snap 2.3 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677787980818283848586878889909192939495969798
  1. .TH SNAP 6
  2. .SH NAME
  3. snap \- process snapshots
  4. .SH DESCRIPTION
  5. Process snapshots are used to
  6. save a process image for debugging on
  7. another machine or at another time.
  8. They are like old Unix core dumps but
  9. can hold multiple process images and
  10. are smaller.
  11. .PP
  12. The first line of a snapshot begins with the prefix
  13. ``process snapshot'' and often contains
  14. other information as well, such as creation time,
  15. user name, system name, cpu type, and kernel type.
  16. This information is intended for humans, not programs.
  17. Programs reading snapshots should only
  18. check that this line begins with the specified prefix.
  19. .PP
  20. Throughout the rest of the snapshot, decimal strings are
  21. always right-justified, blank-padded to at least 11 characters,
  22. and followed by a single space character.
  23. .PP
  24. The rest of the snapshot is one or more records,
  25. each of which begins with a one-line header.
  26. This header is a decimal process id followed by
  27. an identification string, which denotes the type of
  28. data in the record.
  29. .PP
  30. Records of type
  31. .BR fd ,
  32. .BR fpregs ,
  33. .BR kregs ,
  34. .BR noteid ,
  35. .BR ns ,
  36. .BR proc ,
  37. .BR regs ,
  38. .BR segment ,
  39. and
  40. .BR status
  41. are all formatted as a decimal number
  42. .I n
  43. followed by
  44. .I n
  45. bytes of data.
  46. This data is the contents of the file
  47. of the same name found in
  48. .BR /proc .
  49. .PP
  50. The format of the
  51. .B mem
  52. and
  53. .B text
  54. sections is not as simple.
  55. These sections contain one or more page descriptions.
  56. Each describes a one kilobyte page of data.
  57. If the section is not a multiple of a kilobyte in size,
  58. the last page will be shorter.
  59. Each description begins with a one-byte
  60. flag.
  61. If the flag is
  62. .BR r ,
  63. then it is followed by
  64. a page of binary data.
  65. If the flag is
  66. .BR z ,
  67. then the data is understood to be zeros,
  68. and is omitted.
  69. If the flag is
  70. .B m
  71. or
  72. .BR t ,
  73. then it is followed by two decimal strings
  74. .I p
  75. and
  76. .IR o ,
  77. indicating that this page is the same
  78. as the page at offset
  79. .I o
  80. of the memory or text
  81. segment for process
  82. .IR p .
  83. This data must have been previously
  84. described in the snapshot, and the offset
  85. must be a multiple of a kilobyte.
  86. .PP
  87. It is not guaranteed that any of the sections
  88. described above be in a process snapshot,
  89. although the snapshot quickly becomes useless when
  90. too much is missing.
  91. .PP
  92. Memory and text images may be incomplete.
  93. The memory or text file for a given process
  94. may be split across multiple disjoint sections
  95. in the snapshot.
  96. .SH SEE ALSO
  97. .IR proc (3),
  98. .IR snap (4).