What is cross-compilation and how do you know a toolchain is for cross-compiling?

Cross-compilation is building code on one machine (the host, your x86 or ARM laptop) that runs on a different architecture (the target, say a Cortex-M microcontroller). It's necessary because the target can't compile its own code (no OS, limited resources) and its instruction set differs from the host's, so the host toolchain must emit machine code for the target rather than for the host, and you then flash that image to the device since you can't run it on the host. You recognize a cross-toolchain by its prefix, a target triple like arm-none-eabi-gcc: arm is the target CPU family, none means no operating system (bare metal), and eabi is the embedded application binary interface; contrast arm-linux-gnueabihf which targets ARM with Linux. Every tool carries the prefix, arm-none-eabi-gcc, -ld, -objcopy, -gdb, -size. The classic mistake is invoking the plain host gcc to build firmware, which produces x86 code that simply won't run on the MCU; you must use the prefixed cross-toolchain and target flags (-mcpu=cortex-m4 -mthumb, etc.).