[patch 3/7] find_pc_partial_function new parameter

[patch 3/7] find_pc_partial_function new parameter

[Jan Kratochvil]

Hi,

find_pc_partial_function is heavily called inside GDB, therefore I did not
change its prototype and provided find_pc_partial_function_gnu_ifunc instead.

While I have no problem with changing all the callers it means passing
everywhere the unused additional NULL parameter which may not be a win.


Thanks,
Jan


gdb/
2011-03-19  Jan Kratochvil  <jan.kratochvil@redhat.com>

	GDB find_pc_partial_function support for STT_GNU_IFUNC.
	* blockframe.c (cache_pc_function_is_gnu_ifunc): New variable.
	(clear_pc_function_cache): Clear it.
	(find_pc_partial_function): Rename to ...
	(find_pc_partial_function_gnu_ifunc): ... this function.  New
	parameter is_gnu_ifunc_p, describe it.  Set *IS_GNU_IFUNC_P.
	(find_pc_partial_function): New wrapper for this function.
	* symtab.h (find_pc_partial_function_gnu_ifunc): New declaration.

--- a/gdb/blockframe.c
+++ b/gdb/blockframe.c
@@ -163,6 +163,7 @@ static CORE_ADDR cache_pc_function_low = 0;
 static CORE_ADDR cache_pc_function_high = 0;
 static char *cache_pc_function_name = 0;
 static struct obj_section *cache_pc_function_section = NULL;
+static int cache_pc_function_is_gnu_ifunc = 0;
 
 /* Clear cache, e.g. when symbol table is discarded.  */
 
@@ -173,6 +174,7 @@ clear_pc_function_cache (void)
   cache_pc_function_high = 0;
   cache_pc_function_name = (char *) 0;
   cache_pc_function_section = NULL;
+  cache_pc_function_is_gnu_ifunc = 0;
 }
 
 /* Finds the "function" (text symbol) that is smaller than PC but
@@ -180,17 +182,19 @@ clear_pc_function_cache (void)
    *NAME and/or *ADDRESS conditionally if that pointer is non-null.
    If ENDADDR is non-null, then set *ENDADDR to be the end of the
    function (exclusive), but passing ENDADDR as non-null means that
-   the function might cause symbols to be read.  This function either
-   succeeds or fails (not halfway succeeds).  If it succeeds, it sets
-   *NAME, *ADDRESS, and *ENDADDR to real information and returns 1.
-   If it fails, it sets *NAME, *ADDRESS, and *ENDADDR to zero and
-   returns 0.  */
+   the function might cause symbols to be read.  If IS_GNU_IFUNC_P is provided
+   *IS_GNU_IFUNC_P is set to 1 on return if the function is STT_GNU_IFUNC.
+   This function either succeeds or fails (not halfway succeeds).  If it
+   succeeds, it sets *NAME, *ADDRESS, and *ENDADDR to real information and
+   returns 1.  If it fails, it sets *NAME, *ADDRESS, *ENDADDR and
+   *IS_GNU_IFUNC_P to zero and returns 0.  */
 
 /* Backward compatibility, no section argument.  */
 
 int
-find_pc_partial_function (CORE_ADDR pc, char **name, CORE_ADDR *address,
-			  CORE_ADDR *endaddr)
+find_pc_partial_function_gnu_ifunc (CORE_ADDR pc, char **name,
+				    CORE_ADDR *address, CORE_ADDR *endaddr,
+				    int *is_gnu_ifunc_p)
 {
   struct obj_section *section;
   struct symbol *f;
@@ -240,6 +244,7 @@ find_pc_partial_function (CORE_ADDR pc, char **name, CORE_ADDR *address,
 	  cache_pc_function_high = BLOCK_END (SYMBOL_BLOCK_VALUE (f));
 	  cache_pc_function_name = SYMBOL_LINKAGE_NAME (f);
 	  cache_pc_function_section = section;
+	  cache_pc_function_is_gnu_ifunc = TYPE_GNU_IFUNC (SYMBOL_TYPE (f));
 	  goto return_cached_value;
 	}
     }
@@ -262,12 +267,15 @@ find_pc_partial_function (CORE_ADDR pc, char **name, CORE_ADDR *address,
 	*address = 0;
       if (endaddr != NULL)
 	*endaddr = 0;
+      if (is_gnu_ifunc_p != NULL)
+	*is_gnu_ifunc_p = 0;
       return 0;
     }
 
   cache_pc_function_low = SYMBOL_VALUE_ADDRESS (msymbol);
   cache_pc_function_name = SYMBOL_LINKAGE_NAME (msymbol);
   cache_pc_function_section = section;
+  cache_pc_function_is_gnu_ifunc = MSYMBOL_TYPE (msymbol) == mst_text_gnu_ifunc;
 
   /* If the minimal symbol has a size, use it for the cache.
      Otherwise use the lesser of the next minimal symbol in the same
@@ -330,9 +338,22 @@ find_pc_partial_function (CORE_ADDR pc, char **name, CORE_ADDR *address,
 	*endaddr = cache_pc_function_high;
     }
 
+  if (is_gnu_ifunc_p)
+    *is_gnu_ifunc_p = cache_pc_function_is_gnu_ifunc;
+
   return 1;
 }
 
+/* See find_pc_partial_function_gnu_ifunc, only the IS_GNU_IFUNC_P parameter
+   is omitted here for backward API compatibility.  */
+
+int
+find_pc_partial_function (CORE_ADDR pc, char **name, CORE_ADDR *address,
+			  CORE_ADDR *endaddr)
+{
+  return find_pc_partial_function_gnu_ifunc (pc, name, address, endaddr, NULL);
+}
+
 /* Return the innermost stack frame executing inside of BLOCK, or NULL
    if there is no such frame.  If BLOCK is NULL, just return NULL.  */
 
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -956,6 +956,11 @@ extern struct symbol *find_pc_function (CORE_ADDR);
 
 extern struct symbol *find_pc_sect_function (CORE_ADDR, struct obj_section *);
 
+extern int find_pc_partial_function_gnu_ifunc (CORE_ADDR pc, char **name,
+					       CORE_ADDR *address,
+					       CORE_ADDR *endaddr,
+					       int *is_gnu_ifunc_p);
+
 /* lookup function from address, return name, start addr and end addr.  */
 
 extern int find_pc_partial_function (CORE_ADDR, char **, CORE_ADDR *,

Re: [patch 3/7] find_pc_partial_function new parameter

[Tom Tromey]

>>>>> "Jan" == Jan Kratochvil <jan.kratochvil@redhat.com> writes:

Jan> find_pc_partial_function is heavily called inside GDB, therefore I
Jan> did not change its prototype and provided
Jan> find_pc_partial_function_gnu_ifunc instead.

Jan> While I have no problem with changing all the callers it means passing
Jan> everywhere the unused additional NULL parameter which may not be a win.

This approach is fine.  It is already used in many places in GDB.

This patch looked fine to me.

Tom

Re: [patch 3/7] find_pc_partial_function new parameter

[Joel Brobecker]

> find_pc_partial_function is heavily called inside GDB, therefore I did not
> change its prototype and provided find_pc_partial_function_gnu_ifunc instead.
> 
> While I have no problem with changing all the callers it means passing
> everywhere the unused additional NULL parameter which may not be a win.

I would have used that approach as well.

-- 
Joel

[commit] Re: [patch 3/7] find_pc_partial_function new parameter

[Jan Kratochvil]

On Mon, 21 Mar 2011 21:11:03 +0100, Tom Tromey wrote:
> This patch looked fine to me.

Checked in:
	http://sourceware.org/ml/gdb-cvs/2011-03/msg00311.html


Thanks,
Jan