diff --git a/dll/win32/kernel32/client/utils.c b/dll/win32/kernel32/client/utils.c
index 047fbec9fc2..32df2e1a0b6 100644
--- a/dll/win32/kernel32/client/utils.c
+++ b/dll/win32/kernel32/client/utils.c
@@ -588,7 +588,7 @@ BaseInitializeContext(IN PCONTEXT Context,
     /* Setup the Initial Win32 Thread Context */
     Context->Rcx = (ULONG_PTR)StartAddress;
     Context->Rdx = (ULONG_PTR)Parameter;
-    Context->Rsp = (ULONG_PTR)StackAddress - 5 * sizeof(PVOID);
+    Context->Rsp = (ULONG_PTR)StackAddress;
 
     /* Setup the Segments */
     Context->SegGs = KGDT64_R3_DATA | RPL_MASK;
diff --git a/ntoskrnl/ke/amd64/thrdini.c b/ntoskrnl/ke/amd64/thrdini.c
index 612e1252df6..e6518e0bf82 100644
--- a/ntoskrnl/ke/amd64/thrdini.c
+++ b/ntoskrnl/ke/amd64/thrdini.c
@@ -116,10 +116,15 @@ KiInitializeContextThread(IN PKTHREAD Thread,
                              CONTEXT_AMD64 | ContextFlags,
                              UserMode);
 
-        /* Set SS, DS, ES's RPL Mask properly */
-        TrapFrame->SegSs |= RPL_MASK;
-        TrapFrame->SegDs |= RPL_MASK;
-        TrapFrame->SegEs |= RPL_MASK;
+        /* Set user mode segment selectors */
+        TrapFrame->SegDs = KGDT64_R3_DATA | RPL_MASK;
+        TrapFrame->SegEs = KGDT64_R3_DATA | RPL_MASK;
+        TrapFrame->SegFs = KGDT64_R3_CMTEB | RPL_MASK;
+        TrapFrame->SegGs = KGDT64_R3_DATA | RPL_MASK;
+        TrapFrame->SegCs = KGDT64_R3_CODE | RPL_MASK;
+        TrapFrame->SegSs = KGDT64_R3_DATA | RPL_MASK;
+
+        /* Clear DR7 */
         TrapFrame->Dr7 = 0;
 
         /* Set the previous mode as user */
@@ -133,6 +138,9 @@ KiInitializeContextThread(IN PKTHREAD Thread,
 
         /* KiUserThreadStartupExit returns to KiServiceExit3 */
         InitFrame->ExceptionFrame.Return = (ULONG64)KiServiceExit3;
+
+        /* Allocate home space on the stack */
+        TrapFrame->Rsp -= 5 * sizeof(PVOID);
     }
     else
     {
diff --git a/sdk/lib/rtl/amd64/stubs.c b/sdk/lib/rtl/amd64/stubs.c
index 1b77d0d2b15..b68e41e5060 100644
--- a/sdk/lib/rtl/amd64/stubs.c
+++ b/sdk/lib/rtl/amd64/stubs.c
@@ -27,16 +27,15 @@ RtlInitializeContext(
     _In_ PTHREAD_START_ROUTINE ThreadStartAddress,
     _In_ PVOID StackBase)
 {
-    /* Initialize everything to 0 */
-    RtlZeroMemory(ThreadContext, sizeof(*ThreadContext));
+    /* Make sure the stack is aligned */
+    if (((ULONG_PTR)StackBase & 0xF) != 0)
+    {
+        RtlRaiseStatus(STATUS_BAD_INITIAL_STACK);
+    }
 
     /* Initialize StartAddress and Stack */
     ThreadContext->Rip = (ULONG64)ThreadStartAddress;
-    ThreadContext->Rsp = (ULONG64)StackBase - 6 * sizeof(PVOID);
-
-    /* Align stack by 16 and substract 8 (unaligned on function entry) */
-    ThreadContext->Rsp &= ~15;
-    ThreadContext->Rsp -= 8;
+    ThreadContext->Rsp = (ULONG64)StackBase;
 
     /* Enable Interrupts */
     ThreadContext->EFlags = EFLAGS_INTERRUPT_MASK;
@@ -44,35 +43,29 @@ RtlInitializeContext(
     /* Set start parameter */
     ThreadContext->Rcx = (ULONG64)ThreadStartParam;
 
-    /* Set the Selectors */
-    if ((LONG64)ThreadStartAddress < 0)
-    {
-        /* Initialize kernel mode segments */
-        ThreadContext->SegCs = KGDT64_R0_CODE;
-        ThreadContext->SegDs = KGDT64_R3_DATA;
-        ThreadContext->SegEs = KGDT64_R3_DATA;
-        ThreadContext->SegFs = KGDT64_R3_CMTEB;
-        ThreadContext->SegGs = KGDT64_R3_DATA;
-        ThreadContext->SegSs = KGDT64_R0_DATA;
-    }
-    else
-    {
-        /* Initialize user mode segments */
-        ThreadContext->SegCs = KGDT64_R3_CODE |  RPL_MASK;
-        ThreadContext->SegDs = KGDT64_R3_DATA |  RPL_MASK;
-        ThreadContext->SegEs = KGDT64_R3_DATA |  RPL_MASK;
-        ThreadContext->SegFs = KGDT64_R3_CMTEB |  RPL_MASK;
-        ThreadContext->SegGs = KGDT64_R3_DATA |  RPL_MASK;
-        ThreadContext->SegSs = KGDT64_R3_DATA |  RPL_MASK;
-    }
-
+    /* Initialize floating point and SSE state */
+    RtlZeroMemory(&ThreadContext->FltSave, sizeof(ThreadContext->FltSave));
     ThreadContext->MxCsr = INITIAL_MXCSR;
+    ThreadContext->FltSave.MxCsr = INITIAL_MXCSR;
+    ThreadContext->FltSave.ControlWord = INITIAL_FPCSR;
 
-    /* Only the basic Context is initialized */
-    ThreadContext->ContextFlags = CONTEXT_CONTROL |
-                                  CONTEXT_INTEGER |
-                                  CONTEXT_SEGMENTS |
-                                  CONTEXT_FLOATING_POINT;
+    /* Initialize integer registers */
+    ThreadContext->Rbp = 0x0000000000000000ull;
+    ThreadContext->Rax = 0x0000000000000000ull;
+    ThreadContext->Rbx = 0x0000000000000001ull;
+    ThreadContext->Rsi = 0x0000000000000004ull;
+    ThreadContext->Rdi = 0x0000000000000005ull;
+    ThreadContext->R8  = 0x0000000000000008ull;
+    ThreadContext->R9  = 0xF0E0D0C0A0908070ull;
+    ThreadContext->R10 = 0x000000000000000Aull;
+    ThreadContext->R11 = 0x000000000000000Bull;
+    ThreadContext->R12 = 0x000000000000000Cull;
+    ThreadContext->R13 = 0x000000000000000Dull;
+    ThreadContext->R14 = 0x000000000000000Eull;
+    ThreadContext->R15 = 0x000000000000000Full;
+
+    /* Set the context flags */
+    ThreadContext->ContextFlags = CONTEXT_FULL;
 
     return;
 }