INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
- INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0, 2031625 /* reguse:VGPR_32 */, $vgpr1
+ INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
+ INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0, 1835017 /* reguse:VGPR_32 */, $vgpr1
S_ENDPGM 0
...
diff --git a/llvm/test/CodeGen/AMDGPU/hazards-gfx950.mir b/llvm/test/CodeGen/AMDGPU/hazards-gfx950.mir
index 97f8fd6..6b9b77c 100644
--- a/llvm/test/CodeGen/AMDGPU/hazards-gfx950.mir
+++ b/llvm/test/CodeGen/AMDGPU/hazards-gfx950.mir
@@ -1112,11 +1112,11 @@ body: |
; GCN-NEXT: S_WAITCNT 0
; GCN-NEXT: renamable $vgpr2 = V_CVT_SCALEF32_PK_FP4_F16_e64 8, killed $vgpr0, 0, killed $vgpr1, 4, killed $vgpr2, 0, implicit $mode, implicit $exec
; GCN-NEXT: S_NOP 0
- ; GCN-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, killed renamable $vgpr2
+ ; GCN-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, killed renamable $vgpr2
; GCN-NEXT: S_SETPC_B64_return undef $sgpr30_sgpr31
S_WAITCNT 0
renamable $vgpr2 = V_CVT_SCALEF32_PK_FP4_F16_e64 8, killed $vgpr0, 0, killed $vgpr1, 4, killed $vgpr2, 0, implicit $mode, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, killed renamable $vgpr2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, killed renamable $vgpr2
S_SETPC_B64_return undef $sgpr30_sgpr31
...
diff --git a/llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir b/llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir
index 63db24a..2573acd 100644
--- a/llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir
+++ b/llvm/test/CodeGen/AMDGPU/inflate-reg-class-vgpr-mfma-to-av-with-load-source.mir
@@ -486,7 +486,7 @@ body: |
; CHECK-NEXT: S_NOP 0, implicit-def $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55
; CHECK-NEXT: S_NOP 0, implicit-def $vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63
; CHECK-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15
- ; CHECK-NEXT: INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
+ ; CHECK-NEXT: INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: S_ENDPGM 0
bb.0:
S_NOP 0, implicit-def $agpr0
@@ -516,7 +516,7 @@ body: |
S_NOP 0, implicit-def $vgpr40_vgpr41_vgpr42_vgpr43_vgpr44_vgpr45_vgpr46_vgpr47
S_NOP 0, implicit-def $vgpr48_vgpr49_vgpr50_vgpr51_vgpr52_vgpr53_vgpr54_vgpr55
S_NOP 0, implicit-def $vgpr56_vgpr57_vgpr58_vgpr59_vgpr60_vgpr61_vgpr62_vgpr63
- INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, %0:vreg_512_align2
+ INLINEASM &"; use $0 ", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, %0:vreg_512_align2
S_ENDPGM 0
...
@@ -1368,7 +1368,7 @@ body: |
; CHECK-NEXT: renamable $vgpr0_vgpr1 = GLOBAL_LOAD_DWORDX2 undef renamable $vgpr0_vgpr1, 0, 0, implicit $exec :: (load (s64), addrspace 1)
; CHECK-NEXT: early-clobber renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: early-clobber renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33, 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, killed renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15
; CHECK-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; CHECK-NEXT: S_BRANCH %bb.2
; CHECK-NEXT: {{ $}}
@@ -1408,7 +1408,7 @@ body: |
undef %2.sub0_sub1:vreg_512_align2 = GLOBAL_LOAD_DWORDX2 undef %3:vreg_64_align2, 0, 0, implicit $exec :: (load (s64), addrspace 1)
early-clobber %0:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %2, 0, 0, 0, implicit $mode, implicit $exec
early-clobber %4:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, %4
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, %4
S_CBRANCH_VCCNZ %bb.1, implicit $vcc
S_BRANCH %bb.2
@@ -1726,7 +1726,7 @@ body: |
; CHECK-NEXT: renamable $vgpr0_vgpr1 = GLOBAL_LOAD_DWORDX2 undef renamable $vgpr0_vgpr1, 0, 0, implicit $exec :: (load (s64), addrspace 1)
; CHECK-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15 = V_MFMA_F32_32X32X8F16_mac_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: early-clobber renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33 = V_MFMA_F32_32X32X8F16_vgprcd_e64 $vgpr16_vgpr17, $vgpr16_vgpr17, killed $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15, 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, renamable $vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31_vgpr32_vgpr33
; CHECK-NEXT: S_CBRANCH_VCCNZ %bb.1, implicit $vcc
; CHECK-NEXT: S_BRANCH %bb.2
; CHECK-NEXT: {{ $}}
@@ -1763,7 +1763,7 @@ body: |
undef %0.sub0_sub1:vreg_512_align2 = GLOBAL_LOAD_DWORDX2 undef %3:vreg_64_align2, 0, 0, implicit $exec :: (load (s64), addrspace 1)
%0:vreg_512_align2 = V_MFMA_F32_32X32X8F16_mac_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
%4:vreg_512_align2 = V_MFMA_F32_32X32X8F16_vgprcd_e64 %1, %1, %0, 0, 0, 0, implicit $mode, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 40239113 /* reguse:VReg_512_Align2 */, %4
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39976969 /* reguse:VReg_512_Align2 */, %4
S_CBRANCH_VCCNZ %bb.1, implicit $vcc
S_BRANCH %bb.2
diff --git a/llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll b/llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll
index 5b3e486..32b1b9b 100644
--- a/llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll
+++ b/llvm/test/CodeGen/AMDGPU/inline-asm.i128.ll
@@ -8,16 +8,16 @@
define amdgpu_kernel void @s_input_output_i128() {
; GFX908-LABEL: name: s_input_output_i128
; GFX908: bb.0 (%ir-block.0):
- ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 10682378 /* regdef:SGPR_128 */, def %13
+ ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 10420234 /* regdef:SGPR_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY %13
- ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 10682377 /* reguse:SGPR_128 */, [[COPY]]
+ ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 10420233 /* reguse:SGPR_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: s_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
- ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 10682378 /* regdef:SGPR_128 */, def %11
+ ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 10420234 /* regdef:SGPR_128 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:sgpr_128 = COPY %11
- ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 10682377 /* reguse:SGPR_128 */, [[COPY]]
+ ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 10420233 /* reguse:SGPR_128 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = tail call i128 asm sideeffect "; def $0", "=s"()
call void asm sideeffect "; use $0", "s"(i128 %val)
@@ -27,16 +27,16 @@ define amdgpu_kernel void @s_input_output_i128() {
define amdgpu_kernel void @v_input_output_i128() {
; GFX908-LABEL: name: v_input_output_i128
; GFX908: bb.0 (%ir-block.0):
- ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7929866 /* regdef:VReg_128 */, def %13
+ ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7667722 /* regdef:VReg_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:vreg_128 = COPY %13
- ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7929865 /* reguse:VReg_128 */, [[COPY]]
+ ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7667721 /* reguse:VReg_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: v_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
- ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8257546 /* regdef:VReg_128_Align2 */, def %11
+ ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:vreg_128_align2 = COPY %11
- ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8257545 /* reguse:VReg_128_Align2 */, [[COPY]]
+ ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7995401 /* reguse:VReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = tail call i128 asm sideeffect "; def $0", "=v"()
call void asm sideeffect "; use $0", "v"(i128 %val)
@@ -47,16 +47,16 @@ define amdgpu_kernel void @a_input_output_i128() {
; GFX908-LABEL: name: a_input_output_i128
; GFX908: bb.0 (%ir-block.0):
- ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8847370 /* regdef:AReg_128 */, def %13
+ ; GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8585226 /* regdef:AReg_128 */, def %13
; GFX908-NEXT: [[COPY:%[0-9]+]]:areg_128 = COPY %13
- ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8847369 /* reguse:AReg_128 */, [[COPY]]
+ ; GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 8585225 /* reguse:AReg_128 */, [[COPY]]
; GFX908-NEXT: S_ENDPGM 0
;
; GFX90A-LABEL: name: a_input_output_i128
; GFX90A: bb.0 (%ir-block.0):
- ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9568266 /* regdef:AReg_128_Align2 */, def %11
+ ; GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 9306122 /* regdef:AReg_128_Align2 */, def %11
; GFX90A-NEXT: [[COPY:%[0-9]+]]:areg_128_align2 = COPY %11
- ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY]]
+ ; GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY]]
; GFX90A-NEXT: S_ENDPGM 0
%val = call i128 asm sideeffect "; def $0", "=a"()
call void asm sideeffect "; use $0", "a"(i128 %val)
diff --git a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx10.mir b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx10.mir
index 8acf32e..e6a5234 100644
--- a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx10.mir
+++ b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx10.mir
@@ -18,21 +18,21 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_]], 256, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__literal_offsets
; GFX12: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 256, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 512, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e64 %stack.0, 256, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 %stack.0, 512, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -53,27 +53,27 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 256, [[V_ADD_U32_e64_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: [[V_ADD_U32_e64_2:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_ADD_U32_e64_]], -156, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_2]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_2]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__literal_offsets_commute
; GFX12: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 256, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 512, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: [[V_ADD_U32_e64_2:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 100, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_2]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_2]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e64 256, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 512, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
%2:vgpr_32 = V_ADD_U32_e64 %stack.0, 100, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %2
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir
index 1690826..98b7f4a 100644
--- a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir
+++ b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx8.mir
@@ -21,9 +21,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_CO_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
@@ -31,9 +31,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
@@ -41,10 +41,10 @@ body: |
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[COPY1]], implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
@@ -52,9 +52,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets
@@ -62,15 +62,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, [[COPY1]], implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -88,42 +88,42 @@ body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX803: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX900: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX942: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX10: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__literal_offsets_live_vcc
; GFX12: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]], implicit $vcc
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 256, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 512, %stack.0, implicit-def $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1, implicit $vcc
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1, implicit $vcc
SI_RETURN
...
@@ -144,9 +144,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_CO_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
@@ -154,9 +154,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
@@ -164,10 +164,10 @@ body: |
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[COPY1]], implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
@@ -175,9 +175,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[V_ADD_U32_e64_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__inline_imm_offsets
@@ -185,15 +185,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 8, [[COPY1]], implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_CO_U32_e32 8, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 16, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -214,9 +214,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64 = V_ADD_CO_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX803-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_CO_U32_e64_]]
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_CO_U32_e64_]], 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
@@ -224,9 +224,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
@@ -234,9 +234,9 @@ body: |
; GFX942-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[S_ADD_I32_]], 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
@@ -244,9 +244,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets
@@ -254,14 +254,14 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_1:%[0-9]+]]:sreg_32_xm0 = S_MOV_B32 %stack.0
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32_xexec_hi = S_ADD_I32 killed [[S_MOV_B32_]], [[S_MOV_B32_1]], implicit-def dead $scc
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_ADD_I32_]]
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 8, [[S_ADD_I32_]], 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -279,42 +279,42 @@ body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX803: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX900: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX942: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX10: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__inline_imm_offsets_live_vcc
; GFX12: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN implicit [[V_ADD_CO_U32_e64_1]]
%0:vgpr_32, %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, 8, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32, %3:sreg_64_xexec = V_ADD_CO_U32_e64 16, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN implicit %2
...
@@ -332,42 +332,42 @@ body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %0
%1:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %1
SI_RETURN
...
@@ -385,42 +385,42 @@ body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__inline_imm_offsets
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = S_ADD_I32 8, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %0
%1:sreg_32 = S_ADD_I32 16, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %1
SI_RETURN
...
@@ -443,9 +443,9 @@ body: |
; GFX803-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX803-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX803-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
@@ -454,9 +454,9 @@ body: |
; GFX900-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX900-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX900-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
@@ -465,9 +465,9 @@ body: |
; GFX942-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX942-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
@@ -476,9 +476,9 @@ body: |
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX10-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets
@@ -487,17 +487,17 @@ body: |
; GFX12-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX12-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY]], %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 [[COPY1]], %stack.0, implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = COPY $sgpr4
%1:sreg_32 = COPY $sgpr5
%2:sreg_32 = S_ADD_I32 %0, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %2
%3:sreg_32 = S_ADD_I32 %1, %stack.0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %3
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %3
SI_RETURN
...
@@ -520,9 +520,9 @@ body: |
; GFX803-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX803-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX803-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
@@ -531,9 +531,9 @@ body: |
; GFX900-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX900-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX900-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
@@ -542,9 +542,9 @@ body: |
; GFX942-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX942-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX942-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
@@ -553,9 +553,9 @@ body: |
; GFX10-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX10-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX10-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__reg_offsets_commute
@@ -564,17 +564,17 @@ body: |
; GFX12-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr4
; GFX12-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr5
; GFX12-NEXT: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY]], implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 %stack.0, [[COPY1]], implicit-def dead $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN
%0:sreg_32 = COPY $sgpr4
%1:sreg_32 = COPY $sgpr5
%2:sreg_32 = S_ADD_I32 %stack.0, %0, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %2
%3:sreg_32 = S_ADD_I32 %stack.0, %1, implicit-def dead $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %3
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %3
SI_RETURN
...
@@ -592,48 +592,48 @@ body: |
bb.0:
; GFX803-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX803: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX803-NEXT: S_NOP 0, implicit $scc
; GFX803-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX803-NEXT: SI_RETURN implicit $scc
;
; GFX900-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX900: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX900-NEXT: S_NOP 0, implicit $scc
; GFX900-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX900-NEXT: SI_RETURN implicit $scc
;
; GFX942-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX942: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX942-NEXT: S_NOP 0, implicit $scc
; GFX942-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX942-NEXT: SI_RETURN implicit $scc
;
; GFX10-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX10: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX10-NEXT: S_NOP 0, implicit $scc
; GFX10-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX10-NEXT: SI_RETURN implicit $scc
;
; GFX12-LABEL: name: local_stack_alloc__s_add_i32__literal_offsets_live_scc
; GFX12: [[S_ADD_I32_:%[0-9]+]]:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_]]
; GFX12-NEXT: S_NOP 0, implicit $scc
; GFX12-NEXT: [[S_ADD_I32_1:%[0-9]+]]:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, [[S_ADD_I32_1]]
; GFX12-NEXT: SI_RETURN implicit $scc
%0:sreg_32 = S_ADD_I32 256, %stack.0, implicit-def $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %0
S_NOP 0, implicit $scc
%1:sreg_32 = S_ADD_I32 512, %stack.0, implicit-def $scc
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2621449 /* reguse:SReg_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:SReg_32 */, %1
SI_RETURN implicit $scc
...
@@ -656,9 +656,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
@@ -667,9 +667,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
@@ -678,9 +678,9 @@ body: |
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
@@ -689,9 +689,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets
@@ -700,15 +700,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %vgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -731,9 +731,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
@@ -742,9 +742,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
@@ -753,9 +753,9 @@ body: |
; GFX942-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
@@ -764,9 +764,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__vgpr_offsets_commute
@@ -775,15 +775,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 [[S_MOV_B32_]], %vgpr_offset, implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_CO_U32_e32 %stack.0, %vgpr_offset, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %stack.0, %vgpr_offset, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -805,9 +805,9 @@ body: |
; GFX803-NEXT: {{ $}}
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
@@ -815,9 +815,9 @@ body: |
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
@@ -825,9 +825,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
@@ -836,9 +836,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit-def dead $vcc, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e32__sgpr_offsets
@@ -848,16 +848,16 @@ body: |
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[COPY]], implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_]]
; GFX12-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, [[COPY1]], implicit-def dead $vcc, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_CO_U32_e32 %sgpr_offset, %stack.0, implicit-def dead $vcc, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -880,9 +880,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
@@ -891,9 +891,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
@@ -903,10 +903,10 @@ body: |
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[COPY]], 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]]
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[COPY1]], 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
@@ -915,9 +915,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets
@@ -926,15 +926,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[S_MOV_B32_]], 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, [[S_MOV_B32_]], 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -957,9 +957,9 @@ body: |
; GFX803-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX803-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX803-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX803-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX803-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX803-NEXT: SI_RETURN
;
; GFX900-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
@@ -968,9 +968,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX900-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
@@ -980,10 +980,10 @@ body: |
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY %sgpr_offset
; GFX942-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], [[COPY]], 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY %sgpr_offset
; GFX942-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], [[COPY1]], 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
@@ -992,9 +992,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX10-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_co_u32_e64__sgpr_offsets_commute
@@ -1003,15 +1003,15 @@ body: |
; GFX12-NEXT: [[S_MOV_B32_:%[0-9]+]]:sreg_32_xexec_hi = S_MOV_B32 %stack.0
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_CO_U32_e64_:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], %sgpr_offset, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_]]
; GFX12-NEXT: [[V_ADD_CO_U32_e64_2:%[0-9]+]]:vgpr_32, dead [[V_ADD_CO_U32_e64_3:%[0-9]+]]:sreg_64_xexec = V_ADD_CO_U32_e64 [[S_MOV_B32_]], %sgpr_offset, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_CO_U32_e64_2]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32, dead %2:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32, dead %3:sreg_64_xexec = V_ADD_CO_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx9.mir b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx9.mir
index 9183fe5..19ca463 100644
--- a/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx9.mir
+++ b/llvm/test/CodeGen/AMDGPU/local-stack-alloc-add-references.gfx9.mir
@@ -20,16 +20,16 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 256, [[V_ADD_U32_e64_]], implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e32__literal_offsets
; GFX942: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 256, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX942-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 512, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e32__literal_offsets
@@ -37,21 +37,21 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 256, [[V_ADD_U32_e64_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e32__literal_offsets
; GFX12: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 256, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX12-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 512, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e32 256, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e32 512, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -72,16 +72,16 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 8, [[V_ADD_U32_e64_]], implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e32__inline_imm_offsets
; GFX942: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 8, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX942-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 16, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e32__inline_imm_offsets
@@ -89,21 +89,21 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 8, [[V_ADD_U32_e64_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e32__inline_imm_offsets
; GFX12: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 8, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX12-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 16, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e32 8, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e32 16, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -124,16 +124,16 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets
; GFX942: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX942-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets
@@ -141,21 +141,21 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 8, [[V_ADD_U32_e64_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets
; GFX12: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 16, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 16, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -178,9 +178,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX900-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets
@@ -188,9 +188,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX942-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets
@@ -199,9 +199,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX10-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets
@@ -209,15 +209,15 @@ body: |
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX12-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e32 %vgpr_offset, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -240,9 +240,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX900-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX900-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets_commute
@@ -250,9 +250,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX942-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX942-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets_commute
@@ -261,9 +261,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX10-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX10-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 [[V_MOV_B32_e32_]], %vgpr_offset, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e32__vgpr_offsets_commute
@@ -271,15 +271,15 @@ body: |
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: %vgpr_offset:vgpr_32 = COPY $vgpr0
; GFX12-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX12-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%vgpr_offset:vgpr_32 = COPY $vgpr0
%0:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e32 %stack.0, %vgpr_offset, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -301,9 +301,9 @@ body: |
; GFX900-NEXT: {{ $}}
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX900-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e32__sgpr_offsets
@@ -311,9 +311,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX942-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e32__sgpr_offsets
@@ -322,9 +322,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX10-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, [[V_MOV_B32_e32_]], implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e32__sgpr_offsets
@@ -332,15 +332,15 @@ body: |
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_U32_e32_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_]]
; GFX12-NEXT: [[V_ADD_U32_e32_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e32_1]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e32 %sgpr_offset, %stack.0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -363,9 +363,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX900-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets
@@ -373,9 +373,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX942-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets
@@ -384,9 +384,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, [[V_MOV_B32_e32_]], 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets
@@ -394,15 +394,15 @@ body: |
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 %sgpr_offset, %stack.0, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -425,9 +425,9 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX900-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets_commute
@@ -435,9 +435,9 @@ body: |
; GFX942-NEXT: {{ $}}
; GFX942-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX942-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX942-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets_commute
@@ -446,9 +446,9 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 [[V_MOV_B32_e32_]], %sgpr_offset, 0, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__sgpr_offsets_commute
@@ -456,15 +456,15 @@ body: |
; GFX12-NEXT: {{ $}}
; GFX12-NEXT: %sgpr_offset:sreg_32 = COPY $sgpr8
; GFX12-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: SI_RETURN
%sgpr_offset:sreg_32 = COPY $sgpr8
%0:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 %stack.0, %sgpr_offset, 0, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
@@ -486,16 +486,16 @@ body: |
; GFX900-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX900-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX900-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX900-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 8, [[V_ADD_U32_e64_]], 1, implicit $exec
- ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX900-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX900-NEXT: SI_RETURN
;
; GFX942-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets_clamp_modifier
; GFX942: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, 1, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX942-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 16, %stack.0, 1, implicit $exec
- ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX942-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX942-NEXT: SI_RETURN
;
; GFX10-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets_clamp_modifier
@@ -503,21 +503,21 @@ body: |
; GFX10-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 %stack.0, implicit $exec
; GFX10-NEXT: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[S_MOV_B32_]], [[V_MOV_B32_e32_]], 0, implicit $exec
; GFX10-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[V_ADD_U32_e64_]]
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[COPY]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[COPY]]
; GFX10-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 8, [[V_ADD_U32_e64_]], 1, implicit $exec
- ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX10-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX10-NEXT: SI_RETURN
;
; GFX12-LABEL: name: local_stack_alloc__v_add_u32_e64__inline_imm_offsets_clamp_modifier
; GFX12: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, 1, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_]]
; GFX12-NEXT: [[V_ADD_U32_e64_1:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 16, %stack.0, 1, implicit $exec
- ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
+ ; GFX12-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[V_ADD_U32_e64_1]]
; GFX12-NEXT: SI_RETURN
%0:vgpr_32 = V_ADD_U32_e64 %stack.0, 8, /*clamp*/1, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %0
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %0
%1:vgpr_32 = V_ADD_U32_e64 16, %stack.0, /*clamp*/1, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %1
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/machine-scheduler-sink-trivial-remats.mir b/llvm/test/CodeGen/AMDGPU/machine-scheduler-sink-trivial-remats.mir
index 8f228b7..c117473 100644
--- a/llvm/test/CodeGen/AMDGPU/machine-scheduler-sink-trivial-remats.mir
+++ b/llvm/test/CodeGen/AMDGPU/machine-scheduler-sink-trivial-remats.mir
@@ -6429,7 +6429,7 @@ body: |
; GFX908-NEXT: [[V_CVT_I32_F64_e32_19:%[0-9]+]]:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 19, implicit $exec, implicit $mode, implicit-def $m0
; GFX908-NEXT: [[V_CVT_I32_F64_e32_20:%[0-9]+]]:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 20, implicit $exec, implicit $mode, implicit-def $m0
; GFX908-NEXT: [[V_CVT_I32_F64_e32_21:%[0-9]+]]:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 21, implicit $exec, implicit $mode, implicit-def $m0
- ; GFX908-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %22, 2031625 /* reguse:VGPR_32 */, [[V_CVT_I32_F64_e32_4]]
+ ; GFX908-NEXT: INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %22, 1835017 /* reguse:VGPR_32 */, [[V_CVT_I32_F64_e32_4]]
; GFX908-NEXT: {{ $}}
; GFX908-NEXT: bb.1:
; GFX908-NEXT: successors: %bb.2(0x80000000)
@@ -6478,7 +6478,7 @@ body: |
%19:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 19, implicit $exec, implicit $mode, implicit-def $m0
%20:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 20, implicit $exec, implicit $mode, implicit-def $m0
%21:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 21, implicit $exec, implicit $mode, implicit-def $m0
- INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 2031626 /* regdef:VGPR_32 */, def %22:vgpr_32, 2031625 /* reguse:VGPR_32 */, %4
+ INLINEASM &"v_or_b32 $0, 0, $1", 0 /* attdialect */, 1835018 /* regdef:VGPR_32 */, def %22:vgpr_32, 1835017 /* reguse:VGPR_32 */, %4
%23:vgpr_32 = nofpexcept V_CVT_I32_F64_e32 23, implicit $exec, implicit $mode
bb.1:
diff --git a/llvm/test/CodeGen/AMDGPU/mai-hazards.mir b/llvm/test/CodeGen/AMDGPU/mai-hazards.mir
index c19d5a6..8b2c588 100644
--- a/llvm/test/CodeGen/AMDGPU/mai-hazards.mir
+++ b/llvm/test/CodeGen/AMDGPU/mai-hazards.mir
@@ -33,7 +33,7 @@ name: asm_write_vgpr_accvgpr_write_read
body: |
bb.0:
- INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0
+ INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0
$agpr0 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr0, implicit $exec
...
@@ -47,7 +47,7 @@ name: asm_write_vgpr_accvgpr_write_read_partialnop
body: |
bb.0:
- INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr0
+ INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr0
S_NOP 0
$agpr0 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr0, implicit $exec
...
@@ -60,7 +60,7 @@ name: asm_write_vgpr_accvgpr_write_read_otherreg
body: |
bb.0:
liveins: $vgpr0
- INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def $vgpr1
+ INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def $vgpr1
$agpr0 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr0, implicit $exec
...
diff --git a/llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll b/llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll
index f88b1bf..95638c7 100644
--- a/llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll
+++ b/llvm/test/CodeGen/AMDGPU/partial-regcopy-and-spill-missed-at-regalloc.ll
@@ -11,10 +11,10 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; REGALLOC-GFX908: bb.0 (%ir-block.0):
; REGALLOC-GFX908-NEXT: liveins: $sgpr4_sgpr5
; REGALLOC-GFX908-NEXT: {{ $}}
- ; REGALLOC-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %6:agpr_32
- ; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7929866 /* regdef:VReg_128 */, def %25
+ ; REGALLOC-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, undef %6:agpr_32
+ ; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7667722 /* regdef:VReg_128 */, def %25
; REGALLOC-GFX908-NEXT: [[COPY:%[0-9]+]]:av_128 = COPY %25
- ; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64 */, def %27
+ ; REGALLOC-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3473418 /* regdef:VReg_64 */, def %27
; REGALLOC-GFX908-NEXT: SI_SPILL_AV64_SAVE %27, %stack.0, $sgpr32, 0, implicit $exec :: (store (s64) into %stack.0, align 4, addrspace 5)
; REGALLOC-GFX908-NEXT: [[COPY1:%[0-9]+]]:vreg_128 = COPY [[COPY]]
; REGALLOC-GFX908-NEXT: GLOBAL_STORE_DWORDX4 undef %15:vreg_64, [[COPY1]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
@@ -36,10 +36,10 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; PEI-GFX908-NEXT: $sgpr12_sgpr13_sgpr14_sgpr15 = COPY killed $sgpr0_sgpr1_sgpr2_sgpr3
; PEI-GFX908-NEXT: $sgpr12 = S_ADD_U32 $sgpr12, $sgpr9, implicit-def $scc, implicit-def $sgpr12_sgpr13_sgpr14_sgpr15
; PEI-GFX908-NEXT: $sgpr13 = S_ADDC_U32 $sgpr13, 0, implicit-def dead $scc, implicit $scc, implicit-def $sgpr12_sgpr13_sgpr14_sgpr15
- ; PEI-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
- ; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7929866 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
+ ; PEI-GFX908-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, undef renamable $agpr0
+ ; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7667722 /* regdef:VReg_128 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX908-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
- ; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
+ ; PEI-GFX908-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3473418 /* regdef:VReg_64 */, def renamable $vgpr0_vgpr1
; PEI-GFX908-NEXT: BUFFER_STORE_DWORD_OFFSET killed $vgpr0, $sgpr12_sgpr13_sgpr14_sgpr15, 0, 0, 0, 0, implicit $exec, implicit-def $vgpr0_vgpr1, implicit $vgpr0_vgpr1 :: (store (s32) into %stack.0, addrspace 5)
; PEI-GFX908-NEXT: $agpr4 = V_ACCVGPR_WRITE_B32_e64 killed $vgpr1, implicit $exec, implicit killed $vgpr0_vgpr1
; PEI-GFX908-NEXT: renamable $vgpr0_vgpr1_vgpr2_vgpr3 = COPY killed renamable $agpr0_agpr1_agpr2_agpr3, implicit $exec
@@ -60,10 +60,10 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; REGALLOC-GFX90A: bb.0 (%ir-block.0):
; REGALLOC-GFX90A-NEXT: liveins: $sgpr4_sgpr5
; REGALLOC-GFX90A-NEXT: {{ $}}
- ; REGALLOC-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef %6:agpr_32
- ; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8257546 /* regdef:VReg_128_Align2 */, def %23
+ ; REGALLOC-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, undef %6:agpr_32
+ ; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def %23
; REGALLOC-GFX90A-NEXT: [[COPY:%[0-9]+]]:av_128_align2 = COPY %23
- ; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3997706 /* regdef:VReg_64_Align2 */, def %21
+ ; REGALLOC-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64_Align2 */, def %21
; REGALLOC-GFX90A-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY %21
; REGALLOC-GFX90A-NEXT: GLOBAL_STORE_DWORDX4 undef %15:vreg_64_align2, [[COPY]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
; REGALLOC-GFX90A-NEXT: renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
@@ -79,10 +79,10 @@ define amdgpu_kernel void @partial_copy(<4 x i32> %arg) #0 {
; PEI-GFX90A: bb.0 (%ir-block.0):
; PEI-GFX90A-NEXT: liveins: $sgpr4_sgpr5
; PEI-GFX90A-NEXT: {{ $}}
- ; PEI-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, undef renamable $agpr0
- ; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 8257546 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
+ ; PEI-GFX90A-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, undef renamable $agpr0
+ ; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 7995402 /* regdef:VReg_128_Align2 */, def renamable $vgpr0_vgpr1_vgpr2_vgpr3
; PEI-GFX90A-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $vgpr0_vgpr1_vgpr2_vgpr3, implicit $exec
- ; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3997706 /* regdef:VReg_64_Align2 */, def renamable $vgpr2_vgpr3
+ ; PEI-GFX90A-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 3735562 /* regdef:VReg_64_Align2 */, def renamable $vgpr2_vgpr3
; PEI-GFX90A-NEXT: GLOBAL_STORE_DWORDX4 undef renamable $vgpr0_vgpr1, killed renamable $agpr0_agpr1_agpr2_agpr3, 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
; PEI-GFX90A-NEXT: renamable $sgpr0_sgpr1_sgpr2_sgpr3 = S_LOAD_DWORDX4_IMM killed renamable $sgpr4_sgpr5, 0, 0 :: (dereferenceable invariant load (s128) from %ir.arg.kernarg.offset1, addrspace 4)
; PEI-GFX90A-NEXT: renamable $agpr0_agpr1_agpr2_agpr3 = COPY killed renamable $sgpr0_sgpr1_sgpr2_sgpr3, implicit $exec
diff --git a/llvm/test/CodeGen/AMDGPU/regalloc-failure-overlapping-insert-assert.mir b/llvm/test/CodeGen/AMDGPU/regalloc-failure-overlapping-insert-assert.mir
index 80afe7a..94aa6fe 100644
--- a/llvm/test/CodeGen/AMDGPU/regalloc-failure-overlapping-insert-assert.mir
+++ b/llvm/test/CodeGen/AMDGPU/regalloc-failure-overlapping-insert-assert.mir
@@ -43,17 +43,17 @@ machineFunctionInfo:
body: |
bb.0:
- INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2424842 /* regdef:AGPR_32 */, implicit-def $agpr0
+ INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2228234 /* regdef:AGPR_32 */, implicit-def $agpr0
%14:vgpr_32 = COPY killed $agpr0
- INLINEASM &"; def $0 $1 $2 $3 $4", 1 /* sideeffect attdialect */, 40042506 /* regdef:VReg_512 */, def %7, 19464202 /* regdef:VReg_256 */, def %8, 7929866 /* regdef:VReg_128 */, def %9, 5963786 /* regdef:VReg_96 */, def %10, 5963786 /* regdef:VReg_96 */, def %11
+ INLINEASM &"; def $0 $1 $2 $3 $4", 1 /* sideeffect attdialect */, 39780362 /* regdef:VReg_512 */, def %7, 19202058 /* regdef:VReg_256 */, def %8, 7667722 /* regdef:VReg_128 */, def %9, 5701642 /* regdef:VReg_96 */, def %10, 5701642 /* regdef:VReg_96 */, def %11
INLINEASM &"; clobber", 1 /* sideeffect attdialect */, 12 /* clobber */, implicit-def dead early-clobber $agpr0_agpr1_agpr2_agpr3_agpr4_agpr5_agpr6_agpr7_agpr8_agpr9_agpr10_agpr11_agpr12_agpr13_agpr14_agpr15_agpr16_agpr17_agpr18_agpr19_agpr20_agpr21_agpr22_agpr23_agpr24_agpr25_agpr26_agpr27_agpr28_agpr29_agpr30_agpr31, 12 /* clobber */, implicit-def dead early-clobber $vgpr0_vgpr1_vgpr2_vgpr3_vgpr4_vgpr5_vgpr6_vgpr7_vgpr8_vgpr9_vgpr10_vgpr11_vgpr12_vgpr13_vgpr14_vgpr15_vgpr16_vgpr17_vgpr18_vgpr19_vgpr20_vgpr21_vgpr22_vgpr23_vgpr24_vgpr25_vgpr26_vgpr27_vgpr28_vgpr29_vgpr30_vgpr31
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 40042505 /* reguse:VReg_512 */, %7
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 19464201 /* reguse:VReg_256 */, %8
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7929865 /* reguse:VReg_128 */, %9
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5963785 /* reguse:VReg_96 */, %10
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5963785 /* reguse:VReg_96 */, %11
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 39780361 /* reguse:VReg_512 */, %7
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 19202057 /* reguse:VReg_256 */, %8
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 7667721 /* reguse:VReg_128 */, %9
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5701641 /* reguse:VReg_96 */, %10
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 5701641 /* reguse:VReg_96 */, %11
$agpr1 = COPY %14
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, killed $agpr1
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, killed $agpr1
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/rename-independent-subregs.mir b/llvm/test/CodeGen/AMDGPU/rename-independent-subregs.mir
index e1148c4..35be513 100644
--- a/llvm/test/CodeGen/AMDGPU/rename-independent-subregs.mir
+++ b/llvm/test/CodeGen/AMDGPU/rename-independent-subregs.mir
@@ -73,7 +73,7 @@ body: |
# (1) %0.sub0 + %0.sub0 and (2) %0.sub1 + %0.sub1
# Check that renaming (2) does not inadvertently rename (1).
# CHECK-LABEL: name: test2
-# CHECK: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %0.sub0, 2031626 /* regdef:VGPR_32 */, def dead %1.sub1, 2147483657 /* reguse tiedto:$0 */, undef %0.sub0(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %1.sub1(tied-def 5)
+# CHECK: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %0.sub0, 1835018 /* regdef:VGPR_32 */, def dead %1.sub1, 2147483657 /* reguse tiedto:$0 */, undef %0.sub0(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %1.sub1(tied-def 5)
name: test2
body: |
bb.0:
@@ -81,7 +81,7 @@ body: |
bb.1:
undef %0.sub1:vreg_64 = V_ALIGNBIT_B32_e64 %0.sub0:vreg_64, %0.sub0:vreg_64, 16, implicit $exec
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %0.sub0:vreg_64, 2031626 /* regdef:VGPR_32 */, def %0.sub1:vreg_64, 2147483657 /* reguse tiedto:$0 */, undef %0.sub0:vreg_64(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %0.sub1:vreg_64(tied-def 5)
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %0.sub0:vreg_64, 1835018 /* regdef:VGPR_32 */, def %0.sub1:vreg_64, 2147483657 /* reguse tiedto:$0 */, undef %0.sub0:vreg_64(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %0.sub1:vreg_64(tied-def 5)
S_BRANCH %bb.1
...
diff --git a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir
index ad490f8..c58b9da 100644
--- a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir
+++ b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-copy-from.mir
@@ -43,7 +43,7 @@ body: |
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
; CHECK-NEXT: [[COPY3:%[0-9]+]]:vreg_128_align2 = COPY [[GLOBAL_LOAD_DWORDX4_]]
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]]:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[COPY3]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3997705 /* reguse:VReg_64_Align2 */, [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:VReg_64_Align2 */, [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]]
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
%1:av_64_align2 = COPY $vgpr0_vgpr1
@@ -51,7 +51,7 @@ body: |
%3:areg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
%4:vreg_128_align2 = COPY %3
%5:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3997705 /* reguse:VReg_64_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:VReg_64_Align2 */, %5
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir
index 0b4e662..bae84f4 100644
--- a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir
+++ b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-insert-extract.mir
@@ -19,7 +19,7 @@ body: |
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
@@ -30,7 +30,7 @@ body: |
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
undef %5.sub0_sub1:areg_128_align2 = COPY %4
%5.sub2_sub3 = IMPLICIT_DEF
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %5
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
@@ -172,7 +172,7 @@ body: |
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
@@ -183,7 +183,7 @@ body: |
undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
undef %5.sub0_sub1:areg_128_align2 = COPY %4.sub2_sub3
%5.sub2_sub3 = IMPLICIT_DEF
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %5
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
@@ -208,7 +208,7 @@ body: |
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_vgprcd_e64_:%[0-9]+]].sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_vgprcd_e64_]].sub2
; CHECK-NEXT: [[COPY3:%[0-9]+]].sub2_sub3:areg_128_align2 = IMPLICIT_DEF
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]].sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
@@ -219,7 +219,7 @@ body: |
undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
undef %5.sub1:areg_128_align2 = COPY %4.sub2
%5.sub2_sub3 = IMPLICIT_DEF
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %5
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
GLOBAL_STORE_DWORDX2 %0, %5.sub2_sub3, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
diff --git a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir
index dcf3b8b..e9f45e3 100644
--- a/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir
+++ b/llvm/test/CodeGen/AMDGPU/rewrite-vgpr-mfma-to-agpr-subreg-src2-chain.mir
@@ -17,7 +17,7 @@ body: |
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -26,7 +26,7 @@ body: |
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
undef %5.sub0_sub1:areg_128_align2 = COPY %4
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %5
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
...
@@ -47,7 +47,7 @@ body: |
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -56,7 +56,7 @@ body: |
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
undef %5.sub0_sub1:areg_128_align2 = COPY %4
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %5
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %5
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
...
@@ -79,7 +79,7 @@ body: |
; CHECK-NEXT: dead %other_use:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s64), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -90,7 +90,7 @@ body: |
%other_use:vreg_64_align2 = COPY %4.sub0_sub1
%5:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
%6:areg_64_align2 = COPY %5
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %6:areg_64_align2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %6:areg_64_align2
GLOBAL_STORE_DWORDX2 %0, %6, 0, 0, implicit $exec :: (store (s64), addrspace 1)
SI_RETURN
...
@@ -114,7 +114,7 @@ body: |
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]], 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4915209 /* reguse:AReg_64_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4653065 /* reguse:AReg_64_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s64), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -126,7 +126,7 @@ body: |
undef %5.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4, 0, 0, 0, implicit $mode, implicit $exec
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
%7:areg_64_align2 = COPY %6
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4915209 /* reguse:AReg_64_Align2 */, %7
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4653065 /* reguse:AReg_64_Align2 */, %7
GLOBAL_STORE_DWORDX2 %0, %7, 0, 0, implicit $exec :: (store (s64), addrspace 1)
SI_RETURN
@@ -151,7 +151,7 @@ body: |
; CHECK-NEXT: dead %other_use:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -163,7 +163,7 @@ body: |
%other_use:vreg_64_align2 = COPY %5.sub0_sub1
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
undef %8.sub0_sub1:areg_128_align2 = COPY %6
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %8:areg_128_align2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %8:areg_128_align2
GLOBAL_STORE_DWORDX4 %0, %8, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
@@ -189,7 +189,7 @@ body: |
; CHECK-NEXT: dead %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr_32 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORD [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s32), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -202,7 +202,7 @@ body: |
%other_use1:vreg_64_align2 = COPY %5.sub0_sub1
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
%8:agpr_32 = COPY %6
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2424841 /* reguse:AGPR_32 */, %8:agpr_32
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2228233 /* reguse:AGPR_32 */, %8:agpr_32
GLOBAL_STORE_DWORD %0, %8, 0, 0, implicit $exec :: (store (s32), addrspace 1)
SI_RETURN
@@ -231,7 +231,7 @@ body: |
; CHECK-NEXT: dead %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -245,7 +245,7 @@ body: |
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
%other_use2:vreg_64 = COPY %4.sub1_sub2
%6:areg_128_align2 = COPY %4
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9568265 /* reguse:AReg_128_Align2 */, %6:areg_128_align2
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 9306121 /* reguse:AReg_128_Align2 */, %6:areg_128_align2
GLOBAL_STORE_DWORDX4 %0, %6, 0, 0, implicit $exec :: (store (s128), addrspace 1)
SI_RETURN
...
@@ -273,7 +273,7 @@ body: |
; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -287,7 +287,7 @@ body: |
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
%other_use2:vreg_64 = COPY %4.sub1_sub2
%6:areg_64 = COPY %4.sub1_sub2
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4325385 /* reguse:AReg_64 */, %6:areg_64
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4063241 /* reguse:AReg_64 */, %6:areg_64
GLOBAL_STORE_DWORDX2 %0, %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
SI_RETURN
...
@@ -313,7 +313,7 @@ body: |
; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
- ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4325385 /* reguse:AReg_64 */, [[COPY3]]
+ ; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4063241 /* reguse:AReg_64 */, [[COPY3]]
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
; CHECK-NEXT: SI_RETURN
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
@@ -327,7 +327,7 @@ body: |
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
%other_use2:vreg_64 = COPY %4.sub1_sub2
%6:areg_64 = COPY %4.sub1_sub2
- INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4325385 /* reguse:AReg_64 */, %6:areg_64
+ INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 4063241 /* reguse:AReg_64 */, %6:areg_64
GLOBAL_STORE_DWORDX2 %0, %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
SI_RETURN
...
diff --git a/llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir b/llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir
index 3c7dd64..af882c0 100644
--- a/llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir
+++ b/llvm/test/CodeGen/AMDGPU/sched-assert-dead-def-subreg-use-other-subreg.mir
@@ -37,7 +37,7 @@ body: |
; CHECK-NEXT: dead [[DS_READ_B128_gfx9_:%[0-9]+]]:vreg_128 = DS_READ_B128_gfx9 [[V_ADD_U32_e32_]], 0, 0, implicit $exec
; CHECK-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY [[COPY]].sub0
; CHECK-NEXT: undef [[COPY2:%[0-9]+]].sub1:vreg_512 = COPY [[COPY]].sub1
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def dead [[COPY1]], 2031626 /* regdef:VGPR_32 */, def dead [[COPY]].sub1, 2031625 /* reguse:VGPR_32 */, [[COPY1]], 2031625 /* reguse:VGPR_32 */, [[COPY]].sub1
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def dead [[COPY1]], 1835018 /* regdef:VGPR_32 */, def dead [[COPY]].sub1, 1835017 /* reguse:VGPR_32 */, [[COPY1]], 1835017 /* reguse:VGPR_32 */, [[COPY]].sub1
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub0:vreg_512 = COPY [[COPY]].sub0
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub3:vreg_512 = COPY [[COPY]].sub3
; CHECK-NEXT: [[COPY2:%[0-9]+]].sub2:vreg_512 = COPY undef [[V_MOV_B32_e32_1]]
@@ -63,7 +63,7 @@ body: |
undef %11.sub0:vreg_512 = COPY %4.sub0
%12:vgpr_32 = COPY %4.sub0
%11.sub1:vreg_512 = COPY %4.sub1
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def dead %12:vgpr_32, 2031626 /* regdef:VGPR_32 */, def dead %4.sub1:vreg_512, 2031625 /* reguse:VGPR_32 */, %12:vgpr_32, 2031625 /* reguse:VGPR_32 */, %4.sub1:vreg_512
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def dead %12:vgpr_32, 1835018 /* regdef:VGPR_32 */, def dead %4.sub1:vreg_512, 1835017 /* reguse:VGPR_32 */, %12:vgpr_32, 1835017 /* reguse:VGPR_32 */, %4.sub1:vreg_512
%11.sub2:vreg_512 = COPY undef %1
%11.sub3:vreg_512 = COPY %4.sub3
%11.sub5:vreg_512 = COPY undef %1
diff --git a/llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir b/llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir
index 24df4b8..3e8e187 100644
--- a/llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir
+++ b/llvm/test/CodeGen/AMDGPU/sched-handleMoveUp-subreg-def-across-subreg-def.mir
@@ -40,18 +40,18 @@ body: |
; CHECK-NEXT: bb.1:
; CHECK-NEXT: successors: %bb.1(0x80000000)
; CHECK-NEXT: {{ $}}
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def dead %11
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def dead %11
; CHECK-NEXT: GLOBAL_STORE_DWORD undef %12:vreg_64, [[BUFFER_LOAD_DWORD_OFFEN]], 0, 0, implicit $exec :: (store (s32), addrspace 1)
; CHECK-NEXT: [[V_MOV_B32_e32_3:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; CHECK-NEXT: [[V_MOV_B32_e32_4:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
; CHECK-NEXT: [[DS_READ_B64_gfx9_:%[0-9]+]]:vreg_64 = DS_READ_B64_gfx9 undef %14:vgpr_32, 0, 0, implicit $exec :: (load (s64), addrspace 3)
- ; CHECK-NEXT: INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %15, 2031626 /* regdef:VGPR_32 */, def %16
+ ; CHECK-NEXT: INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %15, 1835018 /* regdef:VGPR_32 */, def %16
; CHECK-NEXT: [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_1]], 0, 0, implicit $exec
; CHECK-NEXT: [[DS_READ_B32_gfx9_1:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_2]], 0, 0, implicit $exec
; CHECK-NEXT: [[DS_READ_B32_gfx9_2:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 undef %20:vgpr_32, 0, 0, implicit $exec
- ; CHECK-NEXT: INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %21, 2031626 /* regdef:VGPR_32 */, def %22
+ ; CHECK-NEXT: INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %21, 1835018 /* regdef:VGPR_32 */, def %22
; CHECK-NEXT: [[DS_READ_B32_gfx9_3:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_2]], 0, 0, implicit $exec
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def dead [[V_MOV_B32_e32_3]], 2031626 /* regdef:VGPR_32 */, def dead [[V_MOV_B32_e32_4]], 2031625 /* reguse:VGPR_32 */, [[DS_READ_B64_gfx9_]].sub0, 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_3]](tied-def 3), 2147549193 /* reguse tiedto:$1 */, [[V_MOV_B32_e32_4]](tied-def 5), 2031625 /* reguse:VGPR_32 */, %15, 2031625 /* reguse:VGPR_32 */, %16, 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_1]], 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]], 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_3]], 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_2]]
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def dead [[V_MOV_B32_e32_3]], 1835018 /* regdef:VGPR_32 */, def dead [[V_MOV_B32_e32_4]], 1835017 /* reguse:VGPR_32 */, [[DS_READ_B64_gfx9_]].sub0, 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_3]](tied-def 3), 2147549193 /* reguse tiedto:$1 */, [[V_MOV_B32_e32_4]](tied-def 5), 1835017 /* reguse:VGPR_32 */, %15, 1835017 /* reguse:VGPR_32 */, %16, 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_1]], 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]], 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_3]], 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_2]]
; CHECK-NEXT: [[V_MOV_B32_e32_:%[0-9]+]].sub1:vreg_64 = COPY [[V_MOV_B32_e32_1]]
; CHECK-NEXT: DS_WRITE_B32_gfx9 undef %28:vgpr_32, %21, 0, 0, implicit $exec :: (store (s32), addrspace 3)
; CHECK-NEXT: DS_WRITE_B32_gfx9 undef %29:vgpr_32, %22, 0, 0, implicit $exec :: (store (s32), addrspace 3)
@@ -94,21 +94,21 @@ body: |
%10:vgpr_32 = IMPLICIT_DEF
bb.1:
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %11:vgpr_32
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %11:vgpr_32
GLOBAL_STORE_DWORD undef %12:vreg_64, %1, 0, 0, implicit $exec :: (store (s32), addrspace 1)
%13:vreg_64 = DS_READ_B64_gfx9 undef %14:vgpr_32, 0, 0, implicit $exec :: (load (s64), addrspace 3)
- INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %15:vgpr_32, 2031626 /* regdef:VGPR_32 */, def %16:vgpr_32
+ INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %15:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %16:vgpr_32
%17:vgpr_32 = DS_READ_B32_gfx9 %6, 0, 0, implicit $exec
%18:vgpr_32 = DS_READ_B32_gfx9 %7, 0, 0, implicit $exec
%19:vgpr_32 = DS_READ_B32_gfx9 undef %20:vgpr_32, 0, 0, implicit $exec
- INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %21:vgpr_32, 2031626 /* regdef:VGPR_32 */, def %22:vgpr_32
+ INLINEASM &"def $0 $1", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %21:vgpr_32, 1835018 /* regdef:VGPR_32 */, def %22:vgpr_32
%23:vgpr_32 = DS_READ_B32_gfx9 %7, 0, 0, implicit $exec
%24:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
%5.sub1:vreg_64 = COPY %6
%25:vgpr_32 = V_ADD_U32_e32 1, %10, implicit $exec
%26:sreg_64_xexec = V_CMP_GT_U32_e64 64, %25, implicit $exec
%27:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def dead %24:vgpr_32, 2031626 /* regdef:VGPR_32 */, def dead %27:vgpr_32, 2031625 /* reguse:VGPR_32 */, %13.sub0:vreg_64, 2147483657 /* reguse tiedto:$0 */, %24:vgpr_32(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %27:vgpr_32(tied-def 5), 2031625 /* reguse:VGPR_32 */, %15, 2031625 /* reguse:VGPR_32 */, %16, 2031625 /* reguse:VGPR_32 */, %18, 2031625 /* reguse:VGPR_32 */, %17, 2031625 /* reguse:VGPR_32 */, %23, 2031625 /* reguse:VGPR_32 */, %19
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def dead %24:vgpr_32, 1835018 /* regdef:VGPR_32 */, def dead %27:vgpr_32, 1835017 /* reguse:VGPR_32 */, %13.sub0:vreg_64, 2147483657 /* reguse tiedto:$0 */, %24:vgpr_32(tied-def 3), 2147549193 /* reguse tiedto:$1 */, %27:vgpr_32(tied-def 5), 1835017 /* reguse:VGPR_32 */, %15, 1835017 /* reguse:VGPR_32 */, %16, 1835017 /* reguse:VGPR_32 */, %18, 1835017 /* reguse:VGPR_32 */, %17, 1835017 /* reguse:VGPR_32 */, %23, 1835017 /* reguse:VGPR_32 */, %19
DS_WRITE_B32_gfx9 undef %28:vgpr_32, %21, 0, 0, implicit $exec :: (store (s32), addrspace 3)
DS_WRITE_B32_gfx9 undef %29:vgpr_32, %22, 0, 0, implicit $exec :: (store (s32), addrspace 3)
DS_WRITE_B64_gfx9 undef %30:vgpr_32, %5, 0, 0, implicit $exec :: (store (s64), addrspace 3)
diff --git a/llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll b/llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll
index 4893bff..4286da3 100644
--- a/llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll
+++ b/llvm/test/CodeGen/AMDGPU/spill-vector-superclass.ll
@@ -12,10 +12,10 @@ define amdgpu_kernel void @test_spill_av_class(<4 x i32> %arg) #0 {
; GCN-NEXT: [[AV_MOV_:%[0-9]+]]:vgpr_32 = AV_MOV_B32_IMM_PSEUDO 1, implicit $exec
; GCN-NEXT: [[AV_MOV_1:%[0-9]+]]:vgpr_32 = AV_MOV_B32_IMM_PSEUDO 2, implicit $exec
; GCN-NEXT: [[V_MFMA_I32_4X4X4I8_e64_:%[0-9]+]]:areg_128 = V_MFMA_I32_4X4X4I8_e64 [[AV_MOV_]], [[AV_MOV_1]], [[COPY]], 0, 0, 0, implicit $mode, implicit $exec
- ; GCN-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %14.sub0
+ ; GCN-NEXT: INLINEASM &"; def $0", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %14.sub0
; GCN-NEXT: [[COPY1:%[0-9]+]]:vreg_128 = COPY [[V_MFMA_I32_4X4X4I8_e64_]]
; GCN-NEXT: GLOBAL_STORE_DWORDX4 undef %24:vreg_64, [[COPY1]], 0, 0, implicit $exec :: (volatile store (s128) into `ptr addrspace(1) poison`, addrspace 1)
- ; GCN-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:VReg_64 */, %14
+ ; GCN-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:VReg_64 */, %14
; GCN-NEXT: S_ENDPGM 0
%v0 = call i32 asm sideeffect "; def $0", "=v"()
%tmp = insertelement <2 x i32> poison, i32 %v0, i32 0
diff --git a/llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir b/llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir
index b6d630e..da6b57c 100644
--- a/llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir
+++ b/llvm/test/CodeGen/AMDGPU/subreg-undef-def-with-other-subreg-defs.mir
@@ -28,9 +28,9 @@ body: |
; CHECK-NEXT: successors: %bb.1(0x80000000)
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_1]], 0, 0, implicit $exec :: (load (s32), addrspace 3)
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def [[V_MOV_B32_e32_]], 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_]](tied-def 3)
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]]
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub0, 2031626 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub1
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def [[V_MOV_B32_e32_]], 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_]](tied-def 3)
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]]
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub0, 1835018 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub1
; CHECK-NEXT: S_NOP 0, implicit [[V_MOV_B32_e32_]].sub1
; CHECK-NEXT: $sgpr10 = S_MOV_B32 -1
; CHECK-NEXT: S_BRANCH %bb.1
@@ -41,9 +41,9 @@ body: |
bb.1:
%2:vgpr_32 = DS_READ_B32_gfx9 %1, 0, 0, implicit $exec :: (load (s32), addrspace 3)
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %0, 2147483657 /* reguse tiedto:$0 */, %0(tied-def 3)
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %2
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef %0.sub0, 2031626 /* regdef:VGPR_32 */, def %0.sub1
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %0, 2147483657 /* reguse tiedto:$0 */, %0(tied-def 3)
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %2
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef %0.sub0, 1835018 /* regdef:VGPR_32 */, def %0.sub1
S_NOP 0, implicit %0.sub1
$sgpr10 = S_MOV_B32 -1
S_BRANCH %bb.1
@@ -69,9 +69,9 @@ body: |
; CHECK-NEXT: successors: %bb.1(0x80000000)
; CHECK-NEXT: {{ $}}
; CHECK-NEXT: [[DS_READ_B32_gfx9_:%[0-9]+]]:vgpr_32 = DS_READ_B32_gfx9 [[V_MOV_B32_e32_1]], 0, 0, implicit $exec :: (load (s32), addrspace 3)
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def [[V_MOV_B32_e32_]], 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_]](tied-def 3)
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]]
- ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub1, 2031626 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub0
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def [[V_MOV_B32_e32_]], 2147483657 /* reguse tiedto:$0 */, [[V_MOV_B32_e32_]](tied-def 3)
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, [[DS_READ_B32_gfx9_]]
+ ; CHECK-NEXT: INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub1, 1835018 /* regdef:VGPR_32 */, def undef [[V_MOV_B32_e32_]].sub0
; CHECK-NEXT: S_NOP 0, implicit [[V_MOV_B32_e32_]].sub1
; CHECK-NEXT: $sgpr10 = S_MOV_B32 -1
; CHECK-NEXT: S_BRANCH %bb.1
@@ -82,9 +82,9 @@ body: |
bb.1:
%2:vgpr_32 = DS_READ_B32_gfx9 %1, 0, 0, implicit $exec :: (load (s32), addrspace 3)
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %0, 2147483657 /* reguse tiedto:$0 */, %0(tied-def 3)
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031625 /* reguse:VGPR_32 */, %2
- INLINEASM &"", 1 /* sideeffect attdialect */, 2031626 /* regdef:VGPR_32 */, def %0.sub1, 2031626 /* regdef:VGPR_32 */, def undef %0.sub0
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %0, 2147483657 /* reguse tiedto:$0 */, %0(tied-def 3)
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835017 /* reguse:VGPR_32 */, %2
+ INLINEASM &"", 1 /* sideeffect attdialect */, 1835018 /* regdef:VGPR_32 */, def %0.sub1, 1835018 /* regdef:VGPR_32 */, def undef %0.sub0
S_NOP 0, implicit %0.sub1
$sgpr10 = S_MOV_B32 -1
S_BRANCH %bb.1
diff --git a/llvm/test/TableGen/ResolveSchedClass.td b/llvm/test/TableGen/ResolveSchedClass.td
new file mode 100644
index 0000000..8c9ef1e
--- /dev/null
+++ b/llvm/test/TableGen/ResolveSchedClass.td
@@ -0,0 +1,18 @@
+// RUN: llvm-tblgen -gen-subtarget -I %p/../../include %s -o - | FileCheck %s
+
+include "llvm/Target/Target.td"
+
+def TestTargetInstrInfo : InstrInfo;
+
+def TestTarget : Target {
+ let InstructionSet = TestTargetInstrInfo;
+}
+
+// CHECK: unsigned resolveVariantSchedClassImpl(unsigned SchedClass,
+// CHECK-NEXT: const MCInst *MI, const MCInstrInfo *MCII, const MCSubtargetInfo &STI, unsigned CPUID)
+
+// CHECK: unsigned resolveVariantSchedClass(unsigned SchedClass,
+// CHECK-NEXT: const MCInst *MI, const MCInstrInfo *MCII,
+// CHECK-NEXT: unsigned CPUID) const override {
+// CHECK-NEXT: return TestTarget_MC::resolveVariantSchedClassImpl(SchedClass, MI, MCII, *this, CPUID);
+// CHECK-NEXT: }
diff --git a/llvm/test/ThinLTO/X86/memprof-dups.ll b/llvm/test/ThinLTO/X86/memprof-dups.ll
new file mode 100644
index 0000000..8accc83
--- /dev/null
+++ b/llvm/test/ThinLTO/X86/memprof-dups.ll
@@ -0,0 +1,138 @@
+;; Check that duplicate spurious duplicate (identical) clones are simply
+;; created as aliases to the first identical copy, rather than creating
+;; multiple clones that call the same callee clones or have the same
+;; allocation types. This currently happens in some cases due to additional
+;; cloning performed during function assignment.
+;;
+;; The ThinLTO combined summary was manually modified as described there
+;; to force multiple identical copies of various functions.
+
+;; -stats requires asserts
+; REQUIRES: asserts
+
+; RUN: rm -rf %t && split-file %s %t && cd %t
+; RUN: llvm-as src.ll -o src.o
+; RUN: llvm-as src.o.thinlto.ll -o src.o.thinlto.bc
+; RUN: opt -passes=memprof-context-disambiguation -stats \
+; RUN: -memprof-import-summary=src.o.thinlto.bc \
+; RUN: -pass-remarks=memprof-context-disambiguation \
+; RUN: src.o -S 2>&1 | FileCheck %s
+
+; CHECK: created clone bar.memprof.1
+;; Duplicates of bar are created as declarations since bar is available_externally,
+;; and the compiler does not well support available_externally aliases.
+; CHECK: created clone decl bar.memprof.2
+; CHECK: created clone decl bar.memprof.3
+; CHECK: created clone _Z3foov.memprof.1
+;; Duplicates of _Z3foov are created as aliases to the appropriate materialized
+;; clone of _Z3foov.
+; CHECK: created clone alias _Z3foov.memprof.2
+; CHECK: created clone alias _Z3foov.memprof.3
+
+;--- src.ll
+source_filename = "memprof-distrib-alias.ll"
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+@_Z8fooAliasv = alias ptr (...), ptr @_Z3foov
+
+;; Original alias is unchanged.
+; CHECK: @_Z8fooAliasv = alias ptr (...), ptr @_Z3foov{{$}}
+;; We create an equivalent alias for the cloned def @_Z3foov.memprof.1.
+; CHECK: @_Z8fooAliasv.memprof.1 = alias ptr (...), ptr @_Z3foov.memprof.1
+
+;; We should also create aliases for the duplicate clones of _Z3foov
+;; (_Z3foov.memprof.2 and _Z3foov.memprof.3) to the versions they are duplicates
+;; of, and ditto for the associated @_Z8fooAliasv clones.
+;;
+;; _Z3foov.memprof.2 is a duplicate of original _Z3foov, and thus so is _Z8fooAliasv.memprof.2
+; CHECK: @_Z3foov.memprof.2 = alias ptr (), ptr @_Z3foov{{$}}
+; CHECK: @_Z8fooAliasv.memprof.2 = alias ptr (...), ptr @_Z3foov{{$}}
+;; _Z3foov.memprof.3 is a duplicate of _Z3foov.memprof.1, and thus so is _Z8fooAliasv.memprof.3
+; CHECK: @_Z3foov.memprof.3 = alias ptr (), ptr @_Z3foov.memprof.1
+; CHECK: @_Z8fooAliasv.memprof.3 = alias ptr (...), ptr @_Z3foov.memprof.1
+
+; CHECK-LABEL: define i32 @main()
+define i32 @main() #0 {
+entry:
+ ;; The first call to bar does not allocate cold memory. It should call
+ ;; the original function, which eventually calls the original allocation
+ ;; decorated with a "notcold" attribute.
+ ; CHECK: call {{.*}} @bar()
+ %call = call ptr @bar(), !callsite !0
+ ;; The second call to bar allocates cold memory. It should call the cloned
+ ;; function which eventually calls a cloned allocation decorated with a
+ ;; "cold" attribute.
+ ; CHECK: call {{.*}} @bar.memprof.1()
+ %call1 = call ptr @bar(), !callsite !1
+ ret i32 0
+}
+
+; CHECK-LABEL: define available_externally i32 @bar()
+define available_externally i32 @bar() #0 {
+entry:
+ ; CHECK: call {{.*}} @_Z8fooAliasv()
+ %call = call ptr @_Z8fooAliasv(), !callsite !8
+ ret i32 0
+}
+
+declare ptr @_Znam(i64)
+
+; CHECK-LABEL: define ptr @_Z3foov()
+define ptr @_Z3foov() #0 {
+entry:
+ ; CHECK: call {{.*}} @_Znam(i64 0) #[[NOTCOLD:[0-9]+]]
+ %call = call ptr @_Znam(i64 0), !memprof !2, !callsite !7
+ ret ptr null
+}
+
+; We create actual clone for bar.memprof.1.
+; CHECK: define available_externally i32 @bar.memprof.1()
+; CHECK: call {{.*}} @_Z3foov.memprof.1()
+
+;; bar.memprof.2 and bar.memprof.3 are duplicates (of original bar and
+;; bar.memprof.1, respectively). However, they are available externally,
+;; so rather than create an alias we simply create a declaration, since the
+;; compiler does not fully support available_externally aliases.
+; CHECK: declare i32 @bar.memprof.2
+; CHECK: declare i32 @bar.memprof.3
+
+; We create actual clone for foo.memprof.1.
+; CHECK: define {{.*}} @_Z3foov.memprof.1()
+; CHECK: call {{.*}} @_Znam(i64 0) #[[COLD:[0-9]+]]
+
+; CHECK: attributes #[[NOTCOLD]] = { "memprof"="notcold" }
+; CHECK: attributes #[[COLD]] = { "memprof"="cold" }
+
+; CHECK: 4 memprof-context-disambiguation - Number of function clone duplicates detected during ThinLTO backend
+; CHECK: 2 memprof-context-disambiguation - Number of function clones created during ThinLTO backend
+
+attributes #0 = { noinline optnone }
+
+!0 = !{i64 8632435727821051414}
+!1 = !{i64 -3421689549917153178}
+!2 = !{!3, !5}
+!3 = !{!4, !"notcold"}
+!4 = !{i64 9086428284934609951, i64 1234, i64 8632435727821051414}
+!5 = !{!6, !"cold"}
+!6 = !{i64 9086428284934609951, i64 1234, i64 -3421689549917153178}
+!7 = !{i64 9086428284934609951}
+!8 = !{i64 1234}
+
+;--- src.o.thinlto.ll
+; ModuleID = 'src.o.thinlto.ll'
+source_filename = "src.o.thinlto.bc"
+
+^0 = module: (path: "src.o", hash: (1720506022, 1575514144, 2506794664, 3599359797, 3160884478))
+^1 = gv: (guid: 6583049656999245004, summaries: (alias: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 1, canAutoHide: 0, importType: definition), aliasee: ^2)))
+;; Summary for _Z3foov, where the allocs part has been manually modified to add
+;; two additional clones that are the same as the prior versions:
+;; ... allocs: ((versions: (notcold, cold, notcold, cold), ...
+^2 = gv: (guid: 9191153033785521275, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 1, canAutoHide: 0, importType: definition), insts: 2, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 0, returnDoesNotAlias: 0, noInline: 1, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 0, mustBeUnreachable: 0), allocs: ((versions: (notcold, cold, notcold, cold), memProf: ((type: notcold, stackIds: (1234, 8632435727821051414)), (type: cold, stackIds: (1234, 15025054523792398438))))))))
+^3 = gv: (guid: 15822663052811949562, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 1, canAutoHide: 0, importType: definition), insts: 3, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 0, returnDoesNotAlias: 0, noInline: 1, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 0, mustBeUnreachable: 0), calls: ((callee: ^4)), callsites: ((callee: ^4, clones: (0), stackIds: (8632435727821051414)), (callee: ^4, clones: (1), stackIds: (15025054523792398438))))))
+;; Summary for bar, where the callsites part has been manually modified to add
+;; two additional clones that are the same as the prior clones:
+;; ... callsites: ((callee: ^1, clones: (0, 1, 0, 1), ...
+^4 = gv: (guid: 16434608426314478903, summaries: (function: (module: ^0, flags: (linkage: available_externally, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 1, canAutoHide: 0, importType: definition), insts: 2, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 0, returnDoesNotAlias: 0, noInline: 1, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 0, mustBeUnreachable: 0), calls: ((callee: ^1)), callsites: ((callee: ^1, clones: (0, 1, 0, 1), stackIds: (1234))))))
+^6 = flags: 353
+^7 = blockcount: 0
diff --git a/llvm/test/ThinLTO/X86/memprof_imported_internal.ll b/llvm/test/ThinLTO/X86/memprof_imported_internal.ll
index a6e254c..09784f8 100644
--- a/llvm/test/ThinLTO/X86/memprof_imported_internal.ll
+++ b/llvm/test/ThinLTO/X86/memprof_imported_internal.ll
@@ -63,14 +63,14 @@
; CHECK: tail call void @_ZL9internal1v.llvm.3267420853450984672()
; CHECK: tail call void @_ZL9internal2v.llvm.3267420853450984672.memprof.1()
; CHECK-LABEL: declare void @_ZL9internal2v.llvm.3267420853450984672.memprof.1()
-;; We should have 2 clones of src2.cc's internal1 function, calling a single
-;; clone of external2.
+;; We should have one clone of src2.cc's internal1 function, calling a single
+;; clone of external2, and a second clone that was detected to be a duplicate
+;; of the first that becomes a declaration (since this is available_externally -
+;; in the module with the prevailing copy it would be an alias to clone 1).
; CHECK-LABEL: define available_externally void @_ZL9internal1v.llvm.3267420853450984672.memprof.1()
; CHECK: tail call void @_Z9external2v.memprof.1()
; CHECK: tail call void @_Z9external2v.memprof.1()
-; CHECK-LABEL: define available_externally void @_ZL9internal1v.llvm.3267420853450984672.memprof.2()
-; CHECK: tail call void @_Z9external2v.memprof.1()
-; CHECK: tail call void @_Z9external2v.memprof.1()
+; CHECK: declare void @_ZL9internal1v.llvm.3267420853450984672.memprof.2()
; CHECK-NOT: memprof
;--- src1.ll
diff --git a/llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/amdgpu_isel.ll.expected b/llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/amdgpu_isel.ll.expected
index bc49669d..eaf9f52 100644
--- a/llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/amdgpu_isel.ll.expected
+++ b/llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/amdgpu_isel.ll.expected
@@ -7,10 +7,10 @@ define i64 @i64_test(i64 %i) nounwind readnone {
; CHECK-NEXT: t0: ch,glue = EntryToken
; CHECK-NEXT: t2: i32,ch = CopyFromReg # D:1 t0, Register:i32 %8
; CHECK-NEXT: t4: i32,ch = CopyFromReg # D:1 t0, Register:i32 %9
-; CHECK-NEXT: t50: i64 = REG_SEQUENCE # D:1 TargetConstant:i32<80>, t2, TargetConstant:i32<3>, t4, TargetConstant:i32<11>
+; CHECK-NEXT: t50: i64 = REG_SEQUENCE # D:1 TargetConstant:i32<76>, t2, TargetConstant:i32<3>, t4, TargetConstant:i32<11>
; CHECK-NEXT: t27: i32,ch = BUFFER_LOAD_DWORD_OFFEN<Mem:(dereferenceable load (s32) from %ir.loc, align 8, addrspace 5)> TargetFrameIndex:i32<0>, Register:v4i32 $sgpr0_sgpr1_sgpr2_sgpr3, TargetConstant:i32<0>, TargetConstant:i32<0>, TargetConstant:i32<0>, TargetConstant:i1<0>, t0
; CHECK-NEXT: t30: i32,ch = BUFFER_LOAD_DWORD_OFFEN<Mem:(dereferenceable load (s32) from %ir.loc + 4, basealign 8, addrspace 5)> TargetFrameIndex:i32<0>, Register:v4i32 $sgpr0_sgpr1_sgpr2_sgpr3, TargetConstant:i32<0>, TargetConstant:i32<4>, TargetConstant:i32<0>, TargetConstant:i1<0>, t0
-; CHECK-NEXT: t33: v2i32 = REG_SEQUENCE # D:1 TargetConstant:i32<80>, t27, TargetConstant:i32<3>, t30, TargetConstant:i32<11>
+; CHECK-NEXT: t33: v2i32 = REG_SEQUENCE # D:1 TargetConstant:i32<76>, t27, TargetConstant:i32<3>, t30, TargetConstant:i32<11>
; CHECK-NEXT: t10: i64 = V_ADD_U64_PSEUDO # D:1 t50, t33
; CHECK-NEXT: t24: i32 = EXTRACT_SUBREG # D:1 t10, TargetConstant:i32<3>
; CHECK-NEXT: t17: ch,glue = CopyToReg # D:1 t0, Register:i32 $vgpr0, t24
diff --git a/llvm/unittests/Support/CMakeLists.txt b/llvm/unittests/Support/CMakeLists.txt
index d1dfb1d..25efa00 100644
--- a/llvm/unittests/Support/CMakeLists.txt
+++ b/llvm/unittests/Support/CMakeLists.txt
@@ -52,6 +52,7 @@ add_llvm_unittest(SupportTests
IndexedAccessorTest.cpp
InstructionCostTest.cpp
InterleavedRangeTest.cpp
+ JobserverTest.cpp
JSONTest.cpp
KnownBitsTest.cpp
LEB128Test.cpp
diff --git a/llvm/unittests/Support/JobserverTest.cpp b/llvm/unittests/Support/JobserverTest.cpp
new file mode 100644
index 0000000..ddee023
--- /dev/null
+++ b/llvm/unittests/Support/JobserverTest.cpp
@@ -0,0 +1,442 @@
+//===- llvm/unittest/Support/JobserverTest.cpp ----------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// Jobserver.h unit tests.
+///
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Support/Jobserver.h"
+#include "llvm/Config/llvm-config.h"
+#include "llvm/Support/Debug.h"
+#include "llvm/Support/Parallel.h"
+#include "llvm/Support/ThreadPool.h"
+#include "llvm/Support/raw_ostream.h"
+#include "gtest/gtest.h"
+#include <future>
+#include <random>
+#include <stdlib.h>
+
+#if defined(LLVM_ON_UNIX)
+#include "llvm/ADT/SmallString.h"
+#include "llvm/Support/FileSystem.h"
+#include <atomic>
+#include <condition_variable>
+#include <fcntl.h>
+#include <mutex>
+#include <sys/stat.h>
+#include <thread>
+#include <unistd.h>
+#elif defined(_WIN32)
+#include <windows.h>
+#endif
+
+#define DEBUG_TYPE "jobserver-test"
+
+using namespace llvm;
+
+namespace {
+
+// RAII helper to set an environment variable for the duration of a test.
+class ScopedEnvironment {
+ std::string Name;
+ std::string OldValue;
+ bool HadOldValue;
+
+public:
+ ScopedEnvironment(const char *Name, const char *Value) : Name(Name) {
+#if defined(_WIN32)
+ char *Old = nullptr;
+ size_t OldLen;
+ errno_t err = _dupenv_s(&Old, &OldLen, Name);
+ if (err == 0 && Old != nullptr) {
+ HadOldValue = true;
+ OldValue = Old;
+ free(Old);
+ } else {
+ HadOldValue = false;
+ }
+ _putenv_s(Name, Value);
+#else
+ const char *Old = getenv(Name);
+ if (Old) {
+ HadOldValue = true;
+ OldValue = Old;
+ } else {
+ HadOldValue = false;
+ }
+ setenv(Name, Value, 1);
+#endif
+ }
+
+ ~ScopedEnvironment() {
+#if defined(_WIN32)
+ if (HadOldValue)
+ _putenv_s(Name.c_str(), OldValue.c_str());
+ else
+ // On Windows, setting an environment variable to an empty string
+ // unsets it, making getenv() return NULL.
+ _putenv_s(Name.c_str(), "");
+#else
+ if (HadOldValue)
+ setenv(Name.c_str(), OldValue.c_str(), 1);
+ else
+ unsetenv(Name.c_str());
+#endif
+ }
+};
+
+TEST(Jobserver, Slot) {
+ // Default constructor creates an invalid slot.
+ JobSlot S1;
+ EXPECT_FALSE(S1.isValid());
+ EXPECT_FALSE(S1.isImplicit());
+
+ // Create an implicit slot.
+ JobSlot S2 = JobSlot::createImplicit();
+ EXPECT_TRUE(S2.isValid());
+ EXPECT_TRUE(S2.isImplicit());
+
+ // Create an explicit slot.
+ JobSlot S3 = JobSlot::createExplicit(42);
+ EXPECT_TRUE(S3.isValid());
+ EXPECT_FALSE(S3.isImplicit());
+
+ // Test move construction.
+ JobSlot S4 = std::move(S2);
+ EXPECT_TRUE(S4.isValid());
+ EXPECT_TRUE(S4.isImplicit());
+ EXPECT_FALSE(S2.isValid()); // S2 is now invalid.
+
+ // Test move assignment.
+ S1 = std::move(S3);
+ EXPECT_TRUE(S1.isValid());
+ EXPECT_FALSE(S1.isImplicit());
+ EXPECT_FALSE(S3.isValid()); // S3 is now invalid.
+}
+
+// Test fixture for parsing tests to ensure the singleton state is
+// reset between each test case.
+class JobserverParsingTest : public ::testing::Test {
+protected:
+ void TearDown() override { JobserverClient::resetForTesting(); }
+};
+
+TEST_F(JobserverParsingTest, NoMakeflags) {
+ // No MAKEFLAGS, should be null.
+ ScopedEnvironment Env("MAKEFLAGS", "");
+ // On Unix, setting an env var to "" makes getenv() return an empty
+ // string, not NULL. We must call unsetenv() to test the case where
+ // the variable is truly not present.
+#if !defined(_WIN32)
+ unsetenv("MAKEFLAGS");
+#endif
+ EXPECT_EQ(JobserverClient::getInstance(), nullptr);
+}
+
+TEST_F(JobserverParsingTest, EmptyMakeflags) {
+ // Empty MAKEFLAGS, should be null.
+ ScopedEnvironment Env("MAKEFLAGS", "");
+ EXPECT_EQ(JobserverClient::getInstance(), nullptr);
+}
+
+TEST_F(JobserverParsingTest, DryRunFlag) {
+ // Dry-run flag 'n', should be null.
+ ScopedEnvironment Env("MAKEFLAGS", "n -j --jobserver-auth=fifo:/tmp/foo");
+ EXPECT_EQ(JobserverClient::getInstance(), nullptr);
+}
+
+// Separate fixture for non-threaded client tests.
+class JobserverClientTest : public JobserverParsingTest {};
+
+#if defined(LLVM_ON_UNIX)
+// RAII helper to create and clean up a temporary FIFO file.
+class ScopedFifo {
+ SmallString<128> Path;
+ bool IsValid = false;
+
+public:
+ ScopedFifo() {
+ // To get a unique, non-colliding name for a FIFO, we use the
+ // createTemporaryFile function to reserve a name in the filesystem.
+ std::error_code EC =
+ sys::fs::createTemporaryFile("jobserver-test", "fifo", Path);
+ if (EC)
+ return;
+ // Then we immediately remove the regular file it created, but keep the
+ // unique path.
+ sys::fs::remove(Path);
+ // Finally, we create the FIFO at that safe, unique path.
+ if (mkfifo(Path.c_str(), 0600) != 0)
+ return;
+ IsValid = true;
+ }
+
+ ~ScopedFifo() {
+ if (IsValid)
+ sys::fs::remove(Path);
+ }
+
+ const char *c_str() const { return Path.data(); }
+ bool isValid() const { return IsValid; }
+};
+
+TEST_F(JobserverClientTest, UnixClientFifo) {
+ // This test covers basic FIFO client creation and behavior with an empty
+ // FIFO. No job tokens are available.
+ ScopedFifo F;
+ ASSERT_TRUE(F.isValid());
+
+ // Intentionally inserted \t in environment string.
+ std::string Makeflags = " \t -j4\t \t--jobserver-auth=fifo:";
+ Makeflags += F.c_str();
+ ScopedEnvironment Env("MAKEFLAGS", Makeflags.c_str());
+
+ JobserverClient *Client = JobserverClient::getInstance();
+ ASSERT_NE(Client, nullptr);
+
+ // Get the implicit token.
+ JobSlot S1 = Client->tryAcquire();
+ EXPECT_TRUE(S1.isValid());
+ EXPECT_TRUE(S1.isImplicit());
+
+ // FIFO is empty, next acquire fails.
+ JobSlot S2 = Client->tryAcquire();
+ EXPECT_FALSE(S2.isValid());
+
+ // Release does not write to the pipe for the implicit token.
+ Client->release(std::move(S1));
+
+ // Re-acquire the implicit token.
+ S1 = Client->tryAcquire();
+ EXPECT_TRUE(S1.isValid());
+}
+
+#if LLVM_ENABLE_THREADS
+// Test fixture for tests that use the jobserver strategy. It creates a
+// temporary FIFO, sets MAKEFLAGS, and provides a helper to pre-load the FIFO
+// with job tokens, simulating `make -jN`.
+class JobserverStrategyTest : public JobserverParsingTest {
+protected:
+ std::unique_ptr<ScopedFifo> TheFifo;
+ std::thread MakeThread;
+ std::atomic<bool> StopMakeThread{false};
+ // Save and restore the global parallel strategy to avoid interfering with
+ // other tests in the same process.
+ ThreadPoolStrategy SavedStrategy;
+
+ void SetUp() override {
+ SavedStrategy = parallel::strategy;
+ TheFifo = std::make_unique<ScopedFifo>();
+ ASSERT_TRUE(TheFifo->isValid());
+
+ std::string MakeFlags = "--jobserver-auth=fifo:";
+ MakeFlags += TheFifo->c_str();
+ setenv("MAKEFLAGS", MakeFlags.c_str(), 1);
+ }
+
+ void TearDown() override {
+ if (MakeThread.joinable()) {
+ StopMakeThread = true;
+ MakeThread.join();
+ }
+ unsetenv("MAKEFLAGS");
+ TheFifo.reset();
+ // Restore the original strategy to ensure subsequent tests are unaffected.
+ parallel::strategy = SavedStrategy;
+ }
+
+ // Starts a background thread that emulates `make`. It populates the FIFO
+ // with initial tokens and then recycles tokens released by clients.
+ void startMakeProxy(int NumInitialJobs) {
+ MakeThread = std::thread([this, NumInitialJobs]() {
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Thread started.\n");
+ // Open the FIFO for reading and writing. This call does not block.
+ int RWFd = open(TheFifo->c_str(), O_RDWR);
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Opened FIFO " << TheFifo->c_str()
+ << " with O_RDWR, FD=" << RWFd << "\n");
+ if (RWFd == -1) {
+ LLVM_DEBUG(
+ dbgs()
+ << "[MakeProxy] ERROR: Failed to open FIFO with O_RDWR. Errno: "
+ << errno << "\n");
+ return;
+ }
+
+ // Populate with initial jobs.
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Writing " << NumInitialJobs
+ << " initial tokens.\n");
+ for (int i = 0; i < NumInitialJobs; ++i) {
+ if (write(RWFd, "+", 1) != 1) {
+ LLVM_DEBUG(dbgs()
+ << "[MakeProxy] ERROR: Failed to write initial token " << i
+ << ".\n");
+ close(RWFd);
+ return;
+ }
+ }
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Finished writing initial tokens.\n");
+
+ // Make the read non-blocking so we can periodically check StopMakeThread.
+ int flags = fcntl(RWFd, F_GETFL, 0);
+ fcntl(RWFd, F_SETFL, flags | O_NONBLOCK);
+
+ while (!StopMakeThread) {
+ char Token;
+ ssize_t Ret = read(RWFd, &Token, 1);
+ if (Ret == 1) {
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Read token '" << Token
+ << "' to recycle.\n");
+ // A client released a token, 'make' makes it available again.
+ std::this_thread::sleep_for(std::chrono::microseconds(100));
+ ssize_t WRet;
+ do {
+ WRet = write(RWFd, &Token, 1);
+ } while (WRet < 0 && errno == EINTR);
+ if (WRet <= 0) {
+ LLVM_DEBUG(
+ dbgs()
+ << "[MakeProxy] ERROR: Failed to write recycled token.\n");
+ break; // Error, stop the proxy.
+ }
+ LLVM_DEBUG(dbgs()
+ << "[MakeProxy] Wrote token '" << Token << "' back.\n");
+ } else if (Ret < 0 && errno != EAGAIN && errno != EWOULDBLOCK) {
+ LLVM_DEBUG(dbgs() << "[MakeProxy] ERROR: Read failed with errno "
+ << errno << ".\n");
+ break; // Error, stop the proxy.
+ }
+ // Yield to prevent this thread from busy-waiting.
+ std::this_thread::sleep_for(std::chrono::milliseconds(1));
+ }
+ LLVM_DEBUG(dbgs() << "[MakeProxy] Thread stopping.\n");
+ close(RWFd);
+ });
+
+ // Give the proxy thread a moment to start and populate the FIFO.
+ // This is a simple way to avoid a race condition where the client starts
+ // before the initial tokens are in the pipe.
+ std::this_thread::sleep_for(std::chrono::milliseconds(50));
+ }
+};
+
+TEST_F(JobserverStrategyTest, ThreadPoolConcurrencyIsLimited) {
+ // This test simulates `make -j3`. We will have 1 implicit job slot and
+ // we will add 2 explicit job tokens to the FIFO, for a total of 3.
+ const int NumExplicitJobs = 2;
+ const int ConcurrencyLimit = NumExplicitJobs + 1; // +1 for the implicit slot
+ const int NumTasks = 8; // More tasks than available slots.
+
+ LLVM_DEBUG(dbgs() << "Calling startMakeProxy with " << NumExplicitJobs
+ << " jobs.\n");
+ startMakeProxy(NumExplicitJobs);
+ LLVM_DEBUG(dbgs() << "MakeProxy is running.\n");
+
+ // Create the thread pool. Its constructor will call jobserver_concurrency()
+ // and create a client that reads from our pre-loaded FIFO.
+ StdThreadPool Pool(jobserver_concurrency());
+
+ std::atomic<int> ActiveTasks{0};
+ std::atomic<int> MaxActiveTasks{0};
+ std::atomic<int> CompletedTasks{0};
+ std::mutex M;
+ std::condition_variable CV;
+
+ // Dispatch more tasks than there are job slots. The pool should block
+ // and only run up to `ConcurrencyLimit` tasks at once.
+ for (int i = 0; i < NumTasks; ++i) {
+ Pool.async([&, i] {
+ // Track the number of concurrently running tasks.
+ int CurrentActive = ++ActiveTasks;
+ LLVM_DEBUG(dbgs() << "Task " << i << ": Active tasks: " << CurrentActive
+ << "\n");
+ int OldMax = MaxActiveTasks.load();
+ while (CurrentActive > OldMax)
+ MaxActiveTasks.compare_exchange_weak(OldMax, CurrentActive);
+
+ std::this_thread::sleep_for(std::chrono::milliseconds(25));
+
+ --ActiveTasks;
+ if (++CompletedTasks == NumTasks) {
+ std::lock_guard<std::mutex> Lock(M);
+ CV.notify_one();
+ }
+ });
+ }
+
+ // Wait for all tasks to complete.
+ std::unique_lock<std::mutex> Lock(M);
+ CV.wait(Lock, [&] { return CompletedTasks == NumTasks; });
+
+ LLVM_DEBUG(dbgs() << "Test finished. Max active tasks was " << MaxActiveTasks
+ << ".\n");
+ // The key assertion: the maximum number of concurrent tasks should
+ // not have exceeded the limit imposed by the jobserver.
+ EXPECT_LE(MaxActiveTasks, ConcurrencyLimit);
+ EXPECT_EQ(CompletedTasks, NumTasks);
+}
+
+TEST_F(JobserverStrategyTest, ParallelForIsLimited) {
+ // This test verifies that llvm::parallelFor respects the jobserver limit.
+ const int NumExplicitJobs = 3;
+ const int ConcurrencyLimit = NumExplicitJobs + 1; // +1 implicit
+ const int NumTasks = 20;
+
+ LLVM_DEBUG(dbgs() << "Calling startMakeProxy with " << NumExplicitJobs
+ << " jobs.\n");
+ startMakeProxy(NumExplicitJobs);
+ LLVM_DEBUG(dbgs() << "MakeProxy is running.\n");
+
+ // Set the global strategy. parallelFor will use this.
+ parallel::strategy = jobserver_concurrency();
+
+ std::atomic<int> ActiveTasks{0};
+ std::atomic<int> MaxActiveTasks{0};
+
+ parallelFor(0, NumTasks, [&](int i) {
+ int CurrentActive = ++ActiveTasks;
+ LLVM_DEBUG(dbgs() << "Task " << i << ": Active tasks: " << CurrentActive
+ << "\n");
+ int OldMax = MaxActiveTasks.load();
+ while (CurrentActive > OldMax)
+ MaxActiveTasks.compare_exchange_weak(OldMax, CurrentActive);
+
+ std::this_thread::sleep_for(std::chrono::milliseconds(20));
+ --ActiveTasks;
+ });
+
+ LLVM_DEBUG(dbgs() << "ParallelFor finished. Max active tasks was "
+ << MaxActiveTasks << ".\n");
+ EXPECT_LE(MaxActiveTasks, ConcurrencyLimit);
+}
+
+TEST_F(JobserverStrategyTest, ParallelSortIsLimited) {
+ // This test serves as an integration test to ensure parallelSort completes
+ // correctly when running under the jobserver strategy. It doesn't directly
+ // measure concurrency but verifies correctness.
+ const int NumExplicitJobs = 3;
+ startMakeProxy(NumExplicitJobs);
+
+ parallel::strategy = jobserver_concurrency();
+
+ std::vector<int> V(1024);
+ // Fill with random data
+ std::mt19937 randEngine;
+ std::uniform_int_distribution<int> dist;
+ for (int &i : V)
+ i = dist(randEngine);
+
+ parallelSort(V.begin(), V.end());
+ ASSERT_TRUE(llvm::is_sorted(V));
+}
+
+#endif // LLVM_ENABLE_THREADS
+
+#endif // defined(LLVM_ON_UNIX)
+
+} // end anonymous namespace
diff --git a/llvm/utils/TableGen/SubtargetEmitter.cpp b/llvm/utils/TableGen/SubtargetEmitter.cpp
index 0f42d49..b0a309c 100644
--- a/llvm/utils/TableGen/SubtargetEmitter.cpp
+++ b/llvm/utils/TableGen/SubtargetEmitter.cpp
@@ -1761,7 +1761,7 @@ void SubtargetEmitter::emitSchedModelHelpers(const std::string &ClassName,
<< "\n::resolveVariantSchedClass(unsigned SchedClass, const MCInst *MI,"
<< " const MCInstrInfo *MCII, unsigned CPUID) const {\n"
<< " return " << Target << "_MC"
- << "::resolveVariantSchedClassImpl(SchedClass, MI, MCII, CPUID);\n"
+ << "::resolveVariantSchedClassImpl(SchedClass, MI, MCII, *this, CPUID);\n"
<< "} // " << ClassName << "::resolveVariantSchedClass\n\n";
STIPredicateExpander PE(Target, /*Indent=*/0);
@@ -1923,7 +1923,8 @@ void SubtargetEmitter::parseFeaturesFunction(raw_ostream &OS) {
void SubtargetEmitter::emitGenMCSubtargetInfo(raw_ostream &OS) {
OS << "namespace " << Target << "_MC {\n"
<< "unsigned resolveVariantSchedClassImpl(unsigned SchedClass,\n"
- << " const MCInst *MI, const MCInstrInfo *MCII, unsigned CPUID) {\n";
+ << " const MCInst *MI, const MCInstrInfo *MCII, "
+ << "const MCSubtargetInfo &STI, unsigned CPUID) {\n";
emitSchedModelHelpersImpl(OS, /* OnlyExpandMCPredicates */ true);
OS << "}\n";
OS << "} // end namespace " << Target << "_MC\n\n";
@@ -1945,7 +1946,7 @@ void SubtargetEmitter::emitGenMCSubtargetInfo(raw_ostream &OS) {
<< " const MCInst *MI, const MCInstrInfo *MCII,\n"
<< " unsigned CPUID) const override {\n"
<< " return " << Target << "_MC"
- << "::resolveVariantSchedClassImpl(SchedClass, MI, MCII, CPUID);\n";
+ << "::resolveVariantSchedClassImpl(SchedClass, MI, MCII, *this, CPUID);\n";
OS << " }\n";
if (TGT.getHwModes().getNumModeIds() > 1) {
OS << " unsigned getHwModeSet() const override;\n";
@@ -2073,7 +2074,8 @@ void SubtargetEmitter::run(raw_ostream &OS) {
OS << "class DFAPacketizer;\n";
OS << "namespace " << Target << "_MC {\n"
<< "unsigned resolveVariantSchedClassImpl(unsigned SchedClass,"
- << " const MCInst *MI, const MCInstrInfo *MCII, unsigned CPUID);\n"
+ << " const MCInst *MI, const MCInstrInfo *MCII, "
+ << "const MCSubtargetInfo &STI, unsigned CPUID);\n"
<< "} // end namespace " << Target << "_MC\n\n";
OS << "struct " << ClassName << " : public TargetSubtargetInfo {\n"
<< " explicit " << ClassName << "(const Triple &TT, StringRef CPU, "
diff --git a/llvm/utils/gn/secondary/llvm/lib/Support/BUILD.gn b/llvm/utils/gn/secondary/llvm/lib/Support/BUILD.gn
index 6ca766ca..38ba466 100644
--- a/llvm/utils/gn/secondary/llvm/lib/Support/BUILD.gn
+++ b/llvm/utils/gn/secondary/llvm/lib/Support/BUILD.gn
@@ -103,6 +103,7 @@ static_library("Support") {
"IntEqClasses.cpp",
"IntervalMap.cpp",
"JSON.cpp",
+ "Jobserver.cpp",
"KnownBits.cpp",
"KnownFPClass.cpp",
"LEB128.cpp",
diff --git a/llvm/utils/gn/secondary/llvm/unittests/Support/BUILD.gn b/llvm/utils/gn/secondary/llvm/unittests/Support/BUILD.gn
index 42c1a15..a25f058 100644
--- a/llvm/utils/gn/secondary/llvm/unittests/Support/BUILD.gn
+++ b/llvm/utils/gn/secondary/llvm/unittests/Support/BUILD.gn
@@ -56,6 +56,7 @@ unittest("SupportTests") {
"InstructionCostTest.cpp",
"InterleavedRangeTest.cpp",
"JSONTest.cpp",
+ "JobserverTest.cpp",
"KnownBitsTest.cpp",
"LEB128Test.cpp",
"LineIteratorTest.cpp",