Looked for symbols of function.

git-svn-id: svn://svn.gna.org/svn/chrysalide/trunk@23 abbe820e-26c8-41b2-8c08-b7b2b41f8b0a

9 files changed, 369 insertions, 5 deletions

diff --git a/ChangeLog b/ChangeLog
index eb4fb7d..3516c02 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,5 +1,27 @@
 2008-08-31  Cyrille Bagard <nocbos@gmail.com>
 
+	* src/binary.c:
+	Display the found [and sorted] symbols. 
+
+	* src/format/elf/e_elf.c:
+	Load the ELF symbols.
+
+	* src/format/elf/elf-int.h:
+	Define an ELF symbol (name and address).
+
+	* src/format/elf/Makefile.am:
+	Add symbol.[ch] to libformatelf_a_SOURCES.
+
+	* src/format/elf/symbol.c:
+	* src/format/elf/symbol.h:
+	New entries: look for symbols of function.
+
+	* src/gtksnippet.c:
+	* src/gtksnippet.h:
+	Add a way to compare two lines of code.
+
+2008-08-31  Cyrille Bagard <nocbos@gmail.com>
+
 	* src/arch/x86/instruction.h:
 	* src/arch/x86/opcodes.h:
 	Handle the new opcodes.
diff --git a/src/binary.c b/src/binary.c
index 862a4d0..b7ee15a 100644
--- a/src/binary.c
+++ b/src/binary.c
@@ -25,6 +25,7 @@
 
 
 #include <fcntl.h>
+#include <stdlib.h>
 #include <unistd.h>
 #include <sys/mman.h>
 #include <sys/stat.h>
@@ -165,14 +166,23 @@ void fill_snippet(GtkSnippet *snippet)
 
 
 
-    comments_count = get_dwarf_comments(dformat, &comments, &offsets);
+    //comments_count = get_dwarf_comments(dformat, &comments, &offsets);
+
+    comments = NULL;
+    offsets = NULL;
+    comments_count = 0;
+
+    get_elf_symbol_comments(format, &comments, &offsets, &comments_count);
 
     comments_list = (code_line_info **)calloc(comments_count, sizeof(code_line_info *));
 
     for (i = 0; i < comments_count; i++)
         comments_list[i] = create_code_line_info(offsets[i], NULL, strdup(comments[i]));
 
-    i = 0;
+
+    qsort(comments_list, comments_count, sizeof(code_line_info *), compare_code_line_info);
+
+
 
     find_exe_section(format, ".text", &pos, &len, &base);
 
@@ -187,6 +197,9 @@ void fill_snippet(GtkSnippet *snippet)
 
     offset = base;
 
+    for (i = 0; i < comments_count; i++)
+        if (comments_list[i]->offset >= base) break;
+
 
     gtk_snippet_set_processor(snippet, proc);
 
@@ -206,7 +219,7 @@ void fill_snippet(GtkSnippet *snippet)
         offset = base + pos;
 
         /* Si on a un commentaire pour cette ligne... */
-        if (comments_count > 0 && offsets[i] == offset)
+        if (i < comments_count && comments_list[i]->offset == offset)
         {
             list = (code_line_info **)realloc(list, ++list_len * sizeof(code_line_info *));
             list[list_len - 1] = comments_list[i++];
@@ -232,8 +245,10 @@ void fill_snippet(GtkSnippet *snippet)
         /* TODO: free() */
     }
 
-    ret = munmap(bin_data, length);
 
+    /****
+    ret = munmap(bin_data, length);
+    ****/
 
     /*
     gtk_snippet_build_content(snippet);
diff --git a/src/format/elf/Makefile.am b/src/format/elf/Makefile.am
index cb5682d..4886115 100644
--- a/src/format/elf/Makefile.am
+++ b/src/format/elf/Makefile.am
@@ -4,7 +4,8 @@ lib_LIBRARIES = libformatelf.a
 libformatelf_a_SOURCES =				\
 	e_elf.h e_elf.c						\
 	elf-int.h							\
-	section.h section.c
+	section.h section.c					\
+	symbol.h symbol.c
 
 libformatelf_a_CFLAGS = $(AM_CFLAGS)
 
diff --git a/src/format/elf/e_elf.c b/src/format/elf/e_elf.c
index 1c08cf3..a6399ab 100644
--- a/src/format/elf/e_elf.c
+++ b/src/format/elf/e_elf.c
@@ -30,6 +30,7 @@
 
 #include "elf-int.h"
 #include "section.h"
+#include "symbol.h"
 
 
 
@@ -66,6 +67,10 @@ elf_format *load_elf(const uint8_t *content, off_t length)
 
     printf("ok ? %d\n", test);
 
+    test = load_elf_symbols(result);
+
+    printf("ok ? %d\n", test);
+
 
     return result;
 
diff --git a/src/format/elf/elf-int.h b/src/format/elf/elf-int.h
index 0fd325f..2ec33c3 100644
--- a/src/format/elf/elf-int.h
+++ b/src/format/elf/elf-int.h
@@ -34,6 +34,15 @@
 
 
 
+/* Symbole trouvé */
+typedef struct _elf_symbol
+{
+    const char *name;                       /* Désignation du symbole      */
+    uint64_t address;                       /* Adresse du symbole          */
+
+} elf_symbol;
+
+
 
 /* Description du format ELF */
 struct _elf_format
@@ -45,6 +54,8 @@ struct _elf_format
     char *sec_names;                        /* Noms des sections           */
     size_t sec_size;                        /* Taille de ces définitions   */
 
+    elf_symbol *symbols;                    /* Liste des symboles          */
+    size_t sym_count;                       /* Taille de cette liste       */
 
 };
 
diff --git a/src/format/elf/symbol.c b/src/format/elf/symbol.c
new file mode 100644
index 0000000..729c154
--- /dev/null
+++ b/src/format/elf/symbol.c
@@ -0,0 +1,243 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * symbol.c - gestion des symboles d'un ELF
+ *
+ * Copyright (C) 2008 Cyrille Bagard
+ *
+ *  This file is part of OpenIDA.
+ *
+ *  OpenIDA is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 3 of the License, or
+ *  (at your option) any later version.
+ *
+ *  OpenIDA is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with Foobar.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+#include "symbol.h"
+
+
+#include <malloc.h>
+#include <elf.h>
+#include <string.h>
+
+
+
+#include "elf-int.h"
+#include "section.h"
+
+
+
+
+
+/* Charge en mémoire la liste humaine des symboles (32 bits). */
+bool load_elf_symbol_table_32(elf_format *, const off_t *, const off_t *, const off_t *, const off_t *);
+
+/* Charge en mémoire la liste humaine des symboles (64 bits). */
+bool load_elf_symbol_table_64(elf_format *, const off_t *, const off_t *, const off_t *, const off_t *);
+
+
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : format = description de l'exécutable à compléter.            *
+*                                                                             *
+*  Description : Charge en mémoire la liste humaine des symboles.             *
+*                                                                             *
+*  Retour      : Bilan de l'opération.                                        *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+bool load_elf_symbols(elf_format *format)
+{
+    bool result;                            /* Bilan à retourner           */
+
+
+    off_t sym_start;                        /* Début de section            */
+    off_t sym_size;                         /* Taille de section           */
+    off_t str_start;                        /* Début de section            */
+    off_t str_size;                         /* Taille de section           */
+
+
+    result = find_elf_section(format, ".symtab", &sym_start, &sym_size, NULL);
+
+    result &= find_elf_section(format, ".strtab", &str_start, &str_size, NULL);
+
+
+    if (result)
+    {
+
+
+
+        result = load_elf_symbol_table_32(format, &sym_start, &sym_size, &str_start, &str_start);
+
+
+
+    }
+
+    return result;
+
+}
+
+
+
+
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : format     = description de l'exécutable à compléter.        *
+*                sym_start  = début de la zone à traiter.                     *
+*                sym_size   = taille de la zone à traiter.                    *
+*                str_start  = début de la zone de chaîne de caractères.       *
+*                str_size   = taille de la zone de chaînes de caractères.     *
+*                                                                             *
+*  Description : Charge en mémoire la liste humaine des symboles (32 bits).   *
+*                                                                             *
+*  Retour      : Bilan de l'opération.                                        *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+bool load_elf_symbol_table_32(elf_format *format, const off_t *sym_start, const off_t *sym_size, const off_t *str_start, const off_t *str_size)
+{
+    off_t iter;                             /* Boucle de parcours          */
+    Elf32_Sym symbol;                       /* Symbole ELF lu              */
+
+    if (*sym_size % sizeof(Elf32_Sym) != 0) return false;
+
+    for (iter = *sym_start; iter < (*sym_start + *sym_size); iter += sizeof(Elf32_Sym))
+    {
+        memcpy(&symbol, &EXE_FORMAT(format)->content[iter], sizeof(Elf32_Sym));
+
+        if (!(ELF32_ST_TYPE(symbol.st_info) == STT_FUNC
+              || (ELF32_ST_TYPE(symbol.st_info) == STT_NOTYPE
+                  && ELF32_ST_BIND(symbol.st_info) == STB_GLOBAL))) continue;
+
+        if (symbol.st_value == 0) continue;
+
+        /* Sécurité anti-débordements */
+        if (symbol.st_name >= *str_size) continue;
+
+        /* Si le symbole possède un nom... */
+        if (strlen(&EXE_FORMAT(format)->content[*str_start + symbol.st_name]) > 0)
+        {
+            format->symbols = (elf_symbol *)realloc(format->symbols, ++format->sym_count * sizeof(elf_symbol));
+
+            format->symbols[format->sym_count - 1].name = &EXE_FORMAT(format)->content[*str_start + symbol.st_name];
+            format->symbols[format->sym_count - 1].address = symbol.st_value;
+
+        }
+
+    }
+
+    return true;
+
+}
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : format     = description de l'exécutable à compléter.        *
+*                sym_start  = début de la zone à traiter.                     *
+*                sym_size   = taille de la zone à traiter.                    *
+*                str_start  = début de la zone de chaîne de caractères.       *
+*                str_size   = taille de la zone de chaînes de caractères.     *
+*                                                                             *
+*  Description : Charge en mémoire la liste humaine des symboles (64 bits).   *
+*                                                                             *
+*  Retour      : Bilan de l'opération.                                        *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+bool load_elf_symbol_table_64(elf_format *format, const off_t *sym_start, const off_t *sym_size, const off_t *str_start, const off_t *str_size)
+{
+    off_t iter;                             /* Boucle de parcours          */
+    Elf64_Sym symbol;                       /* Symbole ELF lu              */
+
+    if (*sym_size % sizeof(Elf64_Sym) != 0) return false;
+
+    for (iter = *sym_start; iter < (*sym_start + *sym_size); iter += sizeof(Elf64_Sym))
+    {
+        memcpy(&symbol, &EXE_FORMAT(format)->content[iter], sizeof(Elf64_Sym));
+
+        if (!(ELF64_ST_TYPE(symbol.st_info) == STT_FUNC
+              || (ELF64_ST_TYPE(symbol.st_info) == STT_NOTYPE
+                  && ELF64_ST_BIND(symbol.st_info) == STB_GLOBAL))) continue;
+
+        if (symbol.st_value == 0) continue;
+
+        /* Sécurité anti-débordements */
+        if (symbol.st_name >= *str_size) continue;
+
+        /* Si le symbole possède un nom... */
+        if (strlen(&EXE_FORMAT(format)->content[*str_start + symbol.st_name]) > 0)
+        {
+            format->symbols = (elf_symbol *)realloc(format->symbols, ++format->sym_count * sizeof(elf_symbol));
+
+            format->symbols[format->sym_count - 1].name = &EXE_FORMAT(format)->content[*str_start + symbol.st_name];
+            format->symbols[format->sym_count - 1].address = symbol.st_value;
+
+        }
+
+    }
+
+    return true;
+
+}
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : format   = informations chargées à consulter.                *
+*                comments = liste des commentaires à insérer. [OUT]           *
+*                offsets  = liste des indices des commentaires. [OUT]         *
+*                count    = taille des listes construites. [OUT]              *
+*                                                                             *
+*  Description : Récupère tous les commentaires à insérer dans le code.       *
+*                                                                             *
+*  Retour      : -                                                            *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+void get_elf_symbol_comments(const elf_format *format, char ***comments, uint64_t **offsets, size_t *count)
+{
+    size_t i;                               /* Boucle de parcours          */
+    size_t len;                             /* Longueur d'une désignation  */
+
+    if (format->sym_count > 0)
+    {
+        *comments = (char **)calloc(*count + format->sym_count, sizeof(char *));
+        *offsets = (uint64_t *)calloc(*count + format->sym_count, sizeof(uint64_t));
+
+        for (i = 0; i < format->sym_count; i++)
+        {
+            len = strlen(format->symbols[i].name);
+
+            (*comments)[i] = (char *)calloc(len + 9, sizeof(char));
+            snprintf((*comments)[i], len + 9, "&lt;%s&gt;", format->symbols[i].name);
+
+            (*offsets)[i] = format->symbols[i].address;
+
+        }
+
+        *count += format->sym_count;
+
+    }
+
+}
diff --git a/src/format/elf/symbol.h b/src/format/elf/symbol.h
new file mode 100644
index 0000000..0a8f203
--- /dev/null
+++ b/src/format/elf/symbol.h
@@ -0,0 +1,40 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * symbol.h - prototypes pour la gestion des symboles d'un ELF
+ *
+ * Copyright (C) 2008 Cyrille Bagard
+ *
+ *  This file is part of OpenIDA.
+ *
+ *  OpenIDA is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 3 of the License, or
+ *  (at your option) any later version.
+ *
+ *  OpenIDA is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with Foobar.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+
+#ifndef _FORMAT_ELF_SYMBOL_H
+#define _FORMAT_ELF_SYMBOL_H
+
+
+#include "e_elf.h"
+
+
+
+/* Charge en mémoire la liste humaine des symboles. */
+bool load_elf_symbols(elf_format *);
+
+/* Récupère tous les commentaires à insérer dans le code. */
+void get_elf_symbol_comments(const elf_format *, char ***, uint64_t **, size_t *);
+
+
+
+#endif  /* _FORMAT_ELF_SYMBOL_H */
diff --git a/src/gtksnippet.c b/src/gtksnippet.c
index 6470fcb..7e5e012 100644
--- a/src/gtksnippet.c
+++ b/src/gtksnippet.c
@@ -90,7 +90,30 @@ void delete_code_line_info(code_line_info *line)
 }
 
 
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : a = premières informations à consulter.                      *
+*                b = secondes informations à consulter.                       *
+*                                                                             *
+*  Description : Etablit la comparaison entre deux lignes de représentation.  *
+*                                                                             *
+*  Retour      : Bilan : -1 (a < b), 0 (a == b) ou 1 (a > b).                 *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+int compare_code_line_info(const code_line_info **a, const code_line_info **b)
+{
+    int result;                             /* Bilan à renvoyer            */
+
+    if ((*a)->offset < (*b)->offset) result = -1;
+    else if((*a)->offset > (*b)->offset) result = 1;
+    else result = 0;
 
+    return result;
+
+}
 
 
 
diff --git a/src/gtksnippet.h b/src/gtksnippet.h
index 1858796..6e89a3c 100644
--- a/src/gtksnippet.h
+++ b/src/gtksnippet.h
@@ -56,6 +56,10 @@ code_line_info *create_code_line_info(uint64_t, asm_instr *, const char *);
 /* Supprime une ligne de représentation. */
 void delete_code_line_info(code_line_info *);
 
+/* Etablit la comparaison entre deux lignes de représentation. */
+int compare_code_line_info(const code_line_info **, const code_line_info **);
+
+