[24/40] Per-language symbol name hashing algorithm

Message ID 1496406158-12663-25-git-send-email-palves@redhat.com
State New, archived
Headers

Commit Message

Pedro Alves June 2, 2017, 12:22 p.m. UTC
  Currently, we have a mess of symbol name hashing/comparison routines.
There's msymbol_hash for mangled names, and dict_hash and
msymbol_hash_iw for demangled names.  Then there's strcmp_iw,
strcmp_iw_ordered and Ada's full_match/wild_match, which all have to
agree with the hashing routines.  That's why dict_hash is really about
Ada names.  From the inconsistency department, minimal symbol hashing
doesn't go via dict_hash, so Ada's wild matching can't ever work with
minimal symbols.

This patch starts fixing this, by doing two things:

#1 - adds a language vector method to let each language decide how to
     compute a symbol name hash.

#2 - makes dictionaries know the language of the symbols they hold,
     and then use the dictionaries language to decide which hashing
     method to use.

For now, this is just scaffolding, since all languages install the
default method.  The series will make C++ install its own hashing
method later on, and will add per-language symbol name comparison
routines too.

This patch is based on a patch that Keith wrote for the libcc1/C++ WIP
support.

gdb/ChangeLog:
yyyy-mm-dd  Keith Seitz  <keiths@redhat.com>
	    Pedro Alves  <palves@redhat.com>

	* ada-lang.c (ada_language_defn): Install
	default_search_name_hash.
	* buildsym.c (struct buildsym_compunit): <language>: New field.
	(finish_block_internal): Pass language when creating dictionaries.
	(start_buildsym_compunit, start_symtab): New language parameters.
	Use them.
	(restart_symtab): Pass down compilation unit's language.
	* buildsym.h (enum language): Forward declare.
	(start_symtab): New 'language' parameter.
	* c-lang.c (c_language_defn, cplus_language_defn)
	(asm_language_defn, minimal_language_defn): Install
	default_search_name_hash.
	* coffread.c (coff_start_symtab): Adjust.
	* d-lang.c (d_language_defn): Install default_search_name_hash.
	* dbxread.c (struct symloc): Add 'pst_language' field.
	(PST_LANGUAGE): Define.
	(start_psymtab, read_ofile_symtab): Use it.
	(process_one_symbol): New 'language' parameter.  Pass it down.
	* dictionary.c (struct dictionary) <language>: New field.
	(DICT_LANGUAGE): Define.
	(dict_create_hashed, dict_create_hashed_expandable)
	(dict_create_linear, dict_create_linear_expandable): New parameter
	'language'.  Set the dictionary's language.
	(iter_match_first_hashed): Adjust to rename.
	(insert_symbol_hashed): Assert we don't see mismatching
	languages.  Adjust to rename.
	(dict_hash): Rename to ...
	(default_search_name_hash): ... this and make extern.
	* dictionary.h (struct language_defn): Forward declare.
	(dict_create_hashed): New parameter 'language'.
	* dwarf2read.c (dwarf2_start_symtab): Pass down language.
	* f-lang.c (f_language_defn): Install default_search_name_hash.
	* go-lang.c (go_language_defn): Install default_search_name_hash.
	* jit.c (finalize_symtab): Pass compunit's language to dictionary
	creation.
	* language.c (unknown_language_defn, auto_language_defn):
	* language.h (language_defn::la_search_name_hash): New field.
	(default_search_name_hash): Declare.
	* m2-lang.c (m2_language_defn): Install default_search_name_hash.
	* mdebugread.c (new_block): New parameter 'language'.
	* mdebugread.c (parse_symbol): Pass symbol language to block
	allocation.
	(psymtab_to_symtab_1): Pass down language.
	(new_symtab): Pass compunit's language to block allocation.
	* objc-lang.c (objc_language_defn): Install
	default_search_name_hash.
	* opencl-lang.c (opencl_language_defn):
	* p-lang.c (pascal_language_defn): Install
	default_search_name_hash.
	* rust-lang.c (rust_language_defn): Install
	default_search_name_hash.
	* stabsread.h (enum language): Forward declare.
	(process_one_symbol): Add 'language' parameter.
	* symtab.c (search_name_hash): New function.
	* symtab.h (search_name_hash): Declare.
	* xcoffread.c (read_xcoff_symtab): Pass language to start_symtab.
---
 gdb/ada-lang.c    |  1 +
 gdb/buildsym.c    | 25 +++++++++++++++++--------
 gdb/buildsym.h    |  4 +++-
 gdb/c-lang.c      |  4 ++++
 gdb/coffread.c    |  4 +++-
 gdb/d-lang.c      |  1 +
 gdb/dbxread.c     | 12 ++++++++----
 gdb/dictionary.c  | 36 +++++++++++++++++++++++-------------
 gdb/dictionary.h  | 11 +++++++----
 gdb/dwarf2read.c  |  2 +-
 gdb/f-lang.c      |  1 +
 gdb/go-lang.c     |  1 +
 gdb/jit.c         |  6 ++++--
 gdb/language.c    |  2 ++
 gdb/language.h    | 12 ++++++++++++
 gdb/m2-lang.c     |  1 +
 gdb/mdebugread.c  | 23 +++++++++++++----------
 gdb/objc-lang.c   |  1 +
 gdb/opencl-lang.c |  1 +
 gdb/p-lang.c      |  1 +
 gdb/rust-lang.c   |  1 +
 gdb/stabsread.h   |  3 ++-
 gdb/symtab.c      |  8 ++++++++
 gdb/symtab.h      |  3 +++
 gdb/xcoffread.c   | 11 +++++++----
 25 files changed, 126 insertions(+), 49 deletions(-)
  

Comments

Keith Seitz July 18, 2017, 5:33 p.m. UTC | #1
On 06/02/2017 05:22 AM, Pedro Alves wrote:
> 
> This patch starts fixing this, by doing two things:
> 
> #1 - adds a language vector method to let each language decide how to
>      compute a symbol name hash.
> 
> #2 - makes dictionaries know the language of the symbols they hold,
>      and then use the dictionaries language to decide which hashing
>      method to use.

Me likey! :-)

Again, just the usual trivial comments.

> 
> diff --git a/gdb/buildsym.c b/gdb/buildsym.c
> index cbad027..15f65a8 100644
> --- a/gdb/buildsym.c
> +++ b/gdb/buildsym.c
> @@ -1047,11 +1055,11 @@ prepare_for_building (const char *name, CORE_ADDR start_addr)
>  
>  struct compunit_symtab *
>  start_symtab (struct objfile *objfile, const char *name, const char *comp_dir,
> -	      CORE_ADDR start_addr)
> +	      CORE_ADDR start_addr, enum language language)

Unlike start_buildsym_compunit, this function is exported. IMO, `language' should be mentioned in the comment.

> diff --git a/gdb/coffread.c b/gdb/coffread.c
> index 9db4792..db0b77a 100644
> --- a/gdb/coffread.c
> +++ b/gdb/coffread.c
> @@ -394,7 +394,9 @@ coff_start_symtab (struct objfile *objfile, const char *name)
>  		 NULL,
>    /* The start address is irrelevant, since we set
>       last_source_start_addr in coff_end_symtab.  */
> -		 0);
> +		 0,
> +  /* Let buildsym.c deduce the language for this symtab.  */
> +		 language_unknown);

re: "deduce the language" if language == language_unknown... That's not obvious from start_symtab's documentation, so I think that deserves a mention there.

> diff --git a/gdb/dictionary.h b/gdb/dictionary.h
> index 4f4f160..ef5fbed 100644
> --- a/gdb/dictionary.h
> +++ b/gdb/dictionary.h
> @@ -45,6 +45,7 @@ struct pending;
>     initialized from SYMBOL_LIST.  */
>  
>  extern struct dictionary *dict_create_hashed (struct obstack *obstack,
> +					      enum language language,
>  					      const struct pending
>  					      *symbol_list);
>  
> @@ -53,7 +54,8 @@ extern struct dictionary *dict_create_hashed (struct obstack *obstack,
>     it, call dict_add_symbol().  Call dict_free() when you're done with
>     it.  */
>  
> -extern struct dictionary *dict_create_hashed_expandable (void);
> +extern struct dictionary *
> +  dict_create_hashed_expandable (enum language language);
>  
>  /* Create a dictionary implemented via a fixed-size array.  All memory
>     it uses is allocated on OBSTACK; the environment is initialized
> @@ -61,6 +63,7 @@ extern struct dictionary *dict_create_hashed_expandable (void);
>     that they're found in SYMBOL_LIST.  */
>  
>  extern struct dictionary *dict_create_linear (struct obstack *obstack,
> +					      enum language language,
>  					      const struct pending
>  					      *symbol_list);
>  
> @@ -69,8 +72,8 @@ extern struct dictionary *dict_create_linear (struct obstack *obstack,
>     it, call dict_add_symbol().  Call dict_free() when you're done with
>     it.  */
>  
> -extern struct dictionary *dict_create_linear_expandable (void);
> -
> +extern struct dictionary *
> +  dict_create_linear_expandable (enum language language);
>  
>  /* The functions providing the interface to dictionaries.  Note that
>     the most common parts of the interface, namely symbol lookup, are

All of the above functions are exported, so I think our custom is to document all formal parameters, no?

> diff --git a/gdb/symtab.h b/gdb/symtab.h
> index fffe0f87..20904e4 100644
> --- a/gdb/symtab.h
> +++ b/gdb/symtab.h
> @@ -277,6 +277,9 @@ extern const char *symbol_search_name (const struct general_symbol_info *);
>  #define SYMBOL_MATCHES_SEARCH_NAME(symbol, name)			\
>    (strcmp_iw (SYMBOL_SEARCH_NAME (symbol), (name)) == 0)
>  
> +extern unsigned int search_name_hash (enum language language,
> +				      const char *search_name);
> +

symtab.c says, "See symtab.h." Missing comment here.

Keith
  

Patch

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index f582c56..1d4cf36 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -14034,6 +14034,7 @@  extern const struct language_defn ada_language_defn = {
   c_watch_location_expression,
   ada_get_symbol_name_cmp,	/* la_get_symbol_name_cmp */
   ada_iterate_over_symbols,
+  default_search_name_hash,
   &ada_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/buildsym.c b/gdb/buildsym.c
index cbad027..15f65a8 100644
--- a/gdb/buildsym.c
+++ b/gdb/buildsym.c
@@ -128,6 +128,9 @@  struct buildsym_compunit
 
   /* The compunit we are building.  */
   struct compunit_symtab *compunit_symtab;
+
+  /* Language of this compunit_symtab.  */
+  enum language language;
 };
 
 /* The work-in-progress of the compunit we are building.
@@ -351,20 +354,23 @@  finish_block_internal (struct symbol *symbol,
 
   if (symbol)
     {
-      BLOCK_DICT (block) = dict_create_linear (&objfile->objfile_obstack,
-					       *listhead);
+      BLOCK_DICT (block)
+	= dict_create_linear (&objfile->objfile_obstack,
+			      buildsym_compunit->language, *listhead);
     }
   else
     {
       if (expandable)
 	{
-	  BLOCK_DICT (block) = dict_create_hashed_expandable ();
+	  BLOCK_DICT (block)
+	    = dict_create_hashed_expandable (buildsym_compunit->language);
 	  dict_add_pending (BLOCK_DICT (block), *listhead);
 	}
       else
 	{
 	  BLOCK_DICT (block) =
-	    dict_create_hashed (&objfile->objfile_obstack, *listhead);
+	    dict_create_hashed (&objfile->objfile_obstack,
+				buildsym_compunit->language, *listhead);
 	}
     }
 
@@ -768,7 +774,8 @@  start_subfile (const char *name)
    (or NULL if not known).  */
 
 static struct buildsym_compunit *
-start_buildsym_compunit (struct objfile *objfile, const char *comp_dir)
+start_buildsym_compunit (struct objfile *objfile, const char *comp_dir,
+			 enum language language)
 {
   struct buildsym_compunit *bscu;
 
@@ -777,6 +784,7 @@  start_buildsym_compunit (struct objfile *objfile, const char *comp_dir)
 
   bscu->objfile = objfile;
   bscu->comp_dir = (comp_dir == NULL) ? NULL : xstrdup (comp_dir);
+  bscu->language = language;
 
   /* Initialize the debug format string to NULL.  We may supply it
      later via a call to record_debugformat.  */
@@ -1047,11 +1055,11 @@  prepare_for_building (const char *name, CORE_ADDR start_addr)
 
 struct compunit_symtab *
 start_symtab (struct objfile *objfile, const char *name, const char *comp_dir,
-	      CORE_ADDR start_addr)
+	      CORE_ADDR start_addr, enum language language)
 {
   prepare_for_building (name, start_addr);
 
-  buildsym_compunit = start_buildsym_compunit (objfile, comp_dir);
+  buildsym_compunit = start_buildsym_compunit (objfile, comp_dir, language);
 
   /* Allocate the compunit symtab now.  The caller needs it to allocate
      non-primary symtabs.  It is also needed by get_macro_table.  */
@@ -1088,7 +1096,8 @@  restart_symtab (struct compunit_symtab *cust,
   prepare_for_building (name, start_addr);
 
   buildsym_compunit = start_buildsym_compunit (COMPUNIT_OBJFILE (cust),
-					       COMPUNIT_DIRNAME (cust));
+					       COMPUNIT_DIRNAME (cust),
+					       compunit_language (cust));
   buildsym_compunit->compunit_symtab = cust;
 }
 
diff --git a/gdb/buildsym.h b/gdb/buildsym.h
index 60109a0..80ca7da 100644
--- a/gdb/buildsym.h
+++ b/gdb/buildsym.h
@@ -23,6 +23,7 @@  struct objfile;
 struct symbol;
 struct addrmap;
 struct compunit_symtab;
+enum language;
 
 /* This module provides definitions used for creating and adding to
    the symbol table.  These routines are called from various symbol-
@@ -254,7 +255,8 @@  extern record_line_ftype record_line;
 extern struct compunit_symtab *start_symtab (struct objfile *objfile,
 					     const char *name,
 					     const char *comp_dir,
-					     CORE_ADDR start_addr);
+					     CORE_ADDR start_addr,
+					     enum language language);
 
 extern void restart_symtab (struct compunit_symtab *cust,
 			    const char *name, CORE_ADDR start_addr);
diff --git a/gdb/c-lang.c b/gdb/c-lang.c
index f86e26e..9749935 100644
--- a/gdb/c-lang.c
+++ b/gdb/c-lang.c
@@ -871,6 +871,7 @@  extern const struct language_defn c_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &c_varobj_ops,
   c_get_compile_context,
   c_compute_program,
@@ -1015,6 +1016,7 @@  extern const struct language_defn cplus_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &cplus_varobj_ops,
   NULL,
   NULL,
@@ -1068,6 +1070,7 @@  extern const struct language_defn asm_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
@@ -1121,6 +1124,7 @@  extern const struct language_defn minimal_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/coffread.c b/gdb/coffread.c
index 9db4792..db0b77a 100644
--- a/gdb/coffread.c
+++ b/gdb/coffread.c
@@ -394,7 +394,9 @@  coff_start_symtab (struct objfile *objfile, const char *name)
 		 NULL,
   /* The start address is irrelevant, since we set
      last_source_start_addr in coff_end_symtab.  */
-		 0);
+		 0,
+  /* Let buildsym.c deduce the language for this symtab.  */
+		 language_unknown);
   record_debugformat ("COFF");
 }
 
diff --git a/gdb/d-lang.c b/gdb/d-lang.c
index 941d3ed..b89b636 100644
--- a/gdb/d-lang.c
+++ b/gdb/d-lang.c
@@ -247,6 +247,7 @@  extern const struct language_defn d_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/dbxread.c b/gdb/dbxread.c
index daa3ce9..ce0eeb0 100644
--- a/gdb/dbxread.c
+++ b/gdb/dbxread.c
@@ -54,7 +54,6 @@ 
 #include "cp-support.h"
 #include "psympriv.h"
 #include "block.h"
-
 #include "aout/aout64.h"
 #include "aout/stab_gnu.h"	/* We always use GNU stabs, not
 				   native, now.  */
@@ -92,6 +91,7 @@  struct symloc
     int symbol_offset;
     int string_offset;
     int file_string_offset;
+    enum language pst_language;
   };
 
 #define LDSYMOFF(p) (((struct symloc *)((p)->read_symtab_private))->ldsymoff)
@@ -101,6 +101,7 @@  struct symloc
 #define SYMBOL_OFFSET(p) (SYMLOC(p)->symbol_offset)
 #define STRING_OFFSET(p) (SYMLOC(p)->string_offset)
 #define FILE_STRING_OFFSET(p) (SYMLOC(p)->file_string_offset)
+#define PST_LANGUAGE(p) (SYMLOC(p)->pst_language)
 
 
 /* The objfile we are currently reading.  */
@@ -1422,6 +1423,7 @@  read_dbx_symtab (minimal_symbol_reader &reader, struct objfile *objfile)
 		    || psymtab_language != language_cplus))
 	      psymtab_language = tmp_language;
 
+
 	    /* In C++, one may expect the same filename to come round many
 	       times, when code is coming alternately from the main file
 	       and from inline functions in other files.  So I check to see
@@ -2017,6 +2019,7 @@  start_psymtab (struct objfile *objfile, const char *filename, CORE_ADDR textlow,
 
   /* Deduce the source language from the filename for this psymtab.  */
   psymtab_language = deduce_language_from_filename (filename);
+  PST_LANGUAGE (result) = psymtab_language;
 
   return result;
 }
@@ -2403,7 +2406,8 @@  read_ofile_symtab (struct objfile *objfile, struct partial_symtab *pst)
 		 positive offsets.  */
 	    nlist.n_value = (nlist.n_value ^ 0x80000000) - 0x80000000;
 	  process_one_symbol (type, nlist.n_desc, nlist.n_value,
-			      namestring, section_offsets, objfile);
+			      namestring, section_offsets, objfile,
+			      PST_LANGUAGE (pst));
 	}
       /* We skip checking for a new .o or -l file; that should never
          happen in this routine.  */
@@ -2503,7 +2507,7 @@  cp_set_block_scope (const struct symbol *symbol,
 void
 process_one_symbol (int type, int desc, CORE_ADDR valu, const char *name,
 		    const struct section_offsets *section_offsets,
-		    struct objfile *objfile)
+		    struct objfile *objfile, enum language language)
 {
   struct gdbarch *gdbarch = get_objfile_arch (objfile);
   struct context_stack *newobj;
@@ -2717,7 +2721,7 @@  process_one_symbol (int type, int desc, CORE_ADDR valu, const char *name,
       function_start_offset = 0;
 
       start_stabs ();
-      start_symtab (objfile, name, NULL, valu);
+      start_symtab (objfile, name, NULL, valu, language);
       record_debugformat ("stabs");
       break;
 
diff --git a/gdb/dictionary.c b/gdb/dictionary.c
index b2cfca2..1be7e48 100644
--- a/gdb/dictionary.c
+++ b/gdb/dictionary.c
@@ -165,6 +165,7 @@  struct dictionary_linear_expandable
 
 struct dictionary
 {
+  const struct language_defn *language;
   const struct dict_vector *vector;
   union
   {
@@ -179,6 +180,7 @@  struct dictionary
 /* Accessor macros.  */
 
 #define DICT_VECTOR(d)			(d)->vector
+#define DICT_LANGUAGE(d)                (d)->language
 
 /* These can be used for DICT_HASHED_EXPANDABLE, too.  */
 
@@ -245,8 +247,6 @@  static struct symbol *iter_match_next_hashed (const char *name,
 					      symbol_compare_ftype *compare,
 					      struct dict_iterator *iterator);
 
-static unsigned int dict_hash (const char *string);
-
 /* Functions only for DICT_HASHED.  */
 
 static int size_hashed (const struct dictionary *dict);
@@ -354,6 +354,7 @@  static void expand_hashtable (struct dictionary *dict);
 
 struct dictionary *
 dict_create_hashed (struct obstack *obstack,
+		    enum language language,
 		    const struct pending *symbol_list)
 {
   struct dictionary *retval;
@@ -363,6 +364,7 @@  dict_create_hashed (struct obstack *obstack,
 
   retval = XOBNEW (obstack, struct dictionary);
   DICT_VECTOR (retval) = &dict_hashed_vector;
+  DICT_LANGUAGE (retval) = language_def (language);
 
   /* Calculate the number of symbols, and allocate space for them.  */
   for (list_counter = symbol_list;
@@ -397,11 +399,12 @@  dict_create_hashed (struct obstack *obstack,
    it.  */
 
 extern struct dictionary *
-dict_create_hashed_expandable (void)
+dict_create_hashed_expandable (enum language language)
 {
   struct dictionary *retval = XNEW (struct dictionary);
 
   DICT_VECTOR (retval) = &dict_hashed_expandable_vector;
+  DICT_LANGUAGE (retval) = language_def (language);
   DICT_HASHED_NBUCKETS (retval) = DICT_EXPANDABLE_INITIAL_CAPACITY;
   DICT_HASHED_BUCKETS (retval) = XCNEWVEC (struct symbol *,
 					   DICT_EXPANDABLE_INITIAL_CAPACITY);
@@ -417,6 +420,7 @@  dict_create_hashed_expandable (void)
 
 struct dictionary *
 dict_create_linear (struct obstack *obstack,
+		    enum language language,
 		    const struct pending *symbol_list)
 {
   struct dictionary *retval;
@@ -426,6 +430,7 @@  dict_create_linear (struct obstack *obstack,
 
   retval = XOBNEW (obstack, struct dictionary);
   DICT_VECTOR (retval) = &dict_linear_vector;
+  DICT_LANGUAGE (retval) = language_def (language);
 
   /* Calculate the number of symbols, and allocate space for them.  */
   for (list_counter = symbol_list;
@@ -461,11 +466,12 @@  dict_create_linear (struct obstack *obstack,
    it.  */
 
 struct dictionary *
-dict_create_linear_expandable (void)
+dict_create_linear_expandable (enum language language)
 {
   struct dictionary *retval = XNEW (struct dictionary);
 
   DICT_VECTOR (retval) = &dict_linear_expandable_vector;
+  DICT_LANGUAGE (retval) = language_def (language);
   DICT_LINEAR_NSYMS (retval) = 0;
   DICT_LINEAR_EXPANDABLE_CAPACITY (retval) = DICT_EXPANDABLE_INITIAL_CAPACITY;
   DICT_LINEAR_SYMS (retval)
@@ -638,7 +644,9 @@  iter_match_first_hashed (const struct dictionary *dict, const char *name,
 			 symbol_compare_ftype *compare,
 			 struct dict_iterator *iterator)
 {
-  unsigned int hash_index = dict_hash (name) % DICT_HASHED_NBUCKETS (dict);
+  unsigned int hash_index
+    = (search_name_hash (DICT_LANGUAGE (dict)->la_language, name)
+       % DICT_HASHED_NBUCKETS (dict));
   struct symbol *sym;
 
   DICT_ITERATOR_DICT (iterator) = dict;
@@ -689,10 +697,15 @@  insert_symbol_hashed (struct dictionary *dict,
 		      struct symbol *sym)
 {
   unsigned int hash_index;
+  unsigned int hash;
   struct symbol **buckets = DICT_HASHED_BUCKETS (dict);
 
-  hash_index = 
-    dict_hash (SYMBOL_SEARCH_NAME (sym)) % DICT_HASHED_NBUCKETS (dict);
+  /* We don't want to insert a symbol into a dictionary of a different
+     language.  The two may not use the same hashing algorithm.  */
+  gdb_assert (SYMBOL_LANGUAGE (sym) == DICT_LANGUAGE (dict)->la_language);
+
+  hash = search_name_hash (SYMBOL_LANGUAGE (sym), SYMBOL_SEARCH_NAME (sym));
+  hash_index = hash % DICT_HASHED_NBUCKETS (dict);
   sym->hash_next = buckets[hash_index];
   buckets[hash_index] = sym;
 }
@@ -765,13 +778,10 @@  expand_hashtable (struct dictionary *dict)
   xfree (old_buckets);
 }
 
-/* Produce an unsigned hash value from STRING0 that is consistent
-   with strcmp_iw, strcmp, and, at least on Ada symbols, wild_match.
-   That is, two identifiers equivalent according to any of those three
-   comparison operators hash to the same value.  */
+/* See dictionary.h.  */
 
-static unsigned int
-dict_hash (const char *string0)
+unsigned int
+default_search_name_hash (const char *string0)
 {
   /* The Ada-encoded version of a name P1.P2...Pn has either the form
      P1__P2__...Pn<suffix> or _ada_P1__P2__...Pn<suffix> (where the Pi
diff --git a/gdb/dictionary.h b/gdb/dictionary.h
index 4f4f160..ef5fbed 100644
--- a/gdb/dictionary.h
+++ b/gdb/dictionary.h
@@ -35,7 +35,7 @@  struct dictionary;
 struct symbol;
 struct obstack;
 struct pending;
-
+struct language_defn;
 
 /* The creation functions for various implementations of
    dictionaries.  */
@@ -45,6 +45,7 @@  struct pending;
    initialized from SYMBOL_LIST.  */
 
 extern struct dictionary *dict_create_hashed (struct obstack *obstack,
+					      enum language language,
 					      const struct pending
 					      *symbol_list);
 
@@ -53,7 +54,8 @@  extern struct dictionary *dict_create_hashed (struct obstack *obstack,
    it, call dict_add_symbol().  Call dict_free() when you're done with
    it.  */
 
-extern struct dictionary *dict_create_hashed_expandable (void);
+extern struct dictionary *
+  dict_create_hashed_expandable (enum language language);
 
 /* Create a dictionary implemented via a fixed-size array.  All memory
    it uses is allocated on OBSTACK; the environment is initialized
@@ -61,6 +63,7 @@  extern struct dictionary *dict_create_hashed_expandable (void);
    that they're found in SYMBOL_LIST.  */
 
 extern struct dictionary *dict_create_linear (struct obstack *obstack,
+					      enum language language,
 					      const struct pending
 					      *symbol_list);
 
@@ -69,8 +72,8 @@  extern struct dictionary *dict_create_linear (struct obstack *obstack,
    it, call dict_add_symbol().  Call dict_free() when you're done with
    it.  */
 
-extern struct dictionary *dict_create_linear_expandable (void);
-
+extern struct dictionary *
+  dict_create_linear_expandable (enum language language);
 
 /* The functions providing the interface to dictionaries.  Note that
    the most common parts of the interface, namely symbol lookup, are
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index 7fade00..c44a76e 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -18810,7 +18810,7 @@  dwarf2_start_symtab (struct dwarf2_cu *cu,
 		     const char *name, const char *comp_dir, CORE_ADDR low_pc)
 {
   struct compunit_symtab *cust
-    = start_symtab (cu->objfile, name, comp_dir, low_pc);
+    = start_symtab (cu->objfile, name, comp_dir, low_pc, cu->language);
 
   record_debugformat ("DWARF 2");
   record_producer (cu->producer);
diff --git a/gdb/f-lang.c b/gdb/f-lang.c
index 903cfd1..cfef64f 100644
--- a/gdb/f-lang.c
+++ b/gdb/f-lang.c
@@ -293,6 +293,7 @@  extern const struct language_defn f_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/go-lang.c b/gdb/go-lang.c
index 60bb3c5..befd937 100644
--- a/gdb/go-lang.c
+++ b/gdb/go-lang.c
@@ -608,6 +608,7 @@  extern const struct language_defn go_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/jit.c b/gdb/jit.c
index ddf1005..c30a06a 100644
--- a/gdb/jit.c
+++ b/gdb/jit.c
@@ -661,12 +661,14 @@  finalize_symtab (struct gdb_symtab *stab, struct objfile *objfile)
   size_t blockvector_size;
   CORE_ADDR begin, end;
   struct blockvector *bv;
+  enum language language;
 
   actual_nblocks = FIRST_LOCAL_BLOCK + stab->nblocks;
 
   cust = allocate_compunit_symtab (objfile, stab->file_name);
   allocate_symtab (cust, stab->file_name);
   add_compunit_symtab_to_objfile (cust);
+  language = compunit_language (cust);
 
   /* JIT compilers compile in memory.  */
   COMPUNIT_DIRNAME (cust) = NULL;
@@ -711,7 +713,7 @@  finalize_symtab (struct gdb_symtab *stab, struct objfile *objfile)
 					   "void");
 
       BLOCK_DICT (new_block) = dict_create_linear (&objfile->objfile_obstack,
-                                                   NULL);
+						   language, NULL);
       /* The address range.  */
       BLOCK_START (new_block) = (CORE_ADDR) gdb_block_iter->begin;
       BLOCK_END (new_block) = (CORE_ADDR) gdb_block_iter->end;
@@ -749,7 +751,7 @@  finalize_symtab (struct gdb_symtab *stab, struct objfile *objfile)
 		   ? allocate_global_block (&objfile->objfile_obstack)
 		   : allocate_block (&objfile->objfile_obstack));
       BLOCK_DICT (new_block) = dict_create_linear (&objfile->objfile_obstack,
-                                                   NULL);
+						   language, NULL);
       BLOCK_SUPERBLOCK (new_block) = block_iter;
       block_iter = new_block;
 
diff --git a/gdb/language.c b/gdb/language.c
index df4f3cd..2e95c9e 100644
--- a/gdb/language.c
+++ b/gdb/language.c
@@ -849,6 +849,7 @@  const struct language_defn unknown_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
@@ -899,6 +900,7 @@  const struct language_defn auto_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/language.h b/gdb/language.h
index f4852c1..9b5062e 100644
--- a/gdb/language.h
+++ b/gdb/language.h
@@ -392,6 +392,10 @@  struct language_defn
       (const struct block *block, const char *name, domain_enum domain,
        gdb::function_view<symbol_found_callback_ftype> callback);
 
+    /* Hash the given STRING.  Use default_search_name_hash if no
+       special treatment is required.  */
+    unsigned int (*la_search_name_hash) (const char *name);
+
     /* Various operations on varobj.  */
     const struct lang_varobj_ops *la_varobj_ops;
 
@@ -611,6 +615,14 @@  void default_print_typedef (struct type *type, struct symbol *new_symbol,
 void default_get_string (struct value *value, gdb_byte **buffer, int *length,
 			 struct type **char_type, const char **charset);
 
+/* Default name hashing function.  */
+
+/* Produce an unsigned hash value from SEARCH_NAME that is consistent
+   with strcmp_iw, strcmp, and, at least on Ada symbols, wild_match.
+   That is, two identifiers equivalent according to any of those three
+   comparison operators hash to the same value.  */
+extern unsigned int default_search_name_hash (const char *search_name);
+
 void c_get_string (struct value *value, gdb_byte **buffer, int *length,
 		   struct type **char_type, const char **charset);
 
diff --git a/gdb/m2-lang.c b/gdb/m2-lang.c
index b9ab2b3..66ac34c 100644
--- a/gdb/m2-lang.c
+++ b/gdb/m2-lang.c
@@ -396,6 +396,7 @@  extern const struct language_defn m2_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/mdebugread.c b/gdb/mdebugread.c
index 8e27194..5df0f48 100644
--- a/gdb/mdebugread.c
+++ b/gdb/mdebugread.c
@@ -236,7 +236,7 @@  static struct type *new_type (char *);
 
 enum block_type { FUNCTION_BLOCK, NON_FUNCTION_BLOCK };
 
-static struct block *new_block (enum block_type);
+static struct block *new_block (enum block_type, enum language);
 
 static struct compunit_symtab *new_symtab (const char *, int, struct objfile *);
 
@@ -811,7 +811,7 @@  parse_symbol (SYMR *sh, union aux_ext *ax, char *ext_sh, int bigend,
 	TYPE_PROTOTYPED (SYMBOL_TYPE (s)) = 1;
 
       /* Create and enter a new lexical context.  */
-      b = new_block (FUNCTION_BLOCK);
+      b = new_block (FUNCTION_BLOCK, SYMBOL_LANGUAGE (s));
       SYMBOL_BLOCK_VALUE (s) = b;
       BLOCK_FUNCTION (b) = s;
       BLOCK_START (b) = BLOCK_END (b) = sh->value;
@@ -1144,7 +1144,7 @@  parse_symbol (SYMR *sh, union aux_ext *ax, char *ext_sh, int bigend,
 	}
 
       top_stack->blocktype = stBlock;
-      b = new_block (NON_FUNCTION_BLOCK);
+      b = new_block (NON_FUNCTION_BLOCK, psymtab_language);
       BLOCK_START (b) = sh->value + top_stack->procadr;
       BLOCK_SUPERBLOCK (b) = top_stack->cur_block;
       top_stack->cur_block = b;
@@ -4026,6 +4026,7 @@  psymtab_to_symtab_1 (struct objfile *objfile,
 	  if (ECOFF_IS_STAB (&sh) || (name[0] == '#'))
 	    {
 	      int type_code = ECOFF_UNMARK_STAB (sh.index);
+	      enum language language = PST_PRIVATE (pst)->pst_language;
 
 	      /* We should never get non N_STAB symbols here, but they
 	         should be harmless, so keep process_one_symbol from
@@ -4053,14 +4054,14 @@  psymtab_to_symtab_1 (struct objfile *objfile,
 		    {
 		      last_symtab_ended = 0;
 		      process_one_symbol (type_code, 0, valu, name,
-					  section_offsets, objfile);
+					  section_offsets, objfile, language);
 		    }
 		}
 	      /* Similarly a hack.  */
 	      else if (name[0] == '#')
 		{
 		  process_one_symbol (N_SLINE, 0, valu, name,
-				      section_offsets, objfile);
+				      section_offsets, objfile, language);
 		}
 	      if (type_code == N_FUN)
 		{
@@ -4721,16 +4722,18 @@  new_symtab (const char *name, int maxlines, struct objfile *objfile)
   struct compunit_symtab *cust = allocate_compunit_symtab (objfile, name);
   struct symtab *symtab;
   struct blockvector *bv;
+  enum language lang;
 
   add_compunit_symtab_to_objfile (cust);
   symtab = allocate_symtab (cust, name);
 
   SYMTAB_LINETABLE (symtab) = new_linetable (maxlines);
+  lang = compunit_language (cust);
 
   /* All symtabs must have at least two blocks.  */
   bv = new_bvect (2);
-  BLOCKVECTOR_BLOCK (bv, GLOBAL_BLOCK) = new_block (NON_FUNCTION_BLOCK);
-  BLOCKVECTOR_BLOCK (bv, STATIC_BLOCK) = new_block (NON_FUNCTION_BLOCK);
+  BLOCKVECTOR_BLOCK (bv, GLOBAL_BLOCK) = new_block (NON_FUNCTION_BLOCK, lang);
+  BLOCKVECTOR_BLOCK (bv, STATIC_BLOCK) = new_block (NON_FUNCTION_BLOCK, lang);
   BLOCK_SUPERBLOCK (BLOCKVECTOR_BLOCK (bv, STATIC_BLOCK)) =
     BLOCKVECTOR_BLOCK (bv, GLOBAL_BLOCK);
   COMPUNIT_BLOCKVECTOR (cust) = bv;
@@ -4818,7 +4821,7 @@  new_bvect (int nblocks)
    hashed.  */
 
 static struct block *
-new_block (enum block_type type)
+new_block (enum block_type type, enum language language)
 {
   /* FIXME: carlton/2003-09-11: This should use allocate_block to
      allocate the block.  Which, in turn, suggests that the block
@@ -4826,9 +4829,9 @@  new_block (enum block_type type)
   struct block *retval = XCNEW (struct block);
 
   if (type == FUNCTION_BLOCK)
-    BLOCK_DICT (retval) = dict_create_linear_expandable ();
+    BLOCK_DICT (retval) = dict_create_linear_expandable (language);
   else
-    BLOCK_DICT (retval) = dict_create_hashed_expandable ();
+    BLOCK_DICT (retval) = dict_create_hashed_expandable (language);
 
   return retval;
 }
diff --git a/gdb/objc-lang.c b/gdb/objc-lang.c
index 6ffe85e..4dbc7f2 100644
--- a/gdb/objc-lang.c
+++ b/gdb/objc-lang.c
@@ -405,6 +405,7 @@  extern const struct language_defn objc_language_defn = {
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/opencl-lang.c b/gdb/opencl-lang.c
index 9b0f015..e5aff0d 100644
--- a/gdb/opencl-lang.c
+++ b/gdb/opencl-lang.c
@@ -1085,6 +1085,7 @@  extern const struct language_defn opencl_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/p-lang.c b/gdb/p-lang.c
index 439a377..2dca923 100644
--- a/gdb/p-lang.c
+++ b/gdb/p-lang.c
@@ -456,6 +456,7 @@  extern const struct language_defn pascal_language_defn =
   c_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/rust-lang.c b/gdb/rust-lang.c
index 817976a..f0d9968 100644
--- a/gdb/rust-lang.c
+++ b/gdb/rust-lang.c
@@ -2190,6 +2190,7 @@  extern const struct language_defn rust_language_defn =
   rust_watch_location_expression,
   NULL,				/* la_get_symbol_name_cmp */
   iterate_over_symbols,
+  default_search_name_hash,
   &default_varobj_ops,
   NULL,
   NULL,
diff --git a/gdb/stabsread.h b/gdb/stabsread.h
index b37be1a..b803cf9 100644
--- a/gdb/stabsread.h
+++ b/gdb/stabsread.h
@@ -17,6 +17,7 @@ 
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
 struct objfile;
+enum language;
 
 /* Definitions, prototypes, etc for stabs debugging format support
    functions.
@@ -169,7 +170,7 @@  extern struct partial_symtab *dbx_end_psymtab
 
 extern void process_one_symbol (int, int, CORE_ADDR, const char *,
 				const struct section_offsets *,
-				struct objfile *);
+				struct objfile *, enum language);
 
 extern void elfstab_build_psymtabs (struct objfile *objfile,
 				    asection *stabsect,
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 7c7ff7f..a875c4d 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -1802,6 +1802,14 @@  demangle_for_lookup (const char *name, enum language lang,
   return name;
 }
 
+/* See symtab.h.  */
+
+unsigned int
+search_name_hash (enum language language, const char *search_name)
+{
+  return language_def (language)->la_search_name_hash (search_name);
+}
+
 /* See symtab.h.
 
    This function (or rather its subordinates) have a bunch of loops and
diff --git a/gdb/symtab.h b/gdb/symtab.h
index fffe0f87..20904e4 100644
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -277,6 +277,9 @@  extern const char *symbol_search_name (const struct general_symbol_info *);
 #define SYMBOL_MATCHES_SEARCH_NAME(symbol, name)			\
   (strcmp_iw (SYMBOL_SEARCH_NAME (symbol), (name)) == 0)
 
+extern unsigned int search_name_hash (enum language language,
+				      const char *search_name);
+
 /* Classification types for a minimal symbol.  These should be taken as
    "advisory only", since if gdb can't easily figure out a
    classification it simply selects mst_unknown.  It may also have to
diff --git a/gdb/xcoffread.c b/gdb/xcoffread.c
index 138f941..0d6f290 100644
--- a/gdb/xcoffread.c
+++ b/gdb/xcoffread.c
@@ -1044,7 +1044,8 @@  read_xcoff_symtab (struct objfile *objfile, struct partial_symtab *pst)
   last_csect_name = 0;
 
   start_stabs ();
-  start_symtab (objfile, filestring, (char *) NULL, file_start_addr);
+  start_symtab (objfile, filestring, (char *) NULL, file_start_addr,
+		language_unknown);
   record_debugformat (debugfmt);
   symnum = ((struct symloc *) pst->read_symtab_private)->first_symnum;
   max_symnum =
@@ -1137,7 +1138,8 @@  read_xcoff_symtab (struct objfile *objfile, struct partial_symtab *pst)
 	    }
 
 	  start_stabs ();
-	  start_symtab (objfile, "_globals_", (char *) NULL, (CORE_ADDR) 0);
+	  start_symtab (objfile, "_globals_", (char *) NULL, (CORE_ADDR) 0,
+			language_unknown);
 	  record_debugformat (debugfmt);
 	  cur_src_end_addr = first_object_file_end;
 	  /* Done with all files, everything from here on is globals.  */
@@ -1227,7 +1229,7 @@  read_xcoff_symtab (struct objfile *objfile, struct partial_symtab *pst)
 			  /* Give all csects for this source file the same
 			     name.  */
 			  start_symtab (objfile, filestring, NULL,
-					(CORE_ADDR) 0);
+					(CORE_ADDR) 0, language_unknown);
 			  record_debugformat (debugfmt);
 			}
 
@@ -1347,7 +1349,8 @@  read_xcoff_symtab (struct objfile *objfile, struct partial_symtab *pst)
 	    filestring = cs->c_name;
 
 	  start_stabs ();
-	  start_symtab (objfile, filestring, (char *) NULL, (CORE_ADDR) 0);
+	  start_symtab (objfile, filestring, (char *) NULL, (CORE_ADDR) 0,
+			language_unknown);
 	  record_debugformat (debugfmt);
 	  last_csect_name = 0;