[Bug c/64805] New: Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[Bug c/64805] New: Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[christian.otterstad at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64805

            Bug ID: 64805
           Summary: Specific use of __attribute ((always_inline)) breaks
                    MPX functionality with -fcheck-pointer-bounds -mmpx
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: christian.otterstad at gmail dot com

Using the latest git pull of GCC the following program fails to compile:

#include <stdio.h>

static inline void __attribute ((always_inline)) functionA(void)
{
        return;
}

static inline void __attribute ((always_inline)) functionB(void)
{
        functionA();
}

int test(void)
{
        functionB();

        return 0;
}


Attempt at compiling:

attribute_bug.c: In function 'functionB':
attribute_bug.c:18:1: error: Analyzed node has no reference to instrumented
version
 }
 ^
functionB/12 (functionB) @0x7f8d783a8dc8
  Type: function definition analyzed
  Visibility: prevailing_def_ironly
  References: 
  Referring: 
  Availability: local
  First run: 0
  Function flags: body local
  Called by: 
  Calls: 
  Has instrumented version.
attribute_bug.c:18:1: internal compiler error: verify_cgraph_node failed
0x749ab0 cgraph_node::verify_node()
        ../../../temp_git/gcc/gcc/cgraph.c:3096
0x73eab7 symtab_node::verify()
        ../../../temp_git/gcc/gcc/symtab.c:1161
0xb5a089 optimize_inline_calls(tree_node*)
        ../../../temp_git/gcc/gcc/tree-inline.c:4936
0x10dadf8 early_inliner(function*)
        ../../../temp_git/gcc/gcc/ipa-inline.c:2577
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.


This is rewritten from code found in glibc, in libc-start.c (it also fails to
compile with the same error).

This may be related to fixes resulting from bug 64363
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64363 , as earlier versions of GCC
supporting -fcheck-pointer-bounds -mmpx do not feature this issue. However
since it is a bug of a different type I'm nevertheless opening a new bug for
this problem.

It appears gcc (GCC) 5.0.0 20150122 and up have this problem (latest version
tested: gcc (GCC) 5.0.0 20150126)

It appears that at least gcc (GCC) 5.0.0 20150107, gcc (GCC) 5.0.0 20150105,
gcc (GCC) 5.0.0 20141218 do not feature this bug.

[Bug middle-end/64805] Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64805

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-01-26
                 CC|                            |mpolacek at gcc dot gnu.org
          Component|c                           |middle-end
     Ever confirmed|0                           |1

--- Comment #1 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Confirmed.

[Bug middle-end/64805] Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[enkovich.gnu at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64805

Ilya Enkovich <enkovich.gnu at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |enkovich.gnu at gmail dot com

--- Comment #2 from Ilya Enkovich <enkovich.gnu at gmail dot com> ---
This might be introduced by the recent changes in always_inline functions
instrumentation.  Now we keep them alive longer and therefore have inline of
the original functionA into the original functionB.  It causes error in a
verifier because inliner clears all references and then calls cgraph node
verification which expects IPA_REF_CHKP reference to instrumented functionB.

I would like to keep IPA_REF_CHKP check in the verifier because this ref is
important for reachability analysis.  Thus we probably should rebuild
IPA_REF_CHKP reference in inliner.  Will tests this patch:

diff --git a/gcc/ipa-inline.c b/gcc/ipa-inline.c
index c0ff329..d341619 100644
--- a/gcc/ipa-inline.c
+++ b/gcc/ipa-inline.c
@@ -2464,6 +2464,13 @@ early_inliner (function *fun)
 #endif
   node->remove_all_references ();

+  /* Rebuild this reference because it dosn't depend on
+     function's body and it's required to pass cgraph_node
+     verification.  */
+  if (node->instrumented_version
+      && !node->instrumentation_clone)
+    node->create_reference (node->instrumented_version, IPA_REF_CHKP, NULL);
+
   /* Even when not optimizing or not inlining inline always-inline
      functions.  */
   inlined = inline_always_inline_functions (node);

[Bug middle-end/64805] Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64805

--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
*** Bug 64804 has been marked as a duplicate of this bug. ***

[Bug middle-end/64805] Specific use of __attribute ((always_inline)) breaks MPX functionality with -fcheck-pointer-bounds -mmpx

[ienkovich at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64805

Ilya Enkovich <ienkovich at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #5 from Ilya Enkovich <ienkovich at gcc dot gnu.org> ---
Fixed