diff --git a/src/xmount.c b/src/xmount.c index f51742b..63004a0 100755 --- a/src/xmount.c +++ b/src/xmount.c @@ -1,4179 +1,4203 @@ /******************************************************************************* * xmount Copyright (c) 2008-2016 by Gillen Daniel * * * * xmount is a small tool to "fuse mount" various harddisk image formats as dd, * * vdi, vhd or vmdk files and enable virtual write access to them. * * * * This program 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. * * * * This program 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 * * this program. If not, see . * *******************************************************************************/ #include #include #include #include #include #include // For PRI* #include #include // For dlopen, dlclose, dlsym #include // For opendir, readdir, closedir #include #include #include // For fstat #include #ifdef HAVE_LINUX_FS_H #include // For SEEK_* ?? #endif #if !defined(__APPLE__) && defined(HAVE_GRP_H) && defined(HAVE_PWD_H) #include // For getgrnam, struct group #include // For getpwuid, struct passwd #endif #include #include // For time #define FUSE_USE_VERSION 26 #include #include #include "xmount.h" #include "md5.h" #include "macros.h" #include "../libxmount/libxmount.h" #define XMOUNT_COPYRIGHT_NOTICE \ "xmount v%s Copyright (c) 2008-2016 by Gillen Daniel " #define LOG_WARNING(...) { \ LIBXMOUNT_LOG_WARNING(__VA_ARGS__); \ } #define LOG_ERROR(...) { \ LIBXMOUNT_LOG_ERROR(__VA_ARGS__); \ } #define LOG_DEBUG(...) { \ LIBXMOUNT_LOG_DEBUG(glob_xmount.debug,__VA_ARGS__); \ } /******************************************************************************* * Global vars ******************************************************************************/ //! Struct that contains various runtime configuration options static ts_XmountData glob_xmount; /******************************************************************************* * Forward declarations ******************************************************************************/ // Helper functions static void PrintUsage(char*); static void CheckFuseSettings(); static int ParseCmdLine(const int, char**); static int ExtractVirtFileNames(char*); static int GetMorphedImageSize(uint64_t*); static int GetVirtImageSize(uint64_t*); static int GetInputImageData(pts_InputImage, char*, off_t, size_t, size_t*); static int GetMorphedImageData(char*, off_t, size_t, size_t*); static int GetVirtImageData(char*, off_t, size_t); static int SetVdiFileHeaderData(char*, off_t, size_t); static int SetVhdFileHeaderData(char*, off_t, size_t); static int SetVirtImageData(const char*, off_t, size_t); static int CalculateInputImageHash(uint64_t*, uint64_t*); static int InitVirtVdiHeader(); static int InitVirtVhdHeader(); static int InitVirtualVmdkFile(); static int InitVirtImageInfoFile(); static int InitCacheFile(); static int LoadLibs(); static int FindInputLib(pts_InputImage); static int FindMorphingLib(); static void InitResources(); static void FreeResources(); static int SplitLibraryParameters(char*, uint32_t*, pts_LibXmountOptions**); // Functions exported to LibXmount_Morphing static int LibXmount_Morphing_ImageCount(uint64_t*); static int LibXmount_Morphing_Size(uint64_t, uint64_t*); static int LibXmount_Morphing_Read(uint64_t, char*, off_t, size_t, size_t*); // Functions implementing FUSE functions static int FuseGetAttr(const char*, struct stat*); static int FuseMkDir(const char*, mode_t); static int FuseMkNod(const char*, mode_t, dev_t); static int FuseReadDir(const char*, void*, fuse_fill_dir_t, off_t, struct fuse_file_info*); static int FuseOpen(const char*, struct fuse_file_info*); static int FuseRead(const char*, char*, size_t, off_t, struct fuse_file_info*); static int FuseRename(const char*, const char*); static int FuseRmDir(const char*); static int FuseUnlink(const char*); //static int FuseStatFs(const char*, struct statvfs*); static int FuseWrite(const char *p_path, const char*, size_t, off_t, struct fuse_file_info*); /******************************************************************************* * Helper functions ******************************************************************************/ //! Print usage instructions (cmdline options etc..) /*! * \param p_prog_name Program name (argv[0]) */ static void PrintUsage(char *p_prog_name) { char *p_buf; int first; int ret; printf("\n" XMOUNT_COPYRIGHT_NOTICE "\n",XMOUNT_VERSION); printf("\nUsage:\n"); printf(" %s [fopts] \n\n",p_prog_name); printf("Options:\n"); printf(" fopts:\n"); printf(" -d : Enable FUSE's and xmount's debug mode.\n"); printf(" -h : Display this help message.\n"); printf(" -s : Run single threaded.\n"); printf(" -o no_allow_other : Disable automatic addition of FUSE's " "allow_other option.\n"); printf(" -o : Specify fuse mount options. Will also disable " "automatic addition of FUSE's allow_other option!\n"); printf("\n"); printf(" xopts:\n"); printf(" --cache : Enable virtual write support.\n"); printf(" specifies the cache file to use.\n"); printf(" --in : Input image format and source file(s). " "May be specified multiple times.\n"); printf(" can be "); // List supported input formats first=1; for(uint32_t i=0;ip_supported_input_types; while(*p_buf!='\0') { if(first==1) { printf("\"%s\"",p_buf); first=0; } else printf(", \"%s\"",p_buf); p_buf+=(strlen(p_buf)+1); } } printf(".\n"); printf(" specifies the source file. If your image is split into " "multiple files, you have to specify them all!\n"); printf(" --inopts : Specify input library specific options.\n"); printf(" specifies a comma separated list of key=value options. " "See below for details.\n"); printf(" --info : Print out infos about used compiler and libraries.\n"); printf(" --morph : Morphing function to apply to input image(s). " "If not specified, defaults to \"combine\".\n"); printf(" can be "); // List supported morphing functions first=1; for(uint32_t i=0;ip_supported_morphing_types; while(*p_buf!='\0') { if(first==1) { printf("\"%s\"",p_buf); first=0; } else printf(", \"%s\"",p_buf); p_buf+=(strlen(p_buf)+1); } } printf(".\n"); printf(" --morphopts : Specify morphing library specific " "options.\n"); printf(" specifies a comma separated list of key=value options. " "See below for details.\n"); printf(" --offset : Move the output image data start bytes " "into the input image(s).\n"); printf(" --out : Output image format. If not specified, " "defaults to "); #ifdef __APPLE__ printf("\"dmg\".\n"); #else printf("\"raw\".\n"); #endif printf(" can be "); // List supported output formats printf("\"raw\", \"dmg\", \"vdi\", \"vhd\", \"vmdk\", \"vmdks\".\n"); printf(" --owcache : Same as --cache but overwrites " "existing cache file.\n"); printf(" --sizelimit : The data end of input image(s) is set to no " "more than bytes after the data start.\n"); printf(" --version : Same as --info.\n"); printf("\n"); printf(" mntp:\n"); printf(" Mount point where output image should be located.\n"); printf("\n"); printf("Infos:\n"); printf(" * One --in option and a mount point are mandatory!\n"); printf(" * If you specify --in multiple times, data from all images is " "morphed into one output image using the specified morphing " "function.\n"); printf(" * For VMDK emulation, you have to uncomment \"user_allow_other\" " "in /etc/fuse.conf or run xmount as root.\n"); printf("\n"); printf("Input / Morphing library specific options:\n"); printf(" Input / Morphing libraries might support an own set of " "options to configure / tune their behaviour.\n"); printf(" Libraries supporting this feature (if any) and their " "options are listed below.\n"); printf("\n"); // List input and morphing lib options for(uint32_t i=0;i lib_functions.OptionsHelp((const char**)&p_buf); if(ret!=0) { LOG_ERROR("Unable to get options help for library '%s': %s!\n", glob_xmount.input.pp_libs[i]->p_name, glob_xmount.input.pp_libs[i]-> lib_functions.GetErrorMessage(ret)); } if(p_buf==NULL) continue; printf(" - %s\n",glob_xmount.input.pp_libs[i]->p_name); printf("%s",p_buf); printf("\n"); ret=glob_xmount.input.pp_libs[i]->lib_functions.FreeBuffer(p_buf); if(ret!=0) { LOG_ERROR("Unable to free options help text from library '%s': %s!\n", glob_xmount.input.pp_libs[i]->p_name, glob_xmount.input.pp_libs[i]-> lib_functions.GetErrorMessage(ret)); } } for(uint32_t i=0;i lib_functions.OptionsHelp((const char**)&p_buf); if(ret!=0) { LOG_ERROR("Unable to get options help for library '%s': %s!\n", glob_xmount.morphing.pp_libs[i]->p_name, glob_xmount.morphing.pp_libs[i]-> lib_functions.GetErrorMessage(ret)); } if(p_buf==NULL) continue; printf(" - %s\n",glob_xmount.morphing.pp_libs[i]->p_name); printf("%s",p_buf); printf("\n"); } } //! Check fuse settings /*! * Check if FUSE allows us to pass the -o allow_other parameter. This only works * if we are root or user_allow_other is set in /etc/fuse.conf. * * In addition, this function also checks if the user is member of the fuse * group which is generally needed to use fuse at all. */ static void CheckFuseSettings() { #if !defined(__APPLE__) && defined(HAVE_GRP_H) && defined(HAVE_PWD_H) struct group *p_group; struct passwd *p_passwd; #endif int found; FILE *h_fuse_conf; char line[256]; glob_xmount.may_set_fuse_allow_other=FALSE; if(geteuid()==0) { // Running as root, there should be no problems glob_xmount.may_set_fuse_allow_other=TRUE; return; } #if !defined(__APPLE__) && defined(HAVE_GRP_H) && defined(HAVE_PWD_H) // Check if a fuse group exists and if so, make sure user is a member of it. // Makes only sense on Linux because as far as I know osxfuse has no own group p_group=getgrnam("fuse"); if(p_group!=NULL) { // Get effective user name p_passwd=getpwuid(geteuid()); if(p_passwd==NULL) { printf("\nWARNING: Unable to determine your effective user name. If " "mounting works, you can ignore this message.\n\n"); return; } // Check if user is member of fuse group found=FALSE; while(*(p_group->gr_mem)!=NULL) { if(strcmp(*(p_group->gr_mem),p_passwd->pw_name)==0) { found=TRUE; break; } p_group->gr_mem++; } if(found==FALSE) { printf("\nWARNING: You are not a member of the \"fuse\" group. This will " "prevent you from mounting images using xmount. Please add " "yourself to the \"fuse\" group using the command " "\"sudo usermod -a -G fuse %s\" and reboot your system or " "execute xmount as root.\n\n", p_passwd->pw_name); return; } } else { printf("\nWARNING: Your system does not seem to have a \"fuse\" group. If " "mounting works, you can ignore this message.\n\n"); } #endif // Read FUSE's config file /etc/fuse.conf and check for set user_allow_other h_fuse_conf=(FILE*)FOPEN("/etc/fuse.conf","r"); if(h_fuse_conf!=NULL) { // Search conf file for set user_allow_others found=FALSE; while(fgets(line,sizeof(line),h_fuse_conf)!=NULL) { // TODO: This works as long as there is no other parameter beginning with // "user_allow_other" :) if(strncmp(line,"user_allow_other",16)==0) { found=TRUE; break; } } fclose(h_fuse_conf); if(found==TRUE) { glob_xmount.may_set_fuse_allow_other=TRUE; } else { printf("\nWARNING: FUSE will not allow other users nor root to access " "your virtual harddisk image. To change this behavior, please " "add \"user_allow_other\" to /etc/fuse.conf or execute xmount " "as root.\n\n"); } } else { printf("\nWARNING: Unable to open /etc/fuse.conf. If mounting works, you " "can ignore this message. If you encounter issues, please create " "the file and add a single line containing the string " "\"user_allow_other\" or execute xmount as root.\n\n"); return; } } //! Parse command line options /*! * \param argc Number of cmdline params * \param pp_argv Array containing cmdline params * \return TRUE on success, FALSE on error */ #define SUPPORT_DEPRECATED_IN 1 static int ParseCmdLine(const int argc, char **pp_argv) { int i=1; int FuseMinusOControl=TRUE; int FuseAllowOther=TRUE; int first; char *p_buf; pts_InputImage p_input_image=NULL; int ret; #ifdef SUPPORT_DEPRECATED_IN int use_old_in_syntax=FALSE; #endif // add pp_argv[0] to FUSE's argv XMOUNT_MALLOC(glob_xmount.pp_fuse_argv,char**,sizeof(char*)); XMOUNT_STRSET(glob_xmount.pp_fuse_argv[0],pp_argv[0]); glob_xmount.fuse_argc=1; // Parse options while(i1 && *(pp_argv[i]+1)!='-') { // Options beginning with one - are mostly FUSE specific if(strcmp(pp_argv[i],"-d")==0) { // Enable FUSE's and xmount's debug mode XMOUNT_REALLOC(glob_xmount.pp_fuse_argv, char**, (glob_xmount.fuse_argc+1)*sizeof(char*)); XMOUNT_STRSET(glob_xmount.pp_fuse_argv[glob_xmount.fuse_argc], pp_argv[i]) glob_xmount.fuse_argc++; glob_xmount.debug=TRUE; } else if(strcmp(pp_argv[i],"-h")==0) { // Print help message PrintUsage(pp_argv[0]); exit(0); } else if(strcmp(pp_argv[i],"-o")==0) { // Next parameter specifies fuse mount options if((i+1)p_type); free(p_input_image); return FALSE; } #endif if((i+2)p_type,pp_argv[i]); p_input_image->pp_files=NULL; p_input_image->p_functions=NULL; p_input_image->p_handle=NULL; // Parse input image filename(s) and add to p_input_image->pp_files i++; p_input_image->files_count=0; while(i<(argc-1) && strncmp(pp_argv[i],"--",2)!=0) { p_input_image->files_count++; XMOUNT_REALLOC(p_input_image->pp_files, char**, p_input_image->files_count*sizeof(char*)); XMOUNT_STRSET(p_input_image->pp_files[p_input_image->files_count-1], pp_argv[i]); i++; } i--; if(p_input_image->files_count==0) { #ifndef SUPPORT_DEPRECATED_IN LOG_ERROR("No input files specified for \"--in %s\"!\n", p_input_image->p_type) free(p_input_image->p_type); free(p_input_image); return FALSE; #else use_old_in_syntax=TRUE; i++; continue; #endif } // Add input image struct to input image array glob_xmount.input.images_count++; XMOUNT_REALLOC(glob_xmount.input.pp_images, pts_InputImage*, glob_xmount.input.images_count* sizeof(pts_InputImage)); glob_xmount.input.pp_images[glob_xmount.input.images_count-1]= p_input_image; } else { LOG_ERROR("You must specify an input image type and source file!\n"); return FALSE; } } else if(strcmp(pp_argv[i],"--inopts")==0) { // Set input lib options if((i+1)p_name); p_buf=glob_xmount.input.pp_libs[ii]->p_supported_input_types; first=TRUE; while(*p_buf!='\0') { if(first) { printf("\"%s\"",p_buf); first=FALSE; } else printf(", \"%s\"",p_buf); p_buf+=(strlen(p_buf)+1); } printf("\n"); } printf(" loaded morphing libraries:\n"); for(uint32_t ii=0;iip_name); p_buf=glob_xmount.morphing.pp_libs[ii]->p_supported_morphing_types; first=TRUE; while(*p_buf!='\0') { if(first) { printf("\"%s\"",p_buf); first=FALSE; } else printf(", \"%s\"",p_buf); p_buf+=(strlen(p_buf)+1); } printf("\n"); } printf("\n"); exit(0); } else { LOG_ERROR("Unknown command line option \"%s\"\n",pp_argv[i]); return FALSE; } } i++; } #ifdef SUPPORT_DEPRECATED_IN if(use_old_in_syntax==TRUE && i<(argc-1)) { LOG_WARNING("You are using a deprecated --in option syntax which will be " "removed in the next release. Please see the man page " "on how to use the new syntax.\n"); while(i<(argc-1)) { p_input_image->files_count++; XMOUNT_REALLOC(p_input_image->pp_files, char**, p_input_image->files_count*sizeof(char*)); XMOUNT_STRSET(p_input_image->pp_files[p_input_image->files_count-1], pp_argv[i]); i++; } // Add input image struct to input image array glob_xmount.input.images_count++; XMOUNT_REALLOC(glob_xmount.input.pp_images, pts_InputImage*, glob_xmount.input.images_count*sizeof(pts_InputImage)); glob_xmount.input.pp_images[glob_xmount.input.images_count-1]=p_input_image; } else if(use_old_in_syntax==TRUE) { free(p_input_image->p_type); free(p_input_image); LOG_ERROR("You must specify an input image type, source file(s) and a " "mount point!\n"); return FALSE; } #endif // Extract mountpoint if(i==(argc-1)) { XMOUNT_STRSET(glob_xmount.p_mountpoint,pp_argv[argc-1]) XMOUNT_REALLOC(glob_xmount.pp_fuse_argv, char**, (glob_xmount.fuse_argc+1)*sizeof(char*)); XMOUNT_STRSET(glob_xmount.pp_fuse_argv[glob_xmount.fuse_argc], glob_xmount.p_mountpoint); glob_xmount.fuse_argc++; } else { LOG_ERROR("No mountpoint specified!\n") return FALSE; } if(FuseMinusOControl==TRUE) { // We control the -o flag, set subtype, fsname and allow_other options glob_xmount.fuse_argc+=2; XMOUNT_REALLOC(glob_xmount.pp_fuse_argv, char**, glob_xmount.fuse_argc*sizeof(char*)); XMOUNT_STRSET(glob_xmount.pp_fuse_argv[glob_xmount.fuse_argc-2],"-o"); XMOUNT_STRSET(glob_xmount.pp_fuse_argv[glob_xmount.fuse_argc-1], "subtype=xmount"); if(glob_xmount.input.images_count!=0) { // Set name of first source file as fsname XMOUNT_STRAPP(glob_xmount.pp_fuse_argv[glob_xmount.fuse_argc-1], ",fsname='"); // If possible, use full path p_buf=realpath(glob_xmount.input.pp_images[0]->pp_files[0],NULL); if(p_buf==NULL) { XMOUNT_STRSET(p_buf,glob_xmount.input.pp_images[0]->pp_files[0]); } // Make sure fsname does not include some forbidden chars for(uint32_t i=0;iSize(glob_xmount.morphing.p_handle, p_size); if(ret!=0) { LOG_ERROR("Unable to get morphed image size: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); return FALSE; } return TRUE; } //! Get size of virtual image /*! * \param p_size Pointer to an uint64_t to which the size will be written to * \return TRUE on success, FALSE on error */ static int GetVirtImageSize(uint64_t *p_size) { if(glob_xmount.output.image_size!=0) { *p_size=glob_xmount.output.image_size; return TRUE; } switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: case VirtImageType_VMDK: case VirtImageType_VMDKS: // Virtual image is a DD, DMG or VMDK file. Just return the size of the // original image if(!GetMorphedImageSize(p_size)) { LOG_ERROR("Couldn't get size of input image!\n") return FALSE; } break; case VirtImageType_VDI: // Virtual image is a VDI file. Get size of original image and add size // of VDI header etc. if(!GetMorphedImageSize(p_size)) { LOG_ERROR("Couldn't get size of input image!\n") return FALSE; } (*p_size)+=(sizeof(ts_VdiFileHeader)+ glob_xmount.output.vdi.vdi_block_map_size); break; case VirtImageType_VHD: // Virtual image is a VHD file. Get size of original image and add size // of VHD footer. if(!GetMorphedImageSize(p_size)) { LOG_ERROR("Couldn't get size of input image!\n") return FALSE; } (*p_size)+=sizeof(ts_VhdFileHeader); break; default: LOG_ERROR("Unsupported image type!\n") return FALSE; } glob_xmount.output.image_size=*p_size; return TRUE; } //! Read data from input image /*! * \param p_image Image from which to read data * \param p_buf Pointer to buffer to write read data to (must be preallocated!) * \param offset Offset at which data should be read * \param size Size of data which should be read (size of buffer) * \param p_read Number of read bytes on success * \return 0 on success, negated error code on error */ static int GetInputImageData(pts_InputImage p_image, char *p_buf, off_t offset, size_t size, size_t *p_read) { int ret; size_t to_read=0; int read_errno=0; LOG_DEBUG("Reading %zu bytes at offset %zu from input image '%s'\n", size, offset, p_image->pp_files[0]); // Make sure we aren't reading past EOF of image file if(offset>=p_image->size) { // Offset is beyond image size LOG_DEBUG("Offset %zu is at / beyond size of input image '%s'\n", offset, p_image->pp_files[0]); *p_read=0; return 0; } if(offset+size>p_image->size) { // Attempt to read data past EOF of image file to_read=p_image->size-offset; LOG_DEBUG("Attempt to read data past EOF of input image '%s'. " "Correcting size from %zu to %zu\n", p_image->pp_files[0], size, to_read); } else to_read=size; // Read data from image file (adding input image offset if one was specified) ret=p_image->p_functions->Read(p_image->p_handle, p_buf, offset+glob_xmount.input.image_offset, to_read, p_read, &read_errno); if(ret!=0) { LOG_ERROR("Couldn't read %zu bytes at offset %zu from input image " "'%s': %s!\n", to_read, offset, p_image->pp_files[0], p_image->p_functions->GetErrorMessage(ret)); if(read_errno==0) return -EIO; else return (read_errno*(-1)); } return 0; } //! Read data from morphed image /*! * \param p_buf Pointer to buffer to write read data to (must be preallocated!) * \param offset Offset at which data should be read * \param size Size of data which should be read (size of buffer) * \param p_read Number of read bytes on success * \return TRUE on success, negated error code on error */ static int GetMorphedImageData(char *p_buf, off_t offset, size_t size, size_t *p_read) { int ret; size_t to_read=0; size_t read; uint64_t image_size=0; // Make sure we aren't reading past EOF of image file if(GetMorphedImageSize(&image_size)!=TRUE) { LOG_ERROR("Couldn't get size of morphed image!\n"); return -EIO; } if(offset>=image_size) { // Offset is beyond image size LOG_DEBUG("Offset %zu is at / beyond size of morphed image.\n",offset); *p_read=0; return 0; } if(offset+size>image_size) { // Attempt to read data past EOF of morphed image file to_read=image_size-offset; LOG_DEBUG("Attempt to read data past EOF of morphed image. Corrected size " "from %zu to %zu.\n", size, to_read); } else to_read=size; // Read data from morphed image ret=glob_xmount.morphing.p_functions->Read(glob_xmount.morphing.p_handle, p_buf, offset, to_read, &read); if(ret!=0) { LOG_ERROR("Couldn't read %zu bytes at offset %zu from morphed image: %s!\n", to_read, offset, glob_xmount.morphing.p_functions->GetErrorMessage(ret)); return -EIO; } *p_read=to_read; return TRUE; } //! Read data from virtual image /*! * \param p_buf Pointer to buffer to write read data to * \param offset Offset at which data should be read * \param size Size of data which should be read * \return Number of read bytes on success or negated error code on error */ static int GetVirtImageData(char *p_buf, off_t offset, size_t size) { uint32_t cur_block=0; uint64_t morphed_image_size, virt_image_size; size_t read, to_read=0, cur_to_read=0; off_t file_off=offset, block_off=0; size_t to_read_later=0; int ret; // Get virtual image size if(GetVirtImageSize(&virt_image_size)!=TRUE) { LOG_ERROR("Couldn't get size of virtual image!\n") return -EIO; } if(offset>=virt_image_size) { LOG_DEBUG("Offset %zu is at / beyond size of virtual image.\n",offset); return 0; } if(offset+size>virt_image_size) { LOG_DEBUG("Attempt to read data past EOF of virtual image. Corrected size " "from %zu to %zu.\n", size, virt_image_size-offset); size=virt_image_size-offset; } to_read=size; // Get morphed image size if(GetMorphedImageSize(&morphed_image_size)!=TRUE) { LOG_ERROR("Couldn't get morphed image size!") return -EIO; } // Read virtual image type specific data preceeding morphed image data switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: case VirtImageType_VMDK: case VirtImageType_VMDKS: break; case VirtImageType_VDI: if(file_offglob_xmount.output.vdi.vdi_header_size) { cur_to_read=glob_xmount.output.vdi.vdi_header_size-file_off; } else { cur_to_read=to_read; } if(glob_xmount.output.writable==TRUE && glob_xmount.cache.p_cache_header->VdiFileHeaderCached==TRUE) { // VDI header was already cached if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_header->pVdiFileHeader+file_off, SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to cached VDI header at offset %" PRIu64 "\n", glob_xmount.cache.p_cache_header->pVdiFileHeader+file_off) return -EIO; } if(fread(p_buf,cur_to_read,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't read %zu bytes from cache file at offset %" PRIu64 "\n", cur_to_read, glob_xmount.cache.p_cache_header->pVdiFileHeader+file_off) return -EIO; } LOG_DEBUG("Read %zd bytes from cached VDI header at offset %" PRIu64 " at cache file offset %" PRIu64 "\n", cur_to_read, file_off, glob_xmount.cache.p_cache_header->pVdiFileHeader+file_off) } else { // VDI header isn't cached memcpy(p_buf, ((char*)glob_xmount.output.vdi.p_vdi_header)+file_off, cur_to_read); LOG_DEBUG("Read %zd bytes at offset %" PRIu64 " from virtual VDI header\n",cur_to_read, file_off) } if(to_read==cur_to_read) return to_read; else { // Adjust values to read from morphed image to_read-=cur_to_read; p_buf+=cur_to_read; file_off=0; } } else file_off-=glob_xmount.output.vdi.vdi_header_size; break; case VirtImageType_VHD: // When emulating VHD, make sure the while loop below only reads data // available in the morphed image. Any VHD footer data must be read // afterwards. if(file_off>=morphed_image_size) { to_read_later=to_read; to_read=0; } else if((file_off+to_read)>morphed_image_size) { to_read_later=(file_off+to_read)-morphed_image_size; to_read-=to_read_later; } break; } // Calculate block to read data from cur_block=file_off/CACHE_BLOCK_SIZE; block_off=file_off%CACHE_BLOCK_SIZE; // Read image data while(to_read!=0) { // Calculate how many bytes we have to read from this block if(block_off+to_read>CACHE_BLOCK_SIZE) { cur_to_read=CACHE_BLOCK_SIZE-block_off; } else cur_to_read=to_read; if(glob_xmount.output.writable==TRUE && glob_xmount.cache.p_cache_blkidx[cur_block].Assigned==TRUE) { // Write support enabled and need to read altered data from cachefile if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_blkidx[cur_block].off_data+block_off, SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to offset %" PRIu64 " in cache file\n") return -EIO; } if(fread(p_buf,cur_to_read,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't read data from cache file!\n") return -EIO; } LOG_DEBUG("Read %zd bytes at offset %" PRIu64 " from cache file\n",cur_to_read,file_off) } else { // No write support or data not cached ret=GetMorphedImageData(p_buf,file_off,cur_to_read,&read); if(ret!=TRUE || read!=cur_to_read) { LOG_ERROR("Couldn't read data from virtual image!\n") return -EIO; } LOG_DEBUG("Read %zu bytes at offset %zu from virtual image file\n", cur_to_read, file_off); } cur_block++; block_off=0; p_buf+=cur_to_read; to_read-=cur_to_read; file_off+=cur_to_read; } if(to_read_later!=0) { // Read virtual image type specific data following morphed image data switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: case VirtImageType_VMDK: case VirtImageType_VMDKS: case VirtImageType_VDI: break; case VirtImageType_VHD: // Micro$oft has choosen to use a footer rather then a header. if(glob_xmount.output.writable==TRUE && glob_xmount.cache.p_cache_header->VhdFileHeaderCached==TRUE) { // VHD footer was already cached if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_header->pVhdFileHeader+ (file_off-morphed_image_size), SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to cached VHD footer at offset %" PRIu64 "\n", glob_xmount.cache.p_cache_header->pVhdFileHeader+ (file_off-morphed_image_size)) return -EIO; } if(fread(p_buf,to_read_later,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't read %zu bytes from cache file at offset %" PRIu64 "\n", to_read_later, glob_xmount.cache.p_cache_header->pVhdFileHeader+ (file_off-morphed_image_size)) return -EIO; } LOG_DEBUG("Read %zd bytes from cached VHD footer at offset %" PRIu64 " at cache file offset %" PRIu64 "\n", to_read_later, (file_off-morphed_image_size), glob_xmount.cache.p_cache_header->pVhdFileHeader+ (file_off-morphed_image_size)) } else { // VHD header isn't cached memcpy(p_buf, ((char*)glob_xmount.output.vhd.p_vhd_header)+ (file_off-morphed_image_size), to_read_later); LOG_DEBUG("Read %zd bytes at offset %" PRIu64 " from virtual VHD header\n", to_read_later, (file_off-morphed_image_size)) } break; } } return size; } //! Write data to virtual VDI file header /*! * \param p_buf Buffer containing data to write * \param offset Offset of changes * \param size Amount of bytes to write * \return Number of written bytes on success or "-1" on error */ static int SetVdiFileHeaderData(char *p_buf,off_t offset,size_t size) { if(offset+size>glob_xmount.output.vdi.vdi_header_size) { size=glob_xmount.output.vdi.vdi_header_size-offset; } LOG_DEBUG("Need to cache %zu bytes at offset %" PRIu64 " from VDI header\n", size, offset); if(glob_xmount.cache.p_cache_header->VdiFileHeaderCached==1) { // Header was already cached if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_header->pVdiFileHeader+offset, SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to cached VDI header at address %" PRIu64 "\n", glob_xmount.cache.p_cache_header->pVdiFileHeader+offset) return -1; } if(fwrite(p_buf,size,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n", size, glob_xmount.cache.p_cache_header->pVdiFileHeader+offset) return -1; } LOG_DEBUG("Wrote %zd bytes at offset %" PRIu64 " to cache file\n", size, glob_xmount.cache.p_cache_header->pVdiFileHeader+offset) } else { // Header wasn't already cached. if(fseeko(glob_xmount.cache.h_old_cache_file, 0, SEEK_END)!=0) { LOG_ERROR("Couldn't seek to end of cache file!") return -1; } glob_xmount.cache.p_cache_header->pVdiFileHeader= ftello(glob_xmount.cache.h_old_cache_file); LOG_DEBUG("Caching whole VDI header\n") if(offset>0) { // Changes do not begin at offset 0, need to prepend with data from // VDI header if(fwrite((char*)glob_xmount.output.vdi.p_vdi_header, offset, 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Error while writing %" PRIu64 " bytes " "to cache file at offset %" PRIu64 "!\n", offset, glob_xmount.cache.p_cache_header->pVdiFileHeader); return -1; } LOG_DEBUG("Prepended changed data with %" PRIu64 " bytes at cache file offset %" PRIu64 "\n", offset, glob_xmount.cache.p_cache_header->pVdiFileHeader) } // Cache changed data if(fwrite(p_buf,size,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n",size, glob_xmount.cache.p_cache_header->pVdiFileHeader+offset) return -1; } LOG_DEBUG("Wrote %zu bytes of changed data to cache file offset %" PRIu64 "\n", size, glob_xmount.cache.p_cache_header->pVdiFileHeader+offset) if(offset+size!=glob_xmount.output.vdi.vdi_header_size) { // Need to append data from VDI header to cache whole data struct if(fwrite(((char*)glob_xmount.output.vdi.p_vdi_header)+offset+size, glob_xmount.output.vdi.vdi_header_size-(offset+size), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n", glob_xmount.output.vdi.vdi_header_size-(offset+size), (uint64_t)(glob_xmount.cache.p_cache_header->pVdiFileHeader+ offset+size)); return -1; } LOG_DEBUG("Appended %" PRIu32 " bytes to changed data at cache file offset %" PRIu64 "\n", glob_xmount.output.vdi.vdi_header_size-(offset+size), glob_xmount.cache.p_cache_header->pVdiFileHeader+offset+size) } // Mark header as cached and update header in cache file glob_xmount.cache.p_cache_header->VdiFileHeaderCached=1; if(fseeko(glob_xmount.cache.h_old_cache_file,0,SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to offset 0 of cache file!\n") return -1; } if(fwrite((char*)glob_xmount.cache.p_cache_header, sizeof(ts_CacheFileHeader), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write changed cache file header!\n") return -1; } } // All important data has been written, now flush all buffers to make // sure data is written to cache file fflush(glob_xmount.cache.h_old_cache_file); #ifndef __APPLE__ ioctl(fileno(glob_xmount.cache.h_old_cache_file),BLKFLSBUF,0); #endif return size; } //! Write data to virtual VHD file footer /*! * \param p_buf Buffer containing data to write * \param offset Offset of changes * \param size Amount of bytes to write * \return Number of written bytes on success or "-1" on error */ static int SetVhdFileHeaderData(char *p_buf,off_t offset,size_t size) { LOG_DEBUG("Need to cache %zu bytes at offset %" PRIu64 " from VHD footer\n",size,offset) if(glob_xmount.cache.p_cache_header->VhdFileHeaderCached==1) { // Header has already been cached if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_header->pVhdFileHeader+offset, SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to cached VHD header at address %" PRIu64 "\n", glob_xmount.cache.p_cache_header->pVhdFileHeader+offset); return -1; } if(fwrite(p_buf,size,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n", size, glob_xmount.cache.p_cache_header->pVhdFileHeader+offset); return -1; } LOG_DEBUG("Wrote %zd bytes at offset %" PRIu64 " to cache file\n", size, glob_xmount.cache.p_cache_header->pVhdFileHeader+offset); } else { // Header hasn't been cached yet. if(fseeko(glob_xmount.cache.h_old_cache_file, 0, SEEK_END)!=0) { LOG_ERROR("Couldn't seek to end of cache file!") return -1; } glob_xmount.cache.p_cache_header->pVhdFileHeader= ftello(glob_xmount.cache.h_old_cache_file); LOG_DEBUG("Caching whole VHD header\n") if(offset>0) { // Changes do not begin at offset 0, need to prepend with data from // VHD header if(fwrite((char*)glob_xmount.output.vhd.p_vhd_header, offset, 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Error while writing %" PRIu64 " bytes " "to cache file at offset %" PRIu64 "!\n", offset, glob_xmount.cache.p_cache_header->pVhdFileHeader); return -1; } LOG_DEBUG("Prepended changed data with %" PRIu64 " bytes at cache file offset %" PRIu64 "\n", offset, glob_xmount.cache.p_cache_header->pVhdFileHeader); } // Cache changed data if(fwrite(p_buf,size,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n", size, glob_xmount.cache.p_cache_header->pVhdFileHeader+offset); return -1; } LOG_DEBUG("Wrote %zu bytes of changed data to cache file offset %" PRIu64 "\n", size, glob_xmount.cache.p_cache_header->pVhdFileHeader+offset); if(offset+size!=sizeof(ts_VhdFileHeader)) { // Need to append data from VHD header to cache whole data struct if(fwrite(((char*)glob_xmount.output.vhd.p_vhd_header)+offset+size, sizeof(ts_VhdFileHeader)-(offset+size), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write %zu bytes to cache file at offset %" PRIu64 "\n", sizeof(ts_VhdFileHeader)-(offset+size), (uint64_t)(glob_xmount.cache.p_cache_header->pVhdFileHeader+ offset+size)) return -1; } LOG_DEBUG("Appended %" PRIu32 " bytes to changed data at cache file offset %" PRIu64 "\n", sizeof(ts_VhdFileHeader)-(offset+size), glob_xmount.cache.p_cache_header->pVhdFileHeader+offset+size); } // Mark header as cached and update header in cache file glob_xmount.cache.p_cache_header->VhdFileHeaderCached=1; if(fseeko(glob_xmount.cache.h_old_cache_file,0,SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to offset 0 of cache file!\n") return -1; } if(fwrite((char*)glob_xmount.cache.p_cache_header, sizeof(ts_CacheFileHeader), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't write changed cache file header!\n") return -1; } } // All important data has been written, now flush all buffers to make // sure data is written to cache file fflush(glob_xmount.cache.h_old_cache_file); #ifndef __APPLE__ ioctl(fileno(glob_xmount.cache.h_old_cache_file),BLKFLSBUF,0); #endif return size; } //! Write data to virtual image /*! * \param p_buf Buffer containing data to write * \param offset Offset to start writing at * \param size Size of data to be written * \return Number of written bytes on success or "-1" on error */ static int SetVirtImageData(const char *p_buf, off_t offset, size_t size) { uint64_t cur_block=0; uint64_t virt_image_size; uint64_t orig_image_size; size_t to_write=0; size_t to_write_later=0; size_t to_write_now=0; off_t file_offset=offset; off_t block_offset=0; char *p_write_buf=(char*)p_buf; char *p_buf2; int ret; size_t read; // Get virtual image size if(!GetVirtImageSize(&virt_image_size)) { LOG_ERROR("Couldn't get virtual image size!\n") return -1; } if(offset>=virt_image_size) { LOG_ERROR("Attempt to write beyond EOF of virtual image file!\n") return -1; } if(offset+size>virt_image_size) { LOG_DEBUG("Attempt to write past EOF of virtual image file\n") size=virt_image_size-offset; } to_write=size; // Get original image size if(!GetMorphedImageSize(&orig_image_size)) { LOG_ERROR("Couldn't get morphed image size!\n") return -1; } // Cache virtual image type specific data preceeding original image data switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: case VirtImageType_VMDK: case VirtImageType_VMDKS: break; case VirtImageType_VDI: if(file_offset=orig_image_size) { to_write_later=to_write; to_write=0; } else if((file_offset+to_write)>orig_image_size) { to_write_later=(file_offset+to_write)-orig_image_size; to_write-=to_write_later; } break; } // Calculate block to write data to cur_block=file_offset/CACHE_BLOCK_SIZE; block_offset=file_offset%CACHE_BLOCK_SIZE; while(to_write!=0) { // Calculate how many bytes we have to write to this block if(block_offset+to_write>CACHE_BLOCK_SIZE) { to_write_now=CACHE_BLOCK_SIZE-block_offset; } else to_write_now=to_write; if(glob_xmount.cache.p_cache_blkidx[cur_block].Assigned==1) { // Block was already cached // Seek to data offset in cache file if(fseeko(glob_xmount.cache.h_old_cache_file, glob_xmount.cache.p_cache_blkidx[cur_block].off_data+block_offset, SEEK_SET)!=0) { LOG_ERROR("Couldn't seek to cached block at address %" PRIu64 "\n", glob_xmount.cache.p_cache_blkidx[cur_block].off_data+ block_offset); return -1; } if(fwrite(p_write_buf,to_write_now,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Error while writing %zu bytes " "to cache file at offset %" PRIu64 "!\n", to_write_now, glob_xmount.cache.p_cache_blkidx[cur_block].off_data+ block_offset); return -1; } LOG_DEBUG("Wrote %zd bytes at offset %" PRIu64 " to cache file\n",to_write_now, glob_xmount.cache.p_cache_blkidx[cur_block].off_data+ block_offset); } else { // Uncached block. Need to cache entire new block // Seek to end of cache file to append new cache block fseeko(glob_xmount.cache.h_old_cache_file,0,SEEK_END); glob_xmount.cache.p_cache_blkidx[cur_block].off_data= ftello(glob_xmount.cache.h_old_cache_file); if(block_offset!=0) { // Changed data does not begin at block boundry. Need to prepend // with data from virtual image file XMOUNT_MALLOC(p_buf2,char*,block_offset*sizeof(char)); ret=GetMorphedImageData(p_buf2, file_offset-block_offset, block_offset, &read); if(ret!=TRUE || read!=block_offset) { LOG_ERROR("Couldn't read data from morphed image!\n") return -1; } if(fwrite(p_buf2,block_offset,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't writing %" PRIu64 " bytes " "to cache file at offset %" PRIu64 "!\n", block_offset, glob_xmount.cache.p_cache_blkidx[cur_block].off_data); return -1; } LOG_DEBUG("Prepended changed data with %" PRIu64 " bytes from virtual image file at offset %" PRIu64 "\n",block_offset,file_offset-block_offset) free(p_buf2); } if(fwrite(p_write_buf,to_write_now,1,glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Error while writing %zd bytes " "to cache file at offset %" PRIu64 "!\n", to_write_now, glob_xmount.cache.p_cache_blkidx[cur_block].off_data+ block_offset); return -1; } if(block_offset+to_write_now!=CACHE_BLOCK_SIZE) { // Changed data does not end at block boundry. Need to append // with data from virtual image file XMOUNT_MALLOC(p_buf2,char*,(CACHE_BLOCK_SIZE- (block_offset+to_write_now))*sizeof(char)) memset(p_buf2,0,CACHE_BLOCK_SIZE-(block_offset+to_write_now)); if((file_offset-block_offset)+CACHE_BLOCK_SIZE>orig_image_size) { // Original image is smaller than full cache block ret=GetMorphedImageData(p_buf2, file_offset+to_write_now, orig_image_size-(file_offset+to_write_now), &read); if(ret!=TRUE || read!=orig_image_size-(file_offset+to_write_now)) { LOG_ERROR("Couldn't read data from virtual image file!\n") return -1; } } else { ret=GetMorphedImageData(p_buf2, file_offset+to_write_now, CACHE_BLOCK_SIZE-(block_offset+to_write_now), &read); if(ret!=TRUE || read!=CACHE_BLOCK_SIZE-(block_offset+to_write_now)) { LOG_ERROR("Couldn't read data from virtual image file!\n") return -1; } } if(fwrite(p_buf2, CACHE_BLOCK_SIZE-(block_offset+to_write_now), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Error while writing %zd bytes " "to cache file at offset %" PRIu64 "!\n", CACHE_BLOCK_SIZE-(block_offset+to_write_now), glob_xmount.cache.p_cache_blkidx[cur_block].off_data+ block_offset+to_write_now); return -1; } free(p_buf2); } // All important data for this cache block has been written, // flush all buffers and mark cache block as assigned fflush(glob_xmount.cache.h_old_cache_file); #ifndef __APPLE__ ioctl(fileno(glob_xmount.cache.h_old_cache_file),BLKFLSBUF,0); #endif glob_xmount.cache.p_cache_blkidx[cur_block].Assigned=1; // Update cache block index entry in cache file fseeko(glob_xmount.cache.h_old_cache_file, sizeof(ts_CacheFileHeader)+ (cur_block*sizeof(ts_CacheFileBlockIndex)), SEEK_SET); if(fwrite(&(glob_xmount.cache.p_cache_blkidx[cur_block]), sizeof(ts_CacheFileBlockIndex), 1, glob_xmount.cache.h_old_cache_file)!=1) { LOG_ERROR("Couldn't update cache file block index!\n"); return -1; } LOG_DEBUG("Updated cache file block index: Number=%" PRIu64 ", Data offset=%" PRIu64 "\n", cur_block, glob_xmount.cache.p_cache_blkidx[cur_block].off_data); } // Flush buffers fflush(glob_xmount.cache.h_old_cache_file); #ifndef __APPLE__ ioctl(fileno(glob_xmount.cache.h_old_cache_file),BLKFLSBUF,0); #endif block_offset=0; cur_block++; p_write_buf+=to_write_now; to_write-=to_write_now; file_offset+=to_write_now; } if(to_write_later!=0) { // Cache virtual image type specific data preceeding original image data switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: case VirtImageType_VMDK: case VirtImageType_VMDKS: case VirtImageType_VDI: break; case VirtImageType_VHD: // Micro$oft has choosen to use a footer rather then a header. ret=SetVhdFileHeaderData(p_write_buf, file_offset-orig_image_size, to_write_later); if(ret==-1) { LOG_ERROR("Couldn't write data to virtual VHD file footer!\n") return -1; } break; } } return size; } //! Calculates an MD5 hash of the first HASH_AMOUNT bytes of the input image /*! * \param p_hash_low Pointer to the lower 64 bit of the hash * \param p_hash_high Pointer to the higher 64 bit of the hash * \return TRUE on success, FALSE on error */ static int CalculateInputImageHash(uint64_t *p_hash_low, uint64_t *p_hash_high) { char hash[16]; md5_state_t md5_state; char *p_buf; int ret; size_t read_data; XMOUNT_MALLOC(p_buf,char*,HASH_AMOUNT*sizeof(char)); ret=GetMorphedImageData(p_buf,0,HASH_AMOUNT,&read_data); if(ret!=TRUE || read_data==0) { LOG_ERROR("Couldn't read data from morphed image file!\n") free(p_buf); return FALSE; } // Calculate MD5 hash md5_init(&md5_state); md5_append(&md5_state,(const md5_byte_t*)p_buf,read_data); md5_finish(&md5_state,(md5_byte_t*)hash); // Convert MD5 hash into two 64bit integers *p_hash_low=*((uint64_t*)hash); *p_hash_high=*((uint64_t*)(hash+8)); free(p_buf); return TRUE; } //! Build and init virtual VDI file header /*! * \return TRUE on success, FALSE on error */ static int InitVirtVdiHeader() { // See http://forums.virtualbox.org/viewtopic.php?t=8046 for a // "description" of the various header fields uint64_t image_size; off_t offset; uint32_t i,block_entries; // Get input image size if(!GetMorphedImageSize(&image_size)) { LOG_ERROR("Couldn't get morphed image size!\n") return FALSE; } // Calculate how many VDI blocks we need block_entries=image_size/VDI_IMAGE_BLOCK_SIZE; if((image_size%VDI_IMAGE_BLOCK_SIZE)!=0) block_entries++; glob_xmount.output.vdi.vdi_block_map_size=block_entries*sizeof(uint32_t); LOG_DEBUG("BlockMap: %d (%08X) entries, %d (%08X) bytes!\n", block_entries, block_entries, glob_xmount.output.vdi.vdi_block_map_size, glob_xmount.output.vdi.vdi_block_map_size) // Allocate memory for vdi header and block map glob_xmount.output.vdi.vdi_header_size= sizeof(ts_VdiFileHeader)+glob_xmount.output.vdi.vdi_block_map_size; XMOUNT_MALLOC(glob_xmount.output.vdi.p_vdi_header, pts_VdiFileHeader, glob_xmount.output.vdi.vdi_header_size); memset(glob_xmount.output.vdi.p_vdi_header, 0, glob_xmount.output.vdi.vdi_header_size); glob_xmount.output.vdi.p_vdi_block_map= ((void*)glob_xmount.output.vdi.p_vdi_header)+sizeof(ts_VdiFileHeader); // Init header values strncpy(glob_xmount.output.vdi.p_vdi_header->szFileInfo, VDI_FILE_COMMENT, strlen(VDI_FILE_COMMENT)+1); glob_xmount.output.vdi.p_vdi_header->u32Signature=VDI_IMAGE_SIGNATURE; glob_xmount.output.vdi.p_vdi_header->u32Version=VDI_IMAGE_VERSION; // No idea what the following value is for! Testimage had same value glob_xmount.output.vdi.p_vdi_header->cbHeader=0x00000180; glob_xmount.output.vdi.p_vdi_header->u32Type=VDI_IMAGE_TYPE_FIXED; glob_xmount.output.vdi.p_vdi_header->fFlags=VDI_IMAGE_FLAGS; strncpy(glob_xmount.output.vdi.p_vdi_header->szComment, VDI_HEADER_COMMENT, strlen(VDI_HEADER_COMMENT)+1); glob_xmount.output.vdi.p_vdi_header->offData= glob_xmount.output.vdi.vdi_header_size; glob_xmount.output.vdi.p_vdi_header->offBlocks=sizeof(ts_VdiFileHeader); glob_xmount.output.vdi.p_vdi_header->cCylinders=0; // Legacy info glob_xmount.output.vdi.p_vdi_header->cHeads=0; // Legacy info glob_xmount.output.vdi.p_vdi_header->cSectors=0; // Legacy info glob_xmount.output.vdi.p_vdi_header->cbSector=512; // Legacy info glob_xmount.output.vdi.p_vdi_header->u32Dummy=0; glob_xmount.output.vdi.p_vdi_header->cbDisk=image_size; // Seems as VBox is always using a 1MB blocksize glob_xmount.output.vdi.p_vdi_header->cbBlock=VDI_IMAGE_BLOCK_SIZE; glob_xmount.output.vdi.p_vdi_header->cbBlockExtra=0; glob_xmount.output.vdi.p_vdi_header->cBlocks=block_entries; glob_xmount.output.vdi.p_vdi_header->cBlocksAllocated=block_entries; // Use partial MD5 input file hash as creation UUID and generate a random // modification UUID. VBox won't accept immages where create and modify UUIDS // aren't set. glob_xmount.output.vdi.p_vdi_header->uuidCreate_l= glob_xmount.input.image_hash_lo; glob_xmount.output.vdi.p_vdi_header->uuidCreate_h= glob_xmount.input.image_hash_hi; #define rand64(var) { \ *((uint32_t*)&(var))=rand(); \ *(((uint32_t*)&(var))+1)=rand(); \ } rand64(glob_xmount.output.vdi.p_vdi_header->uuidModify_l); rand64(glob_xmount.output.vdi.p_vdi_header->uuidModify_h); #undef rand64 // Generate block map i=0; for(offset=0;offsetcookie=VHD_IMAGE_HVAL_COOKIE; glob_xmount.output.vhd.p_vhd_header->features=VHD_IMAGE_HVAL_FEATURES; glob_xmount.output.vhd.p_vhd_header->file_format_version= VHD_IMAGE_HVAL_FILE_FORMAT_VERSION; glob_xmount.output.vhd.p_vhd_header->data_offset=VHD_IMAGE_HVAL_DATA_OFFSET; glob_xmount.output.vhd.p_vhd_header->creation_time= htobe32(time(NULL)-VHD_IMAGE_TIME_CONVERSION_OFFSET); glob_xmount.output.vhd.p_vhd_header->creator_app= VHD_IMAGE_HVAL_CREATOR_APPLICATION; glob_xmount.output.vhd.p_vhd_header->creator_ver= VHD_IMAGE_HVAL_CREATOR_VERSION; glob_xmount.output.vhd.p_vhd_header->creator_os= VHD_IMAGE_HVAL_CREATOR_HOST_OS; glob_xmount.output.vhd.p_vhd_header->size_original=htobe64(orig_image_size); glob_xmount.output.vhd.p_vhd_header->size_current= glob_xmount.output.vhd.p_vhd_header->size_original; // Convert size to sectors if(orig_image_size>136899993600) { // image is larger then CHS values can address. // Set sectors to max (C65535*H16*S255). geom_tot_s=267382800; } else { // Calculate actual sectors geom_tot_s=orig_image_size/512; if((orig_image_size%512)!=0) geom_tot_s++; } // Calculate CHS values. This is done according to the VHD specs if(geom_tot_s>=66059280) { // C65535 * H16 * S63 geom_s=255; geom_h=16; geom_c_x_h=geom_tot_s/geom_s; } else { geom_s=17; geom_c_x_h=geom_tot_s/geom_s; geom_h=(geom_c_x_h+1023)/1024; if(geom_h<4) geom_h=4; if(geom_c_x_h>=(geom_h*1024) || geom_h>16) { geom_s=31; geom_h=16; geom_c_x_h=geom_tot_s/geom_s; } if(geom_c_x_h>=(geom_h*1024)) { geom_s=63; geom_h=16; geom_c_x_h=geom_tot_s/geom_s; } } geom_c=geom_c_x_h/geom_h; glob_xmount.output.vhd.p_vhd_header->disk_geometry_c=htobe16(geom_c); glob_xmount.output.vhd.p_vhd_header->disk_geometry_h=geom_h; glob_xmount.output.vhd.p_vhd_header->disk_geometry_s=geom_s; glob_xmount.output.vhd.p_vhd_header->disk_type=VHD_IMAGE_HVAL_DISK_TYPE; glob_xmount.output.vhd.p_vhd_header->uuid_l=glob_xmount.input.image_hash_lo; glob_xmount.output.vhd.p_vhd_header->uuid_h=glob_xmount.input.image_hash_hi; glob_xmount.output.vhd.p_vhd_header->saved_state=0x00; // Calculate footer checksum for(i=0;ichecksum=htobe32(~checksum); LOG_DEBUG("VHD header size = %u\n",sizeof(ts_VhdFileHeader)); return TRUE; } //! Init the virtual VMDK file /*! * \return TRUE on success, FALSE on error */ static int InitVirtualVmdkFile() { uint64_t image_size=0; uint64_t image_blocks=0; char buf[500]; // Get original image size if(!GetMorphedImageSize(&image_size)) { LOG_ERROR("Couldn't get morphed image size!\n") return FALSE; } image_blocks=image_size/512; if(image_size%512!=0) image_blocks++; #define VMDK_DESC_FILE "# Disk DescriptorFile\n" \ "version=1\n" \ "CID=fffffffe\n" \ "parentCID=ffffffff\n" \ "createType=\"monolithicFlat\"\n\n" \ "# Extent description\n" \ "RW %" PRIu64 " FLAT \"%s\" 0\n\n" \ "# The Disk Data Base\n" \ "#DDB\n" \ "ddb.virtualHWVersion = \"3\"\n" \ "ddb.adapterType = \"%s\"\n" \ "ddb.geometry.cylinders = \"0\"\n" \ "ddb.geometry.heads = \"0\"\n" \ "ddb.geometry.sectors = \"0\"\n" if(glob_xmount.output.VirtImageType==VirtImageType_VMDK) { // VMDK with IDE bus sprintf(buf, VMDK_DESC_FILE, image_blocks, (glob_xmount.output.p_virtual_image_path)+1, "ide"); } else if(glob_xmount.output.VirtImageType==VirtImageType_VMDKS){ // VMDK with SCSI bus sprintf(buf, VMDK_DESC_FILE, image_blocks, (glob_xmount.output.p_virtual_image_path)+1, "scsi"); } else { LOG_ERROR("Unknown virtual VMDK file format!\n") return FALSE; } #undef VMDK_DESC_FILE // Do not use XMOUNT_STRSET here to avoid adding '\0' to the buffer! XMOUNT_MALLOC(glob_xmount.output.vmdk.p_vmdk_file,char*,strlen(buf)) strncpy(glob_xmount.output.vmdk.p_vmdk_file,buf,strlen(buf)); glob_xmount.output.vmdk.vmdk_file_size=strlen(buf); return TRUE; } //! Create virtual image info file /*! * \return TRUE on success, FALSE on error */ static int InitVirtImageInfoFile() { int ret; char *p_buf; // Start with static input header XMOUNT_MALLOC(glob_xmount.output.p_info_file, char*, strlen(IMAGE_INFO_INPUT_HEADER)+1); strncpy(glob_xmount.output.p_info_file, IMAGE_INFO_INPUT_HEADER, strlen(IMAGE_INFO_INPUT_HEADER)+1); // Get and add infos from input lib(s) for(uint64_t i=0;ip_functions-> GetInfofileContent(glob_xmount.input.pp_images[i]->p_handle,(const char**)&p_buf); if(ret!=0) { LOG_ERROR("Unable to get info file content for image '%s': %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); return FALSE; } // Add infos to main buffer and free p_buf XMOUNT_STRAPP(glob_xmount.output.p_info_file,"\n--> "); XMOUNT_STRAPP(glob_xmount.output.p_info_file, glob_xmount.input.pp_images[i]->pp_files[0]); XMOUNT_STRAPP(glob_xmount.output.p_info_file," <--\n"); if(p_buf!=NULL) { XMOUNT_STRAPP(glob_xmount.output.p_info_file,p_buf); glob_xmount.input.pp_images[i]->p_functions->FreeBuffer(p_buf); } else { XMOUNT_STRAPP(glob_xmount.output.p_info_file,"None\n"); } } // Add static morphing header XMOUNT_STRAPP(glob_xmount.output.p_info_file,IMAGE_INFO_MORPHING_HEADER); // Get and add infos from morphing lib ret=glob_xmount.morphing.p_functions-> GetInfofileContent(glob_xmount.morphing.p_handle,(const char**)&p_buf); if(ret!=0) { LOG_ERROR("Unable to get info file content from morphing lib: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); return FALSE; } if(p_buf!=NULL) { XMOUNT_STRAPP(glob_xmount.output.p_info_file,p_buf); glob_xmount.morphing.p_functions->FreeBuffer(p_buf); } else { XMOUNT_STRAPP(glob_xmount.output.p_info_file,"None\n"); } return TRUE; } //! Create / load cache file to enable virtual write support /*! * \return TRUE on success, FALSE on error */ static int InitCacheFile() { uint64_t blockindex_size=0; uint64_t image_size=0; uint64_t written=0; uint32_t needed_blocks=0; uint8_t is_new_cache_file=0; teGidaFsError gidafs_ret=eGidaFsError_None; t_CacheFileBlockIndex *p_index_buf=NULL; // Get input image size for later use if(!GetMorphedImageSize(&image_size)) { LOG_ERROR("Couldn't get morphed image size!\n") return FALSE; } if(!glob_xmount.cache.overwrite_cache) { // Try to open an existing cache file or create a new one gidafs_ret=GidaFsLib_OpenFs(&(glob_xmount.cache.h_cache_file), glob_xmount.cache.p_cache_file); if(gidafs_ret!=eGidaFsError_None && gidafs_ret!=eGidaFsError_FailedOpeningFsFile) { // TODO: Check for old cache file type and inform user it isn't supported // anymore! LOG_ERROR("Couldn't open cache file '%s': Error code %u!\n", glob_xmount.cache.p_cache_file, gidafs_ret) return FALSE; } else if(gidafs_ret==eGidaFsError_FailedOpeningFsFile) { // Unable to open cache file. It might simply not exist. LOG_DEBUG("Cache file '%s' does not exist. Creating new one\n", glob_xmount.cache.p_cache_file) gidafs_ret=GidaFsLib_NewFs(&(glob_xmount.cache.h_cache_file), glob_xmount.cache.p_cache_file, 0); if(gidafs_ret!=eGidaFsError_None) { // There is really a problem opening/creating the file LOG_ERROR("Couldn't open cache file '%s': Error code %u!\n", glob_xmount.cache.p_cache_file, gidafs_ret) return FALSE; } is_new_cache_file=1; } } else { // Overwrite existing cache file or create a new one gidafs_ret=GidaFsLib_NewFs(&(glob_xmount.cache.h_cache_file), glob_xmount.cache.p_cache_file, 0); if(gidafs_ret!=eGidaFsError_None) { // There is really a problem opening/creating the file LOG_ERROR("Couldn't open cache file '%s': Error code %u!\n", glob_xmount.cache.p_cache_file, gidafs_ret) return FALSE; } is_new_cache_file=1; } #define INITCACHEFILE__CLOSE_CACHE do { \ gidafs_ret=GidaFsLib_CloseFs(&(glob_xmount.cache.h_cache_file)); \ if(gidafs_ret!=eGidaFsError_None) { \ LOG_ERROR("Unable to close cache file: Error code %u: Ignoring!\n", \ gidafs_ret) \ } \ } while(0) #define INITCACHEFILE__CLOSE_BLOCK_CACHE do { \ gidafs_ret=GidaFsLib_CloseFile(glob_xmount.cache.h_cache_file, \ &(glob_xmount.cache.h_block_cache)); \ if(gidafs_ret!=eGidaFsError_None) { \ LOG_ERROR("Unable to close block cache file: Error code %u: Ignoring!\n", \ gidafs_ret) \ } \ } while(0) #define INITCACHEFILE__CLOSE_BLOCK_CACHE_INDEX do { \ gidafs_ret=GidaFsLib_CloseFile(glob_xmount.cache.h_cache_file, \ &(glob_xmount.cache.h_block_cache_index)); \ if(gidafs_ret!=eGidaFsError_None) { \ LOG_ERROR("Unable to close block cache index file: Error code %u: " \ "Ignoring!\n", \ gidafs_ret) \ } \ } while(0) if(is_new_cache_file==1) { // New cache file, create needed xmount subdirectory gidafs_ret=GidaFsLib_CreateDir(glob_xmount.cache.h_cache_file, XMOUNT_CACHE_FOLDER, eGidaFsNodeFlag_RWXu); if(gidafs_ret!=eGidaFsError_None) { LOG_ERROR("Unable to create cache file directory '%s': Error code %u!\n", XMOUNT_CACHE_FOLDER, gidafs_ret) INITCACHEFILE__CLOSE_CACHE; return FALSE; } } // Open / Create block cache file gidafs_ret=GidaFsLib_OpenFile(glob_xmount.cache.h_cache_file, XMOUNT_CACHE_BLOCK_FILE, &(glob_xmount.cache.h_block_cache), eGidaFsOpenFileFlag_ReadWrite | (is_new_cache_file==1 ? eGidaFsOpenFileFlag_CreateAlways : 0), eGidaFsNodeFlag_Rall | eGidaFsNodeFlag_Wusr); if(gidafs_ret!=eGidaFsError_None) { LOG_ERROR("Unable to open / create block cache file '%s': Error code %u!\n", XMOUNT_CACHE_BLOCK_FILE, gidafs_ret) INITCACHEFILE__CLOSE_CACHE; return FALSE; } // Create block cache index file gidafs_ret=GidaFsLib_OpenFile(glob_xmount.cache.h_cache_file, XMOUNT_CACHE_BLOCK_INDEX_FILE, &(glob_xmount.cache.h_block_cache_index), eGidaFsOpenFileFlag_ReadWrite | (is_new_cache_file==1 ? eGidaFsOpenFileFlag_CreateAlways : 0), eGidaFsNodeFlag_Rall | eGidaFsNodeFlag_Wusr); if(gidafs_ret!=eGidaFsError_None) { LOG_ERROR("Unable to open / create block cache index file '%s': " "Error code %u!\n", XMOUNT_CACHE_BLOCK_FILE, gidafs_ret) INITCACHEFILE__CLOSE_BLOCK_CACHE; INITCACHEFILE__CLOSE_CACHE; return FALSE; } // Calculate how many cache blocks are needed and how big the cache block // index must be needed_blocks=image_size/CACHE_BLOCK_SIZE; if((image_size%CACHE_BLOCK_SIZE)!=0) needed_blocks++; LOG_DEBUG("Cache blocks: %u (0x%04X) entries using %zd (0x%08zX) bytes\n", needed_blocks, needed_blocks, needed_blocks*sizeof(t_CacheFileBlockIndex), needed_blocks*sizeof(t_CacheFileBlockIndex)) if(is_new_cache_file==1) { // Generate initial block cache index blockindex_size=needed_blocks*sizeof(t_CacheFileBlockIndex); XMOUNT_MALLOC(p_index_buf,t_CacheFileBlockIndex*,blockindex_size) for(uint64_t i=0;id_name); // Construct full path to found object p_library_path=realloc(p_library_path, base_library_path_len+strlen(p_dirent->d_name)+1); if(p_library_path==NULL) { LOG_ERROR("Couldn't allocate memory!\n"); exit(1); } strcpy(p_library_path+base_library_path_len,p_dirent->d_name); if(strncmp(p_dirent->d_name,"libxmount_input_",16)==0) { // Found possible input lib. Try to load it LIBXMOUNT_LOAD(p_library_path); // Load library symbols LIBXMOUNT_LOAD_SYMBOL("LibXmount_Input_GetApiVersion", pfun_input_GetApiVersion); // Check library's API version if(pfun_input_GetApiVersion()!=LIBXMOUNT_INPUT_API_VERSION) { LOG_DEBUG("Failed! Wrong API version.\n"); LOG_ERROR("Unable to load input library '%s'. Wrong API version\n", p_library_path); dlclose(p_libxmount); continue; } LIBXMOUNT_LOAD_SYMBOL("LibXmount_Input_GetSupportedFormats", pfun_input_GetSupportedFormats); LIBXMOUNT_LOAD_SYMBOL("LibXmount_Input_GetFunctions", pfun_input_GetFunctions); // Construct new entry for our library list XMOUNT_MALLOC(p_input_lib,pts_InputLib,sizeof(ts_InputLib)); // Initialize lib_functions structure to NULL memset(&(p_input_lib->lib_functions), 0, sizeof(ts_LibXmountInputFunctions)); // Set name and handle XMOUNT_STRSET(p_input_lib->p_name,p_dirent->d_name); p_input_lib->p_lib=p_libxmount; // Get and set supported formats p_supported_formats=pfun_input_GetSupportedFormats(); supported_formats_len=0; p_buf=p_supported_formats; while(*p_buf!='\0') { supported_formats_len+=(strlen(p_buf)+1); p_buf+=(strlen(p_buf)+1); } supported_formats_len++; XMOUNT_MALLOC(p_input_lib->p_supported_input_types, char*, supported_formats_len); memcpy(p_input_lib->p_supported_input_types, p_supported_formats, supported_formats_len); // Get, set and check lib_functions pfun_input_GetFunctions(&(p_input_lib->lib_functions)); if(p_input_lib->lib_functions.CreateHandle==NULL || p_input_lib->lib_functions.DestroyHandle==NULL || p_input_lib->lib_functions.Open==NULL || p_input_lib->lib_functions.Close==NULL || p_input_lib->lib_functions.Size==NULL || p_input_lib->lib_functions.Read==NULL || p_input_lib->lib_functions.OptionsHelp==NULL || p_input_lib->lib_functions.OptionsParse==NULL || p_input_lib->lib_functions.GetInfofileContent==NULL || p_input_lib->lib_functions.GetErrorMessage==NULL || p_input_lib->lib_functions.FreeBuffer==NULL) { LOG_DEBUG("Missing implemention of one or more functions in lib %s!\n", p_dirent->d_name); free(p_input_lib->p_supported_input_types); free(p_input_lib->p_name); free(p_input_lib); dlclose(p_libxmount); continue; } // Add entry to the input library list XMOUNT_REALLOC(glob_xmount.input.pp_libs, pts_InputLib*, sizeof(pts_InputLib)*(glob_xmount.input.libs_count+1)); glob_xmount.input.pp_libs[glob_xmount.input.libs_count++]=p_input_lib; LOG_DEBUG("Input library '%s' loaded successfully\n",p_dirent->d_name); } if(strncmp(p_dirent->d_name,"libxmount_morphing_",19)==0) { // Found possible morphing lib. Try to load it LIBXMOUNT_LOAD(p_library_path); // Load library symbols LIBXMOUNT_LOAD_SYMBOL("LibXmount_Morphing_GetApiVersion", pfun_morphing_GetApiVersion); // Check library's API version if(pfun_morphing_GetApiVersion()!=LIBXMOUNT_MORPHING_API_VERSION) { LOG_DEBUG("Failed! Wrong API version.\n"); LOG_ERROR("Unable to load morphing library '%s'. Wrong API version\n", p_library_path); dlclose(p_libxmount); continue; } LIBXMOUNT_LOAD_SYMBOL("LibXmount_Morphing_GetSupportedTypes", pfun_morphing_GetSupportedTypes); LIBXMOUNT_LOAD_SYMBOL("LibXmount_Morphing_GetFunctions", pfun_morphing_GetFunctions); // Construct new entry for our library list XMOUNT_MALLOC(p_morphing_lib,pts_MorphingLib,sizeof(ts_MorphingLib)); // Initialize lib_functions structure to NULL memset(&(p_morphing_lib->lib_functions), 0, sizeof(ts_LibXmountMorphingFunctions)); // Set name and handle XMOUNT_STRSET(p_morphing_lib->p_name,p_dirent->d_name); p_morphing_lib->p_lib=p_libxmount; // Get and set supported types p_supported_formats=pfun_morphing_GetSupportedTypes(); supported_formats_len=0; p_buf=p_supported_formats; while(*p_buf!='\0') { supported_formats_len+=(strlen(p_buf)+1); p_buf+=(strlen(p_buf)+1); } supported_formats_len++; XMOUNT_MALLOC(p_morphing_lib->p_supported_morphing_types, char*, supported_formats_len); memcpy(p_morphing_lib->p_supported_morphing_types, p_supported_formats, supported_formats_len); // Get, set and check lib_functions pfun_morphing_GetFunctions(&(p_morphing_lib->lib_functions)); if(p_morphing_lib->lib_functions.CreateHandle==NULL || p_morphing_lib->lib_functions.DestroyHandle==NULL || p_morphing_lib->lib_functions.Morph==NULL || p_morphing_lib->lib_functions.Size==NULL || p_morphing_lib->lib_functions.Read==NULL || p_morphing_lib->lib_functions.OptionsHelp==NULL || p_morphing_lib->lib_functions.OptionsParse==NULL || p_morphing_lib->lib_functions.GetInfofileContent==NULL || p_morphing_lib->lib_functions.GetErrorMessage==NULL || p_morphing_lib->lib_functions.FreeBuffer==NULL) { LOG_DEBUG("Missing implemention of one or more functions in lib %s!\n", p_dirent->d_name); free(p_morphing_lib->p_supported_morphing_types); free(p_morphing_lib->p_name); free(p_morphing_lib); dlclose(p_libxmount); continue; } // Add entry to the input library list XMOUNT_REALLOC(glob_xmount.morphing.pp_libs, pts_MorphingLib*, sizeof(pts_MorphingLib)* (glob_xmount.morphing.libs_count+1)); glob_xmount.morphing.pp_libs[glob_xmount.morphing.libs_count++]= p_morphing_lib; LOG_DEBUG("Morphing library '%s' loaded successfully\n",p_dirent->d_name); } else { LOG_DEBUG("Ignoring '%s'.\n",p_dirent->d_name); continue; } } #undef LIBXMOUNT_LOAD_SYMBOL #undef LIBXMOUNT_LOAD LOG_DEBUG("A total of %u input libs and %u morphing libs were loaded.\n", glob_xmount.input.libs_count, glob_xmount.morphing.libs_count); free(p_library_path); closedir(p_dir); return ((glob_xmount.input.libs_count>0 && glob_xmount.morphing.libs_count>0) ? TRUE : FALSE); } //! Search an appropriate input lib for specified input type /*! * \param p_input_image Input image to search input lib for * \return TRUE on success, FALSE on error */ static int FindInputLib(pts_InputImage p_input_image) { char *p_buf; LOG_DEBUG("Trying to find suitable library for input type '%s'.\n", p_input_image->p_type); // Loop over all loaded libs for(uint32_t i=0;ip_name); p_buf=glob_xmount.input.pp_libs[i]->p_supported_input_types; while(*p_buf!='\0') { if(strcmp(p_buf,p_input_image->p_type)==0) { // Library supports input type, set lib functions LOG_DEBUG("Input library '%s' pretends to handle that input type.\n", glob_xmount.input.pp_libs[i]->p_name); p_input_image->p_functions= &(glob_xmount.input.pp_libs[i]->lib_functions); return TRUE; } p_buf+=(strlen(p_buf)+1); } } LOG_DEBUG("Couldn't find any suitable library.\n"); // No library supporting input type found return FALSE; } //! Search an appropriate morphing lib for the specified morph type /*! * \return TRUE on success, FALSE on error */ static int FindMorphingLib() { char *p_buf; LOG_DEBUG("Trying to find suitable library for morph type '%s'.\n", glob_xmount.morphing.p_morph_type); // Loop over all loaded libs for(uint32_t i=0;ip_name); p_buf=glob_xmount.morphing.pp_libs[i]->p_supported_morphing_types; while(*p_buf!='\0') { if(strcmp(p_buf,glob_xmount.morphing.p_morph_type)==0) { // Library supports morph type, set lib functions LOG_DEBUG("Morphing library '%s' pretends to handle that morph type.\n", glob_xmount.morphing.pp_libs[i]->p_name); glob_xmount.morphing.p_functions= &(glob_xmount.morphing.pp_libs[i]->lib_functions); return TRUE; } p_buf+=(strlen(p_buf)+1); } } LOG_DEBUG("Couldn't find any suitable library.\n"); // No library supporting input type found return FALSE; } static void InitResources() { // Input glob_xmount.input.libs_count=0; glob_xmount.input.pp_libs=NULL; glob_xmount.input.lib_params_count=0; glob_xmount.input.pp_lib_params=NULL; glob_xmount.input.images_count=0; glob_xmount.input.pp_images=NULL; glob_xmount.input.image_offset=0; glob_xmount.input.image_size_limit=0; glob_xmount.input.image_hash_lo=0; glob_xmount.input.image_hash_hi=0; // Morphing glob_xmount.morphing.libs_count=0; glob_xmount.morphing.pp_libs=NULL; glob_xmount.morphing.p_morph_type=NULL; glob_xmount.morphing.lib_params_count=0; glob_xmount.morphing.pp_lib_params=NULL; glob_xmount.morphing.p_handle=NULL; glob_xmount.morphing.p_functions=NULL; glob_xmount.morphing.input_image_functions.ImageCount= &LibXmount_Morphing_ImageCount; glob_xmount.morphing.input_image_functions.Size=&LibXmount_Morphing_Size; glob_xmount.morphing.input_image_functions.Read=&LibXmount_Morphing_Read; // Cache + glob_xmount.cache.overwrite_cache=FALSE; glob_xmount.cache.p_cache_file=NULL; + glob_xmount.cache.h_cache_file=NULL; + glob_xmount.cache.h_block_cache=NULL; + glob_xmount.cache.h_block_cache_index=NULL; + // TODO: Remove glob_xmount.cache.h_old_cache_file=NULL; glob_xmount.cache.p_cache_header=NULL; glob_xmount.cache.p_cache_blkidx=NULL; - glob_xmount.cache.overwrite_cache=FALSE; // Output #ifndef __APPLE__ glob_xmount.output.VirtImageType=VirtImageType_DD; #else glob_xmount.output.VirtImageType=VirtImageType_DMG; #endif glob_xmount.output.image_size=0; glob_xmount.output.writable=FALSE; glob_xmount.output.p_virtual_image_path=NULL; glob_xmount.output.p_info_path=NULL; glob_xmount.output.p_info_file=NULL; glob_xmount.output.vdi.vdi_header_size=0; glob_xmount.output.vdi.p_vdi_header=NULL; glob_xmount.output.vdi.vdi_block_map_size=0; glob_xmount.output.vdi.p_vdi_block_map=NULL; glob_xmount.output.vhd.p_vhd_header=NULL; glob_xmount.output.vmdk.p_virtual_vmdk_path=NULL; glob_xmount.output.vmdk.p_vmdk_file=NULL; glob_xmount.output.vmdk.vmdk_file_size=0; glob_xmount.output.vmdk.p_vmdk_lockdir1=NULL; glob_xmount.output.vmdk.p_vmdk_lockdir2=NULL; glob_xmount.output.vmdk.p_vmdk_lockfile_data=NULL; glob_xmount.output.vmdk.vmdk_lockfile_size=0; glob_xmount.output.vmdk.p_vmdk_lockfile_name=NULL; // Misc data glob_xmount.debug=FALSE; glob_xmount.may_set_fuse_allow_other=FALSE; glob_xmount.fuse_argc=0; glob_xmount.pp_fuse_argv=NULL; glob_xmount.p_mountpoint=NULL; } /* * FreeResources */ static void FreeResources() { int ret; + teGidaFsError gidafs_ret=eGidaFsError_None; LOG_DEBUG("Freeing all resources\n"); // Misc if(glob_xmount.pp_fuse_argv!=NULL) { for(int i=0;i DestroyHandle(&(glob_xmount.morphing.p_handle)); if(ret!=0) { LOG_ERROR("Unable to destroy morphing handle: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); } } } if(glob_xmount.morphing.pp_lib_params!=NULL) { for(uint32_t i=0;ip_supported_morphing_types!=NULL) free(glob_xmount.morphing.pp_libs[i]->p_supported_morphing_types); if(glob_xmount.morphing.pp_libs[i]->p_lib!=NULL) dlclose(glob_xmount.morphing.pp_libs[i]->p_lib); if(glob_xmount.morphing.pp_libs[i]->p_name!=NULL) free(glob_xmount.morphing.pp_libs[i]->p_name); free(glob_xmount.morphing.pp_libs[i]); } free(glob_xmount.morphing.pp_libs); } // Input if(glob_xmount.input.pp_images!=NULL) { // Close all input images for(uint64_t i=0;ip_functions!=NULL) { if(glob_xmount.input.pp_images[i]->p_handle!=NULL) { ret=glob_xmount.input.pp_images[i]->p_functions-> Close(glob_xmount.input.pp_images[i]->p_handle); if(ret!=0) { LOG_ERROR("Unable to close input image: %s\n", glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); } ret=glob_xmount.input.pp_images[i]->p_functions-> DestroyHandle(&(glob_xmount.input.pp_images[i]->p_handle)); if(ret!=0) { LOG_ERROR("Unable to destroy input image handle: %s\n", glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); } } } if(glob_xmount.input.pp_images[i]->pp_files!=NULL) { for(uint64_t ii=0;iifiles_count;ii++) { if(glob_xmount.input.pp_images[i]->pp_files[ii]!=NULL) free(glob_xmount.input.pp_images[i]->pp_files[ii]); } free(glob_xmount.input.pp_images[i]->pp_files); } if(glob_xmount.input.pp_images[i]->p_type!=NULL) free(glob_xmount.input.pp_images[i]->p_type); free(glob_xmount.input.pp_images[i]); } free(glob_xmount.input.pp_images); } if(glob_xmount.input.pp_lib_params!=NULL) { for(uint32_t i=0;ip_supported_input_types!=NULL) free(glob_xmount.input.pp_libs[i]->p_supported_input_types); if(glob_xmount.input.pp_libs[i]->p_lib!=NULL) dlclose(glob_xmount.input.pp_libs[i]->p_lib); if(glob_xmount.input.pp_libs[i]->p_name!=NULL) free(glob_xmount.input.pp_libs[i]->p_name); free(glob_xmount.input.pp_libs[i]); } free(glob_xmount.input.pp_libs); } // Before we return, initialize everything in case ReleaseResources would be // called again. InitResources(); } //! Function to split given library options static int SplitLibraryParameters(char *p_params, uint32_t *p_ret_opts_count, pts_LibXmountOptions **ppp_ret_opt) { pts_LibXmountOptions p_opts=NULL; pts_LibXmountOptions *pp_opts=NULL; uint32_t params_len; uint32_t opts_count=0; uint32_t sep_pos=0; char *p_buf=p_params; if(p_params==NULL) return FALSE; // Get params length params_len=strlen(p_params); // Return if no params specified if(params_len==0) { *ppp_ret_opt=NULL; p_ret_opts_count=0; return TRUE; } // Split params while(*p_buf!='\0') { XMOUNT_MALLOC(p_opts,pts_LibXmountOptions,sizeof(ts_LibXmountOptions)); p_opts->valid=0; #define FREE_PP_OPTS() { \ if(pp_opts!=NULL) { \ for(uint32_t i=0;ip_key,p_buf,sep_pos); p_buf+=(sep_pos+1); // Search next separator sep_pos=0; while(p_buf[sep_pos]!='\0' && p_buf[sep_pos]!=',') sep_pos++; if(sep_pos==0) { LOG_ERROR("Library parameter '%s' is not of format key=value!\n", p_opts->p_key); free(p_opts->p_key); free(p_opts); FREE_PP_OPTS(); return FALSE; } // Save option value XMOUNT_STRNSET(p_opts->p_value,p_buf,sep_pos); p_buf+=sep_pos; LOG_DEBUG("Extracted library option: '%s' = '%s'\n", p_opts->p_key, p_opts->p_value); #undef FREE_PP_OPTS // Add current option to return array XMOUNT_REALLOC(pp_opts, pts_LibXmountOptions*, sizeof(pts_LibXmountOptions)*(opts_count+1)); pp_opts[opts_count++]=p_opts; // If we're not at the end of p_params, skip over separator for next run if(*p_buf!='\0') p_buf++; } LOG_DEBUG("Extracted a total of %" PRIu32 " library options\n",opts_count); *p_ret_opts_count=opts_count; *ppp_ret_opt=pp_opts; return TRUE; } /******************************************************************************* * LibXmount_Morphing function implementation ******************************************************************************/ //! Function to get the amount of input images /*! * \param p_count Count of input images * \return 0 on success */ static int LibXmount_Morphing_ImageCount(uint64_t *p_count) { *p_count=glob_xmount.input.images_count; return 0; } //! Function to get the size of the morphed data /*! * \param image Image number * \param p_size Pointer to store input image's size to * \return 0 on success */ static int LibXmount_Morphing_Size(uint64_t image, uint64_t *p_size) { if(image>=glob_xmount.input.images_count) return -1; *p_size=glob_xmount.input.pp_images[image]->size; return 0; } //! Function to read data from input image /*! * \param image Image number * \param p_buf Buffer to store read data to * \param offset Position at which to start reading * \param count Amount of bytes to read * \param p_read Number of read bytes on success * \return 0 on success or negated error code on error */ static int LibXmount_Morphing_Read(uint64_t image, char *p_buf, off_t offset, size_t count, size_t *p_read) { if(image>=glob_xmount.input.images_count) return -EIO; return GetInputImageData(glob_xmount.input.pp_images[image], p_buf, offset, count, p_read); } /******************************************************************************* * FUSE function implementation ******************************************************************************/ //! FUSE access implementation /*! * \param p_path Path of file to get attributes from * \param perm Requested permissisons * \return 0 on success, negated error code on error */ /* static int FuseAccess(const char *path, int perm) { // TODO: Implement propper file permission handling // http://www.cs.cf.ac.uk/Dave/C/node20.html // Values for the second argument to access. // These may be OR'd together. //#define R_OK 4 // Test for read permission. //#define W_OK 2 // Test for write permission. //#define X_OK 1 // Test for execute permission. //#define F_OK 0 // Test for existence. return 0; } */ //! FUSE getattr implementation /*! * \param p_path Path of file to get attributes from * \param p_stat Pointer to stat structure to save attributes to * \return 0 on success, negated error code on error */ static int FuseGetAttr(const char *p_path, struct stat *p_stat) { memset(p_stat,0,sizeof(struct stat)); if(strcmp(p_path,"/")==0) { // Attributes of mountpoint p_stat->st_mode=S_IFDIR | 0777; p_stat->st_nlink=2; } else if(strcmp(p_path,glob_xmount.output.p_virtual_image_path)==0) { // Attributes of virtual image if(!glob_xmount.output.writable) p_stat->st_mode=S_IFREG | 0444; else p_stat->st_mode=S_IFREG | 0666; p_stat->st_nlink=1; // Get virtual image file size if(!GetVirtImageSize((uint64_t*)&(p_stat->st_size))) { LOG_ERROR("Couldn't get image size!\n"); return -ENOENT; } if(glob_xmount.output.VirtImageType==VirtImageType_VHD) { // Make sure virtual image seems to be fully allocated (not sparse file). // Without this, Windows won't attach the vhd file! p_stat->st_blocks=p_stat->st_size/512; if(p_stat->st_size%512!=0) p_stat->st_blocks++; } } else if(strcmp(p_path,glob_xmount.output.p_info_path)==0) { // Attributes of virtual image info file p_stat->st_mode=S_IFREG | 0444; p_stat->st_nlink=1; // Get virtual image info file size if(glob_xmount.output.p_info_file!=NULL) { p_stat->st_size=strlen(glob_xmount.output.p_info_file); } else p_stat->st_size=0; } else if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { // Some special files only present when emulating VMDK files if(strcmp(p_path,glob_xmount.output.vmdk.p_virtual_vmdk_path)==0) { // Attributes of virtual vmdk file if(!glob_xmount.output.writable) p_stat->st_mode=S_IFREG | 0444; else p_stat->st_mode=S_IFREG | 0666; p_stat->st_nlink=1; // Get virtual image info file size if(glob_xmount.output.vmdk.p_vmdk_file!=NULL) { p_stat->st_size=glob_xmount.output.vmdk.vmdk_file_size; } else p_stat->st_size=0; } else if(glob_xmount.output.vmdk.p_vmdk_lockdir1!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir1)==0) { p_stat->st_mode=S_IFDIR | 0777; p_stat->st_nlink=2; } else if(glob_xmount.output.vmdk.p_vmdk_lockdir2!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir2)==0) { p_stat->st_mode=S_IFDIR | 0777; p_stat->st_nlink=2; } else if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0) { p_stat->st_mode=S_IFREG | 0666; if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL) { p_stat->st_size=strlen(glob_xmount.output.vmdk.p_vmdk_lockfile_name); } else p_stat->st_size=0; } else return -ENOENT; } else return -ENOENT; // Set uid and gid of all files to uid and gid of current process p_stat->st_uid=getuid(); p_stat->st_gid=getgid(); return 0; } //! FUSE mkdir implementation /*! * \param p_path Directory path * \param mode Directory permissions * \return 0 on success, negated error code on error */ static int FuseMkDir(const char *p_path, mode_t mode) { // Only allow creation of VMWare's lock directories if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { if(glob_xmount.output.vmdk.p_vmdk_lockdir1==NULL) { char aVmdkLockDir[strlen(glob_xmount.output.vmdk.p_virtual_vmdk_path)+5]; sprintf(aVmdkLockDir, "%s.lck", glob_xmount.output.vmdk.p_virtual_vmdk_path); if(strcmp(p_path,aVmdkLockDir)==0) { LOG_DEBUG("Creating virtual directory \"%s\"\n",aVmdkLockDir) XMOUNT_STRSET(glob_xmount.output.vmdk.p_vmdk_lockdir1,aVmdkLockDir) return 0; } else { LOG_ERROR("Attempt to create illegal directory \"%s\"!\n",p_path) LOG_DEBUG("Supposed: %s\n",aVmdkLockDir) return -1; } } else if(glob_xmount.output.vmdk.p_vmdk_lockdir2==NULL && strncmp(p_path, glob_xmount.output.vmdk.p_vmdk_lockdir1, strlen(glob_xmount.output.vmdk.p_vmdk_lockdir1))==0) { LOG_DEBUG("Creating virtual directory \"%s\"\n",p_path) XMOUNT_STRSET(glob_xmount.output.vmdk.p_vmdk_lockdir2,p_path) return 0; } else { LOG_ERROR("Attempt to create illegal directory \"%s\"!\n",p_path) LOG_DEBUG("Compared to first %u chars of \"%s\"\n", strlen(glob_xmount.output.vmdk.p_vmdk_lockdir1), glob_xmount.output.vmdk.p_vmdk_lockdir1) return -1; } } LOG_ERROR("Attempt to create directory \"%s\" " "on read-only filesystem!\n",p_path) return -1; } //! FUSE create implementation. /*! * Currently only allows to create VMWare's lock file * * \param p_path File to create * \param mode File mode * \param dev ??? but not used * \return 0 on success, negated error code on error */ static int FuseMkNod(const char *p_path, mode_t mode, dev_t dev) { if((glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) && glob_xmount.output.vmdk.p_vmdk_lockdir1!=NULL && glob_xmount.output.vmdk.p_vmdk_lockfile_name==NULL) { LOG_DEBUG("Creating virtual file \"%s\"\n",p_path) XMOUNT_STRSET(glob_xmount.output.vmdk.p_vmdk_lockfile_name,p_path); return 0; } else { LOG_ERROR("Attempt to create illegal file \"%s\"\n",p_path) return -1; } } //! FUSE readdir implementation /*! * \param p_path Path from where files should be listed * \param p_buf Buffer to write file entrys to * \param filler Function to write dir entrys to buffer * \param offset ??? but not used * \param p_fi File info struct * \return 0 on success, negated error code on error */ static int FuseReadDir(const char *p_path, void *p_buf, fuse_fill_dir_t filler, off_t offset, struct fuse_file_info *p_fi) { // Ignore some params (void)offset; (void)p_fi; if(strcmp(p_path,"/")==0) { // Add std . and .. entrys filler(p_buf,".",NULL,0); filler(p_buf,"..",NULL,0); // Add our virtual files (p+1 to ignore starting "/") filler(p_buf,glob_xmount.output.p_virtual_image_path+1,NULL,0); filler(p_buf,glob_xmount.output.p_info_path+1,NULL,0); if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { // For VMDK's, we use an additional descriptor file filler(p_buf,glob_xmount.output.vmdk.p_virtual_vmdk_path+1,NULL,0); // And there could also be a lock directory if(glob_xmount.output.vmdk.p_vmdk_lockdir1!=NULL) { filler(p_buf,glob_xmount.output.vmdk.p_vmdk_lockdir1+1,NULL,0); } } } else if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { // For VMDK emulation, there could be a lock directory if(glob_xmount.output.vmdk.p_vmdk_lockdir1!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir1)==0) { filler(p_buf,".",NULL,0); filler(p_buf,"..",NULL,0); if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL) { filler(p_buf, glob_xmount.output.vmdk.p_vmdk_lockfile_name+ strlen(glob_xmount.output.vmdk.p_vmdk_lockdir1)+1, NULL, 0); } } else if(glob_xmount.output.vmdk.p_vmdk_lockdir2!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir2)==0) { filler(p_buf,".",NULL,0); filler(p_buf,"..",NULL,0); } else return -ENOENT; } else return -ENOENT; return 0; } //! FUSE open implementation /*! * \param p_path Path to file to open * \param p_fi File info struct * \return 0 on success, negated error code on error */ static int FuseOpen(const char *p_path, struct fuse_file_info *p_fi) { #define CHECK_OPEN_PERMS() { \ if(!glob_xmount.output.writable && (p_fi->flags & 3)!=O_RDONLY) { \ LOG_DEBUG("Attempt to open the read-only file \"%s\" for writing.\n", \ p_path) \ return -EACCES; \ } \ return 0; \ } if(strcmp(p_path,glob_xmount.output.p_virtual_image_path)==0 || strcmp(p_path,glob_xmount.output.p_info_path)==0) { CHECK_OPEN_PERMS(); } else if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { if(strcmp(p_path,glob_xmount.output.vmdk.p_virtual_vmdk_path)==0 || (glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0)) { CHECK_OPEN_PERMS(); } } #undef CHECK_OPEN_PERMS LOG_DEBUG("Attempt to open inexistant file \"%s\".\n",p_path); return -ENOENT; } //! FUSE read implementation /*! * \param p_path Path (relative to mount folder) of file to read data from * \param p_buf Pre-allocated buffer where read data should be written to * \param size Number of bytes to read * \param offset Offset to start reading at * \param p_fi: File info struct * \return Read bytes on success, negated error code on error */ static int FuseRead(const char *p_path, char *p_buf, size_t size, off_t offset, struct fuse_file_info *p_fi) { (void)p_fi; int ret; uint64_t len; #define READ_MEM_FILE(filebuf,filesize,filetypestr,mutex) { \ len=filesize; \ if(offsetlen) { \ LOG_DEBUG("Attempt to read past EOF of virtual " filetypestr " file\n"); \ LOG_DEBUG("Adjusting read size from %u to %u\n",size,len-offset); \ size=len-offset; \ } \ pthread_mutex_lock(&mutex); \ memcpy(p_buf,filebuf+offset,size); \ pthread_mutex_unlock(&mutex); \ LOG_DEBUG("Read %" PRIu64 " bytes at offset %" PRIu64 \ " from virtual " filetypestr " file\n",size,offset); \ ret=size; \ } else { \ LOG_DEBUG("Attempt to read behind EOF of virtual " filetypestr " file\n"); \ ret=0; \ } \ } if(strcmp(p_path,glob_xmount.output.p_virtual_image_path)==0) { // Read data from virtual output file // Wait for other threads to end reading/writing data pthread_mutex_lock(&(glob_xmount.mutex_image_rw)); // Get requested data if((ret=GetVirtImageData(p_buf,offset,size))<0) { LOG_ERROR("Couldn't read data from virtual image file!\n") } // Allow other threads to read/write data again pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); } else if(strcmp(p_path,glob_xmount.output.p_info_path)==0) { // Read data from virtual info file READ_MEM_FILE(glob_xmount.output.p_info_file, strlen(glob_xmount.output.p_info_file), "info", glob_xmount.mutex_info_read); } else if(strcmp(p_path,glob_xmount.output.vmdk.p_virtual_vmdk_path)==0) { // Read data from virtual vmdk file READ_MEM_FILE(glob_xmount.output.vmdk.p_vmdk_file, glob_xmount.output.vmdk.vmdk_file_size, "vmdk", glob_xmount.mutex_image_rw); } else if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0) { // Read data from virtual lock file READ_MEM_FILE(glob_xmount.output.vmdk.p_vmdk_lockfile_data, glob_xmount.output.vmdk.vmdk_lockfile_size, "vmdk lock", glob_xmount.mutex_image_rw); } else { // Attempt to read non existant file LOG_DEBUG("Attempt to read from non existant file \"%s\"\n",p_path) ret=-ENOENT; } #undef READ_MEM_FILE // TODO: Return size of read data!!!!! return ret; } //! FUSE rename implementation /*! * \param p_path File to rename * \param p_npath New filename * \return 0 on error, negated error code on error */ static int FuseRename(const char *p_path, const char *p_npath) { if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0) { LOG_DEBUG("Renaming virtual lock file from \"%s\" to \"%s\"\n", glob_xmount.output.vmdk.p_vmdk_lockfile_name, p_npath) XMOUNT_REALLOC(glob_xmount.output.vmdk.p_vmdk_lockfile_name,char*, (strlen(p_npath)+1)*sizeof(char)); strcpy(glob_xmount.output.vmdk.p_vmdk_lockfile_name,p_npath); return 0; } } return -ENOENT; } //! FUSE rmdir implementation /*! * \param p_path Directory to delete * \return 0 on success, negated error code on error */ static int FuseRmDir(const char *p_path) { // Only VMWare's lock directories can be deleted if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { if(glob_xmount.output.vmdk.p_vmdk_lockdir1!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir1)==0) { LOG_DEBUG("Deleting virtual lock dir \"%s\"\n", glob_xmount.output.vmdk.p_vmdk_lockdir1) free(glob_xmount.output.vmdk.p_vmdk_lockdir1); glob_xmount.output.vmdk.p_vmdk_lockdir1=NULL; return 0; } else if(glob_xmount.output.vmdk.p_vmdk_lockdir2!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockdir2)==0) { LOG_DEBUG("Deleting virtual lock dir \"%s\"\n", glob_xmount.output.vmdk.p_vmdk_lockdir1) free(glob_xmount.output.vmdk.p_vmdk_lockdir2); glob_xmount.output.vmdk.p_vmdk_lockdir2=NULL; return 0; } } return -1; } //! FUSE unlink implementation /*! * \param p_path File to delete * \return 0 on success, negated error code on error */ static int FuseUnlink(const char *p_path) { // Only VMWare's lock file can be deleted if(glob_xmount.output.VirtImageType==VirtImageType_VMDK || glob_xmount.output.VirtImageType==VirtImageType_VMDKS) { if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0) { LOG_DEBUG("Deleting virtual file \"%s\"\n", glob_xmount.output.vmdk.p_vmdk_lockfile_name) free(glob_xmount.output.vmdk.p_vmdk_lockfile_name); free(glob_xmount.output.vmdk.p_vmdk_lockfile_data); glob_xmount.output.vmdk.p_vmdk_lockfile_name=NULL; glob_xmount.output.vmdk.p_vmdk_lockfile_data=NULL; glob_xmount.output.vmdk.vmdk_lockfile_size=0; return 0; } } return -1; } //! FUSE statfs implementation /*! * \param p_path Get stats for fs that the specified file resides in * \param stats Stats * \return 0 on success, negated error code on error */ /* static int FuseStatFs(const char *p_path, struct statvfs *stats) { struct statvfs CacheFileFsStats; int ret; if(glob_xmount.writable==TRUE) { // If write support is enabled, return stats of fs upon which cache file // resides in if((ret=statvfs(glob_xmount.p_cache_file,&CacheFileFsStats))==0) { memcpy(stats,&CacheFileFsStats,sizeof(struct statvfs)); return 0; } else { LOG_ERROR("Couldn't get stats for fs upon which resides \"%s\"\n", glob_xmount.p_cache_file) return ret; } } else { // TODO: Return read only return 0; } } */ // FUSE write implementation /*! * \param p_buf Buffer containing data to write * \param size Number of bytes to write * \param offset Offset to start writing at * \param p_fi: File info struct * * Returns: * Written bytes on success, negated error code on error */ static int FuseWrite(const char *p_path, const char *p_buf, size_t size, off_t offset, struct fuse_file_info *p_fi) { (void)p_fi; uint64_t len; if(strcmp(p_path,glob_xmount.output.p_virtual_image_path)==0) { // Wait for other threads to end reading/writing data pthread_mutex_lock(&(glob_xmount.mutex_image_rw)); // Get virtual image file size if(!GetVirtImageSize(&len)) { LOG_ERROR("Couldn't get virtual image size!\n") pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); return 0; } if(offsetlen) size=len-offset; if(SetVirtImageData(p_buf,offset,size)!=size) { LOG_ERROR("Couldn't write data to virtual image file!\n") pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); return 0; } } else { LOG_DEBUG("Attempt to write past EOF of virtual image file\n") pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); return 0; } // Allow other threads to read/write data again pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); } else if(strcmp(p_path,glob_xmount.output.vmdk.p_virtual_vmdk_path)==0) { pthread_mutex_lock(&(glob_xmount.mutex_image_rw)); len=glob_xmount.output.vmdk.vmdk_file_size; if((offset+size)>len) { // Enlarge or create buffer if needed if(len==0) { len=offset+size; XMOUNT_MALLOC(glob_xmount.output.vmdk.p_vmdk_file, char*, len*sizeof(char)); } else { len=offset+size; XMOUNT_REALLOC(glob_xmount.output.vmdk.p_vmdk_file, char*, len*sizeof(char)); } glob_xmount.output.vmdk.vmdk_file_size=offset+size; } // Copy data to buffer memcpy(glob_xmount.output.vmdk.p_vmdk_file+offset,p_buf,size); pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); } else if(glob_xmount.output.vmdk.p_vmdk_lockfile_name!=NULL && strcmp(p_path,glob_xmount.output.vmdk.p_vmdk_lockfile_name)==0) { pthread_mutex_lock(&(glob_xmount.mutex_image_rw)); if((offset+size)>glob_xmount.output.vmdk.vmdk_lockfile_size) { // Enlarge or create buffer if needed if(glob_xmount.output.vmdk.vmdk_lockfile_size==0) { glob_xmount.output.vmdk.vmdk_lockfile_size=offset+size; XMOUNT_MALLOC(glob_xmount.output.vmdk.p_vmdk_lockfile_data,char*, glob_xmount.output.vmdk.vmdk_lockfile_size*sizeof(char)) } else { glob_xmount.output.vmdk.vmdk_lockfile_size=offset+size; XMOUNT_REALLOC(glob_xmount.output.vmdk.p_vmdk_lockfile_data,char*, glob_xmount.output.vmdk.vmdk_lockfile_size*sizeof(char)) } } // Copy data to buffer memcpy(glob_xmount.output.vmdk.p_vmdk_lockfile_data+offset,p_buf,size); pthread_mutex_unlock(&(glob_xmount.mutex_image_rw)); } else if(strcmp(p_path,glob_xmount.output.p_info_path)==0) { // Attempt to write data to read only image info file LOG_DEBUG("Attempt to write data to virtual info file\n"); return -ENOENT; } else { // Attempt to write to non existant file LOG_DEBUG("Attempt to write to the non existant file \"%s\"\n",p_path) return -ENOENT; } return size; } /******************************************************************************* * Main ******************************************************************************/ int main(int argc, char *argv[]) { struct stat file_stat; int ret; int fuse_ret; char *p_err_msg; // Set implemented FUSE functions struct fuse_operations xmount_operations = { //.access=FuseAccess, .getattr=FuseGetAttr, .mkdir=FuseMkDir, .mknod=FuseMkNod, .open=FuseOpen, .readdir=FuseReadDir, .read=FuseRead, .rename=FuseRename, .rmdir=FuseRmDir, //.statfs=FuseStatFs, .unlink=FuseUnlink, .write=FuseWrite }; // Disable std output / input buffering setbuf(stdout,NULL); setbuf(stderr,NULL); // Init glob_xmount InitResources(); // Load input and morphing libs if(!LoadLibs()) { LOG_ERROR("Unable to load any libraries!\n") return 1; } // Check FUSE settings CheckFuseSettings(); // Parse command line options if(ParseCmdLine(argc,argv)!=TRUE) { PrintUsage(argv[0]); FreeResources(); return 1; } // Check command line options if(glob_xmount.input.images_count==0) { LOG_ERROR("No --in command line option specified!\n") PrintUsage(argv[0]); FreeResources(); return 1; } if(glob_xmount.fuse_argc<2) { LOG_ERROR("Couldn't parse command line options!\n") PrintUsage(argv[0]); FreeResources(); return 1; } if(glob_xmount.morphing.p_morph_type==NULL) { XMOUNT_STRSET(glob_xmount.morphing.p_morph_type,"combine"); } // Check if mountpoint is a valid dir if(stat(glob_xmount.p_mountpoint,&file_stat)!=0) { LOG_ERROR("Unable to stat mount point '%s'!\n",glob_xmount.p_mountpoint); PrintUsage(argv[0]); FreeResources(); return 1; } if(!S_ISDIR(file_stat.st_mode)) { LOG_ERROR("Mount point '%s' is not a directory!\n", glob_xmount.p_mountpoint); PrintUsage(argv[0]); FreeResources(); return 1; } if(glob_xmount.debug==TRUE) { LOG_DEBUG("Options passed to FUSE: ") for(int i=0;ifiles_count==1) { LOG_DEBUG("Loading image file \"%s\"...\n", glob_xmount.input.pp_images[i]->pp_files[0]) } else { LOG_DEBUG("Loading image files \"%s .. %s\"...\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]-> pp_files[glob_xmount.input.pp_images[i]->files_count-1]) } } // Find input lib if(!FindInputLib(glob_xmount.input.pp_images[i])) { LOG_ERROR("Unknown input image type '%s' for input image '%s'!\n", glob_xmount.input.pp_images[i]->p_type, glob_xmount.input.pp_images[i]->pp_files[0]) PrintUsage(argv[0]); FreeResources(); return 1; } // Init input image handle ret=glob_xmount.input.pp_images[i]->p_functions-> CreateHandle(&(glob_xmount.input.pp_images[i]->p_handle), glob_xmount.input.pp_images[i]->p_type, glob_xmount.debug); if(ret!=0) { LOG_ERROR("Unable to init input handle for input image '%s': %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); FreeResources(); return 1; } // Parse input lib specific options if(glob_xmount.input.pp_lib_params!=NULL) { ret=glob_xmount.input.pp_images[i]->p_functions-> OptionsParse(glob_xmount.input.pp_images[i]->p_handle, glob_xmount.input.lib_params_count, glob_xmount.input.pp_lib_params, (const char**)&p_err_msg); if(ret!=0) { if(p_err_msg!=NULL) { LOG_ERROR("Unable to parse input library specific options for image " "'%s': %s: %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret), p_err_msg); glob_xmount.input.pp_images[i]->p_functions->FreeBuffer(p_err_msg); FreeResources(); return 1; } else { LOG_ERROR("Unable to parse input library specific options for image " "'%s': %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); FreeResources(); return 1; } } } // Open input image ret= glob_xmount.input.pp_images[i]-> p_functions-> Open(glob_xmount.input.pp_images[i]->p_handle, (const char**)(glob_xmount.input.pp_images[i]->pp_files), glob_xmount.input.pp_images[i]->files_count); if(ret!=0) { LOG_ERROR("Unable to open input image file '%s': %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]->p_functions-> GetErrorMessage(ret)); FreeResources(); return 1; } // Determine input image size ret=glob_xmount.input.pp_images[i]-> p_functions-> Size(glob_xmount.input.pp_images[i]->p_handle, &(glob_xmount.input.pp_images[i]->size)); if(ret!=0) { LOG_ERROR("Unable to determine size of input image '%s': %s!\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.pp_images[i]-> p_functions->GetErrorMessage(ret)); FreeResources(); return 1; } // If an offset was specified, check it against offset and change size if(glob_xmount.input.image_offset!=0) { if(glob_xmount.input.image_offset>glob_xmount.input.pp_images[i]->size) { LOG_ERROR("The specified offset is larger than the size of the input " "image '%s'! (%" PRIu64 " > %" PRIu64 ")\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.image_offset, glob_xmount.input.pp_images[i]->size); FreeResources(); return 1; } glob_xmount.input.pp_images[i]->size-=glob_xmount.input.image_offset; } // If a size limit was specified, check it and change size if(glob_xmount.input.image_size_limit!=0) { if(glob_xmount.input.pp_images[i]->size< glob_xmount.input.image_size_limit) { LOG_ERROR("The specified size limit is larger than the size of the " "input image '%s'! (%" PRIu64 " > %" PRIu64 ")\n", glob_xmount.input.pp_images[i]->pp_files[0], glob_xmount.input.image_size_limit, glob_xmount.input.pp_images[i]->size); FreeResources(); return 1; } glob_xmount.input.pp_images[i]->size=glob_xmount.input.image_size_limit; } LOG_DEBUG("Input image loaded successfully\n") } // Find morphing lib if(FindMorphingLib()!=TRUE) { LOG_ERROR("Unable to find a library supporting the morphing type '%s'!\n", glob_xmount.morphing.p_morph_type); FreeResources(); return 1; } // Init morphing ret=glob_xmount.morphing.p_functions-> CreateHandle(&glob_xmount.morphing.p_handle, glob_xmount.morphing.p_morph_type, glob_xmount.debug); if(ret!=0) { LOG_ERROR("Unable to create morphing handle: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); FreeResources(); return 1; } // Parse morphing lib specific options if(glob_xmount.morphing.pp_lib_params!=NULL) { p_err_msg=NULL; ret=glob_xmount.morphing.p_functions-> OptionsParse(glob_xmount.morphing.p_handle, glob_xmount.morphing.lib_params_count, glob_xmount.morphing.pp_lib_params, (const char**)&p_err_msg); if(ret!=0) { if(p_err_msg!=NULL) { LOG_ERROR("Unable to parse morphing library specific options: %s: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret), p_err_msg); glob_xmount.morphing.p_functions->FreeBuffer(p_err_msg); FreeResources(); return 1; } else { LOG_ERROR("Unable to parse morphing library specific options: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); FreeResources(); return 1; } } } // Morph image ret=glob_xmount.morphing.p_functions-> Morph(glob_xmount.morphing.p_handle, &(glob_xmount.morphing.input_image_functions)); if(ret!=0) { LOG_ERROR("Unable to start morphing: %s!\n", glob_xmount.morphing.p_functions->GetErrorMessage(ret)); FreeResources(); return 1; } // Init random generator srand(time(NULL)); // Calculate partial MD5 hash of input image file if(CalculateInputImageHash(&(glob_xmount.input.image_hash_lo), &(glob_xmount.input.image_hash_hi))==FALSE) { LOG_ERROR("Couldn't calculate partial hash of morphed image!\n") return 1; } if(glob_xmount.debug==TRUE) { LOG_DEBUG("Partial MD5 hash of morphed image: ") for(int i=0;i<8;i++) printf("%02hhx",*(((char*)(&(glob_xmount.input.image_hash_lo)))+i)); for(int i=0;i<8;i++) printf("%02hhx",*(((char*)(&(glob_xmount.input.image_hash_hi)))+i)); printf("\n"); } if(!ExtractVirtFileNames(glob_xmount.input.pp_images[0]->pp_files[0])) { LOG_ERROR("Couldn't extract virtual file names!\n"); FreeResources(); return 1; } LOG_DEBUG("Virtual file names extracted successfully\n") // Gather infos for info file if(!InitVirtImageInfoFile()) { LOG_ERROR("Couldn't gather infos for virtual image info file!\n") FreeResources(); return 1; } LOG_DEBUG("Virtual image info file build successfully\n") // Do some virtual image type specific initialisations switch(glob_xmount.output.VirtImageType) { case VirtImageType_DD: case VirtImageType_DMG: break; case VirtImageType_VDI: // When mounting as VDI, we need to construct a vdi header if(!InitVirtVdiHeader()) { LOG_ERROR("Couldn't initialize virtual VDI file header!\n") FreeResources(); return 1; } LOG_DEBUG("Virtual VDI file header build successfully\n") break; case VirtImageType_VHD: // When mounting as VHD, we need to construct a vhd footer if(!InitVirtVhdHeader()) { LOG_ERROR("Couldn't initialize virtual VHD file footer!\n") FreeResources(); return 1; } LOG_DEBUG("Virtual VHD file footer build successfully\n") break; case VirtImageType_VMDK: case VirtImageType_VMDKS: // When mounting as VMDK, we need to construct the VMDK descriptor file if(!InitVirtualVmdkFile()) { LOG_ERROR("Couldn't initialize virtual VMDK file!\n") FreeResources(); return 1; } break; } if(glob_xmount.output.writable) { // Init cache file and cache file block index if(!InitCacheFile()) { LOG_ERROR("Couldn't initialize cache file!\n") FreeResources(); return 1; } LOG_DEBUG("Cache file initialized successfully\n") } // Call fuse_main to do the fuse magic fuse_ret=fuse_main(glob_xmount.fuse_argc, glob_xmount.pp_fuse_argv, &xmount_operations, NULL); // Destroy mutexes pthread_mutex_destroy(&(glob_xmount.mutex_image_rw)); pthread_mutex_destroy(&(glob_xmount.mutex_info_read)); // Free allocated memory FreeResources(); return fuse_ret; } /* ----- Change log ----- 20090131: v0.1.0 released * Some minor things have still to be done. * Mounting ewf as dd: Seems to work. Diff didn't complain about changes between original dd and emulated dd. * Mounting ewf as vdi: Seems to work too. VBox accepts the emulated vdi as valid vdi file and I was able to mount the containing fs under Debian. INFO: Debian freezed when not using mount -r !! 20090203: v0.1.1 released * Multiple code improvements. For ex. cleaner vdi header allocation. * Fixed severe bug in image block calculation. Didn't check for odd input in conversion from bytes to megabytes. * Added more debug output 20090210: v0.1.2 released * Fixed compilation problem (Typo in image_init_info() function). * Fixed some problems with the debian scripts to be able to build packages. * Added random generator initialisation (Makes it possible to use more than one image in VBox at a time). 20090215: * Added function init_cache_blocks which creates / loads a cache file used to implement virtual write capability. 20090217: * Implemented the fuse write function. Did already some basic tests with dd and it seems to work. But there are certainly still some bugs left as there are also still some TODO's left. 20090226: * Changed program name from mountewf to xmount. * Began with massive code cleanups to ease full implementation of virtual write support and to be able to support multiple input image formats (DD, EWF and AFF are planned for now). * Added defines for supported input formats so it should be possible to compile xmount without supporting all input formats. (DD input images are always supported as these do not require any additional libs). Input formats should later be en/disabled by the configure script in function to which libs it detects. * GetOrigImageSize function added to get the size of the original image whatever type it is in. * GetOrigImageData function added to retrieve data from original image file whatever type it is in. * GetVirtImageSize function added to get the size of the virtual image file. * Cleaned function mountewf_getattr and renamed it to GetVirtFileAttr * Cleaned function mountewf_readdir and renamed it to GetVirtFiles * Cleaned function mountewf_open and renamed it to OpenVirtFile 20090227: * Cleaned function init_info_file and renamed it to InitVirtImageInfoFile 20090228: * Cleaned function init_cache_blocks and renamed it to InitCacheFile * Added LogMessage function to ease error and debug logging (See also LOG_ERROR and LOG_DEBUG macros in xmount.h) * Cleaned function init_vdi_header and renamed it to InitVirtVdiHeader * Added PrintUsage function to print out xmount usage informations * Cleaned function parse_cmdline and renamed it to ParseCmdLine * Cleaned function main * Added ExtractVirtFileNames function to extract virtual file names from input image name * Added function GetVirtImageData to retrieve data from the virtual image file. This includes reading data from cache file if virtual write support is enabled. * Added function ReadVirtFile to replace mountewf_read 20090229: * Fixed a typo in virtual file name creation * Added function SetVirtImageData to write data to virtual image file. This includes writing data to cache file and caching entire new blocks * Added function WriteVirtFile to replace mountewf_write 20090305: * Solved a problem that made it impossible to access offsets >32bit 20090308: * Added SetVdiFileHeaderData function to handle virtual image type specific data to be cached. This makes cache files independent from virtual image type 20090316: v0.2.0 released 20090327: v0.2.1 released * Fixed a bug in virtual write support. Checking whether data is cached didn't use semaphores. This could corrupt cache files when running multi-threaded. * Added IsVdiFileHeaderCached function to check whether VDI file header was already cached * Added IsBlockCached function to check whether a block was already cached 20090331: v0.2.2 released (Internal release) * Further changes to semaphores to fix write support bug. 20090410: v0.2.3 released * Reverted most of the fixes from v0.2.1 and v0.2.2 as those did not solve the write support bug. * Removed all semaphores * Added two pthread mutexes to protect virtual image and virtual info file. 20090508: * Configure script will now exit when needed libraries aren't found * Added support for newest libewf beta version 20090506 as it seems to reduce memory usage when working with EWF files by about 1/2. * Added LIBEWF_BETA define to adept source to new libewf API. * Added function InitVirtualVmdkFile to build a VmWare virtual disk descriptor file. 20090519: * Added function CreateVirtDir implementing FUSE's mkdir to allow VMWare to create his .vmdk.lck lock folder. Function does not allow to create other folders! * Changed cache file handling as VMDK caching will need new cache file structure incompatible to the old one. 20090522: v0.3.0 released * Added function DeleteVirtFile and DeleteVirtDir so VMWare can remove his lock directories and files. * Added function RenameVirtFile because VMWare needs to rename his lock files. * VMDK support should work now but descriptor file won't get cached as I didn't implement it yet. 20090604: * Added --cache commandline parameter doing the same as --rw. * Added --owcache commandline parameter doing the same as --rw but overwrites any existing cache data. This can be handy for debugging and testing purposes. * Added "vmdks" output type. Same as "vmdk" but generates a disk connected to the SCSI bus rather than the IDE bus. 20090710: v0.3.1 released 20090721: * Added function CheckFuseAllowOther to check wether FUSE supports the "-o allow_other" option. It is supported when "user_allow_other" is set in /etc/fuse.conf or when running xmount as root. * Automatic addition of FUSE's "-o allow_other" option if it is supported. * Added special "-o no_allow_other" command line parameter to disable automatic addition of the above option. * Reorganisation of FUSE's and xmount's command line options processing. * Added LogWarnMessage function to output a warning message. 20090722: * Added function CalculateInputImageHash to calculate an MD5 hash of the first input image's HASH_AMOUNT bytes of data. This hash is used as VDI creation UUID and will later be used to match cache files to input images. 20090724: v0.3.2 released 20090725: v0.4.0 released * Added AFF input image support. * Due to various problems with libewf and libaff packages (Mainly in Debian and Ubuntu), I decided to include them into xmount's source tree and link them in statically. This has the advantage that I can use whatever version I want. 20090727: v0.4.1 released * Added again the ability to compile xmount with shared libs as the Debian folks don't like the static ones :) 20090812: * Added TXMountConfData.OrigImageSize and TXMountConfData.VirtImageSize to save the size of the input and output image in order to avoid regetting it always from disk. 20090814: * Replaced all malloc and realloc occurences with the two macros XMOUNT_MALLOC and XMOUNT_REALLOC. 20090816: * Replaced where applicable all occurences of str(n)cpy or alike with their corresponding macros XMOUNT_STRSET, XMOUNT_STRCPY and XMOUNT_STRNCPY pendants. 20090907: v0.4.2 released * Fixed a bug in VMDK lock file access. glob_vmdk_lockfile_size wasn't reset to 0 when the file was deleted. * Fixed a bug in VMDK descriptor file access. Had to add glob_vmdk_file_size to track the size of this file as strlen was a bad idea :). 20100324: v0.4.3 released * Changed all header structs to prevent different sizes on i386 and amd64. See xmount.h for more details. 20100810: v0.4.4 released * Found a bug in InitVirtVdiHeader(). The 64bit values were addressed incorrectly while filled with rand(). This leads to an error message when trying to add a VDI file to VirtualBox 3.2.8. 20110210: * Adding subtype and fsname FUSE options in order to display mounted source in mount command output. 20110211: v0.4.5 released 20111011: * Changes to deal with libewf v2 API (thx to Joachim Metz) 20111109: v0.4.6 released * Added support for DMG output type (actually a DD with .dmg file extension). This type is used as default output type when using xmount under Mac OS X. 20120130: v0.4.7 released * Made InitVirtImageInfoFile less picky about missing EWF infos. 20120507: * Added support for VHD output image as requested by various people. * Statically linked libs updated to 20120504 (libewf) and 3.7.0 (afflib). 20120510: v0.5.0 released * Added stbuf->st_blocks calculation for VHD images in function GetVirtFileAttr. This makes Windows not think the emulated file would be a sparse file. Sparse vhd files are not attachable in Windows. 20130726: v0.6.0 released * Added libaaff to replace libaff (thx to Guy Voncken). * Added libdd to replace raw dd input file handling and finally support split dd files (thx to Guy Voncken). 20140311: * Added libaewf (thx to Guy Voncken). 20140726: * Added support for dynamically loading of input libs. This should ease adding support for new input image formats in the future. * Moved input image functions to their corresponding dynamically loadable libs. * Prepended "glob_" to all global vars for better identification. 20140731: * Added --offset option as requested by HPM. * Began massive code cleanup. 20140803: * Added correct return code handling when calling input lib functions including getting error messages using GetErrorMessage. * Added input lib specific option parsing. * Re-implemented InitVirtImageInfoFile() using input lib's GetInfofileContent() function. * Further code cleanups. 20140807: * Further code cleanups. * Renamed GetVirtFileAttr() to FuseGetAttr() * Renamed CreateVirtDir() to FuseMkDir() * Renamed CreateVirtDir() to FuseMkNod() * Renamed OpenVirtFile() to FuseOpen() * Renamed GetVirtFiles() to FuseReadDir() * Renamed ReadVirtFile() to FuseRead() * Renamed RenameVirtFile() to FuseRename() * Renamed DeleteVirtDir() to FuseRmDir() * Renamed DeleteVirtFile() to FuseUnlink() * Renamed WriteVirtFile() to FuseWrite() * Fixed bug in CalculateInputImageHash() where always HASH_AMOUNT bytes were hased even if input image is smaller. * Fixed a newly introduced bug in FuseRead() and GetVirtImageData() returning -EIO when trying to read behind EOF. The correct return value is 0. 20140811: * Renamed CheckFuseAllowOther() to CheckFuseSettings() and added a check to see if user is part of the fuse group. 20140814: * Replaced cuserid() with getpwuid(geteuid()) in CheckFuseSettings() as it is deprecated on Linux and not available on OSx. * Only build fuse group checks from CheckFuseSettings() on Linux. 20140820: * Added libxmount_morphing handling. * Renamed LoadInputLibs() to LoadLibs() and added code to find and load morphing libs. * Added FindMorphingLib() function. 20140821: * Changed command line syntax to support specification of multiple input images of different types. * Added code in main() to initialize and open multiple input images. * Added GetMorphedImageSize() and GetMorphedImageData(). 20140822: * Removed GetInputImageSize() as it isn't needed anymore. * Quick test showed morphing using libxmount_morphing_combine seems to work. 20140825: * Added InitResources() and FreeResources() functions. 20140926: v0.7.0 released 20140930: v0.7.1 released * Fixed bug with --sizelimit command line option. 20141001: v0.7.2 released * Fixed bug in FreeResources(). Do not free vdi.p_vdi_block_map as it is part of vdi.p_vdi_header. 20141113: v0.7.3 released * Fixed minor bug in ParseCmdLine() which was not checking upper argc bounds correctly. 20150820: v0.7.4 released 20150901: * Improved the way fsname is built */