Memory Dumps as Posets

Last week I was comparing the existing collection of memory dump analysis patterns to the collection of trace analysis patterns (in formation) in the search of isomorphism (or more correctly, general morphism) similar to Missing Component pattern. It is not a coincidence that such pattern pairs can be formed. For example, it is possible to discern deadlocks from both crash dumps and software traces (if appropriate information is available there). Fundamentally, it is implied by the definition of a software trace as some sort of a memory dump. And we can see traces in memory dumps too, for example, Execution Residue pattern. Because raw stack data resides in stack pages and in contemporary operating systems they are created from zero pages (metaphorically, out of the void) we can say that stack regions of threads are sorted by their creation time, for example, in this process user memory dump:

0:017> !runaway 4
 Elapsed Time
  Thread       Time
   0:49c       0 days 5:16:31.076
   4:4d8       0 days 5:16:30.967
   3:4d0       0 days 5:16:30.967
   2:4cc       0 days 5:16:30.967
   1:4c8       0 days 5:16:30.967
   5:4e8       0 days 5:16:30.936
   6:b6c       0 days 5:16:15.695
   7:b70       0 days 5:16:15.679
   9:b88       0 days 5:16:15.586
   8:b84       0 days 5:16:15.586
  11:348       0 days 5:16:12.934
  10:bfc       0 days 5:16:12.934
  12:1200      0 days 5:15:16.528
  15:1298      0 days 5:15:15.220
  14:1290      0 days 5:15:15.220
  13:128c      0 days 5:15:15.220
  17:12e4      0 days 5:15:13.257
  16:12dc      0 days 5:15:13.257
  18:12ec      0 days 5:15:13.117
  20:12f4      0 days 5:15:13.085
  19:12f0      0 days 5:15:13.085
  21:17a0      0 days 5:13:16.321
  22:1628      0 days 5:13:15.729
  24:1778      0 days 1:35:50.773
  23:17ec      0 days 1:35:50.773
  25:1570      0 days 1:27:54.190
  26:1724      0 days 1:27:10.151
  27:1490      0 days 0:05:46.732
  28:1950      0 days 0:02:28.153
  29:19b4      0 days 0:00:58.108
  30:177c      0 days 0:00:38.358
  31:1798      0 days 0:00:23.351
  32:1a7c      0 days 0:00:08.343

If we have complete memory dumps we can also account for other processes and their elapsed time. Within stack pages we have partial stack traces but do not have exact timing information between them except for stack frames from the current frozen thread stack trace or, if we are lucky, from a partial stack trace from the past execution. However, the timing between frames from different stacks is undefined and we can only guess it from higher level considerations like semantics of procedure calls and other information.

These considerations and the notion of a poset (partially ordered set) let me thinking about memory dumps as posets. I even created my interpretation of POSET abbreviation for this occasion:


Partially Ordered Software Execution Trace   

- Dmitry Vostokov @ -

Leave a Reply

You must be logged in to post a comment.