--- /dev/null
+// $Id: vhashify.c,v 1.6 2005/03/24 12:46:59 ensc Exp $ --*- c -*--
+
+// Copyright (C) 2005 Enrico Scholz <enrico.scholz@informatik.tu-chemnitz.de>
+//
+// 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; version 2 of the License.
+//
+// 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, write to the Free Software
+// Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "vhashify.h"
+#include "util.h"
+
+#include "lib/internal.h"
+#include "lib_internal/matchlist.h"
+#include "lib_internal/unify.h"
+#include "ensc_vector/vector.h"
+
+#include <beecrypt/beecrypt.h>
+
+#include <setjmp.h>
+#include <unistd.h>
+#include <getopt.h>
+#include <string.h>
+#include <assert.h>
+#include <stdlib.h>
+#include <fcntl.h>
+#include <dirent.h>
+#include <errno.h>
+#include <signal.h>
+#include <limits.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+
+#define ENSC_WRAPPERS_STDLIB 1
+#define ENSC_WRAPPERS_UNISTD 1
+#define ENSC_WRAPPERS_FCNTL 1
+#define ENSC_WRAPPERS_DIRENT 1
+#define ENSC_WRAPPERS_IO 1
+#include <wrappers.h>
+
+
+#define HASH_BLOCKSIZE 0x10000000u
+#define HASH_MINSIZE 0x10
+#define HASH_MAXBITS 256 // we have to take care about
+ // max filename-length...
+
+#if HASH_MINSIZE<=0
+# error HASH_MINSIZE must be not '0'
+#endif
+
+
+#define CMD_HELP 0x8000
+#define CMD_VERSION 0x8001
+
+#define CMD_DESTINATION 0x1000
+#define CMD_INSECURE 0x1001
+#define CMD_SLEDGE 0x1002
+#define CMD_MANUALLY 0x1003
+#define CMD_REFRESH 0x1004
+
+struct option const
+CMDLINE_OPTIONS[] = {
+ { "help", no_argument, 0, CMD_HELP },
+ { "version", no_argument, 0, CMD_VERSION },
+ { "destination", required_argument, 0, CMD_DESTINATION },
+ { "insecure", no_argument, 0, CMD_INSECURE },
+ { "sledgehammer", no_argument, 0, CMD_SLEDGE },
+ { "manually", no_argument, 0, CMD_MANUALLY },
+ { "refresh", no_argument, 0, CMD_REFRESH },
+ { "dry-run", no_argument, 0, 'n' },
+ { "verbose", no_argument, 0, 'v' },
+ { 0,0,0,0 }
+};
+
+ // hash digest grouped by 2 digits + hash-collision counter + 2* '/' + NULL
+typedef char HashPath[HASH_MAXBITS/4 + (HASH_MAXBITS/4/2) +
+ sizeof(unsigned int)*2 + 3];
+
+struct HashDirConfiguration
+{
+ hashFunction const *method;
+ enum { hshALL=0, hshSTART = 1, hshMIDDLE=2,
+ hshEND = 4, hshINVALID = -1 } blocks;
+ size_t blocksize;
+};
+
+struct WalkdownInfo
+{
+ PathInfo state;
+ struct MatchList dst_list;
+ struct HashDirConfiguration hash_conf;
+ HashDirCollection hash_dirs;
+ size_t hash_dirs_max_size;
+
+ hashFunctionContext hash_context;
+};
+
+int wrapper_exit_code = 1;
+struct Arguments const *global_args;
+static struct SkipReason skip_reason;
+
+struct WalkdownInfo global_info = {
+ .hash_conf = { .method = 0,
+ .blocks = hshALL,
+ .blocksize = 0x10000 }
+};
+
+#include "vhashify-init.hc"
+
+int Global_getVerbosity() {
+ return global_args->verbosity;
+}
+
+int Global_doRenew() {
+ return true;
+}
+
+int Global_isVserverRunning() {
+ // TODO
+ return global_args->insecure<2;
+}
+
+static void
+showHelp(char const *cmd)
+{
+ WRITE_MSG(1, "Usage:\n ");
+ WRITE_STR(1, cmd);
+ WRITE_MSG(1,
+ " [-nv] [--refresh] <vserver>\n or\n ");
+ WRITE_STR(1, cmd);
+ WRITE_MSG(1,
+ " --manually [-nv] [--] <hashdir> <path> <excludelist>\n\n"
+ " --manually ... hashify generic paths; excludelists must be generated\n"
+ " manually\n"
+ " --refresh ... hashify already hashified files also\n"
+ " -n ... do not modify anything; just show what there will be\n"
+ " done (in combination with '-v')\n"
+ " -v ... verbose mode\n"
+ "Please report bugs to " PACKAGE_BUGREPORT "\n");
+
+ exit(0);
+}
+
+static void
+showVersion()
+{
+ WRITE_MSG(1,
+ "vhashify " VERSION " -- hashifies vservers and/or directories\n"
+ "This program is part of " PACKAGE_STRING "\n\n"
+ "Copyright (C) 2005 Enrico Scholz\n"
+ VERSION_COPYRIGHT_DISCLAIMER);
+ exit(0);
+}
+
+int
+HashDirInfo_compareDevice(void const *lhs_v, void const *rhs_v)
+{
+ struct HashDirInfo const * const lhs = lhs_v;
+ dev_t const * const rhs = rhs_v;
+
+ assert(lhs!=0 && rhs!=0);
+ return lhs->device - *rhs;
+}
+
+PathInfo const *
+HashDirInfo_findDevice(HashDirCollection const *coll, dev_t dev)
+{
+ struct HashDirInfo const *res;
+
+ res = Vector_searchSelfOrg_const(coll, &dev,
+ HashDirInfo_compareDevice, vecSHIFT_ONCE);
+
+ if (res!=0) return &res->path;
+ else return 0;
+}
+
+#include "vserver-visitdir.hc"
+
+static bool
+checkFstat(PathInfo const * const basename,
+ struct stat * const st)
+{
+ assert(basename->d[0] != '/');
+
+ // local file does not exist... strange
+ // TODO: message
+ skip_reason.r = rsFSTAT;
+ if (lstat(basename->d, st)==-1) return false;
+
+ // this is a directory and succeeds everytime
+ if (S_ISDIR(st->st_mode))
+ return true;
+
+ // ignore symlinks
+ skip_reason.r = rsSYMLINK;
+ if (S_ISLNK(st->st_mode)) return false;
+
+ // ignore special files
+ skip_reason.r = rsSPECIAL;
+ if (!S_ISREG(st->st_mode) &&
+ !S_ISDIR(st->st_mode)) return false;
+
+ // ignore small files
+ skip_reason.r = rsTOOSMALL;
+ if (st->st_size < HASH_MINSIZE) return false;
+
+ switch (Unify_isIUnlinkable(basename->d)) {
+ case unifyUNSUPPORTED : skip_reason.r = rsUNSUPPORTED; return false;
+ case unifyBUSY :
+ // do an implicit refresh on busy files when there are no active links
+ if (st->st_nlink>1 && !global_args->do_refresh) {
+ // TODO: message
+ skip_reason.r = rsUNIFIED;
+ return false;
+ }
+ break;
+ default : break;
+ }
+
+ return true;
+}
+
+static sigjmp_buf bus_error_restore;
+static volatile sig_atomic_t bus_error;
+
+static void
+handlerSIGBUS(int UNUSED num)
+{
+ bus_error = 1;
+ siglongjmp(bus_error_restore, 1);
+}
+
+static bool
+convertDigest(HashPath d_path)
+{
+ static char const HEX_DIGIT[] = "0123456789abcdef";
+ hashFunctionContext * const h_ctx = &global_info.hash_context;
+ size_t d_size = h_ctx->algo->digestsize;
+
+ unsigned char digest[d_size];
+ size_t out = 0;
+
+ if (hashFunctionContextDigest(h_ctx, digest)==-1)
+ return false;
+
+ for (size_t in=0;
+ out+1<sizeof(HashPath)-(sizeof(unsigned int)*2 + 2) && in<d_size;
+ ++in) {
+ if ((in+254)%(in<=2 ? 1 : 256) == 0 && in>0)
+ d_path[out++]='/';
+ d_path[out++] = HEX_DIGIT[digest[in] >> 4];
+ d_path[out++] = HEX_DIGIT[digest[in] & 0x0f];
+ }
+ d_path[out++] = '\0';
+
+ return true;
+}
+
+static bool
+addStatHash(hashFunctionContext *h_ctx, struct stat const * const st)
+{
+#define DECL_ATTR(X) __typeof__(st->st_##X) X
+#define SET_ATTR(X) .X = st->st_##X
+
+ struct __attribute__((__packed__)) {
+ DECL_ATTR(mode);
+ DECL_ATTR(uid);
+ DECL_ATTR(gid);
+ DECL_ATTR(rdev);
+ DECL_ATTR(size);
+ DECL_ATTR(mtime);
+ } tmp = {
+ SET_ATTR(mode),
+ SET_ATTR(uid),
+ SET_ATTR(gid),
+ SET_ATTR(rdev),
+ SET_ATTR(size),
+ SET_ATTR(mtime)
+ };
+
+ return hashFunctionContextUpdate(h_ctx, (void *)&tmp, sizeof tmp)!=-1;
+}
+
+static bool
+calculateHashFromFD(int fd, HashPath d_path, struct stat const * const st)
+{
+ hashFunctionContext * const h_ctx = &global_info.hash_context;
+ void const * volatile buf = 0;
+ loff_t volatile buf_size = 0;
+ bool volatile res = false;
+
+
+ if (hashFunctionContextReset(h_ctx)==-1 ||
+ !addStatHash(h_ctx, st))
+ return false;
+
+ bus_error = 0;
+ if (sigsetjmp(bus_error_restore,1)==0) {
+ loff_t offset = 0;
+ off_t size = st->st_size;
+
+ while (offset < size) {
+ buf_size = size-offset;
+ if (buf_size>HASH_BLOCKSIZE) buf_size = HASH_BLOCKSIZE;
+
+ if ((buf=mmap(0, buf_size, PROT_READ, MAP_SHARED, fd, offset))==0) {
+ perror("mmap(<hash>)");
+ goto out;
+ }
+
+ offset += buf_size;
+ madvise(const_cast(void *)(buf), buf_size, MADV_SEQUENTIAL); // ignore error...
+
+ if (hashFunctionContextUpdate(h_ctx, buf, buf_size)==-1) goto out;
+
+ munmap(const_cast(void *)(buf), buf_size);
+ buf = 0;
+ }
+
+ res = convertDigest(d_path);
+ }
+
+ out:
+ if (buf!=0) munmap(const_cast(void *)(buf), buf_size);
+ return res;
+}
+
+static bool
+calculateHash(PathInfo const *filename, HashPath d_path, struct stat const * const st)
+{
+ int fd = open(filename->d, O_NOFOLLOW|O_NONBLOCK|O_RDONLY|O_NOCTTY);
+ struct stat fst;
+ bool res = false;
+
+ do {
+ if (fd==-1) {
+ int old_errno = errno;
+ WRITE_MSG(2, "Failed to open '");
+ WRITE_STR(2, filename->d);
+ errno = old_errno;
+ perror("'");
+ break;;
+ }
+
+ if (fstat(fd, &fst)==-1 ||
+ fst.st_dev!=st->st_dev || fst.st_ino!=st->st_ino) {
+ WRITE_MSG(2, "An unexpected event occured while stating '");
+ WRITE_STR(2, filename->d);
+ WRITE_MSG(2, "'.\n");
+ break;
+ }
+
+ if (!calculateHashFromFD(fd, d_path, st)) {
+ WRITE_MSG(2, "Failed to calculate hash for '");
+ WRITE_STR(2, filename->d);
+ WRITE_MSG(2, "'.\n");
+ break;
+ }
+
+ res = true;
+ } while (false);
+
+ if (fd!=-1) close(fd);
+ return res;
+}
+
+static enum { mkdirFAIL, mkdirSUCCESS, mkdirSKIP }
+mkdirSingle(char const *path, char *end_ptr, int good_err)
+{
+ *end_ptr = '\0';
+ if (mkdir(path, 0700)!=-1 || errno==EEXIST) {
+ *end_ptr = '/';
+ return mkdirSUCCESS;
+ }
+ else if (errno==good_err) {
+ *end_ptr = '/';
+ return mkdirSKIP;
+ }
+ else {
+ int old_errno = errno;
+ WRITE_MSG(2, "mkdir('");
+ WRITE_STR(2, path);
+ errno = old_errno;
+ perror("')");
+ return mkdirFAIL;
+ }
+}
+
+static char *
+rstrchr(char *str, char c)
+{
+ while (*str!=c) --str;
+ return str;
+}
+
+static bool
+mkdirRecursive(char const *path)
+{
+ if (path[0]!='/') return false; // only absolute paths
+
+ char buf[strlen(path)+1];
+ char * ptr = buf + sizeof(buf) - 2;
+
+ strcpy(buf, path);
+
+ while (ptr>buf && (ptr = rstrchr(ptr, '/'))!=0) {
+ switch (mkdirSingle(buf, ptr, ENOENT)) {
+ case mkdirSUCCESS : break;
+ case mkdirSKIP : --ptr; continue;
+ case mkdirFAIL : return false;
+ }
+
+ break; // implied by mkdirSUCCESS
+ }
+
+ assert(ptr!=0);
+ ++ptr;
+
+ while ((ptr=strchr(ptr, '/'))!=0) {
+ switch (mkdirSingle(buf, ptr, 0)) {
+ case mkdirSKIP :
+ case mkdirFAIL : return false;
+ case mkdirSUCCESS : ++ptr; continue;
+ }
+ }
+
+ return true;
+}
+
+static bool
+resolveCollisions(char *result, PathInfo const *root, HashPath d_path,
+ struct stat *st, struct stat *hash_st)
+{
+ strcpy(result, root->d); // 'root' ends on '/' already (see initHashList())
+ strcat(result, d_path);
+
+ char *ptr = result + strlen(result);
+ unsigned int idx = 0;
+ char buf[sizeof(int)*2 + 1];
+ size_t len;
+
+ *ptr = '-';
+ ptr[sizeof(int)*2+1] = '\0';
+
+ for (;; ++idx) {
+ len = utilvserver_fmt_xuint(buf, idx);
+ memset(ptr+1, '0', sizeof(int)*2 - len);
+ memcpy(ptr+1 + sizeof(int)*2 - len, buf, len);
+
+ if (lstat(result, hash_st)==-1) {
+ if (global_args->dry_run && errno!=ENOENT) {
+ int old_errno = errno;
+ WRITE_MSG(2, "lstat('");
+ WRITE_STR(2, buf);
+ errno = old_errno;
+ perror("')");
+ return false;
+ }
+ }
+ else if (Unify_isUnified(st, hash_st)) {
+ skip_reason.r = rsUNIFIED;
+ return false;
+ }
+ else if (!Unify_isUnifyable(st, hash_st))
+ continue; // continue with next number*****
+ else
+ break; // ok, we finish here
+
+ if (!global_args->dry_run) {
+ *ptr = '\0';
+ if (!mkdirRecursive(result))
+ return false;
+ *ptr = '-';
+
+ int fd = open(result, O_NOFOLLOW|O_EXCL|O_CREAT|O_WRONLY, 0200);
+
+ if (fd==-1) {
+ int old_errno = errno;
+ WRITE_MSG(2, "open('");
+ WRITE_STR(2, buf);
+ errno = old_errno;
+ perror("')");
+ return false;
+ }
+
+ close(fd);
+ }
+
+ // HACK: avoid an additional lstat on the resulting hash-file
+ hash_st->st_size = 0;
+ break;
+ }
+
+ return true;
+}
+
+static char const *
+checkDirEntry(PathInfo const *path, PathInfo const *basename,
+ bool *is_dir,
+ struct stat *st, struct stat *hash_st,
+ char *result_buf)
+{
+ //printf("checkDirEntry(%s, %s, %u)\n", path->d, d_path, is_dir);
+
+ struct WalkdownInfo const * const info = &global_info;
+
+ // Check if it is in the exclude/include list of the destination vserver and
+ // abort when it is not matching an allowed entry
+ skip_reason.r = rsEXCL;
+ if (MatchList_compare(&info->dst_list, path->d)!=stINCLUDE) return 0;
+
+ if (checkFstat(basename, st)) {
+ PathInfo const *hash_root_path;
+ HashPath d_path;
+
+ *is_dir = S_ISDIR(st->st_mode);
+
+ if (!*is_dir &&
+ !((skip_reason.r = rsWRONGDEV,
+ (hash_root_path = HashDirInfo_findDevice(&info->hash_dirs, st->st_dev))!=0) &&
+ (skip_reason.r = rsGENERAL,
+ calculateHash(basename, d_path, st)) &&
+ resolveCollisions(result_buf, hash_root_path, d_path, st, hash_st)))
+ return 0;
+
+ return result_buf;
+ }
+
+ return 0;
+}
+
+static void
+printSkipReason()
+{
+ WRITE_MSG(1, " (");
+ switch (skip_reason.r) {
+ case rsDOTFILE : WRITE_MSG(1, "dotfile"); break;
+ case rsEXCL : WRITE_MSG(1, "excluded"); break;
+ case rsTOOSMALL : WRITE_MSG(1, "too small"); break;
+ case rsUNSUPPORTED : WRITE_MSG(1, "operation not supported"); break;
+ case rsFSTAT : WRITE_MSG(1, "fstat error"); break;
+ case rsSYMLINK : WRITE_MSG(1, "symlink"); break;
+ case rsUNIFIED : WRITE_MSG(1, "already unified"); break;
+ case rsSPECIAL : WRITE_MSG(1, "non regular file"); break;
+ case rsWRONGDEV : WRITE_MSG(1, "no matching device"); break;
+ case rsGENERAL : WRITE_MSG(1, "general error"); break;
+ default : assert(false); abort();
+ }
+ WRITE_MSG(1, ")");
+}
+
+static bool
+doit(char const *src, char const *dst,
+ struct stat const *src_st, struct stat const *dst_st,
+ PathInfo const *path)
+{
+ if (global_args->dry_run || Global_getVerbosity()>=2) {
+ WRITE_MSG(1, "unifying '");
+ Vwrite(1, path->d, path->l);
+ WRITE_MSG(1, "'");
+
+ if (Global_getVerbosity()>=4) {
+ WRITE_MSG(1, " (to '");
+ WRITE_STR(1, dst);
+ WRITE_MSG(1, "')");
+ }
+
+ WRITE_MSG(1, "\n");
+ }
+
+ // abort here in dry-run mode
+ if (global_args->dry_run) return true;
+
+ if (dst_st->st_size==0) {
+ // file was not unified yet
+
+ if (Global_isVserverRunning()) {
+ (void)unlink(dst);
+ if (Unify_copy (src, src_st, dst) &&
+ // the mixed 'dst' and 'src_st' params are intentionally...
+ Unify_unify(dst, src_st, src, false))
+ return true;
+ }
+ else if (Unify_unify(src, src_st, dst, true))
+ return true;
+
+ (void)unlink(dst); // cleanup in error-case
+ }
+ // there exists already a reference-file
+ else if (Unify_unify(dst, dst_st, src, false))
+ return true;
+
+ return false;
+}
+
+static uint64_t
+visitDirEntry(struct dirent const *ent)
+{
+ uint64_t res = 0;
+ char const * dirname = ent->d_name;
+ PathInfo path = global_info.state;
+ PathInfo tmp_path = {
+ .d = dirname,
+ .l = strlen(dirname)
+ };
+ char path_buf[ENSC_PI_APPSZ(path, tmp_path)];
+ char const *match = 0;
+
+
+ PathInfo_append(&path, &tmp_path, path_buf);
+
+ bool is_dotfile = isDotfile(dirname);
+ bool is_dir;
+ struct stat src_stat = { .st_mode=0 };
+ struct stat hash_stat;
+ char tmpbuf[global_info.hash_dirs_max_size +
+ sizeof(HashPath) + 2];
+
+ skip_reason.r = rsDOTFILE;
+
+ if (is_dotfile ||
+ (match=checkDirEntry(&path, &tmp_path,
+ &is_dir, &src_stat, &hash_stat,
+ tmpbuf))==0) {
+
+ bool is_link = !is_dotfile && S_ISLNK(src_stat.st_mode);
+
+ if (Global_getVerbosity()>=1 &&
+ (Global_getVerbosity()>=3 || skip_reason.r!=rsUNIFIED) &&
+ ((!is_dotfile && !is_link) ||
+ (Global_getVerbosity()>=6 && is_dotfile) ||
+ (Global_getVerbosity()>=6 && is_link)) ) {
+ WRITE_MSG(1, " skipping '");
+ Vwrite(1, path.d, path.l);
+ WRITE_MSG(1, "'");
+ if (Global_getVerbosity()>=2) printSkipReason();
+ WRITE_MSG(1, "\n");
+ }
+
+ return 0;
+ }
+
+ if (is_dir) {
+ res = visitDir(dirname, &src_stat);
+ }
+ else if (doit(dirname, match, &src_stat, &hash_stat, &path))
+ res = 1;
+ else {
+ // TODO: message
+ res = 0;
+ }
+
+ return res;
+
+}
+
+int main(int argc, char *argv[])
+{
+ struct Arguments args = {
+ .mode = mdVSERVER,
+ .hash_dir = 0,
+ .verbosity = 0,
+ .insecure = 0,
+ .dry_run = false,
+ .do_refresh = false,
+ };
+
+ Vector_init(&global_info.hash_dirs, sizeof(struct HashDirInfo));
+
+ global_args = &args;
+ while (1) {
+ int c = getopt_long(argc, argv, "+nv",
+ CMDLINE_OPTIONS, 0);
+ if (c==-1) break;
+
+ switch (c) {
+ case CMD_HELP : showHelp(argv[0]);
+ case CMD_VERSION : showVersion();
+ case CMD_DESTINATION : args.hash_dir = optarg; break;
+ case CMD_MANUALLY : args.mode = mdMANUALLY; break;
+ case CMD_INSECURE : args.insecure = 1; break;
+ case CMD_SLEDGE : args.insecure = 2; break;
+ case CMD_REFRESH : args.do_refresh = true; break;
+ case 'n' : args.dry_run = true; break;
+ case 'v' : ++args.verbosity; break;
+ default :
+ WRITE_MSG(2, "Try '");
+ WRITE_STR(2, argv[0]);
+ WRITE_MSG(2, " --help\" for more information.\n");
+ return EXIT_FAILURE;
+ break;
+ }
+ }
+
+ if (argc==optind) {
+ WRITE_MSG(2, "No directory/vserver given\n");
+ return EXIT_FAILURE;
+ }
+
+ if (args.hash_dir==0 && args.mode==mdMANUALLY) {
+ WRITE_MSG(2, "'--manually' requires '--destination'\n");
+ return EXIT_FAILURE;
+ }
+
+ switch (args.mode) {
+ case mdMANUALLY : initModeManually(&args, argc-optind, argv+optind); break;
+ case mdVSERVER : initModeVserver (&args, argc-optind, argv+optind); break;
+ default : assert(false); return EXIT_FAILURE;
+ };
+
+ if (hashFunctionContextInit(&global_info.hash_context,
+ global_info.hash_conf.method)==-1) {
+ WRITE_MSG(2, "Failed to initialize hash-context\n");
+ return EXIT_FAILURE;
+ }
+
+ if (Global_getVerbosity()>=1)
+ WRITE_MSG(1, "Starting to traverse directories...\n");
+
+ signal(SIGBUS, handlerSIGBUS);
+
+ Echdir(global_info.dst_list.root.d);
+ visitDir("/", 0);
+
+#ifndef NDEBUG
+ MatchList_destroy(&global_info.dst_list);
+ freeHashList(&global_info.hash_dirs);
+ hashFunctionContextFree(&global_info.hash_context);
+#endif
+}
git://git.onelab.eu / util-vserver.git / blobdiff