dfhack/library/DFProcessEnumerator.cpp

/*
www.sourceforge.net/projects/dfhack
Copyright (c) 2009 Petr Mrázek (peterix), Kenneth Ferland (Impaler[WrG]), dorf

This software is provided 'as-is', without any express or implied
warranty. In no event will the authors be held liable for any
damages arising from the use of this software.

Permission is granted to anyone to use this software for any
purpose, including commercial applications, and to alter it and
redistribute it freely, subject to the following restrictions:

1. The origin of this software must not be misrepresented; you must
not claim that you wrote the original software. If you use this
software in a product, an acknowledgment in the product documentation
would be appreciated but is not required.

2. Altered source versions must be plainly marked as such, and
must not be misrepresented as being the original software.

3. This notice may not be removed or altered from any source
distribution.
*/

#include "DFCommonInternal.h"
using namespace DFHack;

/// HACK: global variables (only one process can be attached at the same time.)
Process * DFHack::g_pProcess; ///< current process. non-NULL when picked
ProcessHandle DFHack::g_ProcessHandle; ///< cache of handle to current process. used for speed reasons
int DFHack::g_ProcessMemFile; ///< opened /proc/PID/mem, valid when attached

class DFHack::ProcessEnumerator::Private
{
    public:
        Private(){};
        // memory info entries loaded from a file
        std::vector<memory_info> meminfo;
        Process * currentProcess;
        ProcessHandle currentProcessHandle;
        std::vector<Process *> processes;
        bool loadDescriptors( string path_to_xml);
        void ParseVTable(TiXmlElement* vtable, memory_info& mem);
        void ParseEntry (TiXmlElement* entry, memory_info& mem, map <string ,TiXmlElement *>& knownEntries);
        #ifdef LINUX_BUILD
        Process* addProcess(const string & exe,ProcessHandle PH,const string & memFile);
        #endif
};

#ifdef LINUX_BUILD
/*
 *  LINUX version of the process finder.
 */

bool ProcessEnumerator::findProcessess()
{
    DIR *dir_p;
    struct dirent *dir_entry_p;
    string dir_name;
    string exe_link;
    string cwd_link;
    string cmdline_path;
    string cmdline;

    // ALERT: buffer overrun potential

    int errorcount;
    int result;

    errorcount=0;
    result=0;
    // Open /proc/ directory
    dir_p = opendir("/proc/");
    // Reading /proc/ entries
    while(NULL != (dir_entry_p = readdir(dir_p)))
    {
        // Only PID folders (numbers)
        if (strspn(dir_entry_p->d_name, "0123456789") != strlen(dir_entry_p->d_name))
        {
            continue;
        }
        Process *p = new Process(atoi(dir_entry_p->d_name),d->meminfo);
        if(p->isIdentified())
        {
            d->processes.push_back(p);
        }
        else
        {
            delete p;
        }
    }
    closedir(dir_p);
    // return value depends on if we found some DF processes
    if(d->processes.size())
    {
        return true;
    }
    return false;
}

#else

// some magic - will come in handy when we start doing debugger stuff on Windows
bool EnableDebugPriv()
{
    bool               bRET = FALSE;
    TOKEN_PRIVILEGES   tp;
    HANDLE             hToken;

    if (LookupPrivilegeValue(NULL, SE_DEBUG_NAME, &tp.Privileges[0].Luid))
    {
        if (OpenProcessToken(GetCurrentProcess(), TOKEN_ADJUST_PRIVILEGES, &hToken))
        {
            if (hToken != INVALID_HANDLE_VALUE)
            {
                tp.Privileges[0].Attributes = SE_PRIVILEGE_ENABLED;
                tp.PrivilegeCount = 1;
                if (AdjustTokenPrivileges(hToken, FALSE, &tp, 0, 0, 0))
                {
                    bRET = TRUE;
                }
                CloseHandle(hToken);
            }
        }
    }
    return bRET;
}

// WINDOWS version of the process finder
bool ProcessEnumerator::findProcessess()
{
    // Get the list of process identifiers.
    DWORD ProcArray[2048], memoryNeeded, numProccesses;

    EnableDebugPriv();
    if ( !EnumProcesses( ProcArray, sizeof(ProcArray), &memoryNeeded ) )
    {
        return false;
    }

    // Calculate how many process identifiers were returned.
    numProccesses = memoryNeeded / sizeof(DWORD);

    // iterate through processes
    for ( int i = 0; i < (int)numProccesses; i++ )
    {
        Process *p = new Process(ProcArray[i],d->meminfo);
        if(p->isIdentified())
        {
            d->processes.push_back(p);
        }
        else
        {
            delete p;
        }
    }
    if(d->processes.size())
        return true;
    return false;
}
#endif


void ProcessEnumerator::Private::ParseVTable(TiXmlElement* vtable, memory_info& mem)
{
    TiXmlElement* pClassEntry;
    TiXmlElement* pClassSubEntry;
    // check for rebase, do rebase if check positive
    const char * rebase = vtable->Attribute("rebase");
    if(rebase)
    {
        int32_t rebase_offset = strtol(rebase, NULL, 16);
        mem.RebaseVTable(rebase_offset);
    }
    // parse vtable entries
    pClassEntry = vtable->FirstChildElement();
    for(;pClassEntry;pClassEntry=pClassEntry->NextSiblingElement())
    {
        string type = pClassEntry->Value();
        const char *cstr_name = pClassEntry->Attribute("name");
        const char *cstr_vtable = pClassEntry->Attribute("vtable");
        // it's a simple class
        if(type== "class")
        {
            mem.setClass(cstr_name, cstr_vtable);
        }
        // it's a multi-type class
        else if (type == "multiclass")
        {
            // get offset of the type variable
            const char *cstr_typeoffset = pClassEntry->Attribute("typeoffset");
            int mclass = mem.setMultiClass(cstr_name, cstr_vtable, cstr_typeoffset);
            // parse class sub-entries
            pClassSubEntry = pClassEntry->FirstChildElement();
            for(;pClassSubEntry;pClassSubEntry=pClassSubEntry->NextSiblingElement())
            {
                type = pClassSubEntry->Value();
                if(type== "class")
                {
                    // type is a value loaded from type offset
                    cstr_name = pClassSubEntry->Attribute("name");
                    const char *cstr_value = pClassSubEntry->Attribute("type");
                    mem.setMultiClassChild(mclass,cstr_name,cstr_value);
                }
            }
        }
    }
}



void ProcessEnumerator::Private::ParseEntry (TiXmlElement* entry, memory_info& mem, map <string ,TiXmlElement *>& knownEntries)
{
    TiXmlElement* pMemEntry;
    const char *cstr_version = entry->Attribute("version");
    const char *cstr_os = entry->Attribute("os");
    const char *cstr_base = entry->Attribute("base");
    const char *cstr_rebase = entry->Attribute("rebase");
    if(cstr_base)
    {
        string base = cstr_base;
        ParseEntry(knownEntries[base], mem, knownEntries);
    }
    
    // mandatory attributes missing?
    if(!(cstr_version && cstr_os))
    {
        cerr << "Bad entry in memory.xml detected, version or os attribute is missing.";
        // skip if we don't have valid attributes
        return;
    }
    string os = cstr_os;
    mem.setVersion(cstr_version);
    mem.setOS(cstr_os);

    // offset inherited addresses by 'rebase'.
    int32_t rebase = 0;
    if(cstr_rebase)
    {
        rebase = mem.getBase() + strtol(cstr_rebase, NULL, 16);
        mem.RebaseAddresses(rebase);
    }

    //set base to default, we're overwriting this because the previous rebase could cause havoc on Vista/7
    if(os == "windows")
    {
        // set default image base. this is fixed for base relocation later
        mem.setBase(0x400000);
    }
    else if(os == "linux")
    {
        // this is wrong... I'm not going to do base image relocation on linux though.
        // users are free to use a sane kernel that doesn't do this kind of **** by default
        mem.setBase(0x0);
    }
    else if ( os == "all")
    {
        // yay
    }
    else
    {
        cerr << "unknown operating system " << os << endl;
        return;
    }
    
    // process additional entries
    //cout << "Entry " << cstr_version << " " <<  cstr_os << endl;
    pMemEntry = entry->FirstChildElement()->ToElement();
    for(;pMemEntry;pMemEntry=pMemEntry->NextSiblingElement())
    {
        // only elements get processed
        const char *cstr_type = pMemEntry->Value();
        const char *cstr_name = pMemEntry->Attribute("name");
        const char *cstr_value = pMemEntry->GetText();
        // check for missing parts
        string type, name, value;
        type = cstr_type;
        if(type == "VTable")
        {
            ParseVTable(pMemEntry, mem);
            continue;
        }
        if( !(cstr_name && cstr_value))
        {
            cerr << "underspecified MemInfo entry" << endl;
            continue;
        }
        name = cstr_name;
        value = cstr_value;
        if (type == "HexValue")
        {
            mem.setHexValue(name, value);
        }
        else if (type == "Address")
        {
            mem.setAddress(name, value);
        }
        else if (type == "Offset")
        {
            mem.setOffset(name, value);
        }
        else if (type == "String")
        {
            mem.setString(name, value);
        }
		else if (type == "Profession")
		{
			mem.setProfession(value,name);
		}
		else if (type == "Job")
		{
			mem.setJob(value,name);
		}
        else if (type == "Skill")
		{
			mem.setSkill(value,name);
		}
		else if (type == "Trait")
		{
			mem.setTrait(value, name,pMemEntry->Attribute("level_0"),pMemEntry->Attribute("level_1"),pMemEntry->Attribute("level_2"),pMemEntry->Attribute("level_3"),pMemEntry->Attribute("level_4"),pMemEntry->Attribute("level_5"));
		}
        else if (type == "Labor")
        {
            mem.setLabor(value,name);
        }
        else
        {
            cerr << "Unknown MemInfo type: " << type << endl;
        }
    } // for
} // method


// load the XML file with offsets
bool ProcessEnumerator::Private::loadDescriptors(string path_to_xml)
{
    TiXmlDocument doc( path_to_xml.c_str() );
    bool loadOkay = doc.LoadFile();
	TiXmlHandle hDoc(&doc);
    TiXmlElement* pElem;
    TiXmlHandle hRoot(0);
    memory_info mem;

    if ( loadOkay )
    {
        // block: name
        {
            pElem=hDoc.FirstChildElement().Element();
            // should always have a valid root but handle gracefully if it does
            if (!pElem)
            {
                cerr << "no pElem found" << endl;
                return false;
            }
            string m_name=pElem->Value();
            if(m_name != "DFExtractor")
            {
                cerr << "DFExtractor != " << m_name << endl;
                return false;
            }
            //cout << "got DFExtractor XML!" << endl;
            // save this for later
            hRoot=TiXmlHandle(pElem);
        }
        // transform elements
        {
            // trash existing list
            meminfo.clear();
            TiXmlElement* pMemInfo=hRoot.FirstChild( "MemoryDescriptors" ).FirstChild( "Entry" ).Element();
            map <string ,TiXmlElement *> map_pNamedEntries;
            vector <TiXmlElement *> v_pEntries;
            for( ; pMemInfo; pMemInfo=pMemInfo->NextSiblingElement("Entry"))
            {
                v_pEntries.push_back(pMemInfo);
                const char *id;
                if(id= pMemInfo->Attribute("id"))
                {
                    string str_id = id;
                    map_pNamedEntries[str_id] = pMemInfo;
                }
            }
            for(uint32_t i = 0; i< v_pEntries.size();i++)
            {
                memory_info mem;
                //FIXME: add a set of entries processed in a step of this cycle, use it to check for infinite loops
                /* recursive */ParseEntry( v_pEntries[i] , mem , map_pNamedEntries);
                meminfo.push_back(mem);
            }
            // process found things here
        }
        return true;
    }
    else
    {
        // load failed
        cerr << "Can't load memory offsets from memory.xml" << endl;
        return false;
    }
}


uint32_t ProcessEnumerator::size()
{
    return d->processes.size();
};


Process * ProcessEnumerator::operator[](uint32_t index)
{
    assert(index < d->processes.size());
    return d->processes[index];
};


ProcessEnumerator::ProcessEnumerator( string path_to_xml )
: d(new Private())
{
    d->currentProcess = NULL;
    d->currentProcessHandle = 0;
    d->loadDescriptors( path_to_xml );
}


ProcessEnumerator::~ProcessEnumerator()
{
    // delete all processes
    for(uint32_t i = 0;i < d->processes.size();i++)
    {
        delete d->processes[i];
    }
    delete d;
}