Hi.

The patch is about prevention of LTO section name clashing.
Now we have a situation where body of 2 functions is streamed
into the same ELF section. Then we'll end up with smashed data.

Patch can bootstrap on x86_64-linux-gnu and survives regression tests.

Ready to be installed?
Thanks,
Martin

gcc/ChangeLog:

2019-08-09  Martin Liska  <mli...@suse.cz>

        PR lto/91393
        PR lto/88220
        * lto-streamer.c (lto_get_section_name): Replace '*' leading
        character with '0'.

gcc/testsuite/ChangeLog:

2019-08-09  Martin Liska  <mli...@suse.cz>

        PR lto/91393
        PR lto/88220
        * gcc.dg/lto/pr91393_0.c: New test.
---
 gcc/lto-streamer.c                   | 15 ++++++++++++---
 gcc/testsuite/gcc.dg/lto/pr91393_0.c | 11 +++++++++++
 2 files changed, 23 insertions(+), 3 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/lto/pr91393_0.c


diff --git a/gcc/lto-streamer.c b/gcc/lto-streamer.c
index bd0126faebb..ffcaae516a5 100644
--- a/gcc/lto-streamer.c
+++ b/gcc/lto-streamer.c
@@ -124,9 +124,18 @@ lto_get_section_name (int section_type, const char *name, struct lto_file_decl_d
     {
       gcc_assert (name != NULL);
       if (name[0] == '*')
-	name++;
-      add = name;
-      sep = "";
+	{
+	  /* Symbols starting with '*' can clash with a symbol
+	     that has the same name.  Use then zero as one can't
+	     use digits at the beginning of identifiers.  */
+	  sep = "0";
+	  add = name + 1;
+	}
+      else
+	{
+	  add = name;
+	  sep = "";
+	}
     }
   else if (section_type < LTO_N_SECTION_TYPES)
     {
diff --git a/gcc/testsuite/gcc.dg/lto/pr91393_0.c b/gcc/testsuite/gcc.dg/lto/pr91393_0.c
new file mode 100644
index 00000000000..43b2426c86b
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/lto/pr91393_0.c
@@ -0,0 +1,11 @@
+void __open_alias(int, ...) __asm__("open");
+void __open_alias(int flags, ...) {}
+extern __inline __attribute__((__gnu_inline__)) int open() {}
+struct {
+  void *func;
+} a = {open};
+
+int main()
+{
+  return 0;
+}

Reply via email to