[PATCH] D81403: [IndirectThunks] Make generated MF structure as expected by all instruction selectors.

[Kristof Beyls via Phabricator via llvm-commits]

kristof.beyls updated this revision to Diff 270195.
kristof.beyls retitled this revision from "Work around GlobalISel limitation on Indirect Thunks." to "[IndirectThunks] Make generated MF structure as expected  by all instruction selectors.".
kristof.beyls edited the summary of this revision.
kristof.beyls added a reviewer: chandlerc.
kristof.beyls added a comment.

This new revision fixes the root cause of GlobalISel asserting on an IndirectThunk: the IndirectThunk creation code was producing a MachineFunction with a slightly different structure than what gets produced from a naked function implemented in C code and processed by clang.
This fixes this by letting the IndirectThunk generator produce a MachineFunction structure that is the same as what gets produced by clang for an empty naked function written in C.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81403/new/

https://reviews.llvm.org/D81403

Files:
  llvm/include/llvm/CodeGen/IndirectThunks.h
  llvm/lib/Target/AArch64/AArch64SLSHardening.cpp
  llvm/lib/Target/X86/X86IndirectThunks.cpp
  llvm/test/CodeGen/AArch64/speculation-hardening-sls.ll


Index: llvm/test/CodeGen/AArch64/speculation-hardening-sls.ll
===================================================================
--- llvm/test/CodeGen/AArch64/speculation-hardening-sls.ll
+++ llvm/test/CodeGen/AArch64/speculation-hardening-sls.ll
@@ -1,7 +1,8 @@
 ; RUN: llc -mattr=harden-sls-retbr,harden-sls-blr -verify-machineinstrs -mtriple=aarch64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,HARDEN,ISBDSB
 ; RUN: llc -mattr=harden-sls-retbr,harden-sls-blr -mattr=+sb -verify-machineinstrs -mtriple=aarch64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,HARDEN,SB
 ; RUN: llc -verify-machineinstrs -mtriple=aarch64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,NOHARDEN
-
+; RUN: llc -global-isel -global-isel-abort=0 -mattr=harden-sls-retbr,harden-sls-blr -verify-machineinstrs -mtriple=aarch64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,HARDEN,ISBDSB
+; RUN: llc -global-isel -global-isel-abort=0 -mattr=harden-sls-retbr,harden-sls-blr -mattr=+sb -verify-machineinstrs -mtriple=aarch64-none-linux-gnu < %s | FileCheck %s --check-prefixes=CHECK,HARDEN,SB
 
 ; Function Attrs: norecurse nounwind readnone
 define dso_local i32 @double_return(i32 %a, i32 %b) local_unnamed_addr {
Index: llvm/lib/Target/X86/X86IndirectThunks.cpp
===================================================================
--- llvm/lib/Target/X86/X86IndirectThunks.cpp
+++ llvm/lib/Target/X86/X86IndirectThunks.cpp
@@ -79,12 +79,9 @@
     createThunkFunction(MMI, R11LVIThunkName);
   }
   void populateThunk(MachineFunction &MF) {
-    // Grab the entry MBB and erase any other blocks. O0 codegen appears to
-    // generate two bbs for the entry block.
+    assert (MF.size() == 1);
     MachineBasicBlock *Entry = &MF.front();
     Entry->clear();
-    while (MF.size() > 1)
-      MF.erase(std::next(MF.begin()));
 
     // This code mitigates LVI by replacing each indirect call/jump with a
     // direct call/jump to a thunk that looks like:
@@ -209,12 +206,9 @@
   }
 
   const TargetInstrInfo *TII = MF.getSubtarget<X86Subtarget>().getInstrInfo();
-  // Grab the entry MBB and erase any other blocks. O0 codegen appears to
-  // generate two bbs for the entry block.
+  assert (MF.size() == 1);
   MachineBasicBlock *Entry = &MF.front();
   Entry->clear();
-  while (MF.size() > 1)
-    MF.erase(std::next(MF.begin()));
 
   MachineBasicBlock *CaptureSpec =
       MF.CreateMachineBasicBlock(Entry->getBasicBlock());
Index: llvm/lib/Target/AArch64/AArch64SLSHardening.cpp
===================================================================
--- llvm/lib/Target/AArch64/AArch64SLSHardening.cpp
+++ llvm/lib/Target/AArch64/AArch64SLSHardening.cpp
@@ -208,12 +208,9 @@
 
   const TargetInstrInfo *TII =
       MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
-  // Grab the entry MBB and erase any other blocks. O0 codegen appears to
-  // generate two bbs for the entry block.
+  assert (MF.size() == 1);
   MachineBasicBlock *Entry = &MF.front();
   Entry->clear();
-  while (MF.size() > 1)
-    MF.erase(std::next(MF.begin()));
 
   //  These thunks need to consist of the following instructions:
   //  __llvm_slsblr_thunk_xN:
Index: llvm/include/llvm/CodeGen/IndirectThunks.h
===================================================================
--- llvm/include/llvm/CodeGen/IndirectThunks.h
+++ llvm/include/llvm/CodeGen/IndirectThunks.h
@@ -68,10 +68,11 @@
   // IR-level constructs we already made. Create them and insert them into the
   // module.
   MachineFunction &MF = MMI.getOrCreateMachineFunction(*F);
-  MachineBasicBlock *EntryMBB = MF.CreateMachineBasicBlock(Entry);
+  // A MachineBasicBlock must not be created for the Entry block; code
+  // generation from an empty naked function in C source code also does not
+  // generate one.  At least GlobalISel asserts if this invariant isn't
+  // respected.
 
-  // Insert EntryMBB into MF. It's not in the module until we do this.
-  MF.insert(MF.end(), EntryMBB);
   // Set MF properties. We never use vregs...
   MF.getProperties().set(MachineFunctionProperties::Property::NoVRegs);
 }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D81403.270195.patch
Type: text/x-patch
Size: 4095 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200611/1bc79df2/attachment.bin>