hash.cxx

   1 // Copyright (C) Andrew Tridgell 2002 (original file)
   2 // Copyright (C) 2006-2019 Red Hat Inc. (systemtap changes)
   3 //
   4 // This program is free software; you can redistribute it and/or modify
   5 // it under the terms of the GNU General Public License as published by
   6 // the Free Software Foundation; either version 2 of the License, or
   7 // (at your option) any later version.
   8 //
   9 // This program is distributed in the hope that it will be useful,
  10 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 // GNU General Public License for more details.
  13 //
  14 // You should have received a copy of the GNU General Public License
  15 // along with this program.  If not, see <http://www.gnu.org/licenses/>.
  16
  17 #include "config.h"
  18 #include "session.h"
  19 #include "hash.h"
  20 #include "util.h"
  21
  22 #include <cstdlib>
  23 #include <cstring>
  24 #include <fstream>
  25 #include <sstream>
  26 #include <iomanip>
  27 #include <cerrno>
  28
  29 extern "C" {
  30 #include <sys/types.h>
  31 #include <sys/stat.h>
  32 #include <unistd.h>
  33 #include "mdfour.h"
  34 }
  35
  36 using namespace std;
  37
  38
  39 class stap_hash
  40 {
  41 private:
  42   struct mdfour md4;
  43   std::ostringstream parm_stream;
  44
  45 public:
  46   stap_hash() { start(); }
  47   stap_hash(const stap_hash &base) { md4 = base.md4; parm_stream << base.parm_stream.str(); }
  48
  49   void start();
  50
  51   void add(const std::string& description, const unsigned char *buffer, size_t size);
  52   template<typename T> void add(const std::string& d, const T& x);
  53   void add(const std::string& d, const char *s) { add((const std::string&)d, (const unsigned char *)s, strlen(s)); }
  54   void add(const std:: string& d, const std::string& s) { add(d, (const unsigned char *)s.c_str(), s.length()); }
  55
  56   void add_path(const std::string& description, const std::string& path);
  57
  58   void result(std::string& r);
  59   std::string get_parms() { return parm_stream.str(); }
  60 };
  61
  62
  63 void
  64 stap_hash::start()
  65 {
  66   mdfour_begin(&md4);
  67 }
  68
  69
  70 void
  71 stap_hash::add(const std::string& description, const unsigned char *buffer, size_t size)
  72 {
  73   parm_stream << description << buffer << endl;
  74   mdfour_update(&md4, buffer, size);
  75 }
  76
  77
  78 template <typename T> void
  79 stap_hash::add(const std::string& d, const T& x)
  80 {
  81   parm_stream << d << x << endl;
  82   mdfour_update(&md4, (const unsigned char *)&x, sizeof(x));
  83 }
  84
  85
  86 void
  87 stap_hash::add_path(const std::string& description, const std::string& path)
  88 {
  89   struct stat st;
  90   memset (&st, 0, sizeof(st));
  91
  92   if (stat(path.c_str(), &st) != 0)
  93     st.st_size = st.st_mtime = -1;
  94
  95   add(description + "Path: ", path);
  96   add(description + "Size: ", st.st_size);
  97   add(description + "Timestamp: ", st.st_mtime);
  98 }
  99
 100
 101 void
 102 stap_hash::result(string& r)
 103 {
 104   ostringstream rstream;
 105   unsigned char sum[16];
 106
 107   mdfour_update(&md4, NULL, 0);
 108   mdfour_result(&md4, sum);
 109
 110   for (int i=0; i<16; i++)
 111     {
 112       rstream << hex << setfill('0') << setw(2) << (unsigned)sum[i];
 113     }
 114   rstream << "_" << setw(0) << dec << (unsigned)md4.totalN;
 115   r = rstream.str();
 116 }
 117
 118 void create_hash_log(const string &type_str, const string &parms, const string &result, const string &hash_log_path)
 119 {
 120   ofstream log_file;
 121   time_t rawtime;
 122   time ( &rawtime );
 123   string time_str(ctime (&rawtime));
 124
 125   log_file.open(hash_log_path.c_str());
 126   log_file << "[" << time_str.substr(0,time_str.length()-1); // erase terminated '\n'
 127   log_file << "] " << type_str << ":" << endl;
 128   log_file << parms << endl;
 129   log_file << _("result:") << result << endl;
 130   log_file.close();
 131 }
 132
 133 static const stap_hash&
 134 get_base_hash (systemtap_session& s)
 135 {
 136   if (s.base_hash)
 137     return *s.base_hash;
 138
 139   s.base_hash = new stap_hash();
 140   stap_hash& h = *s.base_hash;
 141
 142   // Hash systemtap version
 143   h.add("Systemtap version: ", s.version_string());
 144
 145   // Hash kernel release and arch.
 146   h.add("Kernel Release: ", s.kernel_release);
 147   h.add_path("Kernel Build Tree ", s.kernel_build_tree);
 148   h.add("Architecture: ", s.architecture);
 149
 150   // Hash a few kernel version/build-id files too
 151   // (useful for kernel developers reusing a single source tree)
 152   h.add_path("Kernel Build Tree .config ", s.kernel_build_tree + "/.config");
 153   h.add_path("Kernel Build Tree .version ", s.kernel_build_tree + "/.version");
 154   h.add_path("Kernel Build Tree compile.h ", s.kernel_build_tree + "/include/linux/compile.h");
 155   h.add_path("Kernel Build Tree version.h ", s.kernel_build_tree + "/include/linux/version.h");
 156   h.add_path("Kernel Build Tree utsrelease.h ", s.kernel_build_tree + "/include/linux/utsrelease.h");
 157
 158   // Also hash guru mode flag, since behaviour can diverge.
 159   h.add("Guru mode ", s.guru_mode);
 160
 161   // Hash runtime path (that gets added in as "-R path").
 162   h.add_path("Runtime ", s.runtime_path);
 163   h.add_path("Runtime transport ", s.runtime_path + "/transport");
 164   h.add_path("Runtime unwind ", s.runtime_path + "/unwind");
 165   h.add_path("Runtime sub ", s.runtime_path +
 166              (s.runtime_usermode_p() ? "/dyninst" : "/linux"));
 167
 168   // Hash compiler path, size, and mtime.  We're just going to assume
 169   // we'll be using gcc. XXX: getting kbuild to spit out out would be
 170   // better, especially since this is fooled by ccache.
 171   h.add_path("Compiler ", find_executable("gcc"));
 172
 173   // Hash the systemtap size and mtime.  We could use VERSION/DATE,
 174   // but when developing systemtap that doesn't work well (since you
 175   // can compile systemtap multiple times in 1 day).  Since we don't
 176   // know exactly where we're getting run from, we'll use
 177   // /proc/self/exe (and we resolve it ourselves to help valgrind).
 178   h.add_path("Systemtap ", get_self_path());
 179
 180   return h;
 181 }
 182
 183
 184 static bool
 185 create_hashdir (systemtap_session& s, const string& result, string& hashdir)
 186 {
 187   int nlevels = 1;
 188
 189   // Use a N level subdir for the cache path to reduce the impact on
 190   // filesystems which are slow for large directories.  Let N be adjustable.
 191   const char *s_n = getenv("SYSTEMTAP_NLEVELS");
 192   if (s_n)
 193     {
 194       nlevels = atoi(s_n);
 195       if (nlevels < 1) nlevels = 1;
 196       if (nlevels > 8) nlevels = 8;
 197     }
 198
 199   hashdir = s.cache_path;
 200
 201   for (int i = 0; i < nlevels; i++)
 202     {
 203       hashdir += string("/") + result[i*2] + result[i*2 + 1];
 204       if (create_dir(hashdir.c_str()) != 0)
 205         {
 206           s.print_warning("failed to create cache directory (\"" + hashdir + "\") " + strerror(errno) + ", disabling cache support");
 207           s.use_cache = s.use_script_cache = false;
 208           return false;
 209         }
 210     }
 211   return true;
 212 }
 213
 214
 215 void
 216 find_script_hash (systemtap_session& s, const string& script)
 217 {
 218   stap_hash h(get_base_hash(s));
 219
 220   // Hash getuid.  This really shouldn't be necessary (since who you
 221   // are doesn't change the generated output), but the hash gets used
 222   // as the module name.  If two different users try to run the same
 223   // script at the same time, we need something to differentiate the
 224   // module name.
 225   h.add("UID: ", getuid());
 226
 227   // Hash user-specified arguments (that change the generated module).
 228   h.add("Bulk Mode (-b): ", s.bulk_mode);
 229   h.add("Timing (-t): ", s.timing);
 230   h.add("Skip Badvars (--skip-badvars): ", s.skip_badvars);
 231   h.add("Privilege (--privilege): ", s.privilege);
 232   h.add("Compatible (--compatible): ", s.compatible);
 233   h.add("Error suppression (--suppress-handler-errors): ", s.suppress_handler_errors);
 234   h.add("Suppress Time Limits (--suppress-time-limits): ", s.suppress_time_limits);
 235   h.add("Prologue Searching (--prologue-searching[=WHEN]): ", int(s.prologue_searching_mode));
 236
 237   for (unsigned i = 0; i < s.c_macros.size(); i++)
 238     h.add("Macros: ", s.c_macros[i]);
 239
 240   // Add any custom kbuild flags (-B)
 241   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 242     h.add("Kbuildflags: ", s.kbuildflags[i]);
 243
 244   // Add any custom --modinfo strings
 245   for (unsigned i = 0; i < s.modinfos.size(); i++)
 246     h.add("MODULE_INFO: ", s.modinfos[i]);
 247
 248   // -d MODULE
 249   for (set<string>::iterator it = s.unwindsym_modules.begin();
 250        it != s.unwindsym_modules.end();
 251        it++)
 252     h.add_path("Unwindsym Modules ", *it);
 253
 254   // Add the build id of each module
 255   for(vector<string>::iterator it = s.build_ids.begin();
 256       it != s.build_ids.end();
 257       it++)
 258     h.add("Build ID: ", *it);
 259
 260   // Add in pass 2 script output.
 261   h.add("Script:\n", script);
 262
 263   // Get the directory path to store our cached script
 264   string result, hashdir;
 265   h.result(result);
 266   if (!create_hashdir(s, result, hashdir))
 267     return;
 268
 269   // Update module name to be 'stap_{hash start}'.  '{hash start}'
 270   // must not be too long.  This shouldn't happen, since the maximum
 271   // size of a hash is 32 fixed chars + 1 (for the '_') + a max of 11.
 272   s.module_name = "stap_" + result;
 273   if (s.module_name.size() >= (MODULE_NAME_LEN - 1))
 274     s.module_name.resize(MODULE_NAME_LEN - 1);
 275
 276   // 'ccache' would use a hash path of something like:
 277   //    s.hash_path = hashdir + "/" + result.substr(nlevels);
 278   // which would look like:
 279   //    ~/.stap_cache/A/B/CDEFGHIJKLMNOPQRSTUVWXYZABCDEF_XXX
 280   //
 281   // We're using the following so that the module can be used straight
 282   // from the cache if desired.  This ends up looking like this:
 283   //    ~/.stap_cache/A/B/stap_ABCDEFGHIJKLMNOPQRSTUVWXYZABCDEF_XXX.ko
 284   s.hash_path = hashdir + "/" + s.module_filename();
 285
 286   // Update C source name with new module_name.
 287   s.translated_source = string(s.tmpdir) + "/" + s.module_name + "_src.c";
 288   create_hash_log(string("script_hash"), h.get_parms(), result,
 289                   hashdir + "/" + s.module_name + "_hash.log");
 290 }
 291
 292
 293 void
 294 find_stapconf_hash (systemtap_session& s)
 295 {
 296   stap_hash h(get_base_hash(s));
 297
 298   // Add any custom kbuild flags
 299   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 300     h.add("Kbuildflags: ", s.kbuildflags[i]);
 301
 302   // Get the directory path to store our cached stapconf parameters
 303   string result, hashdir;
 304   h.result(result);
 305   if (!create_hashdir(s, result, hashdir))
 306     return;
 307
 308   s.stapconf_name = "stapconf_" + result + ".h";
 309   s.stapconf_path = hashdir + "/" + s.stapconf_name;
 310   create_hash_log(string("stapconf_hash"), h.get_parms(), result,
 311                   hashdir + "/stapconf_" + result + "_hash.log");
 312 }
 313
 314
 315 string
 316 find_tracequery_hash (systemtap_session& s, const string& header)
 317 {
 318   stap_hash h(get_base_hash(s));
 319
 320   // Add the tracepoint header to the computed hash
 321   h.add_path("Header ", header);
 322
 323   // Add any custom kbuild flags
 324   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 325     h.add("Kbuildflags: ", s.kbuildflags[i]);
 326
 327   // Get the directory path to store our cached module
 328   string result, hashdir;
 329   h.result(result);
 330   if (!create_hashdir(s, result, hashdir))
 331     return ""; // XXX: as opposed to throwing an exception?
 332
 333   create_hash_log(string("tracequery_hash"), h.get_parms(), result,
 334                   hashdir + "/tracequery_" + result + "_hash.log");
 335   return hashdir + "/tracequery_" + result + ".o";
 336 }
 337
 338
 339 string
 340 find_typequery_hash (systemtap_session& s, const string& name)
 341 {
 342   stap_hash h(get_base_hash(s));
 343
 344   // Add the typequery name to distinguish the hash
 345   h.add("Typequery Name: ", name);
 346
 347   if (name[0] == 'k')
 348     // Add any custom kbuild flags
 349     for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 350       h.add("Kbuildflags: ", s.kbuildflags[i]);
 351
 352   // Get the directory path to store our cached module
 353   string result, hashdir;
 354   h.result(result);
 355   if (!create_hashdir(s, result, hashdir))
 356     return "";
 357
 358   create_hash_log(string("typequery_hash"), h.get_parms(), result,
 359                   hashdir + "/typequery_" + result + "_hash.log");
 360   return hashdir + "/typequery_" + result
 361     + (name[0] == 'k' ? ".ko" : ".so");
 362 }
 363
 364
 365 string
 366 find_uprobes_hash (systemtap_session& s)
 367 {
 368   stap_hash h(get_base_hash(s));
 369
 370   // Hash runtime uprobes paths
 371   h.add_path("Uprobes Runtime Path /uprobes ", s.runtime_path + "/linux/uprobes");
 372   h.add_path("Uprobes Runtime Path /uprobes2 ", s.runtime_path + "/linux/uprobes2");
 373
 374   // Add any custom kbuild flags
 375   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 376     h.add("Kbuildflags: ", s.kbuildflags[i]);
 377
 378   // Add any custom --modinfo strings
 379   for (unsigned i = 0; i < s.modinfos.size(); i++)
 380     h.add("MODULE_INFO: ", s.modinfos[i]);
 381
 382   // Get the directory path to store our cached module
 383   string result, hashdir;
 384   h.result(result);
 385   if (!create_hashdir(s, result, hashdir))
 386     return "";
 387
 388   create_hash_log(string("uprobes_hash"), h.get_parms(), result,
 389                   hashdir + "/uprobes_" + result + "_hash.log");
 390   return hashdir + "/uprobes_" + result;
 391 }
 392
 393 /* vim: set sw=2 ts=8 cino=>4,n-2,{2,^-2,t0,(0,u0,w1,M1 : */