count occurrences of events
An example of counting the occurrence of page-faults
Last updated
An example of counting the occurrence of page-faults
Last updated
One of the use cases of HyperDbg is counting the occurrence of different events. However, incrementing and decrementing values from different cores is problematic as you can't add everything using mathematical operators. It is because the result of the computation might not be accurate.
The problem arises because if you want to access a value to add or subtract the value, other cores might do the same operation simultaneously. Each of the cores is computing the new values at the same time, so we might (and will) lose some of the operations as other core's results might replace the current core's computation results.
In order to prevent this problem, you can use interlocked functions in HyperDbg's script engine.
Imagine we want to count the number of times that system process (process id = 4) will access a paged-out memory (or invalid access) leading to a page-fault.
For this purpose, first, we define a global variable.
After that, we'll use the !exception command with the 0xe parameter, which indicates that we want to intercept page-faults and pid 4, which shows that only the system process should trigger this event.
In the event's script, we use the interlocked_increment function to increment the value by 1.
Now we let the debuggee continue its normal execution using the 'g' command.
After some time, we can read the value of the .my_counter global variable using the 'print' command.
You can see the final results:
