genksyms: Track changes to enum constants

Enum constants can be used as array sizes; if the enum itself does not
appear in the symbol expansion, a change in the enum constant will go
unnoticed. Example patch that changes the ABI but does not change the
checksum with current genksyms:

| enum e {
|	E1,
|	E2,
|+	E3,
|	E_MAX
| };
|
| struct s {
|	int a[E_MAX];
| }
|
| int f(struct s *s) { ... }
| EXPORT_SYMBOL(f)

Therefore, remember the value of each enum constant and
expand each occurence to <constant> <value>. The value is not actually
computed, but instead an expression in the form
(last explicitly assigned value) + N
is used. This avoids having to parse and semantically understand whole
of C.

Note: The changes won't take effect until the lexer and parser are
rebuilt by the next patch.

Signed-off-by: Michal Marek <mmarek@suse.cz>
Acked-by: Sam Ravnborg <sam@ravnborg.org>

1 files changed, 27 insertions, 3 deletions

diff --git a/scripts/genksyms/lex.l b/scripts/genksyms/lex.l
index c125d06..e4ddd49 100644
--- a/scripts/genksyms/lex.l
+++ b/scripts/genksyms/lex.l
@@ -99,12 +99,23 @@ MC_TOKEN		([~%^&*+=|<>/-]=)|(&&)|("||")|(->)|(<<)|(>>)
 
 /* Macros to append to our phrase collection list.  */
 
+/*
+ * We mark any token, that that equals to a known enumerator, as
+ * SYM_ENUM_CONST. The parser will change this for struct and union tags later,
+ * the only problem is struct and union members:
+ *    enum e { a, b }; struct s { int a, b; }
+ * but in this case, the only effect will be, that the ABI checksums become
+ * more volatile, which is acceptable. Also, such collisions are quite rare,
+ * so far it was only observed in include/linux/telephony.h.
+ */
 #define _APP(T,L)	do {						   \
 			  cur_node = next_node;				   \
 			  next_node = xmalloc(sizeof(*next_node));	   \
 			  next_node->next = cur_node;			   \
 			  cur_node->string = memcpy(xmalloc(L+1), T, L+1); \
-			  cur_node->tag = SYM_NORMAL;			   \
+			  cur_node->tag =				   \
+			    find_symbol(cur_node->string, SYM_ENUM_CONST, 1)?\
+			    SYM_ENUM_CONST : SYM_NORMAL ;		   \
 			} while (0)
 
 #define APP		_APP(yytext, yyleng)
@@ -182,8 +193,8 @@ repeat:
 
 		  case STRUCT_KEYW:
 		  case UNION_KEYW:
-		    dont_want_brace_phrase = 3;
 		  case ENUM_KEYW:
+		    dont_want_brace_phrase = 3;
 		    suppress_type_lookup = 2;
 		    goto fini;
 
@@ -312,7 +323,20 @@ repeat:
 	  ++count;
 	  APP;
 	  goto repeat;
-	case ')': case ']': case '}':
+	case '}':
+	  /* is this the last line of an enum declaration? */
+	  if (count == 0)
+	    {
+	      /* Put back the token we just read so's we can find it again
+		 after registering the expression.  */
+	      unput(token);
+
+	      lexstate = ST_NORMAL;
+	      token = EXPRESSION_PHRASE;
+	      break;
+	    }
+	  /* FALLTHRU */
+	case ')': case ']':
 	  --count;
 	  APP;
 	  goto repeat;