From 069555a807ef4d8d027ac3efa314c20398a2ed61 Mon Sep 17 00:00:00 2001
From: Lei Zhang <antiagainst@gmail.com>
Date: Sat, 6 Dec 2025 05:37:39 +0000
Subject: [PATCH] [Backend] Bump to llvm/llvm-project@a992f294

This pulls in the following fixes:

* https://github.com/llvm/llvm-project/pull/168801
* https://github.com/llvm/llvm-project/pull/168805
* https://github.com/llvm/llvm-project/pull/169851

Along the way, we need to adjust `TMEMLoadOpPattern`
and `RelayoutTritonGPU` due to

* https://github.com/llvm/llvm-project/pull/169606
---
 cmake/llvm-hash.txt                               |  2 +-
 .../TritonToTritonGPU/RelayoutTritonGPU.cpp       | 15 ++++++++++++---
 third_party/amd/backend/compiler.py               |  5 -----
 3 files changed, 13 insertions(+), 9 deletions(-)

diff --git a/cmake/llvm-hash.txt b/cmake/llvm-hash.txt
index be0c83e16a88..6bdc3d72df1f 100644
--- a/cmake/llvm-hash.txt
+++ b/cmake/llvm-hash.txt
@@ -1 +1 @@
-ebf5d9ef7de29b55fd9e9d504f83689b4013e0de
+a992f29451b9e140424f35ac5e20177db4afbdc0
diff --git a/lib/Conversion/TritonToTritonGPU/RelayoutTritonGPU.cpp b/lib/Conversion/TritonToTritonGPU/RelayoutTritonGPU.cpp
index 7e8aeb24b7a7..7ea2e94a9914 100644
--- a/lib/Conversion/TritonToTritonGPU/RelayoutTritonGPU.cpp
+++ b/lib/Conversion/TritonToTritonGPU/RelayoutTritonGPU.cpp
@@ -34,14 +34,20 @@ struct TMEMLoadOpPattern : public OpConversionPattern<ttng::TMEMLoadOp> {
   LogicalResult
   matchAndRewrite(ttng::TMEMLoadOp op, OpAdaptor adaptor,
                   ConversionPatternRewriter &rewriter) const override {
+    Type resultType = getTypeConverter()->convertType(op.getType());
     RankedTensorType type = getTMEMTensorLayout(
         typeConverter, op.getType(), op.getSrc().getType(), lookupNumWarps(op));
     rewriter.modifyOpInPlace(op, [&] { op.getResult().setType(type); });
-    Type resultType = getTypeConverter()->convertType(op.getType());
+    if (type == resultType)
+      return success();
+
     rewriter.setInsertionPointAfter(op);
     auto cvt = ConvertLayoutOp::create(rewriter, op.getLoc(), resultType,
                                        op.getResult());
-    rewriter.replaceAllUsesExcept(op.getResult(), cvt, cvt);
+    // Bypass the rewriter to avoid issues with the conversion framework's
+    // tracking of conditional replacements.
+    // See https://github.com/llvm/llvm-project/commit/504b50789602
+    op.getResult().replaceAllUsesExcept(cvt, cvt);
     return success();
   }
 };
@@ -115,7 +121,10 @@ class RelayoutTritonGPU
         // clang-format on
         >(typeConverter, context);
 
-    if (failed(applyPartialConversion(mod, target, std::move(patterns))))
+    ConversionConfig config;
+    config.allowPatternRollback = false;
+    if (failed(
+            applyPartialConversion(mod, target, std::move(patterns), config)))
       return signalPassFailure();
   }
 };
diff --git a/third_party/amd/backend/compiler.py b/third_party/amd/backend/compiler.py
index 1db2b2cb30a1..bd7a2cc537d1 100644
--- a/third_party/amd/backend/compiler.py
+++ b/third_party/amd/backend/compiler.py
@@ -451,11 +451,6 @@ def make_amdgcn(src, metadata, options):
                             dump_file_id)
         amdgcn = llvm.translate_to_asm(src, amd.TARGET_TRIPLE, options.arch, features, flags, options.enable_fp_fusion,
                                        False)
-        # TODO: Remove the following workaround once LLVM is bumped to include: https://github.com/llvm/llvm-project/pull/169851
-        # Workaround for LLVM ERROR: cannot evaluate equated symbol 'amdgcn.device.init.num_named_barrier'
-        if knobs.compilation.enable_asan and 'gfx1250' not in options.arch:
-            amdgcn = amdgcn.replace('.amdgpu_metadata',
-                                    '\t.set\tamdgcn.device.init.num_named_barrier, 0\n.amdgpu_metadata')
         if knobs.amd.dump_amdgcn:
             print("// -----// AMDGCN Dump //----- //")
             print(amdgcn)