diff --git a/mlir/include/mlir/IR/MLIRContext.h b/mlir/include/mlir/IR/MLIRContext.h
index e2014954e9fb11..659cd712655927 100644
--- a/mlir/include/mlir/IR/MLIRContext.h
+++ b/mlir/include/mlir/IR/MLIRContext.h
@@ -147,6 +147,13 @@ class MLIRContext {
   /// this call in this case.
   void setThreadPool(llvm::ThreadPool &pool);
 
+  /// Return the number of threads used by the thread pool in this context. The
+  /// number of computed hardware threads can change over the lifetime of a
+  /// process based on affinity changes, so users should use the number of
+  /// threads actually in the thread pool for dispatching work. Returns 1 if
+  /// multithreading is disabled.
+  unsigned getNumThreads();
+
   /// Return the thread pool used by this context. This method requires that
   /// multithreading be enabled within the context, and should generally not be
   /// used directly. Users should instead prefer the threading utilities within
diff --git a/mlir/lib/IR/MLIRContext.cpp b/mlir/lib/IR/MLIRContext.cpp
index 96fe76eaac28da..7e811316c4e6f0 100644
--- a/mlir/lib/IR/MLIRContext.cpp
+++ b/mlir/lib/IR/MLIRContext.cpp
@@ -518,6 +518,16 @@ void MLIRContext::setThreadPool(llvm::ThreadPool &pool) {
   enableMultithreading();
 }
 
+unsigned MLIRContext::getNumThreads() {
+  if (isMultithreadingEnabled()) {
+    assert(impl->threadPool &&
+           "multi-threading is enabled but threadpool not set");
+    return impl->threadPool->getThreadCount();
+  }
+  // No multithreading or active thread pool. Return 1 thread.
+  return 1;
+}
+
 llvm::ThreadPool &MLIRContext::getThreadPool() {
   assert(isMultithreadingEnabled() &&
          "expected multi-threading to be enabled within the context");
diff --git a/mlir/lib/Transforms/Inliner.cpp b/mlir/lib/Transforms/Inliner.cpp
index cbeb94980dccaa..f080b4d112ba5f 100644
--- a/mlir/lib/Transforms/Inliner.cpp
+++ b/mlir/lib/Transforms/Inliner.cpp
@@ -679,8 +679,7 @@ InlinerPass::optimizeSCCAsync(MutableArrayRef<CallGraphNode *> nodesToVisit,
   // Note: The number of pass managers here needs to remain constant
   // to prevent issues with pass instrumentations that rely on having the same
   // pass manager for the main thread.
-  llvm::ThreadPool &threadPool = ctx->getThreadPool();
-  size_t numThreads = threadPool.getThreadCount();
+  size_t numThreads = ctx->getNumThreads();
   if (opPipelines.size() < numThreads) {
     // Reserve before resizing so that we can use a reference to the first
     // element.
diff --git a/mlir/test/Transforms/inlining.mlir b/mlir/test/Transforms/inlining.mlir
index 98692853bf24ef..00a0445f5a88d9 100644
--- a/mlir/test/Transforms/inlining.mlir
+++ b/mlir/test/Transforms/inlining.mlir
@@ -1,4 +1,5 @@
 // RUN: mlir-opt %s -inline='default-pipeline=''' | FileCheck %s
+// RUN: mlir-opt %s --mlir-disable-threading -inline='default-pipeline=''' | FileCheck %s
 // RUN: mlir-opt %s -inline='default-pipeline=''' -mlir-print-debuginfo -mlir-print-local-scope | FileCheck %s --check-prefix INLINE-LOC
 // RUN: mlir-opt %s -inline | FileCheck %s --check-prefix INLINE_SIMPLIFY