We know interfaces based on "vtable" principle. Once you have a pointer to an object, you can narrow-cast it to an interface and the new object is the same object but is very limited to the interface. I always thought hardware firmware is somewhat similar. For example, for block devices (HDDs or SSDs), this interface is like read, write, status and similar. So driver is a user of such a device interface.
No, not really. Object-orientated programming is unrelated to personal computer hardware and your impression that virtual
calls are relatable to device drivers is misguided. They're completely unrelated.
As it turns out, any storage device has firmware and a special area of its storage marked internal where firmware is saved. Manufacturers release programs that allow to "flash" their specific devices, e.g. by writing new program to its internal space, hidden from the OS.
This is not true. Not all storage devices have firmware - and whatever firmware they have (if any) is not necessarily stored on rewritable flash-storage. ROM chips exist, for example, which are not rewritable.
My question is: on a software level, how do they perform this read-write operations to the "hidden" area of a drive? How dead "COM ports" are related?
If you're referring to firmware updates of modern (post-2004) SATA and NVMe storage devices, then those devices' firmware can be updated using SATA and NVMe's built-in commands.
This is documented in places like and t13.org ATA/ATAPI Command Set - 4
If HDDs work across all OSes, why do firmware upgrade software is only released for Windows? In open source world of linux, what do i need to read to understand "debugging firmware" better?
why do firmware upgrade software is only released for Window
Because Windows is the predominant operating-system used by users of those kinds of hardware. While the firmware can be updated using raw SATA/NVMe commands, you still need a host operating system to run the program that will issue those SATA/NVMe commands. Supposing it costs $100k to build a firmware update for an SSD for Windows and another $100k for Linux (for $200k for both Linux and Windows) - but 90% of all Linux users also run Windows - so why spend $200k for 100% coverage when you can spend $100k on 90% coverage, then spend the extra $100k buying a Ferrari or Tesla Model X P100D on yourself, and blame the users for not booting from a Windows USB stick to upgrade their firmware? (Side note: I chose the latter, and yes, I really do love my Tesla Model X)
You cannot have a program that just magically runs on any computer platform (Windows, BSD, Linux, macOS, QNX, etc) and updates periphial device firmware: it always needs to be a program that can be executed by a host OS (you can argue that UEFI/EFI is a platform-agnostic approach, but in reality UEFI/EFI is still its own platform)
In open source world of linux, what do i need to read to understand "debugging firmware" better?
200mg of Adderall and a pirated copy of IDA Pro.
...or 500mg of Dexedrine and NSA Ghidra.