UEFI

UEFI
In this video from ITFreeTraining I will look at Unified Extensible Firmware Interface or UEFI. Traditionally BIOS performed the role of the initial boot process for a computer; UEFI is replacing BIOS and addresses a number of limitations of the BIOS.

What is UEFI?
0:0:17 UEFI was first developed in 2005. It was designed to replace BIOS. BIOS or the Basic Input Output System has been around since the 70s. There have been a lot of improvements in computing during this time and BIOS has been able to address some of these but not others. UEFI addresses the limitations of BIOS and also adds additional features that were not available in BIOS. The UEFI is a single chip located on the motherboard. You can see in this example, the left motherboard has one UEFI chip and the right motherboard has two. In the case of the right motherboard there are two chips in case one was to fail. The chip can vary in size and shape but generally nowadays is quite small. The UEFI chip contains the software that is used when the computer first starts up. You may also hear it referred to as firmware or even BIOS. Often hardware devices will have software embedded in them which is used to operate the device. For example, a video camera. Think of it as software for hardware. This software for hardware is often referred to as firmware. As the UEFI is software to make the hardware of the computer operate, this is why it is often called firmware. You may also hear UEFI referred to as BIOS. Whilst technically this is incorrect, BIOS has been around for so long that people, especially IT technicians, are just more accustomed to using this name. It may also be called UEFI BIOS. Next, I will take a look at some of the differences between UEFI and BIOS.

UEFI Direct Architecture Support
0:1:53 The first big difference is that BIOS supports only 16bit instructions, regardless of what the CPU supports. UEFI supports the same instructions as the CPU. CPUs on the market today are generally 32bit or 64bit. Since the first Intel CPUs ran in 16bit mode, it made sense for the BIOS to operate on 16bit instructions. However, as CPU’s improved, for backward compatibility reasons, BIOS kept running in 16bit. For a long time, since the BIOS was used for initial start-up and setup this was not a problem. With modern 32bit and 64bit CPUs, the CPU will start in pseudo 16bit mode. This mode allows the BIOS to operate with 16bit instructions. This has a lot of limitations, for example the BIOS will not be able to access all the memory in the computer. The idea behind having a pseudo 16bit mode is to allow BIOS to start the computer up and then switch to either 32 or 64bit mode. Thus, 16bit is designed essentially just to allow the operating system to boot and is very limited in what it can do. UEFI on the other hand, can run code that is the same as the CPU. This allows UEFI to access all the RAM on the computer. UEFI can also run its own software and device drivers without an operating system being installed.

Larger Storage Support
0:3:20 The next big difference is that UEFI supports larger storage devices. UEFI supports storage devices over two terabytes in size. It does this by using the GUID Partition Table or GPT partition table. BIOS uses a master boot record or MBR. MBR has the greatest compatibility since it has been around since the first personal computers were developed. However, MBR has a limit of only being able to address two terabytes of space. You will find however, that some operating systems and BIOS combinations will be able to use GPT drives as data drives and in some cases may be able to boot from them. The difference with UEFI is that it will always support booting from a drive with GPT. BIOS will not always support booting from a GPT drive, it depends on which operating system is running. Linux will generally support it whereas Windows will generally not. Most UEFI will also have backward compatibility options. These options will allow UEFI to use a storage device with an MBR partition.

References
None

Credits

[brucegavin]

Kudos.. very comprehensive explanation. Like all "innovations" the increased complexity leads to increased failure. Those who have had to repair UEFI boot failures and corrupted configurations resulting from Microsoft update baloney, will understand this. BIOS was the rope-and-tire-swing method... all it did was work.

It would be nice to see the author do a second take on UEFI... bug fixing. The steps required for resetting, clearing, repairing and fixing UEFI problems.
BCDEDIT /export is your friend.... make a backup before your UEFI system takes a dive.

[adrianandrews]

FOR INTEREST ONLY: (1) All GPT drives still have an MBR ! You can make a dual boot disk which uses both methods and will boot to Windows 11 on a new PC and Windows 10 on an old one. (2) I have an Intel i3-4130 motherboard running original IBM-PC software and using all the original ports by never switching out of 16 bit mode. (3) I have an 8-BIT BIOS chip from 1978. Because EPROM was so expensive I had to get the BIOS in 512 bytes. That's why early controllers for networks and HDD used to have their own BIOS "extension" chips on board.

[tuyetnhungnguyenthi]

This tells us that enabling CSM increases the booting time, because it allows you to boot via UEFI even if the system has MBR partition.

[adrianandrews]

Just a hobby-horse of mine. I wrote the firmware for an early IBM PC clone.

MBR drives are NOT limited to any capacity. It depends on both the SECTOR size of the drive and the version of FAT used by the operating system. I have devices with 128 bit, 512 bit and 4096 bit sectors and formatted in FAT8, FAT12, FAT16 and FAT32. The smallest is a 40 year old 35 track single sided 5.25" FDD and the largest is a 2023 Seagate drive with 4K sectors.The drive capacities vary from 37KB to 16TB ! All the HDD and USB drives BOOT on a 2015 SuperMicro Dual Xeon server board with unmodified BIOS (latest update).

[raghutr]

outstanding... All drivers to be digitally signed, for secure boot to be successful

[leichenstapler]

that bios ship is labeled as bios but its meant that its the uevi nvram, and this chip doesnt even content the uefi but the uefi boot manager which boots an uefi on an uefi partition of a hard drive. right?