Jump to content

137GB limit - ESDI_506.PDR and other limits


Recommended Posts

I can't speak for 98SE but with Windows ME and the native Esdi_506.pdr 4.90.3000, there is no problem for handling more than 128/137 GB of data on a single disk provided the partitions are smaller than 128/137 GB. I just tested it. See screenshots :

1) Filling the drive with data

2) Result

I've noticed that some people on these forums say that they can use large hard drives without dealing with the 137GB limit. Windows in many cases will say everything is fine. It is only when they have loaded up their drive with a lot of data, and have the most to lose, that all of a sudden their directories turn to gibberish. I have seen a lot of postings in other forums where people have suddenly found their systems to be seriously corrupted. Using small partitions will not solve this problem.

It would seem that despite apparences I might be wrong (I haven't posted earlier, sorry I was very sick in bed for several days). I have found out that many of the tar files are internally corrupted. Their size is alright, they can be copied on other volumes or deleted but their content is crap. I will test again as I believe the damaged files may correspond to the moment I began to simultaneously copy three files at once on the volumes with Total Copy. I did so with the aim of generating heavy fragmententation. I wanted to test the performance of Diskeeper versus Defrag on such volumes.

Link to comment
Share on other sites


I have now erased the files which appeared to be corrupt and it resulted in the apparition of absurd folders and files with gigantic size on the first partition of the same drive as well as the destruction of all the data that was on it. Sorry I could not edit my previous post to add that.

Link to comment
Share on other sites

VIA chose an even higher switchover point which makes their driver faulty.

Could you please give more info about this bug?

Is the bug in VIADSK.MPD? VIAIDE.VXD? Somewhere else? In particular versions or all versions?

Can it be fixed?

Does the bug occur in practice? I don't understand what's wrong with a switchover point of 10000000. Are there hard drives supporting 48-bit LBA but dumb enough to "wrap" when receiving 28-bit commands? Lower switchover points are guaranteed to produce bugs with the right size of hard drive.

The VIA Drives you mentioned only works on some Chipsets and has a Bug that can affect some Hard Drives.

Aha, so it's only some hard drives. Maybe some hard drives are buggy. BTW, have you tested your drivers with every kind of hard drive?

I think the best thing is to not have a switchover hack and simply query the drive for 48-bit support. I think linux's ide.c does this (note: this is probably a good example code to look at). I guess this is difficult to fix when hacking an existing binary though :(

Link to comment
Share on other sites

@rloew:

Thanks for the bit of info :)

There are two approaches to overcoming the 4GB limit. I wrote one that patches the drivers and works with AmigaOS 1.3, 2.0, 3.0 and 3.1 and the X-Surf Card. The Amiga community developed a number of extensions to the driver interface standard. One of these is the Trackdisk64 protocol that provides 64 Bit Byte offsets. Support for larger hard drives was integrated into the AmigaOS 3.5 and above. Many of the files are on Aminet but I am not sure if enough of them are there for a complete solution.

Now if i could only just get a lan card, I'd give it a go.

Just for info, if you could overcome the 4GB limit by your suggestions, what would then be the limit?

Sorry about the off-topic thing here - I will indeed shut-up now :)

Regards,

N.

Link to comment
Share on other sites

VIA chose an even higher switchover point which makes their driver faulty.

BTW, VIA + Western Digital seems to be an OK combination. I have data written from 120 GiB - 160 GiB, 0 - 10 GiB, and a few other areas. All seem fine. Scandisk of either area detects no problem.

Link to comment
Share on other sites

Repeating my previous warning:

I've noticed that some people on these forums say that they can use large hard drives without dealing with the 137GB limit. Windows in many cases will say everything is fine. It is only when they have loaded up their drive with a lot of data, and have the most to lose, that all of a sudden their directories turn to gibberish. I have seen a lot of postings in other forums where people have suddenly found their systems to be seriously corrupted. Using small partitions will not solve this problem.
Dear Eidenk,

Sorry, you were sick the last few days.

Your experiment demonstrates how people can be fooled into thinking that everything is fine while their data is being turned into mush. The corruption is not always obvious and may not even show up until you access some vital data that has been overwritten. I hope that people who read your last two posts will take my warning seriously and take the necessary measures before they lose irreplaceable data.

VIA chose an even higher switchover point which makes their driver faulty.

Could you please give more info about this bug?

Is the bug in VIADSK.MPD? VIAIDE.VXD? Somewhere else? In particular versions or all versions?

Can it be fixed?

Dear Azaqahl,

The bug is in the VIADSK.MPD file. I have only seen one version so far.

I have a Patch for this file. I have not advertised it on my website since no one had shown an interest in it.

Interestingly enough the bug appears only when using hard drives that actually follow the ATA-7 Standard. Many hard drives do not follow the Standard and work with the VIA driver, which is probably why they never found the bug themselves. The Western Digital drives I have tested fall into the latter category. This may change in the future if they decide to correct their non-compliance.

Note: SCANDISK may not detect this problem. Depending how the partitions are laid out, SCANDISK may not access the sector in a way that triggers the error.

Just for info, if you could overcome the 4GB limit by your suggestions, what would then be the limit?

Dear it_ybd,

My approach allows up to 2TB which is also the limit for SCSI drives.

The Trackdisk64 protocol effectively provides 55-Bit Support.

I bought the X-Surf Card for the LAN. The IDE Port was a bonus, and a challenge.

It supports Trackdisk64. I added my protocol to it and added the 48-Bit LBA Support it lacked.

Yes, this is off-topic. You may want to correspond directly with me at rloew@hotmail.com.

Rudolph R. Loew

http://rloew1.no-ip.com

Edited by rloew
Link to comment
Share on other sites

Maybe I will find some time this weekend and try put 160G disk into W98SE box and try all possible combinations of drivers, partitions and tools to bring some more light into the darkness of contradictory statements regarding various tools.

So I've found short time only so here are the results. WD2000JB disk was used.

GA-586HX (Intel 430HX chipset) motherboard with Award BIOS: No boot

GA-6OXE-1 (Intel 815E chipset) + IAA2.3: everything OK

GA-8PE667Ultra2 (Intel 845PE chipset) + IAA2.3: everything OK

Kouwell KW-571A ATA/100 disk controller (With CMD 649 chip): not functional

Windows 98 SE FDISK: no problem with 137GB, cosmetical errors for >=100 GB

Windows 98 SE FORMAT: no problem with 137GB, small cosmetical error for >xxGB?

Windows 98 SE CHKDSK: display correctly

Windows 98 SE SCANDISK (DOS): OK

Windows 98 SE formatting (GUI): OK

Windows 98 SE SCANDISK (GUI): won't start (insufficient memory)

Windows 98 SE DEFRAG (GUI): won't start (insufficient memory)

The test was not extensive, but it seems it is not necessary to replace DOS FDISK and FORMAT utilities, it just may be interesting to modify the display (add one column in FDISK and correct the Formatting xxxxxx MB message).

Short look at PCI controller cards shows the following:

Silicon Image/CMD 646, 648 (ATA/66), 649 (ATA/100) No 48-bit LBA support

Silicon Image SiI0680/0680A (ATA/133): 48-bit LBA supported

Promise Ultra 66: BIOS No 48-bit LBA support, Windows driver: 48-bit LBA supported

Promise Ultra 100, 133: 48-bit LBA supported

HighPoint HPT370 (ATA/100), HPT302, HPT372, HPT374 (ATA/133): 48-bit LBA supported

This seems to be support for Intel disk controllers:

Standard ESDI_506.PDR:

Intel® 82371FB PCI Bus Master IDE Controller (PIIX)

Intel® 82371SB PCI Bus Master IDE Controller (PIIX3)

Intel® 82371AB/EB/MB PCI Bus Master IDE Controller (PIIX4/4E/4M)

Intel Application Accelerator 2.3 has support for

Intel® 82801AA Ultra ATA Controller (ICH)

Intel® 82801AB Ultra ATA Controller (ICH-0)

Intel® 82801BA Ultra ATA Controller (ICH-2)

Intel® 82801DB Ultra ATA Controller (ICH-4)

(IntelATA.mpd version 2.3.0.2160 15-Oct-2002)

Older Ultra ATA Storage Driver 6.20.2008 has support for:

Intel® 82801AA Ultra ATA Controller (ICH)

Intel® 82801AB Ultra ATA Controller (ICH-0)

Intel® 82801BA Ultra ATA Controller (ICH-2)

Intel® 82801CA Ultra ATA Controller (ICH-3)

(IntelATA.mpd version 5.2.2008.0, 20-Aug-2001 - not sure if 48-bit capable)

The following newer Intel IDE controllers are serverd by ESDI_506.PDR only:

Intel® 82801EB Ultra ATA Storage Controllers (ICH-5)

Intel® 82801FB Ultra ATA Storage Controllers (ICH-6)

Probably no support for this newest IDE controller:

Intel® 82801GB Ultra ATA Storage Controllers (ICH-7)

Petr

Link to comment
Share on other sites

Windows 98 SE FDISK: no problem with 137GB, cosmetical errors for >=100 GB

Windows 98 SE FORMAT: no problem with 137GB, small cosmetical error for >xxGB?

Windows 98 SE CHKDSK: display correctly

Windows 98 SE SCANDISK (DOS): OK

Windows 98 SE formatting (GUI): OK

Windows 98 SE SCANDISK (GUI): won't start (insufficient memory)

Windows 98 SE DEFRAG (GUI): won't start (insufficient memory)

Thanks!

Did you try Updated FDISK (from Microsoft) and Tihiy's SCANDISK (1.0) ?

Link to comment
Share on other sites

Did you try Updated FDISK (from Microsoft) and Tihiy's SCANDISK (1.0) ?

Yes, I have tested fully patched system, i.e. with updated FDISK.

No, the only advantage of Tihiy's Scandisk seems to be skinning.

Petr

Link to comment
Share on other sites

Few new observations:

GA-586HX (Intel 430HX chipset) BIOS won't boot with WD2000JB enabled, it freezes during disk detection. If I disable the disk, big surprise, Windows 98 SE will detect it and shows the right capacity! Strange thing - how can Windows find disabled disk?

Maybe the XBIOS is installed on the HDD? I have no idea. Some information is here: http://support.microsoft.com/kb/126855/

One problem - since the disk is disabled, BIOS does not set the right timing of the disk. There may be the solution to limit the maximum LBA by WD LifeGuard tools, I have not tested it.

I have also found an explanation of the scandisk for Windows limit: http://support.microsoft.com/default.aspx?...kb;en-us;184006

The ScanDisk tool included with Microsoft Windows 95 and Microsoft Windows 98 is a 16-bit program. Such programs have a single memory block maximum allocation size of 16 MB less 64 KB. Therefore, The Windows 95 or Windows 98 ScanDisk tool cannot process volumes using the FAT32 file system that have a FAT larger than 16 MB less 64 KB in size. A FAT entry on a volume using the FAT32 file system uses 4 bytes, so ScanDisk cannot process the FAT on a volume using the FAT32 file system that defines more than 4,177,920 clusters (including the two reserved clusters). Including the FATs themselves, this works out, at the maximum of 32 KB per cluster, to a volume size of 127.53 gigabytes (GB).

Dskmaint.dll is 16-bit so probably no chance to overcome this limit.

Petr

Link to comment
Share on other sites

So, from this, I gather am better off using 120gb hdds and below to avoid possible data corruption from scandisk or defrag?

If your BIOS and disk controller driver support 48-bit LBA, you should just use partition size below 136 GB (127 GiB).

Petr

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...