In my “Tutorial: Catching Rogue Memory Accesses with Eclipse and GDB Watchpoints” I have used Eclipse/CDT and GDB watchpoints. I used a conditional watchpoint, but this comes with a performance hit. In this article I show how to use the ARM Cortex trace hardware to catch specific writes to a memory location. Without severe performance degradation. But for this I need a little helper: the DEADBEEF catcher!
0xdeadbeef catcher
Eclipse is great: it gives me the tools and capabilities to solve the really hard bugs to find. An example of that ‘hard’ category are ‘rogue’ memory accesses: something in the application is accessing an unwanted memory location and corrupts the data. This might be very sporadic, or takes a long while until it happens. With normal ‘stop-mode’ debugging (setting a normal breakpoint) and stepping usually won’t let me find that bug, as it might be coming from a pointer somewhere. Maybe from an interrupt routine. Or maybe an unitialized or corrupted pointer corrupts to my memory. Usually all what I know is the memory adddress of the data, maybe what is written, but not what or who is writing to that location.
In this article I’m using one of the ‘less-known’ debugging techniques available in Eclipse and CDT and how it works: watchpoints!
Watchpoint with Condition
In this article I’m using one of the ‘less-known’ debugging techniques available in Eclipse and CDT and how it works: watchpoints!
