path: root/src/analysis/disass/disassembler.c

/* Chrysalide - Outil d'analyse de fichiers binaires
 * disassembler.c - encadrement des phases de désassemblage
 *
 * Copyright (C) 2010-2014 Cyrille Bagard
 *
 *  This file is part of Chrysalide.
 *
 *  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 "disassembler.h"


#include <malloc.h>
#include <stdio.h>
#include <string.h>


#include <i18n.h>


#include "fetch.h"
#include "limit.h"
#include "links.h"
#include "loop.h"
#include "macro.h"
#include "output.h"
#include "rank.h"
#include "../../decomp/lang/asm.h"
#include "../../format/format.h"
#include "../../glibext/delayed-int.h"
#include "../../gui/panels/log.h"
#include "../../plugins/pglist.h"



/* ------------------------ DESASSEMBLAGE DE BINAIRE DIFFERE ------------------------ */


/* Ensembles binaires à désassembler (instance) */
struct _GDelayedDisassembly
{
    GDelayedWork parent;                    /* A laisser en premier        */

    GLoadedBinary *binary;                  /* Destinataire final          */
    GExeFormat *format;                     /* Format du binaire représenté*/

    GArchInstruction **instrs;              /* Instructions résultantes    */
    GCodeBuffer *buffer;                    /* Tampon pour le rendu        */

};

/* Ensembles binaires à désassembler (classe) */
struct _GDelayedDisassemblyClass
{
    GDelayedWorkClass parent;               /* A laisser en premier        */

};


/* Initialise la classe des tâches de désassemblage différé. */
static void g_delayed_disassembly_class_init(GDelayedDisassemblyClass *);

/* Initialise une tâche de désassemblage différé. */
static void g_delayed_disassembly_init(GDelayedDisassembly *);

/* Crée une tâche de désassemblage différé. */
static GDelayedDisassembly *g_delayed_disassembly_new(GLoadedBinary *, GArchInstruction **, GCodeBuffer *);

/* Assure le désassemblage en différé. */
static void g_delayed_disassembly_process(GDelayedDisassembly *, GtkExtStatusBar *);



/* -------------------------- GESTION GLOBALE DE PROCEDURE -------------------------- */


/* Construit la description d'introduction du désassemblage. */
static void build_disass_prologue(GCodeBuffer *, const char *, const char *);



/* ---------------------------------------------------------------------------------- */
/*                          DESASSEMBLAGE DE BINAIRE DIFFERE                          */
/* ---------------------------------------------------------------------------------- */


/* Indique le type défini pour les tâches de désassemblage différé. */
G_DEFINE_TYPE(GDelayedDisassembly, g_delayed_disassembly, G_TYPE_DELAYED_WORK);


/******************************************************************************
*                                                                             *
*  Paramètres  : klass = classe à initialiser.                                *
*                                                                             *
*  Description : Initialise la classe des tâches de désassemblage différé.    *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static void g_delayed_disassembly_class_init(GDelayedDisassemblyClass *klass)
{
    GDelayedWorkClass *work;                /* Version en classe parente   */

    work = G_DELAYED_WORK_CLASS(klass);

    work->run = (run_task_fc)g_delayed_disassembly_process;

}


/******************************************************************************
*                                                                             *
*  Paramètres  : disass = instance à initialiser.                             *
*                                                                             *
*  Description : Initialise une tâche de désassemblage différé.               *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static void g_delayed_disassembly_init(GDelayedDisassembly *disass)
{

}


/******************************************************************************
*                                                                             *
*  Paramètres  : binary = binaire chargé en attente des résultats.            *
*                format = format du binaire représenté.                       *
*                instrs = emplacement pour la liste d'instructions.           *
*                buffer = tampon de sortie pour les instructions.             *
*                                                                             *
*  Description : Crée une tâche de désassemblage différé.                     *
*                                                                             *
*  Retour      : Tâche créée.                                                 *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static GDelayedDisassembly *g_delayed_disassembly_new(GLoadedBinary *binary, GArchInstruction **instrs, GCodeBuffer *buffer)
{
    GDelayedDisassembly *result;            /* Tâche à retourner           */

    result = g_object_new(G_TYPE_DELAYED_DISASSEMBLY, NULL);

    result->binary = binary;
    result->format = g_loaded_binary_get_format(binary);

    result->instrs = instrs;
    result->buffer = buffer;

    return result;

}


/******************************************************************************
*                                                                             *
*  Paramètres  : disass    = analyse à mener.                                 *
*                statusbar = barre de statut à tenir informée.                *
*                                                                             *
*  Description : Assure le désassemblage en différé.                          *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static void g_delayed_disassembly_process(GDelayedDisassembly *disass, GtkExtStatusBar *statusbar)
{
    wgroup_id_t gid;                        /* Identifiant pour les tâches */

    //GBinFormat *format;                     /* Format du fichier binaire   */
    GArchProcessor *proc;                   /* Architecture du binaire     */


    unsigned int valid;                     /* Instructions traduites      */
    unsigned int db;                        /* Instructions non décodées   */
    unsigned int valid_sum;                 /* Instructions traduites      */
    unsigned int instr_sum;                 /* Instructions totales        */
    size_t i;                               /* Boucle de parcours          */

    GBinRoutine **routines;                 /* Liste des routines trouvées */
    size_t routines_count;                  /* Nombre de ces routines      */
    bstatus_id_t id;                        /* Identifiant de statut       */


    //GArchProcessor *proc;                   /* Architecture du binaire     */



    gid = g_work_queue_define_work_group(get_work_queue());





    //format = G_BIN_FORMAT(g_loaded_binary_get_format(binary));
    proc = g_loaded_binary_get_processor(disass->binary);


    routines = g_binary_format_get_routines(G_BIN_FORMAT(disass->format), &routines_count);


    /* Première étape */

    //id = gtk_extended_status_bar_push(statusbar, _("Disassembling..."), true);



    process_disassembly_event(PGA_DISASSEMBLY_STARTED, disass->binary);



    *disass->instrs = disassemble_binary_content(disass->binary, gid, statusbar);


    g_arch_processor_set_disassembled_instructions(proc, *disass->instrs);


    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_RAW, disass->binary);



    /*
    *disass->instrs = disassemble_binary_parts(disass->binary, disass->parts, disass->count,
                                               statusbar, id);
    */





    do
    {
        GBinFormat *format;                     /* Format du fichier binaire   */
        GArchInstruction *iter;                 /* Boucle de parcours          */



        format = G_BIN_FORMAT(g_loaded_binary_get_format(disass->binary));

        for (iter = *disass->instrs;
             iter != NULL;
             iter = g_arch_instruction_get_next_iter(*disass->instrs, iter, 0))
        {

            g_arch_instruction_call_hook(iter, IPH_LINK, proc, /*ctx*/NULL, format);



        }



    } while (0);



    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_HOOKED_LINK, disass->binary);



    //gtk_extended_status_bar_remove(statusbar, id);


    //run_plugins_on_binary(disass->binary, PGA_BINARY_DISASSEMBLED, true);


    do
    {
        GBinFormat *format;                     /* Format du fichier binaire   */
        GArchInstruction *iter;                 /* Boucle de parcours          */



        format = G_BIN_FORMAT(g_loaded_binary_get_format(disass->binary));

        for (iter = *disass->instrs;
             iter != NULL;
             iter = g_arch_instruction_get_next_iter(*disass->instrs, iter, 0))
        {

            g_arch_instruction_call_hook(iter, IPH_POST, proc, /*ctx*/NULL, format);



        }



    } while (0);




    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_HOOKED_POST, disass->binary);



    /**
     * TODO : établir les couvertures de fonctions,
     * pour être en mesure de disposer de résolution de type XXX+yyy lors
     * de l'établissement des liens.
     */


    /* Seconde étape */

    routines = g_binary_format_get_routines(G_BIN_FORMAT(disass->format), &routines_count);




    id = gtk_extended_status_bar_push(statusbar, _("Finding remaining limits..."), true);

    //qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    limit_all_routines(disass->format, proc, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    //run_plugins_on_binary(disass->binary, PGA_BINARY_BOUNDED, true);



    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_LIMITED, disass->binary);






    /* Troisième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Establishing links..."), true);


    /**
     *
     * Lequel choisir ???

G_BIN_FORMAT(disass->format)
G_BIN_FORMAT(g_loaded_binary_get_format(disass->binary)

     */


    establish_links_between_instructions(*disass->instrs, G_BIN_FORMAT(disass->format), statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    //run_plugins_on_binary(disass->binary, PGA_BINARY_LINKED, true);



    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_LINKED, disass->binary);





    /* Quatrième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Detecting loops..."), true);

    detect_loops_in_code(proc, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    ///

    // plugins //////////////////////////
    process_disassembly_event(PGA_DISASSEMBLY_LOOPS, disass->binary);





    /* Cinquième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Grouping routines instructions..."), true);

    //qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    group_routines_instructions(proc, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    //run_plugins_on_binary(disass->binary, PGA_BINARY_GROUPED, true);


    process_disassembly_event(PGA_DISASSEMBLY_GROUPED, disass->binary);





    /* Sixième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Ranking each instructions block..."), true);

    //qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    rank_routines_blocks(routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    //run_plugins_on_binary(disass->binary, PGA_BINARY_GROUPED, true);


    process_disassembly_event(PGA_DISASSEMBLY_RANKED, disass->binary);



    /* Septième étape */

    //id = gtk_extended_status_bar_push(statusbar, _("Printing disassembled code..."), true);

    qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_compare);


    proc = g_loaded_binary_get_processor(disass->binary);

    print_disassembled_instructions(disass->buffer, disass->format, proc, *disass->instrs,
                                    routines, routines_count, statusbar, id);

    g_object_unref(G_OBJECT(proc));






    process_disassembly_event(PGA_DISASSEMBLY_ENDED, disass->binary);


    printf("---fin\n");

    //gtk_extended_status_bar_remove(statusbar, id);

    //run_plugins_on_binary(disass->binary, PGA_BINARY_PRINTED, true);



}

#if 0
/******************************************************************************
*                                                                             *
*  Paramètres  : disass    = analyse à mener.                                 *
*                statusbar = barre de statut à tenir informée.                *
*                                                                             *
*  Description : Assure le désassemblage en différé.                          *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static void g_delayed_disassembly_process_old(GDelayedDisassembly *disass, GtkExtStatusBar *statusbar)
{
#ifdef DEBUG
    unsigned int valid;                     /* Instructions traduites      */
    unsigned int db;                        /* Instructions non décodées   */
    unsigned int valid_sum;                 /* Instructions traduites      */
    unsigned int instr_sum;                 /* Instructions totales        */
    size_t i;                               /* Boucle de parcours          */
#endif
    GBinRoutine **routines;                 /* Liste des routines trouvées */
    size_t routines_count;                  /* Nombre de ces routines      */
    bstatus_id_t id;                        /* Identifiant de statut       */

    routines = g_binary_format_get_routines(G_BIN_FORMAT(disass->format), &routines_count);

    /* Première étape */

    id = gtk_extended_status_bar_push(statusbar, _("Disassembling..."), true);

    *disass->instrs = disassemble_binary_parts(disass->binary, disass->parts, disass->count,
                                               statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

#ifdef DEBUG

    valid_sum = 0;
    instr_sum = 0;

    for (i = 0; i < disass->count; i++)
    {
        g_binary_part_get_checkup(disass->parts[i], &valid, &db);
        valid_sum += valid;
        instr_sum += (valid + db);
    }

    log_variadic_message(LMT_WARNING, _("Disassembled instructions : %u %% (%u / %d)"),
                         (valid_sum * 100) / instr_sum,
                         valid_sum, instr_sum);

#endif

    run_plugins_on_binary(disass->binary, PGA_BINARY_DISASSEMBLED, true);

    /* Seconde étape */

    id = gtk_extended_status_bar_push(statusbar, _("Establishing links..."), true);

    establish_links_between_lines(*disass->instrs, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    run_plugins_on_binary(disass->binary, PGA_BINARY_LINKED, true);

    /* Troisième  étape */

    id = gtk_extended_status_bar_push(statusbar, _("Finding remaining limits..."), true);

    qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    limit_all_routines(*disass->instrs, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    run_plugins_on_binary(disass->binary, PGA_BINARY_BOUNDED, true);

    /* Quatrième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Detecting loops..."), true);

    detect_loops_in_code(*disass->instrs, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    /* Cinquième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Grouping routines instructions..."), true);

    qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    group_routines_instructions(*disass->instrs, routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    run_plugins_on_binary(disass->binary, PGA_BINARY_GROUPED, true);

    /* Sixième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Ranking each instructions block..."), true);

    qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_rcompare);

    rank_routines_blocks(routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    run_plugins_on_binary(disass->binary, PGA_BINARY_GROUPED, true);

    /* Septième étape */

    id = gtk_extended_status_bar_push(statusbar, _("Printing disassembled code..."), true);

    qsort(routines, routines_count, sizeof(GBinRoutine *), (__compar_fn_t)g_binary_routine_compare);

    print_disassembled_instructions(disass->buffer, disass->format, *disass->instrs,
                                    routines, routines_count, statusbar, id);

    gtk_extended_status_bar_remove(statusbar, id);

    run_plugins_on_binary(disass->binary, PGA_BINARY_PRINTED, true);

}
#endif


/* ---------------------------------------------------------------------------------- */
/*                            GESTION GLOBALE DE PROCEDURE                            */
/* ---------------------------------------------------------------------------------- */


/******************************************************************************
*                                                                             *
*  Paramètres  : buffer   = tampon de destination pour le texte.              *
*                filename = nom du fichier ciblé à décompiler.                *
*                data     = données en mémoire pour l'empreinte.              *
*                length   = quantité de données à prendre en compte.          *
*                                                                             *
*  Description : Construit la description d'introduction du désassemblage.    *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

static void build_disass_prologue(GCodeBuffer *buffer, const char *filename, const char *checksum)
{
    GLangOutput *output;                    /* Modèle de sortie adéquat    */
    GBufferLine *line;                      /* Ligne de destination        */
    bool managed;                           /* Groupe déjà défini ?        */
    size_t len;                             /* Taille du texte             */
    char *content;                          /* Contenu textuel d'une ligne */

    output = g_asm_output_new();

    line = g_lang_output_start_comments(output, buffer);
    if (line != NULL)
    {
        g_buffer_line_start_merge_at(line, BLC_PHYSICAL);
        g_buffer_line_add_flag(line, BLF_WIDTH_MANAGER);
    }

    managed = (line != NULL);

    /* Introduction */

    line = g_lang_output_continue_comments(output, buffer,
                                           SL(_("Disassembly generated by Chrysalide")));
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    if (!managed)
        g_buffer_line_add_flag(line, BLF_WIDTH_MANAGER);

    line = g_lang_output_continue_comments(output, buffer,
                                           SL(_("Chrysalide is free software - © 2008-2015 Cyrille Bagard")));
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    line = g_lang_output_continue_comments(output, buffer, NULL, 0);
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    /* Fichier */

    len = strlen(_("File: ")) + strlen(filename) + 1;
    content = (char *)calloc(len, sizeof(char));

    snprintf(content, len, "%s%s", _("File: "), filename);

    line = g_lang_output_continue_comments(output, buffer, content, len - 1);
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    free(content);

    /* Checksum SHA256 */

    len = strlen(_("Sha256: ")) + strlen(checksum);
    content = (char *)calloc(len + 1, sizeof(char));

    snprintf(content, len + 1, "%s%s", _("Sha256: "), checksum);

    line = g_lang_output_continue_comments(output, buffer, content, len - 1);
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    free(content);

    /* Ligne de séparation */

    line = g_lang_output_continue_comments(output, buffer, NULL, 0);
    g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    /* Conclusion */

    line = g_lang_output_end_comments(output, buffer);
    if (line != NULL) g_buffer_line_start_merge_at(line, BLC_PHYSICAL);

    g_object_unref(G_OBJECT(output));

}


/******************************************************************************
*                                                                             *
*  Paramètres  : binary = représentation de binaire chargé.                   *
*                parts  = parties binaires à désassembler.                    *
*                count  = nombre de parties à traiter.                        *
*                instrs = liste des instructions chargées. [OUT]              *
*                buffer = tampon de code mis en place. [OUT]                  *
*                ack    = fonction à appeler une fois l'opération terminée.   *
*                                                                             *
*  Description : Procède au désassemblage d'un contenu binaire donné.         *
*                                                                             *
*  Retour      : -                                                            *
*                                                                             *
*  Remarques   : -                                                            *
*                                                                             *
******************************************************************************/

void disassemble_binary(GLoadedBinary *binary, GArchInstruction **instrs, GCodeBuffer **buffer, disassembly_ack_fc ack)
{
    GBinFormat *format;                     /* Format associé au binaire   */
    GBinContent *content;                   /* Contenu bianire manipulé    */
    const gchar *checksum;                  /* Identifiant de binaire      */
    GDelayedDisassembly *disass;            /* Désassemblage à mener       */
    GWorkQueue *queue;                      /* Gestionnaire de différés    */

    *buffer = g_code_buffer_new(BLC_ASSEMBLY);

    format = G_BIN_FORMAT(g_loaded_binary_get_format(binary));

    content = g_binary_format_get_content(format);
    checksum = g_binary_content_get_cheksum(content);
    g_object_unref(G_OBJECT(content));

    build_disass_prologue(*buffer, g_binary_content_describe(content, true), checksum);

    disass = g_delayed_disassembly_new(binary, instrs, *buffer);
    g_signal_connect(disass, "work-completed", G_CALLBACK(ack), binary);

    queue = get_work_queue();
    g_work_queue_schedule_work(queue, G_DELAYED_WORK(disass), DEFAULT_WORK_GROUP);

}