Followed the excution flow to decompile instructions.

git-svn-id: svn://svn.gna.org/svn/chrysalide/trunk@293 abbe820e-26c8-41b2-8c08-b7b2b41f8b0a

6 files changed, 536 insertions, 4 deletions

diff --git a/src/analysis/decomp/Makefile.am b/src/analysis/decomp/Makefile.am
index 7b797bd..27388d6 100755
--- a/src/analysis/decomp/Makefile.am
+++ b/src/analysis/decomp/Makefile.am
@@ -2,7 +2,9 @@
 noinst_LTLIBRARIES = libanalysisdecomp.la
 
 libanalysisdecomp_la_SOURCES =			\
-	decompiler.h decompiler.c
+	decompiler.h decompiler.c			\
+	il.h il.c							\
+	reduce.h reduce.c
 
 libanalysisdecomp_la_LDFLAGS = 
 
diff --git a/src/analysis/decomp/decompiler.c b/src/analysis/decomp/decompiler.c
index 8e574c0..91a9163 100644
--- a/src/analysis/decomp/decompiler.c
+++ b/src/analysis/decomp/decompiler.c
@@ -2,7 +2,7 @@
 /* OpenIDA - Outil d'analyse de fichiers binaires
  * decompiler.c - encadrement des phases de décompilation
  *
- * Copyright (C) 2010 Cyrille Bagard
+ * Copyright (C) 2010-2012 Cyrille Bagard
  *
  *  This file is part of OpenIDA.
  *
@@ -32,7 +32,9 @@
 #include <i18n.h>
 
 
+#include "il.h"
 #include "../../decomp/output.h"
+#include "../../decomp/expr/block.h"
 #include "../../decomp/lang/java.h"    /* FIXME : remme ! */
 #include "../../format/format.h"
 
@@ -138,7 +140,11 @@ static void prepare_all_routines_for_decomp(const GLoadedBinary *binary, const c
     size_t i;
 
     GDecContext *context;                   /* Contexte pour la décompil.  */
-    GDecInstruction *instr;
+    GDecInstruction *dinstrs;
+
+    GArchInstruction *instrs;                /* Instructions natives        */
+
+    vmpa_t max;                             /* Première adresse à écarter  */
 
     format = g_loaded_binary_get_format(binary);
     proc = get_arch_processor_from_format(G_EXE_FORMAT(format));
@@ -152,7 +158,28 @@ static void prepare_all_routines_for_decomp(const GLoadedBinary *binary, const c
     {
         context = g_arch_processor_get_decomp_context(proc);
 
-        instr = g_binary_format_decompile_routine(G_BIN_FORMAT(format), routines[i], context);
+        g_object_set_data(G_OBJECT(context), "format", format);
+        g_object_set_data(G_OBJECT(context), "routine", routines[i]);
+        g_dec_context_set_max_address(context, max);
+
+        instrs = g_binary_routine_get_instructions(routines[i]);
+
+        max = g_binary_routine_get_address(routines[i])
+            + g_binary_routine_get_size(routines[i]);
+
+        printf("##### DECOMPILE '%s' #####\n", g_binary_routine_to_string(routines[i]));
+
+        dinstrs = build_decompiled_block(instrs,
+                                         g_binary_routine_get_address(routines[i]),
+                                         max, VMPA_MAX, context);
+
+        //instr = g_binary_format_decompile_routine(G_BIN_FORMAT(format), routines[i], context);
+
+
+
+        g_binary_routine_set_decomp_instructions(routines[i], dinstrs);
+
+
 
         if (context != NULL)
             g_object_unref(context);
diff --git a/src/analysis/decomp/il.c b/src/analysis/decomp/il.c
new file mode 100644
index 0000000..d9b9588
--- /dev/null
+++ b/src/analysis/decomp/il.c
@@ -0,0 +1,382 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * il.h - mise en place d'un langage intermédiaire
+ *
+ * Copyright (C) 2012 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 "il.h"
+
+
+#include <malloc.h>
+#include <stdlib.h>
+#include <string.h>
+
+
+#include "../../decomp/expr/block.h"
+#include "../../decomp/instr/ite.h"
+
+
+
+/* Indications sur une branche */
+typedef struct _branch_info
+{
+    vmpa_t *jumps;                          /* Jalons de la branche        */
+    size_t count;                           /* Quantité de ces jalons      */
+
+} branch_info;
+
+
+/* Indique si une adresse est retenue comme point de passage. */
+static bool is_addr_in_branch(const branch_info *, const vmpa_t *, bool);
+
+/* Identifie les différents points de passage d'une branche. */
+static void find_next_jumps(GArchInstruction *, vmpa_t, vmpa_t, branch_info *);
+
+/* Retrouve le point de ralliement entre deux branches. */
+static vmpa_t compute_first_common_addr(branch_info *, branch_info *);
+
+
+
+
+
+
+
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : info = informations à consulter.                             *
+*                addr = adresse à rechercher.                                 *
+*                fast = autorise une recherche rapide.                        *
+*                                                                             *
+*  Description : Indique si une adresse est retenue comme point de passage.   *
+*                                                                             *
+*  Retour      : true si le jalon est déjà dans la liste, false sinon.        *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+static bool is_addr_in_branch(const branch_info *info, const vmpa_t *addr, bool fast)
+{
+    bool result;                            /* Bilan à retourner           */
+    size_t i;                               /* Boucle de parcours          */
+    void *ptr;                              /* Résultat des recherches     */
+
+    result = false;
+
+    if (!fast)
+        for (i = 0; i < info->count && !result; i++)
+            result = (info->jumps[i] == *addr);
+
+    else
+    {
+        ptr = bsearch(addr, info->jumps, info->count, sizeof(vmpa_t), (__compar_fn_t)compare_vmpa);
+        result = (ptr != NULL);
+    }
+
+    return result;
+
+}
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : instrs = ensemble des instructions d'assemblage.             *
+*                start  = adresse de début du bloc.                           *
+*                end    = adresse de fin du bloc (exclusive).                 *
+*                count  = nombre de sauts détectés. [OUT]                     *
+*                                                                             *
+*  Description : Identifie les différents points de passage d'une branche.    *
+*                                                                             *
+*  Retour      : Jalons dans le flot d'exécution.                             *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+static void find_next_jumps(GArchInstruction *instrs, vmpa_t start, vmpa_t end, branch_info *info)
+{
+    GArchInstruction *iter;                 /* Boucle de parcours #1       */
+    GArchInstruction **dests;               /* Instr. visée par une autre  */
+    InstructionLinkType *types;             /* Type de lien entre lignes   */
+    size_t dcount;                          /* Nombre de liens de dest.    */
+    size_t i;                               /* Boucle de parcours #2       */
+    vmpa_t addr;                            /* Adresse de la destination   */
+
+    /* On évite de boucler... */
+    if (is_addr_in_branch(info, &start, false))
+        return;
+
+    info->jumps = (vmpa_t *)realloc(info->jumps, ++(info->count) * sizeof(vmpa_t));
+    info->jumps[info->count - 1] = start;
+
+    /* On suit le flot jusqu'à la prochaine bifurcation */
+    for (iter = g_arch_instruction_find_by_address(instrs, start, true);
+         iter != NULL;
+         iter = g_arch_instruction_get_next_iter(instrs, iter, end))
+    {
+        if (!g_arch_instruction_has_destinations(iter))
+            continue;
+
+        dcount = g_arch_instruction_get_destinations(iter, &dests, &types);
+
+        for (i = 0; i < dcount; i++)
+            switch (types[i])
+            {
+                case ILT_EXEC_FLOW:
+                case ILT_JUMP:
+                case ILT_JUMP_IF_TRUE:
+                case ILT_JUMP_IF_FALSE:
+                    g_arch_instruction_get_location(dests[i], NULL, NULL, &addr);
+                    find_next_jumps(instrs, addr, end, info);
+                    break;
+
+                default:
+                    break;
+
+            }
+
+        break;
+
+    }
+
+    /* Si on termine... */
+    if (iter != NULL && !is_addr_in_branch(info, &end, false))
+    {
+        info->jumps = (vmpa_t *)realloc(info->jumps, ++(info->count) * sizeof(vmpa_t));
+        info->jumps[info->count - 1] = end;
+    }
+
+}
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : a = premier ensemble de jalons à parcourir.                  *
+*                b = second ensemble de jalons à parcourir.                   *
+*                                                                             *
+*  Description : Retrouve le point de ralliement entre deux branches.         *
+*                                                                             *
+*  Retour      : Adresse commune à deux branches.                             *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+static vmpa_t compute_first_common_addr(branch_info *a, branch_info *b)
+{
+    vmpa_t result;                          /* Adresse trouvée à retourner */
+    size_t i;                               /* Boucle de parcours          */
+
+    /* Valeur conceptuellement impossible à renvoyer */
+    result = VMPA_MAX;
+
+    //qsort(a->jumps, a->count, sizeof(vmpa_t), (__compar_fn_t)compare_vmpa);
+    //qsort(b->jumps, b->count, sizeof(vmpa_t), (__compar_fn_t)compare_vmpa);
+
+    for (i = 0; i < a->count && result == VMPA_MAX; i++)
+        if (is_addr_in_branch(b, &a->jumps[i], false))
+            result = a->jumps[i];
+
+    return result;
+
+}
+
+
+
+#include "../../arch/processor.h"
+
+
+/******************************************************************************
+*                                                                             *
+*  Paramètres  : instrs = ensemble des instructions d'assemblage.             *
+*                start  = adresse de début du bloc.                           *
+*                end    = adresse de fin du bloc (exclusive).                 *
+*                stop   = adresse d'arrêt en cas de saut ou VMPA_MAX.         *
+*                ctx    = contexte de soutien à associer à l'opération.       *
+*                                                                             *
+*  Description : Procède à la décompilation basique d'un bloc déterminé.      *
+*                                                                             *
+*  Retour      : Instructions créées et enregistrées, ou NULL si erreur.      *
+*                                                                             *
+*  Remarques   : -                                                            *
+*                                                                             *
+******************************************************************************/
+
+GDecInstruction *build_decompiled_block(GArchInstruction *instrs, vmpa_t start, vmpa_t end, vmpa_t stop, GDecContext *ctx)
+{
+    GDecInstruction *result;                /* Instructions décompilées    */
+    GArchInstruction *iter;                 /* Boucle de parcours          */
+
+    GDecInstruction *pite;                  /* IfThenElse potientiel...    */
+
+    GArchInstruction **dests;               /* Instr. visée par une autre  */
+    InstructionLinkType *types;             /* Type de lien entre lignes   */
+    size_t dcount;                          /* Nombre de liens de dest.    */
+
+    size_t i;                               /* Boucle de parcours          */
+    vmpa_t addr;                            /* Adresse de la destination   */
+
+    branch_info true_branch;                /* Branche 'condition vraie'   */
+    branch_info false_branch;               /* Branche 'condition fausse'  */
+    GDecInstruction *true_dinstr;           /* Décompilation 'cond vraie'  */
+    GDecInstruction *false_dinstr;          /* Décompilation 'cond fausse' */
+
+    vmpa_t next_addr;                       /* Prochaine instruction visée */
+
+    GDecInstruction *first;                 /* Première décompilation      */
+    GDecInstruction *dinstr;                /* Nouvelle décompilation      */
+
+
+    GExeFormat *format;                     /* Format du binaire fourni    */
+    GArchProcessor *proc;                   /* Architecture du binaire     */
+    GDecContext *context;                   /* Contexte pour la décompil.  */
+
+
+    result = NULL;
+
+    printf("[+] processing 0x%08llx -> 0x%08llx... stop @ 0x%08llx\n", start, end, stop);
+
+    for (iter = g_arch_instruction_find_by_address(instrs, start, true);
+         iter != NULL;
+         )
+    {
+        /* On s'arrêter si l'instruction est déjà décompilée */
+        if (g_object_get_data(G_OBJECT(iter), "decomp_done") != NULL) break;
+        g_object_set_data(G_OBJECT(iter), "decomp_done", iter);
+
+        pite = g_arch_instruction_decompile(iter, ctx);
+
+        g_arch_instruction_get_location(iter, NULL, NULL, &addr);
+        printf(" --- decomp %p @ 0x%08llx\n", pite, addr);
+
+        /* On n'approfondit que les chemins qui se séparent */
+        if (!g_arch_instruction_has_destinations(iter))
+        {
+            iter = g_arch_instruction_get_next_iter(instrs, iter, end);
+            continue;
+        }
+
+        /* Adaptations en fonction du type de bifurcation */
+
+        dcount = g_arch_instruction_get_destinations(iter, &dests, &types);
+
+        next_addr = 0;
+        memset(&true_branch, 0, sizeof(branch_info));
+        memset(&false_branch, 0, sizeof(branch_info));
+
+        for (i = 0; i < dcount; i++)
+            switch (types[i])
+            {
+                case ILT_EXEC_FLOW:
+                case ILT_JUMP:
+                    g_arch_instruction_get_location(dests[i], NULL, NULL, &next_addr);
+                    break;
+
+                case ILT_JUMP_IF_TRUE:
+                    g_arch_instruction_get_location(dests[i], NULL, NULL, &addr);
+                    find_next_jumps(instrs, addr, end, &true_branch);
+                    break;
+
+                case ILT_JUMP_IF_FALSE:
+                    g_arch_instruction_get_location(dests[i], NULL, NULL, &addr);
+                    find_next_jumps(instrs, addr, end, &false_branch);
+                    break;
+
+                default:
+                    next_addr = VMPA_MAX;
+                    break;
+
+            }
+
+        if (next_addr == VMPA_MAX)
+        {
+            iter = g_arch_instruction_get_next_iter(instrs, iter, end);
+            continue;
+        }
+
+        else if (true_branch.count > 0 || false_branch.count > 0)
+        {
+            next_addr = compute_first_common_addr(&true_branch, &false_branch);
+            next_addr = MIN(next_addr, end);
+
+            format = g_object_get_data(G_OBJECT(ctx), "format");
+            proc = get_arch_processor_from_format(G_EXE_FORMAT(format));
+
+            context = g_arch_processor_get_decomp_context(proc);
+
+            g_object_set_data(G_OBJECT(context), "format", g_object_get_data(G_OBJECT(ctx), "format"));
+            g_object_set_data(G_OBJECT(context), "routine", g_object_get_data(G_OBJECT(ctx), "routine"));
+            g_dec_context_set_max_address(context, next_addr);
+
+            true_dinstr = build_decompiled_block(instrs, true_branch.jumps[0],
+                                                 end, next_addr, context);
+
+
+            context = g_arch_processor_get_decomp_context(proc);
+
+            g_object_set_data(G_OBJECT(context), "format", g_object_get_data(G_OBJECT(ctx), "format"));
+            g_object_set_data(G_OBJECT(context), "routine", g_object_get_data(G_OBJECT(ctx), "routine"));
+            g_dec_context_set_max_address(context, next_addr);
+
+            false_dinstr = build_decompiled_block(instrs, false_branch.jumps[0],
+                                                  end, next_addr, context);
+
+
+            printf("{branch : %p (0x%08llx) | %p (0x%08llx)\n",
+                   true_dinstr, true_branch.jumps[0],
+                   false_dinstr, false_branch.jumps[0]);
+
+            g_ite_instruction_set_branches(G_ITE_INSTRUCTION(pite), true_dinstr, false_dinstr);
+
+            if (next_addr == end) break;
+
+        }
+
+        /* Détermination du prochain point de chute */
+
+        if (next_addr == stop) break;
+
+        iter = g_arch_instruction_find_by_address(instrs, next_addr, true);
+
+    }
+
+
+
+
+    first = g_dec_context_get_decomp_instrs(ctx);
+
+    printf(" ... context instr : %p\n", first);
+
+    for (dinstr = first;
+         dinstr != NULL;
+         dinstr = g_dec_instruction_get_next_iter(first, dinstr))
+    {
+        if (result == NULL) result = g_expr_block_new(dinstr);
+        else g_expr_block_add_item(G_EXPR_BLOCK(result), dinstr);
+
+    }
+
+    printf(" ... return %p\n", result);
+
+    return result;
+
+}
diff --git a/src/analysis/decomp/il.h b/src/analysis/decomp/il.h
new file mode 100644
index 0000000..4f38b4f
--- /dev/null
+++ b/src/analysis/decomp/il.h
@@ -0,0 +1,38 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * il.h - prototypes pour la mise en place d'un langage intermédiaire
+ *
+ * Copyright (C) 2012 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 _ANALYSIS_DECOMP_IL_H
+#define _ANALYSIS_DECOMP_IL_H
+
+
+#include "../../arch/instruction.h"
+#include "../../decomp/instruction.h"
+
+
+
+/* Procède à la décompilation basique d'un bloc déterminé. */
+GDecInstruction *build_decompiled_block(GArchInstruction *, vmpa_t, vmpa_t, vmpa_t, GDecContext *);
+
+
+
+#endif  /* _ANALYSIS_DECOMP_IL_H */
diff --git a/src/analysis/decomp/reduce.c b/src/analysis/decomp/reduce.c
new file mode 100644
index 0000000..926650d
--- /dev/null
+++ b/src/analysis/decomp/reduce.c
@@ -0,0 +1,53 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * reduce.c - réduction de l'usage des [pseudo]-registres
+ *
+ * Copyright (C) 2012 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 "reduce.h"
+
+
+#if 0
+
+
+
+/* Mémorisation de l'usage de chaque variable */
+typedef struct _var_usage
+{
+    GDecInstruction *var;                   /* Variable manipulée          */
+
+    size_t ref_counter;                     /* Décompte des utilisations   */
+    GDecInstruction *creation;              /* Emplacement de la création  */
+    GDecInstruction *usage;                 /* Lieu de dernère utilisation */
+
+} var_usage;
+
+/* Réduction des usages */
+typedef struct _usage_reduc
+{
+
+
+
+} usage_reduc;
+
+
+#endif
+
+
diff --git a/src/analysis/decomp/reduce.h b/src/analysis/decomp/reduce.h
new file mode 100644
index 0000000..aa13d7f
--- /dev/null
+++ b/src/analysis/decomp/reduce.h
@@ -0,0 +1,30 @@
+
+/* OpenIDA - Outil d'analyse de fichiers binaires
+ * reduce.h - prototypes pour la réduction de l'usage des [pseudo]-registres
+ *
+ * Copyright (C) 2012 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 _ANALYSIS_DECOMP_REDUCE_H
+#define _ANALYSIS_DECOMP_REDUCE_H
+
+
+
+
+#endif  /* _ANALYSIS_DECOMP_REDUCE_H */