public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] Come up with htab_hash_string_vptr and use string-specific if possible.
@ 2018-10-31 17:32 Martin Liška
  2018-11-02  8:03 ` Richard Biener
  0 siblings, 1 reply; 11+ messages in thread
From: Martin Liška @ 2018-10-31 17:32 UTC (permalink / raw)
  To: gcc-patches

[-- Attachment #1: Type: text/plain, Size: 1525 bytes --]

Hi.

As seen in r265663 having htab_hash_string accepting const char * would
report a compilation error. The void * argument is needed for old C-style
htab used in libiberty. I'm suggesting to come up with htab_hash_string_vptr and
change signature of the old one (htab_hash_string). And putting these into
hashtab.h will make it inlinable in C++-style hash table.

Patch survives regression tests on ppc64le-linux-gnu.
Total cc1 change with the patch:
  +0.0% +2.09Ki

Hope it's acceptable.
Thanks,
Martin

gcc/ChangeLog:

2018-10-31  Martin Liska  <mliska@suse.cz>

	* gengtype-state.c (read_state): Use newly added
	htab_hash_string_vptr.
	* gensupport.c (gen_mnemonic_attr): Likewise.
	(check_define_attr_duplicates): Likewise.
	* godump.c (go_finish): Likewise.

include/ChangeLog:

2018-10-31  Martin Liska  <mliska@suse.cz>

	* hashtab.h (htab_hash_string): Change signature
	to const char * and make it static inline.
	(htab_hash_string_vptr): Likewise.

libcpp/ChangeLog:

2018-10-31  Martin Liska  <mliska@suse.cz>

	* files.c (_cpp_init_files): Use htab_hash_string_vptr.

libiberty/ChangeLog:

2018-10-31  Martin Liska  <mliska@suse.cz>

	* hashtab.c:
	(htab_hash_string): Move to header file.
---
 gcc/gengtype-state.c |  2 +-
 gcc/gensupport.c     |  4 ++--
 gcc/godump.c         |  6 +++---
 include/hashtab.h    | 47 ++++++++++++++++++++++++++++++++++++++++++--
 libcpp/files.c       |  2 +-
 libiberty/hashtab.c  | 38 -----------------------------------
 6 files changed, 52 insertions(+), 47 deletions(-)



[-- Attachment #2: 0001-Come-up-with-htab_hash_string_vptr-and-use-string-sp.patch --]
[-- Type: text/x-patch, Size: 6601 bytes --]

diff --git a/gcc/gengtype-state.c b/gcc/gengtype-state.c
index 6a3b981eedc..facff4a29d7 100644
--- a/gcc/gengtype-state.c
+++ b/gcc/gengtype-state.c
@@ -2588,7 +2588,7 @@ read_state (const char *path)
   state_seen_types =
     htab_create (2017, hash_type_number, equals_type_number, NULL);
   state_ident_tab =
-    htab_create (4027, htab_hash_string, string_eq, NULL);
+    htab_create (4027, htab_hash_string_vptr, string_eq, NULL);
   read_state_version (version_string);
   read_state_srcdir ();
   read_state_languages ();
diff --git a/gcc/gensupport.c b/gcc/gensupport.c
index 1d76af7ce5c..2f865d918f3 100644
--- a/gcc/gensupport.c
+++ b/gcc/gensupport.c
@@ -2434,7 +2434,7 @@ gen_mnemonic_attr (void)
   if (!mnemonic_attr)
     return;
 
-  mnemonic_htab = htab_create_alloc (MNEMONIC_HTAB_SIZE, htab_hash_string,
+  mnemonic_htab = htab_create_alloc (MNEMONIC_HTAB_SIZE, htab_hash_string_vptr,
 				     htab_eq_string, 0, xcalloc, free);
 
   for (elem = define_insn_queue; elem; elem = elem->next)
@@ -2492,7 +2492,7 @@ check_define_attr_duplicates ()
   char * attr_name;
   void **slot;
 
-  attr_htab = htab_create (500, htab_hash_string, htab_eq_string, NULL);
+  attr_htab = htab_create (500, htab_hash_string_vptr, htab_eq_string, NULL);
 
   for (elem = define_attr_queue; elem; elem = elem->next)
     {
diff --git a/gcc/godump.c b/gcc/godump.c
index baf21e27eec..7098ac80a91 100644
--- a/gcc/godump.c
+++ b/gcc/godump.c
@@ -1377,11 +1377,11 @@ go_finish (const char *filename)
 
   real_debug_hooks->finish (filename);
 
-  container.type_hash = htab_create (100, htab_hash_string,
+  container.type_hash = htab_create (100, htab_hash_string_vptr,
                                      string_hash_eq, NULL);
-  container.invalid_hash = htab_create (10, htab_hash_string,
+  container.invalid_hash = htab_create (10, htab_hash_string_vptr,
 					string_hash_eq, NULL);
-  container.keyword_hash = htab_create (50, htab_hash_string,
+  container.keyword_hash = htab_create (50, htab_hash_string_vptr,
                                         string_hash_eq, NULL);
   obstack_init (&container.type_obstack);
 
diff --git a/include/hashtab.h b/include/hashtab.h
index 531eb101ff7..44bce8c2a6b 100644
--- a/include/hashtab.h
+++ b/include/hashtab.h
@@ -189,8 +189,51 @@ extern htab_hash htab_hash_pointer;
 /* An equality function for pointers.  */
 extern htab_eq htab_eq_pointer;
 
-/* A hash function for null-terminated strings.  */
-extern hashval_t htab_hash_string (const void *);
+/* Hash STR as a null-terminated string.
+
+   Copied from gcc/hashtable.c.  Zack had the following to say with respect
+   to applicability, though note that unlike hashtable.c, this hash table
+   implementation re-hashes rather than chain buckets.
+
+   http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01021.html
+   From: Zack Weinberg <zackw@panix.com>
+   Date: Fri, 17 Aug 2001 02:15:56 -0400
+
+   I got it by extracting all the identifiers from all the source code
+   I had lying around in mid-1999, and testing many recurrences of
+   the form "H_n = H_{n-1} * K + c_n * L + M" where K, L, M were either
+   prime numbers or the appropriate identity.  This was the best one.
+   I don't remember exactly what constituted "best", except I was
+   looking at bucket-length distributions mostly.
+
+   So it should be very good at hashing identifiers, but might not be
+   as good at arbitrary strings.
+
+   I'll add that it thoroughly trounces the hash functions recommended
+   for this use at http://burtleburtle.net/bob/hash/index.html, both
+   on speed and bucket distribution.  I haven't tried it against the
+   function they just started using for Perl's hashes.  */
+
+static inline hashval_t
+htab_hash_string (const char *str)
+{
+  hashval_t r = 0;
+  unsigned char c;
+
+  while ((c = *str++) != 0)
+    r = r * 67 + c - 113;
+
+  return r;
+}
+
+/* Hash P as a null-terminated string.  */
+
+static inline hashval_t
+htab_hash_string_vptr (const PTR p)
+{
+  return htab_hash_string ((const char *)p);
+}
+
 
 /* An iterative hash function for arbitrary data.  */
 extern hashval_t iterative_hash (const void *, size_t, hashval_t);
diff --git a/libcpp/files.c b/libcpp/files.c
index 08b7c647c91..e4eb34e7d55 100644
--- a/libcpp/files.c
+++ b/libcpp/files.c
@@ -1311,7 +1311,7 @@ _cpp_init_files (cpp_reader *pfile)
   pfile->dir_hash = htab_create_alloc (127, file_hash_hash, file_hash_eq,
 					NULL, xcalloc, free);
   allocate_file_hash_entries (pfile);
-  pfile->nonexistent_file_hash = htab_create_alloc (127, htab_hash_string,
+  pfile->nonexistent_file_hash = htab_create_alloc (127, htab_hash_string_vptr,
 						    nonexistent_file_hash_eq,
 						    NULL, xcalloc, free);
   obstack_specify_allocation (&pfile->nonexistent_file_ob, 0, 0,
diff --git a/libiberty/hashtab.c b/libiberty/hashtab.c
index ca392647d4a..b6821502d15 100644
--- a/libiberty/hashtab.c
+++ b/libiberty/hashtab.c
@@ -803,44 +803,6 @@ htab_collisions (htab_t htab)
   return (double) htab->collisions / (double) htab->searches;
 }
 
-/* Hash P as a null-terminated string.
-
-   Copied from gcc/hashtable.c.  Zack had the following to say with respect
-   to applicability, though note that unlike hashtable.c, this hash table
-   implementation re-hashes rather than chain buckets.
-
-   http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01021.html
-   From: Zack Weinberg <zackw@panix.com>
-   Date: Fri, 17 Aug 2001 02:15:56 -0400
-
-   I got it by extracting all the identifiers from all the source code
-   I had lying around in mid-1999, and testing many recurrences of
-   the form "H_n = H_{n-1} * K + c_n * L + M" where K, L, M were either
-   prime numbers or the appropriate identity.  This was the best one.
-   I don't remember exactly what constituted "best", except I was
-   looking at bucket-length distributions mostly.
-   
-   So it should be very good at hashing identifiers, but might not be
-   as good at arbitrary strings.
-   
-   I'll add that it thoroughly trounces the hash functions recommended
-   for this use at http://burtleburtle.net/bob/hash/index.html, both
-   on speed and bucket distribution.  I haven't tried it against the
-   function they just started using for Perl's hashes.  */
-
-hashval_t
-htab_hash_string (const PTR p)
-{
-  const unsigned char *str = (const unsigned char *) p;
-  hashval_t r = 0;
-  unsigned char c;
-
-  while ((c = *str++) != 0)
-    r = r * 67 + c - 113;
-
-  return r;
-}
-
 /* DERIVED FROM:
 --------------------------------------------------------------------
 lookup2.c, by Bob Jenkins, December 1996, Public Domain.


^ permalink raw reply	[flat|nested] 11+ messages in thread
end of thread, other threads:[~2018-11-12 13:43 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-10-31 17:32 [PATCH] Come up with htab_hash_string_vptr and use string-specific if possible Martin Liška
2018-11-02  8:03 ` Richard Biener
2018-11-02  8:10   ` Martin Liška
2018-11-02  9:25     ` Martin Liška
2018-11-05 18:37       ` Michael Matz
2018-11-07  8:54         ` Martin Liška
2018-11-08 13:15           ` Michael Matz
2018-11-08 13:41             ` Martin Liška
2018-11-09 13:28               ` Michael Matz
2018-11-12  9:35                 ` Martin Liška
2018-11-12 13:43                   ` Michael Matz

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).