hash.cxx

   1 // Copyright (C) Andrew Tridgell 2002 (original file)
   2 // Copyright (C) 2006-2011 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 kernel release and arch.
 143   h.add("Kernel Release: ", s.kernel_release);
 144   h.add_path("Kernel Build Tree ", s.kernel_build_tree);
 145   h.add("Architecture: ", s.architecture);
 146
 147   // Hash a few kernel version/build-id files too
 148   // (useful for kernel developers reusing a single source tree)
 149   h.add_path("Kernel Build Tree .config ", s.kernel_build_tree + "/.config");
 150   h.add_path("Kernel Build Tree .version ", s.kernel_build_tree + "/.version");
 151   h.add_path("Kernel Build Tree compile.h ", s.kernel_build_tree + "/include/linux/compile.h");
 152   h.add_path("Kernel Build Tree version.h ", s.kernel_build_tree + "/include/linux/version.h");
 153   h.add_path("Kernel Build Tree utsrelease.h ", s.kernel_build_tree + "/include/linux/utsrelease.h");
 154
 155   // Hash runtime path (that gets added in as "-R path").
 156   h.add_path("Runtime ", s.runtime_path);
 157   h.add_path("Runtime transport ", s.runtime_path + "/transport");
 158   h.add_path("Runtime unwind ", s.runtime_path + "/unwind");
 159   h.add_path("Runtime sub ", s.runtime_path +
 160              (s.runtime_usermode_p() ? "/dyninst" : "/linux"));
 161
 162   // Hash compiler path, size, and mtime.  We're just going to assume
 163   // we'll be using gcc. XXX: getting kbuild to spit out out would be
 164   // better, especially since this is fooled by ccache.
 165   h.add_path("Compiler ", find_executable("gcc"));
 166
 167   // Hash the systemtap size and mtime.  We could use VERSION/DATE,
 168   // but when developing systemtap that doesn't work well (since you
 169   // can compile systemtap multiple times in 1 day).  Since we don't
 170   // know exactly where we're getting run from, we'll use
 171   // /proc/self/exe (and we resolve it ourselves to help valgrind).
 172   h.add_path("Systemtap ", get_self_path());
 173
 174   return h;
 175 }
 176
 177
 178 static bool
 179 create_hashdir (systemtap_session& s, const string& result, string& hashdir)
 180 {
 181   int nlevels = 1;
 182
 183   // Use a N level subdir for the cache path to reduce the impact on
 184   // filesystems which are slow for large directories.  Let N be adjustable.
 185   const char *s_n = getenv("SYSTEMTAP_NLEVELS");
 186   if (s_n)
 187     {
 188       nlevels = atoi(s_n);
 189       if (nlevels < 1) nlevels = 1;
 190       if (nlevels > 8) nlevels = 8;
 191     }
 192
 193   hashdir = s.cache_path;
 194
 195   for (int i = 0; i < nlevels; i++)
 196     {
 197       hashdir += string("/") + result[i*2] + result[i*2 + 1];
 198       if (create_dir(hashdir.c_str()) != 0)
 199         {
 200           s.print_warning("failed to create cache directory (\"" + hashdir + "\") " + strerror(errno) + ", disabling cache support");
 201           s.use_cache = s.use_script_cache = false;
 202           return false;
 203         }
 204     }
 205   return true;
 206 }
 207
 208
 209 void
 210 find_script_hash (systemtap_session& s, const string& script)
 211 {
 212   stap_hash h(get_base_hash(s));
 213
 214   // Hash getuid.  This really shouldn't be necessary (since who you
 215   // are doesn't change the generated output), but the hash gets used
 216   // as the module name.  If two different users try to run the same
 217   // script at the same time, we need something to differentiate the
 218   // module name.
 219   h.add("UID: ", getuid());
 220
 221   // Hash user-specified arguments (that change the generated module).
 222   h.add("Bulk Mode (-b): ", s.bulk_mode);
 223   h.add("Timing (-t): ", s.timing);
 224   h.add("Prologue Searching (-P): ", s.prologue_searching);
 225   h.add("Skip Badvars (--skip-badvars): ", s.skip_badvars);
 226   h.add("Privilege (--privilege): ", s.privilege);
 227   h.add("Compatible (--compatible): ", s.compatible);
 228   h.add("Omit Werror (undocumented): ", s.omit_werror);
 229   h.add("Prologue Searching (-P): ", s.prologue_searching);
 230   h.add("Error suppression (--suppress-handler-errors): ", s.suppress_handler_errors);
 231   h.add("Suppress Time Limits (--suppress-time-limits): ", s.suppress_time_limits);
 232   for (unsigned i = 0; i < s.c_macros.size(); i++)
 233     h.add("Macros: ", s.c_macros[i]);
 234
 235   // Add any custom kbuild flags (-B)
 236   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 237     h.add("Kbuildflags: ", s.kbuildflags[i]);
 238
 239   // Add any custom --modinfo strings
 240   for (unsigned i = 0; i < s.modinfos.size(); i++)
 241     h.add("MODULE_INFO: ", s.modinfos[i]);
 242
 243   // -d MODULE
 244   for (set<string>::iterator it = s.unwindsym_modules.begin();
 245        it != s.unwindsym_modules.end();
 246        it++)
 247     h.add_path("Unwindsym Modules ", *it);
 248
 249   // Add the build id of each module
 250   for(vector<string>::iterator it = s.build_ids.begin();
 251       it != s.build_ids.end();
 252       it++)
 253     h.add("Build ID: ", *it);
 254
 255   // Add in pass 2 script output.
 256   h.add("Script:\n", script);
 257
 258   // Get the directory path to store our cached script
 259   string result, hashdir;
 260   h.result(result);
 261   if (!create_hashdir(s, result, hashdir))
 262     return;
 263
 264   // Update module name to be 'stap_{hash start}'.  '{hash start}'
 265   // must not be too long.  This shouldn't happen, since the maximum
 266   // size of a hash is 32 fixed chars + 1 (for the '_') + a max of 11.
 267   s.module_name = "stap_" + result;
 268   if (s.module_name.size() >= (MODULE_NAME_LEN - 1))
 269     s.module_name.resize(MODULE_NAME_LEN - 1);
 270
 271   // 'ccache' would use a hash path of something like:
 272   //    s.hash_path = hashdir + "/" + result.substr(nlevels);
 273   // which would look like:
 274   //    ~/.stap_cache/A/B/CDEFGHIJKLMNOPQRSTUVWXYZABCDEF_XXX
 275   //
 276   // We're using the following so that the module can be used straight
 277   // from the cache if desired.  This ends up looking like this:
 278   //    ~/.stap_cache/A/B/stap_ABCDEFGHIJKLMNOPQRSTUVWXYZABCDEF_XXX.ko
 279   s.hash_path = hashdir + "/" + s.module_filename();
 280
 281   // Update C source name with new module_name.
 282   s.translated_source = string(s.tmpdir) + "/" + s.module_name + "_src.c";
 283   create_hash_log(string("script_hash"), h.get_parms(), result,
 284                   hashdir + "/" + s.module_name + "_hash.log");
 285 }
 286
 287
 288 void
 289 find_stapconf_hash (systemtap_session& s)
 290 {
 291   stap_hash h(get_base_hash(s));
 292
 293   // Add any custom kbuild flags
 294   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 295     h.add("Kbuildflags: ", s.kbuildflags[i]);
 296
 297   // Get the directory path to store our cached stapconf parameters
 298   string result, hashdir;
 299   h.result(result);
 300   if (!create_hashdir(s, result, hashdir))
 301     return;
 302
 303   s.stapconf_name = "stapconf_" + result + ".h";
 304   s.stapconf_path = hashdir + "/" + s.stapconf_name;
 305   create_hash_log(string("stapconf_hash"), h.get_parms(), result,
 306                   hashdir + "/stapconf_" + result + "_hash.log");
 307 }
 308
 309
 310 string
 311 find_tracequery_hash (systemtap_session& s, const string& header)
 312 {
 313   stap_hash h(get_base_hash(s));
 314
 315   // Add the tracepoint header to the computed hash
 316   h.add_path("Header ", header);
 317
 318   // Add any custom kbuild flags
 319   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 320     h.add("Kbuildflags: ", s.kbuildflags[i]);
 321
 322   // Get the directory path to store our cached module
 323   string result, hashdir;
 324   h.result(result);
 325   if (!create_hashdir(s, result, hashdir))
 326     return ""; // XXX: as opposed to throwing an exception?
 327
 328   create_hash_log(string("tracequery_hash"), h.get_parms(), result,
 329                   hashdir + "/tracequery_" + result + "_hash.log");
 330   return hashdir + "/tracequery_" + result + ".o";
 331 }
 332
 333
 334 string
 335 find_typequery_hash (systemtap_session& s, const string& name)
 336 {
 337   stap_hash h(get_base_hash(s));
 338
 339   // Add the typequery name to distinguish the hash
 340   h.add("Typequery Name: ", name);
 341
 342   if (name[0] == 'k')
 343     // Add any custom kbuild flags
 344     for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 345       h.add("Kbuildflags: ", s.kbuildflags[i]);
 346
 347   // Get the directory path to store our cached module
 348   string result, hashdir;
 349   h.result(result);
 350   if (!create_hashdir(s, result, hashdir))
 351     return "";
 352
 353   create_hash_log(string("typequery_hash"), h.get_parms(), result,
 354                   hashdir + "/typequery_" + result + "_hash.log");
 355   return hashdir + "/typequery_" + result
 356     + (name[0] == 'k' ? ".ko" : ".so");
 357 }
 358
 359
 360 string
 361 find_uprobes_hash (systemtap_session& s)
 362 {
 363   stap_hash h(get_base_hash(s));
 364
 365   // Hash runtime uprobes paths
 366   h.add_path("Uprobes Runtime Path /uprobes ", s.runtime_path + "/linux/uprobes");
 367   h.add_path("Uprobes Runtime Path /uprobes2 ", s.runtime_path + "/linux/uprobes2");
 368
 369   // Add any custom kbuild flags
 370   for (unsigned i = 0; i < s.kbuildflags.size(); i++)
 371     h.add("Kbuildflags: ", s.kbuildflags[i]);
 372
 373   // Add any custom --modinfo strings
 374   for (unsigned i = 0; i < s.modinfos.size(); i++)
 375     h.add("MODULE_INFO: ", s.modinfos[i]);
 376
 377   // Get the directory path to store our cached module
 378   string result, hashdir;
 379   h.result(result);
 380   if (!create_hashdir(s, result, hashdir))
 381     return "";
 382
 383   create_hash_log(string("uprobes_hash"), h.get_parms(), result,
 384                   hashdir + "/uprobes_" + result + "_hash.log");
 385   return hashdir + "/uprobes_" + result;
 386 }
 387
 388 /* vim: set sw=2 ts=8 cino=>4,n-2,{2,^-2,t0,(0,u0,w1,M1 : */