watchpoints

Watchpoints are a debugging facility that monitors a specific memory location or program variable and suspends execution when that location is accessed or modified. They extend the traditional breakpoint concept by focusing on data rather than control flow, enabling developers to pinpoint when or where a variable changes or is read.

Types of watchpoints include: write watchpoints, which trigger when a value is written to the designated location; read watchpoints, which trigger on reads; and access or read/write watchpoints (often called awatch), which trigger on either reads or writes. Some debuggers also support change watchpoints, which trigger when the value changes, regardless of how it happens.

Operation and implementation: Most modern debuggers implement watchpoints using hardware support provided by the processor, such as dedicated debug registers, or through software instrumentation when hardware support is unavailable. Hardware watchpoints are typically limited in number (often a few) and may be faster but not universally available in all environments. Software-based watchpoints can incur higher overhead and may trap on every access to the monitored address, making them slower in performance-critical code.

Limitations and considerations: The number of concurrent watchpoints is limited by hardware capabilities. Performance can degrade if watchpoints trigger frequently. Some memory regions (such as certain I/O or shared memory areas) may behave differently or be unsuitable for watchpoints. Optimized builds, inlining, or aggressive compiler transformations can complicate the use of watchpoints.

In practice: Watchpoints help identify where a variable is mutated, who mutates it, or when a value is read. In GDB, for example, they are accessed with commands such as watch (write), rwatch (read), and awatch (read or write); other debuggers offer similar functionality.