summaryrefslogtreecommitdiff
path: root/scripts/analyze-migration.py
blob: 0c8b22f2aa9ec5240fe7735c37f955df8613e85d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
#!/usr/bin/env python
#
#  Migration Stream Analyzer
#
#  Copyright (c) 2015 Alexander Graf <agraf@suse.de>
#
# This library is free software; you can redistribute it and/or
# modify it under the terms of the GNU Lesser General Public
# License as published by the Free Software Foundation; either
# version 2 of the License, or (at your option) any later version.
#
# This library 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
# Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public
# License along with this library; if not, see <http://www.gnu.org/licenses/>.

import numpy as np
import json
import os
import argparse
import collections
import pprint

def mkdir_p(path):
    try:
        os.makedirs(path)
    except OSError:
        pass

class MigrationFile(object):
    def __init__(self, filename):
        self.filename = filename
        self.file = open(self.filename, "rb")

    def read64(self):
        return np.asscalar(np.fromfile(self.file, count=1, dtype='>i8')[0])

    def read32(self):
        return np.asscalar(np.fromfile(self.file, count=1, dtype='>i4')[0])

    def read16(self):
        return np.asscalar(np.fromfile(self.file, count=1, dtype='>i2')[0])

    def read8(self):
        return np.asscalar(np.fromfile(self.file, count=1, dtype='>i1')[0])

    def readstr(self, len = None):
        if len is None:
            len = self.read8()
        if len == 0:
            return ""
        return np.fromfile(self.file, count=1, dtype=('S%d' % len))[0]

    def readvar(self, size = None):
        if size is None:
            size = self.read8()
        if size == 0:
            return ""
        value = self.file.read(size)
        if len(value) != size:
            raise Exception("Unexpected end of %s at 0x%x" % (self.filename, self.file.tell()))
        return value

    def tell(self):
        return self.file.tell()

    # The VMSD description is at the end of the file, after EOF. Look for
    # the last NULL byte, then for the beginning brace of JSON.
    def read_migration_debug_json(self):
        QEMU_VM_VMDESCRIPTION = 0x06

        # Remember the offset in the file when we started
        entrypos = self.file.tell()

        # Read the last 10MB
        self.file.seek(0, os.SEEK_END)
        endpos = self.file.tell()
        self.file.seek(max(-endpos, -10 * 1024 * 1024), os.SEEK_END)
        datapos = self.file.tell()
        data = self.file.read()
        # The full file read closed the file as well, reopen it
        self.file = open(self.filename, "rb")

        # Find the last NULL byte, then the first brace after that. This should
        # be the beginning of our JSON data.
        nulpos = data.rfind("\0")
        jsonpos = data.find("{", nulpos)

        # Check backwards from there and see whether we guessed right
        self.file.seek(datapos + jsonpos - 5, 0)
        if self.read8() != QEMU_VM_VMDESCRIPTION:
            raise Exception("No Debug Migration device found")

        jsonlen = self.read32()

        # Seek back to where we were at the beginning
        self.file.seek(entrypos, 0)

        return data[jsonpos:jsonpos + jsonlen]

    def close(self):
        self.file.close()

class RamSection(object):
    RAM_SAVE_FLAG_COMPRESS = 0x02
    RAM_SAVE_FLAG_MEM_SIZE = 0x04
    RAM_SAVE_FLAG_PAGE     = 0x08
    RAM_SAVE_FLAG_EOS      = 0x10
    RAM_SAVE_FLAG_CONTINUE = 0x20
    RAM_SAVE_FLAG_XBZRLE   = 0x40
    RAM_SAVE_FLAG_HOOK     = 0x80

    def __init__(self, file, version_id, ramargs, section_key):
        if version_id != 4:
            raise Exception("Unknown RAM version %d" % version_id)

        self.file = file
        self.section_key = section_key
        self.TARGET_PAGE_SIZE = ramargs['page_size']
        self.dump_memory = ramargs['dump_memory']
        self.write_memory = ramargs['write_memory']
        self.sizeinfo = collections.OrderedDict()
        self.data = collections.OrderedDict()
        self.data['section sizes'] = self.sizeinfo
        self.name = ''
        if self.write_memory:
            self.files = { }
        if self.dump_memory:
            self.memory = collections.OrderedDict()
            self.data['memory'] = self.memory

    def __repr__(self):
        return self.data.__repr__()

    def __str__(self):
        return self.data.__str__()

    def getDict(self):
        return self.data

    def read(self):
        # Read all RAM sections
        while True:
            addr = self.file.read64()
            flags = addr & (self.TARGET_PAGE_SIZE - 1)
            addr &= ~(self.TARGET_PAGE_SIZE - 1)

            if flags & self.RAM_SAVE_FLAG_MEM_SIZE:
                while True:
                    namelen = self.file.read8()
                    # We assume that no RAM chunk is big enough to ever
                    # hit the first byte of the address, so when we see
                    # a zero here we know it has to be an address, not the
                    # length of the next block.
                    if namelen == 0:
                        self.file.file.seek(-1, 1)
                        break
                    self.name = self.file.readstr(len = namelen)
                    len = self.file.read64()
                    self.sizeinfo[self.name] = '0x%016x' % len
                    if self.write_memory:
                        print self.name
                        mkdir_p('./' + os.path.dirname(self.name))
                        f = open('./' + self.name, "wb")
                        f.truncate(0)
                        f.truncate(len)
                        self.files[self.name] = f
                flags &= ~self.RAM_SAVE_FLAG_MEM_SIZE

            if flags & self.RAM_SAVE_FLAG_COMPRESS:
                if flags & self.RAM_SAVE_FLAG_CONTINUE:
                    flags &= ~self.RAM_SAVE_FLAG_CONTINUE
                else:
                    self.name = self.file.readstr()
                fill_char = self.file.read8()
                # The page in question is filled with fill_char now
                if self.write_memory and fill_char != 0:
                    self.files[self.name].seek(addr, os.SEEK_SET)
                    self.files[self.name].write(chr(fill_char) * self.TARGET_PAGE_SIZE)
                if self.dump_memory:
                    self.memory['%s (0x%016x)' % (self.name, addr)] = 'Filled with 0x%02x' % fill_char
                flags &= ~self.RAM_SAVE_FLAG_COMPRESS
            elif flags & self.RAM_SAVE_FLAG_PAGE:
                if flags & self.RAM_SAVE_FLAG_CONTINUE:
                    flags &= ~self.RAM_SAVE_FLAG_CONTINUE
                else:
                    self.name = self.file.readstr()

                if self.write_memory or self.dump_memory:
                    data = self.file.readvar(size = self.TARGET_PAGE_SIZE)
                else: # Just skip RAM data
                    self.file.file.seek(self.TARGET_PAGE_SIZE, 1)

                if self.write_memory:
                    self.files[self.name].seek(addr, os.SEEK_SET)
                    self.files[self.name].write(data)
                if self.dump_memory:
                    hexdata = " ".join("{0:02x}".format(ord(c)) for c in data)
                    self.memory['%s (0x%016x)' % (self.name, addr)] = hexdata

                flags &= ~self.RAM_SAVE_FLAG_PAGE
            elif flags & self.RAM_SAVE_FLAG_XBZRLE:
                raise Exception("XBZRLE RAM compression is not supported yet")
            elif flags & self.RAM_SAVE_FLAG_HOOK:
                raise Exception("RAM hooks don't make sense with files")

            # End of RAM section
            if flags & self.RAM_SAVE_FLAG_EOS:
                break

            if flags != 0:
                raise Exception("Unknown RAM flags: %x" % flags)

    def __del__(self):
        if self.write_memory:
            for key in self.files:
                self.files[key].close()


class HTABSection(object):
    HASH_PTE_SIZE_64       = 16

    def __init__(self, file, version_id, device, section_key):
        if version_id != 1:
            raise Exception("Unknown HTAB version %d" % version_id)

        self.file = file
        self.section_key = section_key

    def read(self):

        header = self.file.read32()

        if (header > 0):
            # First section, just the hash shift
            return

        # Read until end marker
        while True:
            index = self.file.read32()
            n_valid = self.file.read16()
            n_invalid = self.file.read16()

            if index == 0 and n_valid == 0 and n_invalid == 0:
                break

            self.file.readvar(n_valid * self.HASH_PTE_SIZE_64)

    def getDict(self):
        return ""

class VMSDFieldGeneric(object):
    def __init__(self, desc, file):
        self.file = file
        self.desc = desc
        self.data = ""

    def __repr__(self):
        return str(self.__str__())

    def __str__(self):
        return " ".join("{0:02x}".format(ord(c)) for c in self.data)

    def getDict(self):
        return self.__str__()

    def read(self):
        size = int(self.desc['size'])
        self.data = self.file.readvar(size)
        return self.data

class VMSDFieldInt(VMSDFieldGeneric):
    def __init__(self, desc, file):
        super(VMSDFieldInt, self).__init__(desc, file)
        self.size = int(desc['size'])
        self.format = '0x%%0%dx' % (self.size * 2)
        self.sdtype = '>i%d' % self.size
        self.udtype = '>u%d' % self.size

    def __repr__(self):
        if self.data < 0:
            return ('%s (%d)' % ((self.format % self.udata), self.data))
        else:
            return self.format % self.data

    def __str__(self):
        return self.__repr__()

    def getDict(self):
        return self.__str__()

    def read(self):
        super(VMSDFieldInt, self).read()
        self.sdata = np.fromstring(self.data, count=1, dtype=(self.sdtype))[0]
        self.udata = np.fromstring(self.data, count=1, dtype=(self.udtype))[0]
        self.data = self.sdata
        return self.data

class VMSDFieldUInt(VMSDFieldInt):
    def __init__(self, desc, file):
        super(VMSDFieldUInt, self).__init__(desc, file)

    def read(self):
        super(VMSDFieldUInt, self).read()
        self.data = self.udata
        return self.data

class VMSDFieldIntLE(VMSDFieldInt):
    def __init__(self, desc, file):
        super(VMSDFieldIntLE, self).__init__(desc, file)
        self.dtype = '<i%d' % self.size

class VMSDFieldBool(VMSDFieldGeneric):
    def __init__(self, desc, file):
        super(VMSDFieldBool, self).__init__(desc, file)

    def __repr__(self):
        return self.data.__repr__()

    def __str__(self):
        return self.data.__str__()

    def getDict(self):
        return self.data

    def read(self):
        super(VMSDFieldBool, self).read()
        if self.data[0] == 0:
            self.data = False
        else:
            self.data = True
        return self.data

class VMSDFieldStruct(VMSDFieldGeneric):
    QEMU_VM_SUBSECTION    = 0x05

    def __init__(self, desc, file):
        super(VMSDFieldStruct, self).__init__(desc, file)
        self.data = collections.OrderedDict()

        # When we see compressed array elements, unfold them here
        new_fields = []
        for field in self.desc['struct']['fields']:
            if not 'array_len' in field:
                new_fields.append(field)
                continue
            array_len = field.pop('array_len')
            field['index'] = 0
            new_fields.append(field)
            for i in xrange(1, array_len):
                c = field.copy()
                c['index'] = i
                new_fields.append(c)

        self.desc['struct']['fields'] = new_fields

    def __repr__(self):
        return self.data.__repr__()

    def __str__(self):
        return self.data.__str__()

    def read(self):
        for field in self.desc['struct']['fields']:
            try:
                reader = vmsd_field_readers[field['type']]
            except:
                reader = VMSDFieldGeneric

            field['data'] = reader(field, self.file)
            field['data'].read()

            if 'index' in field:
                if field['name'] not in self.data:
                    self.data[field['name']] = []
                a = self.data[field['name']]
                if len(a) != int(field['index']):
                    raise Exception("internal index of data field unmatched (%d/%d)" % (len(a), int(field['index'])))
                a.append(field['data'])
            else:
                self.data[field['name']] = field['data']

        if 'subsections' in self.desc['struct']:
            for subsection in self.desc['struct']['subsections']:
                if self.file.read8() != self.QEMU_VM_SUBSECTION:
                    raise Exception("Subsection %s not found at offset %x" % ( subsection['vmsd_name'], self.file.tell()))
                name = self.file.readstr()
                version_id = self.file.read32()
                self.data[name] = VMSDSection(self.file, version_id, subsection, (name, 0))
                self.data[name].read()

    def getDictItem(self, value):
       # Strings would fall into the array category, treat
       # them specially
       if value.__class__ is ''.__class__:
           return value

       try:
           return self.getDictOrderedDict(value)
       except:
           try:
               return self.getDictArray(value)
           except:
               try:
                   return value.getDict()
               except:
                   return value

    def getDictArray(self, array):
        r = []
        for value in array:
           r.append(self.getDictItem(value))
        return r

    def getDictOrderedDict(self, dict):
        r = collections.OrderedDict()
        for (key, value) in dict.items():
            r[key] = self.getDictItem(value)
        return r

    def getDict(self):
        return self.getDictOrderedDict(self.data)

vmsd_field_readers = {
    "bool" : VMSDFieldBool,
    "int8" : VMSDFieldInt,
    "int16" : VMSDFieldInt,
    "int32" : VMSDFieldInt,
    "int32 equal" : VMSDFieldInt,
    "int32 le" : VMSDFieldIntLE,
    "int64" : VMSDFieldInt,
    "uint8" : VMSDFieldUInt,
    "uint16" : VMSDFieldUInt,
    "uint32" : VMSDFieldUInt,
    "uint32 equal" : VMSDFieldUInt,
    "uint64" : VMSDFieldUInt,
    "int64 equal" : VMSDFieldInt,
    "uint8 equal" : VMSDFieldInt,
    "uint16 equal" : VMSDFieldInt,
    "float64" : VMSDFieldGeneric,
    "timer" : VMSDFieldGeneric,
    "buffer" : VMSDFieldGeneric,
    "unused_buffer" : VMSDFieldGeneric,
    "bitmap" : VMSDFieldGeneric,
    "struct" : VMSDFieldStruct,
    "unknown" : VMSDFieldGeneric,
}

class VMSDSection(VMSDFieldStruct):
    def __init__(self, file, version_id, device, section_key):
        self.file = file
        self.data = ""
        self.vmsd_name = ""
        self.section_key = section_key
        desc = device
        if 'vmsd_name' in device:
            self.vmsd_name = device['vmsd_name']

        # A section really is nothing but a FieldStruct :)
        super(VMSDSection, self).__init__({ 'struct' : desc }, file)

###############################################################################

class MigrationDump(object):
    QEMU_VM_FILE_MAGIC    = 0x5145564d
    QEMU_VM_FILE_VERSION  = 0x00000003
    QEMU_VM_EOF           = 0x00
    QEMU_VM_SECTION_START = 0x01
    QEMU_VM_SECTION_PART  = 0x02
    QEMU_VM_SECTION_END   = 0x03
    QEMU_VM_SECTION_FULL  = 0x04
    QEMU_VM_SUBSECTION    = 0x05
    QEMU_VM_VMDESCRIPTION = 0x06

    def __init__(self, filename):
        self.section_classes = { ( 'ram', 0 ) : [ RamSection, None ],
                                 ( 'spapr/htab', 0) : ( HTABSection, None ) }
        self.filename = filename
        self.vmsd_desc = None

    def read(self, desc_only = False, dump_memory = False, write_memory = False):
        # Read in the whole file
        file = MigrationFile(self.filename)

        # File magic
        data = file.read32()
        if data != self.QEMU_VM_FILE_MAGIC:
            raise Exception("Invalid file magic %x" % data)

        # Version (has to be v3)
        data = file.read32()
        if data != self.QEMU_VM_FILE_VERSION:
            raise Exception("Invalid version number %d" % data)

        self.load_vmsd_json(file)

        # Read sections
        self.sections = collections.OrderedDict()

        if desc_only:
            return

        ramargs = {}
        ramargs['page_size'] = self.vmsd_desc['page_size']
        ramargs['dump_memory'] = dump_memory
        ramargs['write_memory'] = write_memory
        self.section_classes[('ram',0)][1] = ramargs

        while True:
            section_type = file.read8()
            if section_type == self.QEMU_VM_EOF:
                break
            elif section_type == self.QEMU_VM_SECTION_START or section_type == self.QEMU_VM_SECTION_FULL:
                section_id = file.read32()
                name = file.readstr()
                instance_id = file.read32()
                version_id = file.read32()
                section_key = (name, instance_id)
                classdesc = self.section_classes[section_key]
                section = classdesc[0](file, version_id, classdesc[1], section_key)
                self.sections[section_id] = section
                section.read()
            elif section_type == self.QEMU_VM_SECTION_PART or section_type == self.QEMU_VM_SECTION_END:
                section_id = file.read32()
                self.sections[section_id].read()
            else:
                raise Exception("Unknown section type: %d" % section_type)
        file.close()

    def load_vmsd_json(self, file):
        vmsd_json = file.read_migration_debug_json()
        self.vmsd_desc = json.loads(vmsd_json, object_pairs_hook=collections.OrderedDict)
        for device in self.vmsd_desc['devices']:
            key = (device['name'], device['instance_id'])
            value = ( VMSDSection, device )
            self.section_classes[key] = value

    def getDict(self):
        r = collections.OrderedDict()
        for (key, value) in self.sections.items():
           key = "%s (%d)" % ( value.section_key[0], key )
           r[key] = value.getDict()
        return r

###############################################################################

class JSONEncoder(json.JSONEncoder):
    def default(self, o):
        if isinstance(o, VMSDFieldGeneric):
            return str(o)
        return json.JSONEncoder.default(self, o)

parser = argparse.ArgumentParser()
parser.add_argument("-f", "--file", help='migration dump to read from', required=True)
parser.add_argument("-m", "--memory", help='dump RAM contents as well', action='store_true')
parser.add_argument("-d", "--dump", help='what to dump ("state" or "desc")', default='state')
parser.add_argument("-x", "--extract", help='extract contents into individual files', action='store_true')
args = parser.parse_args()

jsonenc = JSONEncoder(indent=4, separators=(',', ': '))

if args.extract:
    dump = MigrationDump(args.file)

    dump.read(desc_only = True)
    print "desc.json"
    f = open("desc.json", "wb")
    f.truncate()
    f.write(jsonenc.encode(dump.vmsd_desc))
    f.close()

    dump.read(write_memory = True)
    dict = dump.getDict()
    print "state.json"
    f = open("state.json", "wb")
    f.truncate()
    f.write(jsonenc.encode(dict))
    f.close()
elif args.dump == "state":
    dump = MigrationDump(args.file)
    dump.read(dump_memory = args.memory)
    dict = dump.getDict()
    print jsonenc.encode(dict)
elif args.dump == "desc":
    dump = MigrationDump(args.file)
    dump.read(desc_only = True)
    print jsonenc.encode(dump.vmsd_desc)
else:
    raise Exception("Please specify either -x, -d state or -d dump")