From b32e435e50ad90a376d33097cb2a660227a6abd7 Mon Sep 17 00:00:00 2001
From: Jan Vorlicek <janvorli@microsoft.com>
Date: Thu, 7 Aug 2025 09:56:36 +0200
Subject: [PATCH] Fix interpreter helper and stack size computation

This change fixes two issues:
* INTOP_CALL_HELPER_P_S was taking arguments from wrong locations in the instruction
* The total stack size computation in the call stub generator was incorrect in case
  arguments smaller than 8 bytes were passed on the stack.
---
 src/coreclr/vm/callstubgenerator.cpp | 6 +-----
 src/coreclr/vm/interpexec.cpp        | 4 ++--
 2 files changed, 3 insertions(+), 7 deletions(-)

diff --git a/src/coreclr/vm/callstubgenerator.cpp b/src/coreclr/vm/callstubgenerator.cpp
index 10855af6bc402f..02f275d9b9506a 100644
--- a/src/coreclr/vm/callstubgenerator.cpp
+++ b/src/coreclr/vm/callstubgenerator.cpp
@@ -1416,7 +1416,7 @@ void CallStubGenerator::ComputeCallStub(MetaSig &sig, PCODE *pRoutines)
     m_s1 = NoRange; // indicates that there is no active range of stack arguments
     m_s2 = 0;
     m_routineIndex = 0;
-    m_totalStackSize = 0;
+    m_totalStackSize = argIt.SizeOfArgStack();
 #if LOG_COMPUTE_CALL_STUB
     printf("ComputeCallStub\n");
 #endif
@@ -1527,7 +1527,6 @@ void CallStubGenerator::ComputeCallStub(MetaSig &sig, PCODE *pRoutines)
     }
     else if (m_s1 != NoRange)
     {
-        m_totalStackSize += m_s2 - m_s1 + 1;
         pRoutines[m_routineIndex++] = GetStackRoutine();
         pRoutines[m_routineIndex++] = ((int64_t)(m_s2 - m_s1 + 1) << 32) | m_s1;
     }
@@ -1571,7 +1570,6 @@ void CallStubGenerator::ProcessArgument(ArgIterator *pArgIt, ArgLocDesc& argLocD
     {
         // No stack argument is used to pass the current argument, but we already have a range of stack arguments,
         // store the routine for the range
-        m_totalStackSize += m_s2 - m_s1 + 1;
         pRoutines[m_routineIndex++] = GetStackRoutine();
         pRoutines[m_routineIndex++] = ((int64_t)(m_s2 - m_s1 + 1) << 32) | m_s1;
         m_s1 = NoRange;
@@ -1650,7 +1648,6 @@ void CallStubGenerator::ProcessArgument(ArgIterator *pArgIt, ArgLocDesc& argLocD
         else
         {
             // Discontinuous range - store a routine for the current and start a new one
-            m_totalStackSize += m_s2 - m_s1 + 1;
             pRoutines[m_routineIndex++] = GetStackRoutine();
             pRoutines[m_routineIndex++] = ((int64_t)(m_s2 - m_s1 + 1) << 32) | m_s1;
             m_s1 = argLocDesc.m_byteStackIndex;
@@ -1699,7 +1696,6 @@ void CallStubGenerator::ProcessArgument(ArgIterator *pArgIt, ArgLocDesc& argLocD
             _ASSERTE(argLocDesc.m_byteStackIndex != -1);
             pRoutines[m_routineIndex++] = GetStackRefRoutine();
             pRoutines[m_routineIndex++] = ((int64_t)pArgIt->GetArgSize() << 32) | argLocDesc.m_byteStackIndex;
-            m_totalStackSize += argLocDesc.m_byteStackSize;
             m_s1 = NoRange;
         }
     }
diff --git a/src/coreclr/vm/interpexec.cpp b/src/coreclr/vm/interpexec.cpp
index 3419a19710d3b3..881b58c4c3a469 100644
--- a/src/coreclr/vm/interpexec.cpp
+++ b/src/coreclr/vm/interpexec.cpp
@@ -1721,8 +1721,8 @@ void InterpExecMethod(InterpreterFrame *pInterpreterFrame, InterpMethodContextFr
 
                 case INTOP_CALL_HELPER_P_S:
                 {
-                    HELPER_FTN_P_P helperFtn = GetPossiblyIndirectHelper<HELPER_FTN_P_P>(pMethod, ip[2]);
-                    void* helperArg = LOCAL_VAR(ip[3], void*);
+                    HELPER_FTN_P_P helperFtn = GetPossiblyIndirectHelper<HELPER_FTN_P_P>(pMethod, ip[3]);
+                    void* helperArg = LOCAL_VAR(ip[2], void*);
 
                     LOCAL_VAR(ip[1], void*) = helperFtn(helperArg);
                     ip += 4;