https://github.com/akkartik/mu/blob/master/subx/010---vm.cc
  1 //: Core data structures for simulating the SubX VM (subset of an x86 processor)
  2 //:
  3 //: At the lowest level ("level 1") of abstraction, SubX executes x86
  4 //: instructions provided in the form of an array of bytes, loaded into memory
  5 //: starting at a specific address.
  6 
  7 //:: registers
  8 //: assume segment registers are hard-coded to 0
  9 //: no floating-point, MMX, etc. yet
 10 
 11 :(before "End Types")
 12 enum {
 13   EAX,
 14   ECX,
 15   EDX,
 16   EBX,
 17   ESP,
 18   EBP,
 19   ESI,
 20   EDI,
 21   NUM_INT_REGISTERS,
 22 };
 23 union reg {
 24   int32_t i;
 25   uint32_t u;
 26 };
 27 :(before "End Globals")
 28 reg Reg[NUM_INT_REGISTERS] = { {0} };
 29 uint32_t EIP = 1;  // preserve null pointer
 30 :(before "End Reset")
 31 bzero(Reg, sizeof(Reg));
 32 EIP = 1;  // preserve null pointer
 33 
 34 :(before "End Help Contents")
 35 cerr << "  registers\n";
 36 :(before "End Help Texts")
 37 put_new(Help, "registers",
 38   "SubX currently supports eight 32-bit integer registers. From 0 to 7, they are:\n"
 39   "  EAX ECX EDX EBX ESP EBP ESI EDI\n"
 40   "ESP contains the top of the stack.\n"
 41   "\n"
 42   "-- 8-bit registers\n"
 43   "Some instructions operate on eight *overlapping* 8-bit registers.\n"
 44   "From 0 to 7, they are:\n"
 45   "  AL CL DL BL AH CH DH BH\n"
 46   "The 8-bit registers overlap with the 32-bit ones. AL is the lowest signicant byte\n"
 47   "of EAX, AH is the second lowest significant byte, and so on.\n"
 48   "\n"
 49   "For example, if EBX contains 0x11223344, then BL contains 0x44, and BH contains 0x33.\n"
 50   "\n"
 51   "There is no way to access bytes within ESP, EBP, ESI or EDI.\n"
 52   "\n"
 53   "For complete details consult the IA-32 software developer's manual, volume 2,\n"
 54   "table 2-2, \"32-bit addressing forms with the ModR/M byte\".\n"
 55   "It is included in this repository as 'modrm.pdf'.\n"
 56   "The register encodings are described in the top row of the table, but you'll need\n"
 57   "to spend some time with it.\n"
 58   "\n"
 59   "-- flag registers\n"
 60   "Various instructions (particularly 'compare') modify one or more of three 1-bit 'flag'\n"
 61   "registers, as a side-effect:\n"
 62   "- the sign flag (SF): usually set if an arithmetic result is negative, or\n"
 63   "  reset if not.\n"
 64   "- the zero flag (ZF): usually set if a result is zero, or reset if not.\n"
 65   "- the carry flag (CF): usually set if an arithmetic result overflows by just one bit.\n"
 66   "  Useful for operating on unsigned numbers.\n"
 67   "- the overflow flag (OF): usually set if an arithmetic result overflows by more than one bit.\n"
 68   "  Useful for operating on signed numbers.\n"
 69   "The flag bits are read by conditional jumps.\n"
 70   "\n"
 71   "For complete details on how different instructions update the flags, consult the IA-32\n"
 72   "manual (volume 2). There's various versions of it online, such as https://c9x.me/x86,\n"
 73   "though of course you'll need to be careful to ignore instructions and flag registers\n"
 74   "that SubX doesn't support.\n"
 75   "\n"
 76   "It isn't simple, but if this is the processor you have running on your computer,\n"
 77   "might as well get good at it.\n"
 78 );
 79 
 80 :(before "End Globals")
 81 // the subset of x86 flag registers we care about
 82 bool SF = false;  // sign flag
 83 bool ZF = false;  // zero flag
 84 bool CF = false;  // carry flag
 85 bool OF = false;  // overflow flag
 86 :(before "End Reset")
 87 SF = ZF = CF = OF = false;
 88 
 89 //:: simulated RAM
 90 
 91 :(before "End Types")
 92 const uint32_t SEGMENT_ALIGNMENT = 0x1000000;  // 16MB
 93 inline uint32_t align_upwards(uint32_t x, uint32_t align) {
 94   return (x+align-1) & -(align);
 95 }
 96 
 97 // Like in real-world Linux, we'll allocate RAM for our programs in disjoint
 98 // slabs called VMAs or Virtual Memory Areas.
 99 struct vma {
100   uint32_t start;  // inclusive
101   uint32_t end;  // exclusive
102   vector<uint8_t> _data;
103   vma(uint32_t s, uint32_t e) :start(s), end(e) {}
104   vma(uint32_t s) :start(s), end(align_upwards(s+1, SEGMENT_ALIGNMENT)) {}
105   bool match(uint32_t a) {
106     return a >= start && a < end;
107   }
108   bool match32(uint32_t a) {
109     return a >= start && a+4 <= end;
110   }
111   uint8_t& data(uint32_t a) {
112     assert(match(a));
113     uint32_t result_index = a-start;
114     if (_data.size() <= result_index) {
115       const int align = 0x1000;
116       uint32_t result_size = result_index + 1;  // size needed for result_index to be valid
117       uint32_t new_size = align_upwards(result_size, align);
118       // grow at least 2x to maintain some amortized complexity guarantees
119       if (new_size < _data.size() * 2)
120         new_size = _data.size() * 2;
121       // never grow past the stated limit
122       if (new_size > end-start)
123         new_size = end-start;
124       _data.resize(new_size);
125     }
126     return _data.at(result_index);
127   }
128   void grow_until(uint32_t new_end_address) {
129     if (new_end_address < end) return;
130     // Ugly: vma knows about the global Memory list of vmas
131     void sanity_check(uint32_t start, uint32_t end);
132     sanity_check(start, new_end_address);
133     end = new_end_address;
134   }
135   // End vma Methods
136 };
137 :(code)
138 void sanity_check(uint32_t start, uint32_t end) {
139   bool dup_found = false;
140   for (int i = 0;  i < SIZE(Mem);  ++i) {
141     const vma& curr = Mem.at(i);
142     if (curr.start == start) {
143       assert(!dup_found);
144       dup_found = true;
145     }
146     else if (curr.start > start) {
147       assert(curr.start > end);
148     }
149     else if (curr.start < start) {
150       assert(curr.end < start);
151     }
152   }
153 }
154 
155 :(before "End Globals")
156 // RAM is made of VMAs.
157 vector<vma> Mem;
158 :(code)
159 // The first 3 VMAs are special. When loading ELF binaries in later layers,
160 // we'll assume that the first VMA is for code, the second is for data
161 // (including the heap), and the third for the stack.
162 void grow_code_segment(uint32_t new_end_address) {
163   assert(!Mem.empty());
164   Mem.at(0).grow_until(new_end_address);
165 }
166 void grow_data_segment(uint32_t new_end_address) {
167   assert(SIZE(Mem) > 1);
168   Mem.at(1).grow_until(new_end_address);
169 }
170 :(before "End Globals")
171 uint32_t End_of_program = 0;  // when the program executes past this address in tests we'll stop the test
172 // The stack grows downward. Can't increase its size for now.
173 :(before "End Reset")
174 Mem.clear();
175 End_of_program = 0;
176 :(code)
177 // These helpers depend on Mem being laid out contiguously (so you can't use a
178 // map, etc.) and on the host also being little-endian.
179 inline uint8_t read_mem_u8(uint32_t addr) {
180   uint8_t* handle = mem_addr_u8(addr);  // error messages get printed here
181   return handle ? *handle : 0;
182 }
183 inline int8_t read_mem_i8(uint32_t addr) {
184   return static_cast<int8_t>(read_mem_u8(addr));
185 }
186 inline uint32_t read_mem_u32(uint32_t addr) {
187   uint32_t* handle = mem_addr_u32(addr);  // error messages get printed here
188   return handle ? *handle : 0;
189 }
190 inline int32_t read_mem_i32(uint32_t addr) {
191   return static_cast<int32_t>(read_mem_u32(addr));
192 }
193 
194 inline uint8_t* mem_addr_u8(uint32_t addr) {
195   uint8_t* result = NULL;
196   for (int i = 0;  i < SIZE(Mem);  ++i) {
197     if (Mem.at(i).match(addr)) {
198       if (result)
199         raise << "address 0x" << HEXWORD << addr << " is in two segments\n" << end();
200       result = &Mem.at(i).data(addr);
201     }
202   }
203   if (result == NULL) {
204     if (Trace_file) Trace_file.flush();
205     raise << "Tried to access uninitialized memory at address 0x" << HEXWORD << addr << '\n' << end();
206     exit(1);
207   }
208   return result;
209 }
210 inline int8_t* mem_addr_i8(uint32_t addr) {
211   return reinterpret_cast<int8_t*>(mem_addr_u8(addr));
212 }
213 inline uint32_t* mem_addr_u32(uint32_t addr) {
214   uint32_t* result = NULL;
215   for (int i = 0;  i < SIZE(Mem);  ++i) {
216     if (Mem.at(i).match32(addr)) {
217       if (result)
218         raise << "address 0x" << HEXWORD << addr << " is in two segments\n" << end();
219       result = reinterpret_cast<uint32_t*>(&Mem.at(i).data(addr));
220     }
221   }
222   if (result == NULL) {
223     if (Trace_file) Trace_file.flush();
224     raise << "Tried to access uninitialized memory at address 0x" << HEXWORD << addr << '\n' << end();
225     raise << "The entire 4-byte word should be initialized and lie in a single segment.\n" << end();
226   }
227   return result;
228 }
229 inline int32_t* mem_addr_i32(uint32_t addr) {
230   return reinterpret_cast<int32_t*>(mem_addr_u32(addr));
231 }
232 // helper for some syscalls. But read-only.
233 inline const char* mem_addr_kernel_string(uint32_t addr) {
234   return reinterpret_cast<const char*>(mem_addr_u8(addr));
235 }
236 inline string mem_addr_string(uint32_t addr, uint32_t size) {
237   ostringstream out;
238   for (size_t i = 0;  i < size;  ++i)
239     out << read_mem_u8(addr+i);
240   return out.str();
241 }
242 
243 
244 inline void write_mem_u8(uint32_t addr, uint8_t val) {
245   uint8_t* handle = mem_addr_u8(addr);
246   if (handle != NULL) *handle = val;
247 }
248 inline void write_mem_i8(uint32_t addr, int8_t val) {
249   int8_t* handle = mem_addr_i8(addr);
250   if (handle != NULL) *handle = val;
251 }
252 inline void write_mem_u32(uint32_t addr, uint32_t val) {
253   uint32_t* handle = mem_addr_u32(addr);
254   if (handle != NULL) *handle = val;
255 }
256 inline void write_mem_i32(uint32_t addr, int32_t val) {
257   int32_t* handle = mem_addr_i32(addr);
258   if (handle != NULL) *handle = val;
259 }
260 
261 inline bool already_allocated(uint32_t addr) {
262   bool result = false;
263   for (int i = 0;  i < SIZE(Mem);  ++i) {
264     if (Mem.at(i).match(addr)) {
265       if (result)
266         raise << "address 0x" << HEXWORD << addr << " is in two segments\n" << end();
267       result = true;
268     }
269   }
270   return result;
271 }
272 
273 //:: core interpreter loop
274 
275 :(code)
276 // skeleton of how x86 instructions are decoded
277 void run_one_instruction() {
278   uint8_t op=0, op2=0, op3=0;
279   // Run One Instruction
280   if (Trace_file) {
281     dump_registers();
282     // End Dump Info for Instruction
283   }
284   uint32_t inst_start_address = EIP;
285   op = next();
286   trace(Callstack_depth+1, "run") << "0x" << HEXWORD << inst_start_address << " opcode: " << HEXBYTE << NUM(op) << end();
287   switch (op) {
288   case 0xf4:  // hlt
289     EIP = End_of_program;
290     break;
291   // End Single-Byte Opcodes
292   case 0x0f:
293     switch(op2 = next()) {
294     // End Two-Byte Opcodes Starting With 0f
295     default:
296       cerr << "unrecognized second opcode after 0f: " << HEXBYTE << NUM(op2) << '\n';
297       exit(1);
298     }
299     break;
300   case 0xf2:
301     switch(op2 = next()) {
302     // End Two-Byte Opcodes Starting With f2
303     case 0x0f:
304       switch(op3 = next()) {
305       // End Three-Byte Opcodes Starting With f2 0f
306       default:
307         cerr << "unrecognized third opcode after f2 0f: " << HEXBYTE << NUM(op3) << '\n';
308         exit(1);
309       }
310       break;
311     default:
312       cerr << "unrecognized second opcode after f2: " << HEXBYTE << NUM(op2) << '\n';
313       exit(1);
314     }
315     break;
316   case 0xf3:
317     switch(op2 = next()) {
318     // End Two-Byte Opcodes Starting With f3
319     case 0x0f:
320       switch(op3 = next()) {
321       // End Three-Byte Opcodes Starting With f3 0f
322       default:
323         cerr << "unrecognized third opcode after f3 0f: " << HEXBYTE << NUM(op3) << '\n';
324         exit(1);
325       }
326       break;
327     default:
328       cerr << "unrecognized second opcode after f3: " << HEXBYTE << NUM(op2) << '\n';
329       exit(1);
330     }
331     break;
332   default:
333     cerr << "unrecognized opcode: " << HEXBYTE << NUM(op) << '\n';
334     exit(1);
335   }
336 }
337 
338 inline uint8_t next() {
339   return read_mem_u8(EIP++);
340 }
341 
342 void dump_registers() {
343   ostringstream out;
344   out << "registers before: ";
345   for (int i = 0;  i < NUM_INT_REGISTERS;  ++i) {
346     if (i > 0) out << "; ";
347     out << "  " << i << ": " << std::hex << std::setw(8) << std::setfill('_') << Reg[i].u;
348   }
349   out << " -- SF: " << SF << "; ZF: " << ZF << "; CF: " << CF << "; OF: " << OF;
350   trace(Callstack_depth+1, "run") << out.str() << end();
351 }
352 
353 //: start tracking supported opcodes
354 :(before "End Globals")
355 map</*op*/string, string> Name;
356 map</*op*/string, string> Name_0f;
357 map</*op*/string, string> Name_f3;
358 map</*op*/string, string> Name_f3_0f;
359 :(before "End One-time Setup")
360 init_op_names();
361 :(code)
362 void init_op_names() {
363   put(Name, "f4", "halt (hlt)");
364   // End Initialize Op Names
365 }
366 
367 :(before "End Help Special-cases(key)")
368 if (key == "opcodes") {
369   cerr << "Opcodes currently supported by SubX:\n";
370   for (map<string, string>::iterator p = Name.begin();  p != Name.end();  ++p)
371     cerr << "  " << p->first << ": " << p->second << '\n';
372   for (map<string, string>::iterator p = Name_0f.begin();  p != Name_0f.end();  ++p)
373     cerr << "  0f " << p->first << ": " << p->second << '\n';
374   for (map<string, string>::iterator p = Name_f3.begin();  p != Name_f3.end();  ++p)
375     cerr << "  f3 " << p->first << ": " << p->second << '\n';
376   for (map<string, string>::iterator p = Name_f3_0f.begin();  p != Name_f3_0f.end();  ++p)
377     cerr << "  f3 0f " << p->first << ": " << p->second << '\n';
378   cerr << "Run `subx help instructions` for details on words like 'r32' and 'disp8'.\n"
379           "For complete details on these instructions, consult the IA-32 manual (volume 2).\n"
380           "There's various versions of it online, such as https://c9x.me/x86.\n"
381           "The mnemonics in brackets will help you locate each instruction.\n";
382   return 0;
383 }
384 :(before "End Help Contents")
385 cerr << "  opcodes\n";
386 
387 //: Helpers for managing trace depths
388 //:
389 //: We're going to use trace depths primarily to segment code running at
390 //: different frames of the call stack. This will make it easy for the trace
391 //: browser to collapse over entire calls.
392 //:
393 //: Errors will be at depth 0.
394 //: Warnings will be at depth 1.
395 //: SubX instructions will occupy depth 2 and up to Max_depth, organized by
396 //: stack frames. Each instruction's internal details will be one level deeper
397 //: than its 'main' depth. So 'call' instruction details will be at the same
398 //: depth as the instructions of the function it calls.
399 :(before "End Globals")
400 extern const int Initial_callstack_depth = 2;
401 int Callstack_depth = Initial_callstack_depth;
402 :(before "End Reset")
403 Callstack_depth = Initial_callstack_depth;
404 
405 :(before "End Includes")
406 #include <iomanip>
407 #define HEXBYTE  std::hex << std::setw(2) << std::setfill('0')
408 #define HEXWORD  std::hex << std::setw(8) << std::setfill('0')
409 // ugly that iostream doesn't print uint8_t as an integer
410 #define NUM(X) static_cast<int>(X)
411 #include <stdint.h>