public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Jan Beulich <jbeulich@suse.com>
To: Binutils <binutils@sourceware.org>
Cc: "H.J. Lu" <hjl.tools@gmail.com>
Subject: [PATCH 3/8] x86: re-use insn mnemonic strings as much as possible
Date: Fri, 13 Jan 2023 12:07:24 +0100	[thread overview]
Message-ID: <d87c0ed5-4678-d8ea-9744-7447ff82a843@suse.com> (raw)
In-Reply-To: <beeaf9c6-7d1e-bcc6-624a-918c4a6c293e@suse.com>

Compact the mnemonic string table such that the tails of longer
mnemonics are re-used for shorter ones, going beyond what compilers
would typically do, but matching what ELF linkers may do when processing
SHF_MERGE|SHF_STRINGS sections. This reduces table size by about 12.5%.

--- a/opcodes/i386-gen.c
+++ b/opcodes/i386-gen.c
@@ -1576,12 +1576,29 @@ expand_templates (char *name, const char
   return idx;
 }
 
+static int mnemonic_cmp(const void *p1, const void *p2)
+{
+  const struct opcode_hash_entry *const *e1 = p1, *const *e2 = p2;
+  const char *s1 = (*e1)->name, *s2 = (*e2)->name;
+  unsigned int i;
+  size_t l1 = strlen (s1), l2 = strlen (s2);
+
+  for (i = 1; i <= l1 && i <= l2; ++i)
+    {
+      if (s1[l1 - i] != s2[l2 - i])
+	return (unsigned char)s1[l1 - i] - (unsigned char)s2[l2 - i];
+    }
+
+  return (int)(l1 - l2);
+}
+
 static void
 process_i386_opcodes (FILE *table)
 {
   FILE *fp;
   char buf[2048];
   unsigned int i, j, nr, offs;
+  size_t l;
   char *str, *p, *last, *name;
   htab_t opcode_hash_table;
   struct opcode_hash_entry **opcode_array = NULL;
@@ -1720,6 +1737,8 @@ process_i386_opcodes (FILE *table)
   fprintf (table, "};\n");
 
   /* Emit mnemonics and associated #define-s.  */
+  qsort (opcode_array, i, sizeof (*opcode_array), mnemonic_cmp);
+
   fp = fopen ("i386-mnem.h", "w");
   if (fp == NULL)
     fail (_("can't create i386-mnem.h, errno = %s\n"),
@@ -1731,14 +1750,28 @@ process_i386_opcodes (FILE *table)
   fprintf (table, "const char i386_mnemonics[] =\n");
   fprintf (fp, "\nextern const char i386_mnemonics[];\n\n");
 
-  for (offs = j = 0; j < i; j++)
+  for (l = strlen (opcode_array[offs = j = 0]->name); j < i; j++)
     {
+      const char *next = NULL;
+      size_t l1 = j + 1 < i ? strlen(next = opcode_array[j + 1]->name) : 0;
+
       name = opcode_array[j]->name;
-      fprintf (table, "  \"\\0\"\"%s\"\n", name);
       str = mkident (name);
-      fprintf (fp, "#define MN_%s %#x\n", str, offs + 1);
+      if (l < l1 && !strcmp(name, next + l1 - l))
+	{
+	  fprintf (fp, "#define MN_%s ", str);
+	  free (str);
+	  str = mkident (next);
+	  fprintf (fp, "(MN_%s + %u)\n", str, l1 - l);
+	}
+      else
+	{
+	  fprintf (table, "  \"\\0\"\"%s\"\n", name);
+	  fprintf (fp, "#define MN_%s %#x\n", str, offs + 1);
+	  offs += strlen (name) + 1;
+	}
       free (str);
-      offs += strlen (name) + 1;
+      l = l1;
     }
 
   fprintf (table, ";\n");


  parent reply	other threads:[~2023-01-13 11:07 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-13 11:04 [PATCH 0/8] x86: mnemonic and register string literals Jan Beulich
2023-01-13 11:05 ` [PATCH 1/8] x86: abstract out obtaining of a template's mnemonic Jan Beulich
2023-01-13 11:06 ` [PATCH 1/8] x86: move insn mnemonics to a separate table Jan Beulich
2023-01-13 11:12   ` Jan Beulich
2023-01-13 11:07 ` Jan Beulich [this message]
2023-01-13 11:07 ` [PATCH 4/8] x86: absorb allocation in i386-gen Jan Beulich
2023-01-13 11:08 ` [PATCH 5/8] x86: avoid strcmp() in a few places Jan Beulich
2023-01-13 11:10 ` [PATCH 6/8] x86: embed register names in reg_entry Jan Beulich
2023-01-13 11:11 ` [PATCH 7/8] x86: embed register and alike names in disassembler Jan Beulich
2023-01-13 11:11 ` [PATCH 8/8] x86: split i386-gen's opcode hash entry struct Jan Beulich

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d87c0ed5-4678-d8ea-9744-7447ff82a843@suse.com \
    --to=jbeulich@suse.com \
    --cc=binutils@sourceware.org \
    --cc=hjl.tools@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).