Weekly Subject -07- BIOS

Just reread the last post that DrMDJ wrote on BIOS. I brought the following forward for just one reason. Digest this and the various relationships discuused and you will be able to improve your troubleshooting of your computer. I would suggest copy and past this to your "Computer Tips" folder in My Documents.

I wish to reiterate something once again. If you have a question on any part of the material presented do ask for clarification. We work at all levels in these threads and has said before this is important to do this.

We not only want to help you in your understanding but challenge you as well. As each and everyone of us gain in knowledge we can review one of the posts on these threads on a particular subject and understand even more that we did the first few times we read it.

"I'd like to clarify a couple things based on some things in (or that I thought were in) some earlier posts. The bios chip contains code and data, that's it. It has no execution or processing capabilities. The bios is dependent on the CPU in the system to execute the code in it. The bios was referred to as the "kickstarter" of the system. And that's basically a fair statement, in a sense. Its code is the first real code that gets executed, the functions performed by this code are critical to getting the system off and running. However, I think it's worth noting that the CPU is and must be active before (and in order for) the bios code can do its thing. Not to go in to great detail, but... When the system is first powered on there is a communication between the motherboard and the power supply to establish there is proper power to (certain) things. The motherboard chipset then sends a signal to the CPU essentially saying "go to work". The processor then starts executing code at a pre-set, pre-determined, memory location, which happens to be in the bios code area. So what this means is that if there is a problem with the CPU (or anything that could effect it's operation initially) things will stop before the bios (or much else) is even in the picture. If one sees indications that the bios code is executing, then you at least know the CPU was functioning. Now it can be rendered non-functional by some bios setting when it kicks in, but… Also, the CPU never relinquishes control to the bios, so the bios never returns (or turns over) control to the CPU. I don't mean to nit-pick. I just wanted to make sure this was clear, as I think it has importance from an overall understanding and a diagnostic standpoint.

Another thing I'll bring up here is something I mentioned before. Namely that the system bios is not (generally/frequently) the only code, only bios code, executed during system boot up and initialization. All the functions to carry out the necessary actions/tasks are not within the bios' capability to provide. It therefore hands things off to other code/bioses. As previously mentioned, the video card/adapter is a good example (one of the most visible as well). Other's can come in to play as well. So I just wanted to reiterate that it is not only the bios that is executing from the time your system starts booting until you see (say) the "Starting Windows" messages.

It might be easy to conclude the way it is often referred to that the BIOS is some "all knowing, all powerful" thing. That it has some ultimate knowledge and control over all that goes on in a PC, or at least from a hardware standpoint. This is not the case. It is not my intention to diminish the importance of the bios in saying this. The bios is indeed very important. It is critical for certain functions, certain aspects of PC operation. But the truth is that it has (by design) a somewhat limited role, and is not terribly sophisticated. It only does (is responsible for) certain (again, as its name implies) basic things. The bios is an integral tie/interface to some of the hardware. But there are a number of components/hardware in a PC the bios knows nothing about, doesn't "talk to" (communicate with) or control, doesn't know how to talk to or control. In some cases there may be some level of knowledge, communication or control, but it's very rudimentary. In some cases the bios can or must be "gone through" to communicate with certain hardware, in other cases it can and is (frequently) completely bypassed. And there are a lot of operations that go on in a PC that the (system) bios has nothing to do with. These are handled by other hardware, other bioses, other software. The examples are too numerous and it would take to long to go in to them here, so I won't. My reason is bring this out is simply to make people aware that the bios is not in the loop in the case of all hardware issues. Sometimes it is, sometimes it isn't. This can be worth knowing when troubleshooting.

Related to the above, and just so people are clear... The code in the bios is not just "in the loop" during boot up (initialization). Some portions of the code are (may) only run at this point. But there are other portions of the bios that are or may be used throughout the time a PC is powered on an in use. Some portions are always used beyond boot up, some are used sometimes or optionally. Again, something worth noting from a troubshooting standpoint.

I referred to the bios above as simple, basic, unsophisticated. It is. And there's a reason I point this out, and a reason to should be kept in mind. Namely, because it is this way (simple, etc) it means it does exactly what it's told, relies on exactly what it's told and "sees", isn't equipped to figure a lot out, isn't very good or communicative when it comes to handling problems. Call it dumb if you want. But the point is that when setting parameters or providing information to the bios, when hooking/setting up hardware that the bios does deal with, one should be somewhat careful, mindful, deliberate. The bios (99% of the time) is going to accept what you give/tell it and go on its merry way (or try to). And if there's a problem, it's not going to correct it, try and figure out what you meant or what should be. If a problem is not too severe, not fatal, it may just go on and function (with the incorrect setting/info/setup). If it is severe enough, the bios may simply just bring everything to a grinding halt, perhaps (likely) throwing out some very general, non-specific, non-descriptive error message. Or it may resort to its ultimate cop-out and restart (reboot). No, the bios isn't very smart. That means we have to be when dealing with it."


------------------
Seek knowledge and all else will follow


Please post back results - Press Ctl D to bookmark

Information

[This message has been edited by greengoose1 (edited 07-01-2001).]

I will edit this thing. Sorry for the Interuption......LOL

[This message has been edited by Dodge (edited 07-01-2001).]

Dodge, you sure interrupted my "asimilating" I was in the middle of. https://discussions.virtualdr.com/

------------------
Seek knowledge and all else will follow


Please post back results - Press Ctl D to bookmark

Information

I'm having trouble assimilating that interruption.... https://discussions.virtualdr.com/

Found this info on flash chips that will add and expand what we have already.

"Flash chips are a relatively new type of non-volatile memory chip that can be erased and reprogrammed without being removed from the system they are used in. Flash chips are sometimes referred to as flash ROMs or flash EEPROMs (Electrically Eraseable Programable Read Only Memory). While this may be acceptable usage, technically flash memory is a distinct and different type of technology.

Flash chips began to be used on PC motherboards in the early-to-mid 1990's to store the system ROM BIOS. This allows motherboard manufacturers to make BIOS updates available electronically (such as through FTP sites, web sites, and BBSs or on diskette), instead of requiring chips to be removed and replaced.

There are many types, packages, and capacities of flash chips made by various manufacturers. The ones that we have seen most commonly used on motherboards are 32-pin PLCC chips, specifically the Intel N28F001BXT. This chip is about 5/8" by 1/2" in size.

A functionally equivalent chip in a DIP package, the Intel P28F001BXT, is also commonly used on PC motherboards.

The Intel N28F001BXT chip has a capacity of 128KB (1 megabit). It is divided into several sections or "blocks". The Main Block is 112KB, the Boot Block is 8KB, and two Parameter Blocks are 4KB each. Programming voltage of 12 volts must be enabled in order to erase the chip, which must be done before it can be programmed.

The boot block is a protected area that is not erased when the main blocks of the chip are reprogrammed in the system. The boot block normally contains special code known as "boot block code" that can be activated to restore a BIOS into the chip if the BIOS code has become corrupt or if the BIOS has been erased or if the chip has been accidentally reprogrammed with the wrong BIOS code.

Here is a diagram of the Intel N28F001BXT:

System address: _______________
FFFFF | | 8KB Boot Block
|______________ |
|______________ | 4KB Parameter Block
|_______________| 4KB Parameter Block
FBFFF | |
| |
| |
| 48KB |
| |
F0000 | _________ |
EFFFF | | 112KB Main Block
| |
| |
| |
| 64KB |
| |
| |
| |
| |
E0000 |_______________|

On many systems, address line A16 to the flash chip is inverted, which causes the chip to be addressed as if the top and bottom halves are interchanged as is shown below:

Non-inverted Inverted
_______________ _______________
FFFFF | |8K Boot Blk FFFFF | |
|______________ | | |
|______________ |4K Parm Blk | |
|_______________|4K Parm Blk | |
| | | | 64K BIOS
| |48K BIOS | |
| | | |
| | | |
| | | |
F0000 |_______________| F0000 |______________|
EFFFF | | EFFFF | | 8K Boot Blk
| | |______________|
| | |______________| 4K Parm Blk
| | |______________| 4K Parm Blk
| |64K BIOS | |
| | | |
| | | | 48K BIOS
| | | |
| | | |
E0000 |_______________| E0000 |______________|

When a PC is powered on, the 80x86 CPU loads its CS:IP registers with F000:FFF0, which tells it where to go from there. Normally, the next step is to copy the BIOS code into RAM (known as shadowing). On most newer PCs, the system ROM BIOS is always shadowed. It is much faster to access the BIOS from RAM than to access it from a ROM chip or flash chip. Also on systems with flash BIOSes, the BIOS code may not be contiguous in the flash chip and may also be compressed, so that it is not ready to do its job until it has been copied to RAM.

The boot block code can be implemented in various ways. In the inverted model above, the two halves of the chip can be reversed with a jumper or switch on the motherboard so that the boot block code appears at the top and will load into memory and execute instead of the system BIOS. Normally the boot block code consists of a small program that will erase the flash chip and reprogram it with code on a diskette in drive A:.

On some systems the flash chip is not ever inverted and the boot block code which appears at the top of the chip will run on every boot and will check to see if the BIOS is corrupt or missing. If the BIOS is OK, the boot block code jumps to the appropriate location to let the BIOS take over. If the boot block code finds that the BIOS is not functional, it should then attempt to reprogram the chip using the contents of a diskette in drive A:.

Some boot block code will program a chip with whatever happens to be on a diskette. Many computer users accidentally reprogram their flash chips with DOS boot disks and various other diskettes. The diskette that is used for the boot block recovery procedure (known as a BIOS recovery disk) must be in a particular format, which varies from system to system. Some recovery disks use a non-DOS format, so that they may not be accessible under DOS or Windows. Some boot block programs will verify that a diskette is a valid recovery disk before proceeding. Some boot block programs will turn control over to a program on the recovery disk.

Flash chips can go bad just as any other type of chip can. Normally flash chips are socketted, in which case they can be easily replaced, but on many motherboards (specifically Intel motherboards) they are soldered on, making replacement difficult and impractical. Some people are concerned about the fact that flash chips can be written to a limited number of times. The standard flash chip used on PCs can be written to 100,000 times, so this really is not a reason for concern. A flash chip can be defective in such a way that it may function just fine with the system's BIOS code being read from it on every boot but will not allow erase or new code to be programmed into it. If it cannot be erased then it cannot be reprogrammed. Sometimes a defective flash chip will erase but cannot be reprogrammed.

Most motherboard manuals and PC system manuals do not adequately (or even at all) address the issues of BIOS updates and BIOS recovery disks. These issues are the responsibility of the motherboard manufacturer and/or system vendor. Many computer technicians are not familiar with issues concerning flash BIOSes and will assume that a motherboard is dead and must be replaced when it really just needs to have a BIOS flashed in with a recovery disk. Many system vendors are not equipped to deal with these issues. It may sometimes be useful to consult with the manufacturer of the motherboard if the system vendor is unable to help. Micro Firmware is able to provide assistance only on motherboards for which we have developed BIOS upgrades. We make BIOS recovery disks for these motherboards available to anyone who needs them through our BBS and FTP site. Note that companies such as AMI, Award, and Phoenix, who license BIOS source code to manufacturers, are not able to provide any assistance to end users as they did not write the BIOS as actually implemented on a particular system.

We have Intel N28F001BXT flash chips available for $19.95 plus shipping. Note that the chip will not be of any use unless code has been programmed into it.

It is very easy to damage a PLCC flash chip or its socket when removing or replacing the chip. There should be notches on two diagonally opposite corners of the socket. While it is possible to use various items such as a strong paperclip to remove the chip, it is advisable to use the correct tool, a PLCC chip extractor, to avoid damaging the socket or the chip. It is also possible to insert a flash chip into the socket backwards. This will normally result in the chip being destroyed when the system is powered on.

Here are some part numbers and sources for PLCC extractors:

AMP 821903-1

OK Industries EX-5

Radio Shack - #276-2101

The AMP and the OK tools are available from Digikey at 800-344-4539

--------------------------------------------------------------------------------

Detailed information on flash memory can be found in a 2-volume set of books from Intel titled "Flash Memory: Volume 1" and "Flash Memory: Volume 2". These books are published by McGraw Hill, whose phone number is 800-722-4726.

Much of the information contained in these books is also available on Intel's web site at this URL:
http://developer.intel.com/design/flash/ "



------------------
Seek knowledge and all else will follow


Please post back results - Press Ctl D to bookmark

Information

God help me.
Please God, help me.
Please help me understand?

I told Greengoose1 I sometimes wish someone get the foot out of my mouth, and that it should be clear what I am writing ain't gospel.

There simply is no way anyone can cover it all.
But lordy, there is such nice interaction here, and it gets to a point where just about everything gets a mention.

Having said that, most of what is being said here is rather general information, NOT pertaining to a particular piece of hardware.
It should therefore also be clear an end user has to find the info relating to the kit in use, and these general tips can help. One gets to know what to look at/for while searching for a solution, and what to watch out for/beware of.

One also gets the impression you cannot ever take anything presented here at face value.
That's basically true, and you shouldn't.
Most of the things are Generally true, though.
And yet there are a few contradicting things found here, and they are true.
VIEW A is correct for machine A
and view B is correct for machine B.

Therefore, when you see contradicting views, don't despair.
Going over this topic and assimilating the things said here, one gains an insight.
It will one day spark you to find a solution to a problem, and armed with a general understanding of what to look for, you will find more relevant information on the kit you have.

That's after you got the manual, and read the manual and the online help pertaining to your kit.
Whatever you find on the CDROM that came with your kit, it usually has some in depth information as well, and it's nearby. The WWW is your next great resource, it's where you found VDr, of course. https://discussions.virtualdr.com/

Please try avoid doing things you cannot undo.
And take notes.

------------------
There's Always "something entirely different"

Kind regards, Jaak.

[This message has been edited by jtdoom (edited 07-02-2001).]

Quote:

---

Originally posted by Jynne62:
Hi, Just a question ... if the Bios is normally in a ROM chip and it starts up first why would it matter if it was copied to RAM each time the computer is booted? RAM is faster but once it is booted (with the ROM chip) isn't that all that is needed, or is the bios actually accessed over and over during the time the computer is on?

I hope this makes sense. Trying to understand this, is like reading Greek.
https://discussions.virtualdr.com/

Thanks

---

yes, it makes a difference.
BIOS is Basic In Out, and it contains the instruction set that makes the components on the motherboard interact.
When you load it into RAM, the instructions for this interaction get loaded in (lets say) PC133 @ 133Mhz memory, and access to it is FAST.
when it is not shadowed, the instruction set is located in ROM @4.7Mhz

exerpt from http://www.firmware.com/support/bios/shadow.htm.

Quote:

---

Shadowing refers to the technique of copying BIOS code from slow ROM chips into faster RAM chips during boot-up so that any access to BIOS routines will be faster. DOS and other operating systems may access BIOS routines frequently. System performance is greatly improved if the BIOS is accessed from RAM rather than from a slower ROM chip.

A PC system's ROM BIOS is stored in a ROM chip on the motherboard. On an older motherboard, the BIOS will reside in one, two or four EPROM chips. On a newer motherboard, the BIOS will probably be kept in a flash chip. In either case, the chip is accessed 8 bits at a time, while RAM on a 486 or 386DX system is accessed 32 bits at a time. Also the access time of ROM chips is slower - 150ns to 200ns compared to 60ns or 70ns for RAM.

---

todays PC133 RAM runs at about 5 nano, and DDR at less than 3 nano.
that's at least 4*30 or 120 times faster...

I have a Question.
What speed is the Video BIOS nowadays?
(let's say on an AGPx4 card)

found something, but it still don't fully answer it http://www.monitorworld.com/faq_pages/q43_page.html
exerpt>>

Quote:

---

Should I have BIOS shadowing on?

The code which tells the computer how to access the video card is stored on the video card itself in a ROM (Read Only Memory) chip. When the computer wants to access the video card, it uses the video BIOS (Basic Input/Output System) routines on the ROM chip(s). The only real problem with this is that ROM chips run more slowly that traditional DRAM which is used for main system RAM. As a result, most (if not all) modern BIOS setup utilities (sometimes referred to as CMOS) allow the video BIOS to be copied to a section of main system DRAM (this is the shadowing). This has the benefit of speeding up video operations between the CPU and video card because the video BIOS 'instructions' can be read more quickly from the shadow RAM, and the disadvantage of using a relatively small block of upper memory (the chunk of memory is located above 640k and below 1 Meg).

---

------------------
Isn't There Always "something entirely different" we'll need to reconsider?
Kind regards, Jaak.

[This message has been edited by jtdoom (edited 07-02-2001).]

Oh my! https://discussions.virtualdr.com/ So much for the "intro" course. Who's gonna be the first to post a machine code listing of their bios? https://discussions.virtualdr.com/



------------------
Bookmark your post to make it easier to find.
And remember, please post back whether the problem is resolved or not.

Good Morning All, DrMDJ since you had time to ask the question, maybe you could do it since the rest of us are still digesting and assimilating what has been presented. https://discussions.virtualdr.com/ https://discussions.virtualdr.com/

Bistro had it right when way back he said something like "when we get into BIOS...."

It would seem we went from the simple to complex but I don't believe it is that difficult but rather a lot to the subject.

So with that in mind maybe we could come up with a simple checklist or troubleshooting procedure that a person can use when a problem comes up. Then if the problem was in BIOS this info could then be utilized. https://discussions.virtualdr.com/



------------------
Seek knowledge and all else will follow


Please post back results - Press Ctl D to bookmark

Information

Repeating and expanding on what I said earlier...

When it comes to shadowing the system bios the issue is not one of whether the system RAM in today's pcs is faster (the access to it is faster) than the memory on the bios chip. The issue is that with today's systems, software and operating systems, after the initial boot process has completed and execution of the OS has started, the bios code is not generally used all that much (or enough to matter). As a result, while there might have been a potential performance gain by having the bios routines in RAM, the gain is generally lost or greatly diminished by the fact that the routines are just not used (or used that much).

Yes, BIOS does stand for the "Basic Input/Output System". But the interpretation of "Basic" as meaning that the bios is at the core (root) of everything or in the loop on everything, or that all roads go through the bios is not correct. It used to be much more the case. But as technology, hardware and software has evolved the (system) bios' role has been greatly diminished (after some of the initial boot processing). Many dealings, many functions that the bios used to be used for (used to needed for) now are handled by other bioses, other firmware and software. Many functions/interactions have to be handled elsewhere because the bios routines aren't equipped to do it or aren't equipped to fully do it (eg. taking advantage of certain hardware features). And Many functions/interactions are handled elsewhere because the bios routines are not efficient, don't (and can't) provide the best performance.

Yes, the bios has evolved (and subsequently grown in size) over time. But it was realized a long time ago that the bios could not possibly be (or continue to be) the "core" of everything, "God" of the PC. When the number of, types of and differences in hardware/devices were fewer, when hardware/devices were simpler, when software (including OSes) were simpler, it was one thing. But given what was, is and will happen in PC technology, to have the bios be god is and would be impractical (and unnecessary). It would require much more code, much more logic, much more sophistication, etc, etc. To have the bios be god is also inefficient. Not just from a size standpoint. It is inefficient from a design standpoint, from a performance standpoint. So with today's PCs, the bios is not god. The bios provides, does certain things. Then other hardware and software (which I use as a general term representing any type of code) do the rest, or do it their way.

I won't belabor the point much more. But I assure you that many, many things are not handled by the bios. As two quick examples... With today's video cards and Windows it is common for video operations to bypass the system bios. And Windows commonly bypasses the bios on disk operations as well. The list goes on. The bios is still very important. It is still in the loop on certain things, or can be. But for the most part, it's most significant role, most significant involvement is (probably) at the time the system is booting when it comes to many of its routines. There was a time when bios was king, even dictator. But now it in many ways it is simply a member of congress. Perhaps a chairperson of a committee, but...

It is for the above reasons that I say that shadowing or not shadowing the system bios on today's systems, running certain operating systems like (say) Windows, using much of the software that is used, is frequently not terribly significant. Now, there is no harm in shadowing it. And people are welcome to experiment to see if it matters on their system.

As for video bios shadowing…

Again, with today's systems and video cards is often of no value to shadow their bioses. And frequently the benefit can be a negative one. Access to the bios code is often faster on these cards than in ram. Furthermore, it is also not uncommon for the OS or drivers to bypass the bios and access the card hardware directly (so it doesn't matter where the bios is). Like with the system bios though, there's little harm is experimenting to see if shadowing makes a difference for you.

Shadowing (in general) is more of a throw-back to days gone by. When hardware and OSes were different.

By the way, bioses generally also offer the option of caching the system and video bioses. The effect of this is to enable the system's (processor's) L2 cache to be used to cache the code of these respective bioses. It should be first noted that in order for this caching to be used one must also have enabled shadowing. Now this being said, I personally advise not enabling either of these. Largely for the same reasons I don't think shadowing is worth it. But also because I believe the L2 cache space and bandwidth is precious. Therefore I don't think it is worth it (from an overall performance standpoint) to risk cluttering up the L2 cache with this stuff. I believe there is more important data and code it should be left for.

------------------
Bookmark your post to make it easier to find.
And remember, please post back whether the problem is resolved or not.

[This message has been edited by DrMDJ (edited 07-02-2001).]

I was just on Capitol Hill yesterday. I can assure you that the BIOS is simply a congressional aide. https://discussions.virtualdr.com/

Politics enters into BIOS? https://discussions.virtualdr.com/ https://discussions.virtualdr.com/

DrMDJ brought up an interesting point on BIOS. In the whole scheme of things concerning a pc its BIOS is a relatively small part. But when you have a problem it can help you. A front door key is a small part of a home until you misplace it. I think it is important to learn to keep the various parts of a computer in a good perspective.

Maybe this. Learn the various components of your computer and their relationships to each other. https://discussions.virtualdr.com/

------------------
Seek knowledge and all else will follow


Please post back results - Press Ctl D to bookmark

Information

DrMDJ confirms what I suspected.
one of the things msinfo32 "told" me a while ago was this...
look in kernel info
you find ebios, extended Bios driver, part of vmm32.vxd w/i the compiled driver...

we also know safe mode in win9x has more DOS compatibility.

Windows memory management is on an other level than when in true DOS. (which explains why some systems keep on running in safe mode, but while running in normal windows modus show memory corruption since advanced cache handling features have kicked in...)

it's gotta have to do with the things DrMDJ points out.
Oh lordy, I'm learning a TON here.
Next thing I'll be shown what to do in a DMI editor, w/i something I heard about, but never investigated since, well, because... Ermmm, well, because I didn't. https://discussions.virtualdr.com/

Greengoose, what do you mean?
BIOS based Diagnostics?


------------------
Isn't There Always "something entirely different" we'll need to reconsider?
Kind regards, Jaak.

Not sure on other cards, but with Nvidia chip cards, it's recommended to turn off the Video Shadowing in the BIOS. Well, just out of curiousity, I tried both ways and benchmarked the card. While there wasn't all that much difference really, with the vid shadowing off, there was a bit more performance benchmarks. Off it goes. As DrMDJ suggested, it won't hurt to experiment a little--just remember the settings you changed. But for the most part, any shadowing is basically a leftover from the BIOS glory days.

GG1,

Like Jaak, I too am a little unclear on the troubleshooting/diagnois question.

------------------
Bookmark your post to make it easier to find.
And remember, please post back whether the problem is resolved or not.