commit f149ca5d774aa81d8b78cfcd5be6f5cf58cbdab7
author Yonghong Song <yhs@fb.com> Thu Feb 06 12:29:16 2020 -0800
committer yonghong-song <ys114321@gmail.com> Thu Feb 06 13:15:42 2020 -0800
tree dad00999153c1cf810ba663fba0f433afe39a84b
parent a4834a6cacd14f251e72695f08766a16216642b2

Do not rewrite array subscripts if invalid sourceloc range

Fix #2739.

The issue exposed an issue that to rewrite
  CODE1: __u8 byte = daddr->s6_addr[4];
will segfault and to rewrite
  CODE2: __u8 byte = (daddr->s6_addr)[4];
will be okay.

For CODE1, the clang did not give enough information to find the text
which contains the left bracket "[", given base "daddr->s6_addr"
and subscript "4". For CODE2, the clang is able to get the information
successfuly.

I think if we really go inside the base "daddr->s6_addr" and gets to
its member field "s6_addr", we can find the needed information
for the text range containing "[". Let us fix the segfault first
and if really desirable, we can try to enhance later for CODE1 patterns.

Signed-off-by: Yonghong Song <yhs@fb.com>

src/cc/frontends/clang/b_frontend_action.cc

diff --git a/src/cc/frontends/clang/b_frontend_action.cc b/src/cc/frontends/clang/b_frontend_action.cc
index 8ed8e69..bee7dd6 100644
--- a/src/cc/frontends/clang/b_frontend_action.cc
+++ b/src/cc/frontends/clang/b_frontend_action.cc
@@ -534,6 +534,20 @@
   LangOptions opts;
   SourceLocation lbracket_start, lbracket_end;
   SourceRange lbracket_range;
+
+  /* For cases like daddr->s6_addr[4], clang encodes the end location of "base"
+   * as "]". This makes it hard to rewrite the expression like
+   * "daddr->s6_addr  [ 4 ]" since we do not know the end location
+   * of "addr->s6_addr". Let us abort the operation if this is the case.
+   */
+  lbracket_start = Lexer::getLocForEndOfToken(GET_ENDLOC(base), 1,
+                                              rewriter_.getSourceMgr(),
+                                              opts).getLocWithOffset(1);
+  lbracket_end = GET_BEGINLOC(idx).getLocWithOffset(-1);
+  lbracket_range = expansionRange(SourceRange(lbracket_start, lbracket_end));
+  if (rewriter_.getRewrittenText(lbracket_range).size() == 0)
+    return true;
+
   pre = "({ typeof(" + E->getType().getAsString() + ") _val; __builtin_memset(&_val, 0, sizeof(_val));";
   pre += " bpf_probe_read(&_val, sizeof(_val), (u64)((";
   if (isMemberDereference(base)) {
@@ -549,11 +563,6 @@
    * a method to retrieve the left bracket, replace everything from the end of
    * the base to the start of the index. */
   lbracket = ") + (";
-  lbracket_start = Lexer::getLocForEndOfToken(GET_ENDLOC(base), 1,
-                                              rewriter_.getSourceMgr(),
-                                              opts).getLocWithOffset(1);
-  lbracket_end = GET_BEGINLOC(idx).getLocWithOffset(-1);
-  lbracket_range = expansionRange(SourceRange(lbracket_start, lbracket_end));
   rewriter_.ReplaceText(lbracket_range, lbracket);
 
   rbracket = "))); _val; })";