[PATCH] D29646: [X86] Add separate-stack-seg feature

[Michael LeMay via Phabricator via llvm-commits]

mlemay-intel created this revision.

Add support for separate-stack-seg feature that restricts access to the stack to
help prevent stack corruption.  The implementation of this feature may vary
depending on available HW support and other factors.


https://reviews.llvm.org/D29646

Files:
  lib/Target/X86/X86.td
  lib/Target/X86/X86Subtarget.cpp
  lib/Target/X86/X86Subtarget.h


Index: lib/Target/X86/X86Subtarget.h
===================================================================
--- lib/Target/X86/X86Subtarget.h
+++ lib/Target/X86/X86Subtarget.h
@@ -289,6 +289,9 @@
   /// Use software floating point for code generation.
   bool UseSoftFloat;
 
+  /// Restrict access to the stack
+  bool UseSeparateStackSeg;
+
   /// The minimum alignment known to hold of the stack frame on
   /// entry to the function and which must be maintained by every function.
   unsigned stackAlignment;
@@ -487,6 +490,7 @@
   bool isAtom() const { return X86ProcFamily == IntelAtom; }
   bool isSLM() const { return X86ProcFamily == IntelSLM; }
   bool useSoftFloat() const { return UseSoftFloat; }
+  bool useSeparateStackSeg() const { return UseSeparateStackSeg; }
 
   /// Use mfence if we have SSE2 or we're on x86-64 (even if we asked for
   /// no-sse2). There isn't any reason to disable it if the target processor
Index: lib/Target/X86/X86Subtarget.cpp
===================================================================
--- lib/Target/X86/X86Subtarget.cpp
+++ lib/Target/X86/X86Subtarget.cpp
@@ -304,6 +304,7 @@
   // FIXME: this is a known good value for Yonah. How about others?
   MaxInlineSizeThreshold = 128;
   UseSoftFloat = false;
+  UseSeparateStackSeg = false;
 }
 
 X86Subtarget &X86Subtarget::initializeSubtargetDependencies(StringRef CPU,
Index: lib/Target/X86/X86.td
===================================================================
--- lib/Target/X86/X86.td
+++ lib/Target/X86/X86.td
@@ -270,6 +270,9 @@
     : SubtargetFeature<
           "fast-lzcnt", "HasFastLZCNT", "true",
           "LZCNT instructions are as fast as most simple integer ops">;
+def FeatureSeparateStackSeg
+    : SubtargetFeature<"separate-stack-seg", "UseSeparateStackSeg", "true",
+                       "Restrict access to the stack">;
 
 //===----------------------------------------------------------------------===//
 // X86 processors supported.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D29646.87444.patch
Type: text/x-patch
Size: 1961 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170207/ecd7f5b2/attachment.bin>