thanks MDGx. can you make an unofficial 48bit LBA 4.00.1119 esdi_506.pdr fix for Win95 SR2?

Will do, soon.

I already fixed 4.00.1119 a long time ago, it's here.

At first, I was wondering about the NOPs as well, thinking that they'd presumably be for I/O delay purposes, but as the original code does not have any other instances of I/O instructions separated by so many NOPs, I decided to ignore them, and the code of the patched version ended up having delays between the OUTs in any case. The driver works fine without them anyway.

My driver does send an extra zero to the sector count register, and this is not needed for LBA28 commands, but is for the LBA48, and at the time I couldn't think of a better place to put it. Yes, this does decrease performance for non-48bitLBA commands, but the additional ~30ns delay is unlikely to be noticeable as the actual command execution time is much greater.

Hello everybody.

I've been following the topic with great interest throughout the threads, as I'm a Win98SE addict. I would hate to hijack this thread though, so I'll only say this: if anybody feels able/willing to help me, please PM or IM me. My problem is that the BIOS wouldn't see the whole capacity of the HDD (only sees 136GB out of 160GB) and there's no other place I can get help from - neither from the manufacturer (replies like "mobo's been out of phase long ago") nor from the guys at Wim's BIOS board. Any details in private, to spare the thread from off-topics.

Thank you in advance, and most of all thank you LLXX for your excellent work.

About the only thing you could do then, is get a PCI IDE controller that has 137GB+ support (LBA 48bit). This patch is designed for systems where this hurtle is no longer an issue, but the OS's native support for LBA 48-bit is non-existant, hence the issue at hand.

Here's a discussion that should offer all the needed details. What I was thinking about is a BIOS mod - nothing more, nothing less.

Any commercial solution is out of discussion. I'd rather set up a Linux partition in the unusable space (which is what I most likely will do anyway).

http://www.msfn.org/board/index.php?s=&...st&p=541943

Same situation?

Are you referring to Petr's post or the "wrong gender assumption" one above?
If it's the latter, I apologize - I only found out after posting there. I will correct it.
As for Petr's post, I couldn't say, because he didn't mention what the respective BIOS reported and furthermore, I haven't tried to install anything on that HDD yet. It's brand new and I'm still trying to find the proper solution, which includes trading it in with a smaller 120GB one that would have no problems with the BIOS.
As far as I've read in all threads around that deal with this, all Win9x versions rely on the BIOS' report regarding HDD capacity, unlike NT versions. So I assumed (correctly, probably) that even with the driver patch installed, any Win9x installed on that HDD would mess the primary boot partition if it would have to deal with anything beyond 137GB.
The weird thing is that the BIOS is an AWARD 6.00PG which - at least theoretically - should support 48bit LBA, so I thought it might be an easy job to find a possible bug in it and patch it accordingly. Unfortunately, I'm no specialist so I can't say how hard it would be to do that, or if at all possible. It may even be a SiS630 chipset limitation.
So if no BIOS patching would be possible, I might settle for a partitioning scheme as follows:
1. 30GB FAT32 - Win98SE
2. 30GB FAT32 - Win2003
3. 30GB FAT32 - data
4. 30GB NTFS - data
5. 40GB (EXT or whatever) Ubuntu Linux
I hope the above will work with no problems.
(This board has a weird way of formatting the posts when smileys are involved. Every CR adds and etra LF or something; it's probably an ANSI-Unicode thing, since I'm using Win98SE. I had to edit the post text in notepad and paste it back.)

I was referring to Petr's post... where he used a BIOS that did not support 48bit LBA and it worked fine.

The BIOS shows as 8Gb, but Windows will use the information stored in the MBR. With some partitioners it is possible to create partitions manually and ignore BIOS information, or you can use a disk editor to do it.

What about some older pc's with Bios's that only report 64GB but you've got an 80GB drive? I just got an Intel SE440BX that's like that. Windows sees and uses the whole drive that I FDISK'd and formated in one big partition. To be certain, I first used the Western Digital cd to prepare the drive. When I noticed it hadn't installed the On-Track thing I figured they must know what they're doing and went and prepared it normally with a 98SE startup floppy.

My old HP Pavilion 4430 also did that. Bios would show 64GB max, but larger hard drives worked fine. Not past the 137GB Windows driver limit mind you, but up to that was fine.

It seems similar to the "cosmetic" error in the Microsoft format program that doesn't report the correct size until it's finished formatting.

So are these bios's actually providing the full drive access and the size reported error is cosmetic? Or will we run into problems when we use the sectors larger than 64GB?

Inquring minds want to know.

Windows uses the partition sizes that are reported in the MBR, I can demonstrate this by using disk editor to change the size of the partition to something that exceeds the physical capacity of the disk (e.g. 1.5 terabytes) and then viewing the drive properties under Windows.

Maybe I'll post a screenshot of that on my next boot to DOS...

So, what reports the partition sizes to the MBR? The Bios? If so, then the Bios is showing me 64,440 (something like that) but the MBR somehow knows there are 80GB's there (or, 74 some odd thousand MB).

See what I'm saying? The Bios shows me a total as if it has the 64GB limit, but then FDISK, format, and Windows wind up using and seeing the whole thing.

These old 1997-1999 Bios's seem to behave this way. But is there really a 64GB limit if Windows seems to be operating correctly with using and seeing the whole drive? I figure if there was a real limit then the WD disk would have installed the On-Track Bios to be able to access the whole drive, but this wasn't done and doesn't seem to be needed.

I have one board that did have the 8GB limit, a Chaintech 9ESA. The WD cd installed the EZ-BIOS automatically when it prepared the drive.

Now, in both of my systems that only show up to 64GB in the Bios they both had Bios updates. But neither mentioned anything about dealing with Logical Block Addressing. Maybe they did somehow fix this but that fix only shows the 64GB even though it will report more to the MBR (like the Windows format command).

So, just a cosmetic thing?

OK LLXX, thanks for some clarifications. So, to get things straight once and for all: if I partition that HDD as described in post #158 with the exception that partition #5 would be FAT32 (not necessarily, as there is possibility to access Linux partitions under Windows) and I would use your patched driver in Win98SE, would I be able to access partition #5 from within Win98SE without any problems whatsoever?

I just need to be 101% sure about this before starting messing with the HDD.

[EDIT]
Oh and Eck, you might wanna check Wim's BIOS forum board I linked to in post #156, as there's chance you might find patched BIOSes for your boards. Check the users' signatures, they'll lead you to a page with patched BIOSes. If you can't find your boards listed, you can make a request in the appropriate forum.

Thank you for making it possible, LLXX! Great work! I've got one question though - I've read everything here but I'm only more confused about it: I'm using Win98SE 4.10.2222 A, my original .pdr was v.2222; should I use your 1.0 (2222) or 1.1 (2225) version? Is 2225 your newer version of the 2222 or is it your patch of Microsoft's newer 2225 version? Am I making some terrible mistake using the patched v.2225 instead of my old v.2222 .pdr ?
Thanks,
Max

Well with an AWARD BIOS Revision 6 your mobo must be a fairly recent one. I have got this BIOS on my Abit NF7S v2 and the patched esdi 506.pdr works like a charm above 137GB in the fourth primary partition under windows ME with my 200GB Western Digital Caviar drive.

If you are in doubt, do like me, change your master drive for a sub 137 GB one and use your big one as slave. Then install the patched esdi and experiment.

I think that it is not possible to create more than 4 primary FAT32 partitions under Windows 98SE. If you want to have 5 partitions, I think you'll need to create extended partitions within one of the four primaries.

Someone will correct me if I am wrong about the latter.

Thanks Drugwash. I have tried over there. The SiS5598 SPAX-M board only uses HP Bios's and the Intel SE440BX isn't on any list I've noticed. I've got the last Intel Bios on it now. So I think that unless I order one of those $30 jobs from another site I can't seem to remember now I'm stuck with the manufacturer's releases.

The site I'm talking about makes their own custom Bios's for just about any board made. I just can't remember the name right now. It's late in the day. Even if I did remember I think that if the Wim bios site hasn't made one (which is just an updated manufacturer's bios) then I'd rather just use what I've got. I really don't need full 48LBA, USB 2.0 etc. on these old things.

I just was interested to find out why the bios said one thing but the system didn't seem to care, and uses the larger hard drives fine as far as I can tell.

If everything was fine with .2222, use the patched .2222. The newer versions are for systems that don't work with the original 2222 version.

The version numbers just refer to M$'s original file versions. The 2225 is a patched 2225, etc.

There are several sites with patched Award BIOSes that had the 32 GiB and/or 64 GiB bug, for bug description see http://www.ryston.cz/petr/bios/award.html . And an example of such site is http://wims.rainbow-software.org/

AFAIK Intel uses customized AMI BIOSes, HP I don't know, so these patches are not for you.

Great information including the Bios Patcher is at http://rom.by/ - unfortunately the English version is not so updated.

Regarding the paid BIOS upgrades, maybe this is the site: http://www.unicore.com/biosupgrades/upgradenow.cfm

Petr

Yes that is the site for the paid bios's I was talking about. I'm just not sure I have a bios problem or not. These computers work with 80GB hard drives. It's just that the Bios page only shows up to 64 GB.

I really don't know if there's anything wrong or not. I do know that I've continued to use format from the 98SE Windows Startup floppy and that shows the wrong amount on larger drives yet formats with no problem's. When it's done it shows the correct amount. That's a cosmetic flaw acknowledged by Microsoft in a KB article.

I'm thinking what I'm seeing in these Bios's is the same type of thing. After all, if Windows winds up seeing the whole thing what could be wrong?

I have read most of the available descriptions of the various limits. So far nothing I've read has even mentioned that a Bios could show a smaller amount of space than is actually there, yet all tools used will see and use the whole thing.

It just seems a bit weird that no one has even mentioned encountering this. As far as I can tell, I can use up to the 137GB limit on 9x (with the standard ESDI driver) with a regular IDE hard drive on these computers even though the Bios will read 64GB.

Not that I would. I don't have a need for 137GB of space on that old of a computer. But 80GB is nice to put my mp3's on there. They take up about 18GB. And I like putting full encylopedia's too, so I don't need to insert cd's.

How about a new driver? A complete replacement for Micro$oft file. Faster, smaller, written in assembler? I'm sure there are a lot of improvements to make, especially in speed.

That was my original goal, but it's a bit too much, and I couldn't figure out how to assemble the driver correctly. M$'s is already written in Asm, and although not maximally optimised, it's quite a bit better than some of the other drivers they've written...

Most IDE drivers (if not all) show real name of the IDE HDD in the device manager, just the default Windows driver shows just TYPE47 or TYPE80. Would be possible to add this feature to this driver?

Petr

For Award BIOS problems use BIOS Patcher on a BIOS update from your mobo manufacturer (never use a BIOS dump, it could ruin your system). Unfortunately the site is down from some time ago, here is the cache from the Internet archive.

I've uploaded the required files to rapidshare. Unzip all files to the same folder, put a BIOS image there, patch it with BP-4_23.EXE (use a DOS prompt and select the apporiate parameters), restart to clean DOS session and flash the patched rom-image to your BIOS. For flashing use the version of AwdFlash suggested by your mobo manufacturer.

These are the files:

http://rapidshare.de/files/33482674/BIOS_Patcher_v4.23.zip
http://rapidshare.de/files/33482675/cbrom_v2.07.zip
http://rapidshare.de/files/33482676/LHA_v2.55.zip
http://rapidshare.de/files/33482677/real_microcodes.zip

P.S.
There may be a newer version of the patcher. The 4.23 version was suggested for Award Bios v6.00 when i last visited the site and that's what i use on my systems.

UPDATE:
http://k6-2.narod.ru/ seems to be the new site. Still most links pointing to the old (non-working) domain.

I never had problems using 120 GB on Win 98 SE alias 4.00.2222 a.

2 months ago i bought a new 300 GB HDD. I copied all my AVI-stuff recorded from TV on it and got lost after storing more then 137 GB.

I was searching for drivers to use my new HDD. The shop assistant who sold this HDD told me i have to use Win XP or Linux to use this HDD.

I have to shout out a big thanks to everyone who developed this esdi_506.pdr. It works. I'm planning to buy a second 300 GB HDD soon. because of you ...

Win 98 rulez

Good to know

I tried to create the patch here: http://www.msfn.org/board/index.php?showtopic=83413

Petr

Hello all,

First of all, thanks for making this patch!
I have installed it on my win98se machine using http://www.mdgx.com/files/48BITLBA.EXE and will be testing it on a 250GB drive.

There seems to be a strange side-effect though:
Using a knoppix cd (to get around the 8GB limit of the standard win98se fdisk) I created two primary partitions, the first of close to 200GB, the second using the rest. With the standard ESDI_506.PDR windows recognised both correctly but of course messed up once data is written over the 128GB limit.

With the patched version there seems to be no data mutilation (so far) but the second partition shows up twice in the explorer. The first instance E: (which is the drive letter I'd expect) appears to be unformatted. The second instance gets G: (in between the SCSI cdrom drive F: and the IDE dvd-burner H:). That one is formatted correctly and shows the proper drive label.
I can format E: without breaking data on G:. Writing data to both E and G will mess things up of course.

Has anyone ever seen this effect before? Better still, is there a known solution?

I think that's just your partition table messed up, or you used a different version of the PDR than the one you originally had.

Use the Free FDISK instead.

The setup I used indeed installed the patch for 4.10.2225. I went back to the one from 4.10.2222F.zip attached to the first post in this topic. No improvement there.

Ran Free FDISK, which showed the two primary partitions I had made (though claiming the drive is less than 9GB and claiming that both partitions each used 100% of the disk capacity).
I deleted the second partition, rebooted and then tried to create it again but free fdisk claimed that there was no space left on the drive.
Deleted teh first partition as well, still no luck creating new partitions of decent size with free fdisk (or other partition tools that should be able to do it under windows).

Can you tell me where I can get your latest/greatest version of the patch. The attachment from the first post might be outdated by now.

Thanx for any help you can give!

In the mean time I'm gonna check whether the patch I have now at least fixes the data corruption problem when using just a single 250GB partition.......to be continued

What's the BIOS version and date?

You should try using a newer version of Free FDISK, or the Windows ME FDISK that can be found somewhere in this thread.

since when does the 98se fdisk have an 8gig size limit?

IMHO there is no limit on Win 98 SE FDISK. I used Win 98 SE FDISK on my 300 GB HDD. I simply created 100% extended partition with a single drive of 100%. Everything without any problems.

AFAIK, 98SE's original FDisk executable has a limit of 64GB. However, I've used it on larger drives with no problems creating partitions over 64GB. Perhaps it's just a rumored limit?

That's not fdisk that's his BIOS. See my post above yours?

IMHO, it's got nothing to do with the patch but everything with the formatting(s) you did and yes I did see this once or twice on my machine long time before I used the LLXX patch.

You should backup your data if any, erase your drive and recreate your partitions with a trusted tool IMO.

This should get you straight normally.

Win98SE
Maxtor DiamondMax 10 200 GB IDE
nForce2 Ultra 400 / nFORCE2 MCP chipset
Soltek 75FRN2 mobo

Made a 2nd partition with Maxblast4.

Everything seems to work fine with the install package which included the ME scandisk and defrag. Tried both with no problems, so I guess they were modified by MS to circumvent the limit in the 98SE ones.

Anyway thanks, it's nice being able to use up the full capacity of my 200 gigger now

Now you must fill the second partition with data and see if it screws the first one because what you show here can be done without a patched esdi_506.pdr I think. Only then can you actually applause.

Please follow validation instruction in first post after filled drive with the data exceeds 128Gb. Thank you.

Is this sufficient?

I filled, rebooted Windows, and then scandisked each partition with no errors of any kind.

Windows Me Scandisk and Defrag can work with PARTITIONS bigger than 128GiB, Windows 98 / SE Scandisk and Defrag can work with PARTITIONS smaller than 128GiB only. So for your setup it is not necessary to use Me versions - and if you don't have Windows Me license you should not use them.

Where you got ESDI_506.pdr 4.10.2230? Both in the first post of this thread and on MDGx site are versions 4.10.2225 and 4.10.2226 only.

Petr

Why's that Petr? I don't know about him but I've been using them in 98se for at least 5 yrs now, they're freely available all over the web & superior to the 98se ones

Have you found the download on the Microsoft website?

You can find the download even of the full Windows 98 or Millennium on the web but it does not mean that it is legal to use it. Yes, you can accept the risk and rely on the fact that Microsoft will not take care about 6 years old operating system. But you should be aware of that.

Petr

I stand corrected then. Just tried the win98 version of scandisk and it seems to work. Hmm, interesting.

Not entirely true. It gives me the freedom to resize my partitions to whatever I want now. The sizes I posted are not necessarily the sizes I plan on keeping. That was just for demonstration purposes after I quickly added the 2nd partition.

What I'd like to do is make the Windows one much smaller than it currently is now (common practice), and thus I do need the ME versions of those apps.



With all due respect, not your concern.



It's included in maximus-decim's install package posted on this page...

http://www.msfn.org/board/index.php?showto...78592&st=60

Regards.

C: is not filled enough


And regarding licenses etc. M$ has completely dropped support for 9x/ME and not even making any profits from it so there is no harm done at all from downloading 9x/ME.

Microsoft should have made both Windows 98SE and Windows ME free upgrades to Windows 98 I think, like you have SP1 and SP2 as free upgrades to Windows XP.

I think we can distribute these Windows ME files now to work around some Windows 98 problems, as far as we don't ditribute complete Windows iso's

Maybe we can get MS to support 9x if we can write the WGA validation and Notification tools for 9x?

With all regards due to you, please refrain with such senseless (and dangerous) statements.

A user who sticks to 98SE/ME is someone who does not buy XP or Vista.

I agree that lack of support for 98/me doesnt mean its free.all oses even win95 require you to have a valid key reguardless of age but on the other hand if a 3rd party application improves a os is it ilegal?does it urk Ms when someone prolonges the life of a unsupported os?If theres no money changing hands does microsoft care other then the money lost in not upgrading to xp.telling me constantly that 9x is unsecure on the net doesnt scare me at all in fact i couldnt care aless if i get a virus i reinstall via ghost etc and all my apps are on cd including updates etc.I think these enchancements like 48bitlba and sp2 are fantastic and i hail their creators .Im forever fixing pcs for older family and friends who just wantta surf check email im and all the stuff comming out of the 98 forum is a godsend no sense arguing over 9x vs nt u cant strong arm grandma into luna she wont bend.lol.

Is the wrapping bug dependant on the size of combined data on all partitions or is it instead dependant on data being filled after the 128 GB point on the first partition(s), regardless if that first 128 GB is completely filled?

Or to put it simply...

Is it the > 128 GB partition size or > 128 GB data size that triggers the bug, once data is written after that point?

I feel this is an important point to clarify, thanks.

The bug is triggered when both the sector with address n and the sector with address n+128GiB is written.

Old LBA is 28-bit and can address 128 GiB only, it means that if you try to write address on position 128GiB + 1, the first sector of the HDD is written.

This behavior is not related to partition size or data size, but if you have the first partition 128 GiB, then even few bytes in the beginning of the second partition can overwrite some data in the second partition.

And what may happen? It depends on the fact what you will overwrite. If partition tavble, boot records, file allocation tables or directory entries are overwritten, then the HDD may become unaccessible or scandisk can find the errors. But if just some data are overwritten, scandisk will show no error and the only possibility to detect is to compare original and copied files.

Petr