设备驱动中MOV CX指令及CX寄存器存双中断类型的疑问

Let's unpack what's happening here step by step—this is classic legacy x86 real-mode interrupt handling, which was ubiquitous in DOS-era device drivers and BIOS routines.

First, Let's Clarify the CX Value #

Your code is setting up the CX register as follows:

• DEVRD = 0000 0100 (04h, overflow interrupt) shifted left by 8 bits becomes 0000 0100 0000 0000, which loads into CH (the high 8 bits of CX)
• DRDWRHL = 0001 0110 (16h, keyboard input interrupt) loads directly into CL (the low 8 bits of CX)
• The end result is CX = 0416h—a 16-bit value packing two 8-bit interrupt type codes.

What's CX Doing Here? #

In legacy x86 system code (like device drivers), CX is commonly used as a parameter register to pass 16-bit data, especially when you need to pack two related 8-bit values into a single register (saves using two separate registers for related data).

In this specific case:

• The high byte (CH = 04h) refers to the CPU's overflow interrupt—this is normally triggered automatically by the CPU when an arithmetic operation overflows.
• The low byte (CL = 16h) is the standard keyboard input interrupt—this is triggered by the keyboard controller when a key is pressed/released.

The Purpose in Device Drivers #

This setup is almost always tied to modifying or extending the Interrupt Vector Table (IVT)—the real-mode table that maps interrupt numbers to their handler functions. Here's the likely intent:

• You're telling the driver to link these two interrupts together—for example, redirecting overflow interrupt handling to reuse the keyboard input handler, or vice versa.
• Another common use case: The driver is registering a custom handler that needs to monitor both interrupts. Packing their type codes into CX lets you pass both to an initialization routine in one go.
• For example, some DOS keyboard drivers would repurpose the overflow interrupt to implement special features (like rapid key repeat, hotkey triggers, or capturing raw keyboard input) by tying it to the standard 16h keyboard interrupt.

Why Pack Them into CX? #

Real-mode x86 has a limited number of general-purpose registers, so packing two 8-bit values into a single 16-bit register is a space-efficient way to pass related parameters. This is a standard pattern in BIOS and DOS system calls, where registers are used heavily for argument passing.

内容的提问来源于stack exchange，提问作者vijaydidmca