Page 2 of 3
1.44MB exFAT/FAT64 floppy disk?
Posted: Sat May 21, 2016 1:23 pm
by ~
Why not an exFAT floppy?
It's supposed to be different from FAT12/FAT16/FAT32 be it CHS or LBA.
It's also supposed to be possible to call it FAT64 (it's supposed to be 64-bit).
It can actually store files bigger than 2 Gigabytes, and there are disks like 1-Terabyte SD cards with exFAT format (although be careful with fake 1TB Micro/SD cards).
Re: The 1.44MB FAT32 floppy disk
Posted: Sat May 21, 2016 2:40 pm
by Techel
NTFS for floppies? Sure....
Re: The 1.44MB FAT32 floppy disk
Posted: Sat May 21, 2016 2:58 pm
by Kazinsal
To hell with it, just put UDF on everything!
Re: The 1.44MB FAT32 floppy disk
Posted: Sat May 21, 2016 5:08 pm
by ~
A floppy image could be a good sample if it could have any filesystem installed and recognized on it as a floppy.
Then it would make things more understandable when dealing with filesystems in their intended devices, but it would serve like an excellent and easy-to-understand tutorial for filesystems.
So as we see, floppy disks are still extremely useful to simplify the understanding of filesystems, translating LBA to CHS, and many other details, as an excellent ad actually the original starting point for mass-storage media.
Re: The 1.44MB FAT32 floppy disk
Posted: Sat May 21, 2016 5:42 pm
by mikegonta
~ wrote:A floppy image could be a good sample if it could have any filesystem installed and recognized on it as a floppy.
However, when used as a USB flash drive image or emulation image properly written code will permit it to be used as a BIOS emulated
hard drive in addition to a BIOS emulated floppy disk, depending on the BIOS setup or BIOS default.
~ wrote:So as we see, floppy disks are still extremely useful to simplify the understanding of filesystems, translating LBA to CHS, and many
other details, as an excellent ad actually the original starting point for mass-storage media.
But not the actual long gone and forgeten disks.
Also, if the PC (or emulator) can boot a USB flash drive the
int 0x13 extensions can be used without CHS.
Re: The 1.44MB FAT32 floppy disk - Win7 update
Posted: Tue Jun 07, 2016 1:28 pm
by mikegonta
I just received the USB floppy drive that I had ordered.
The 1.44MB FAT32 floppy disk works just fine on Win7 (chkdsk shows no errors) and of course it also boots and runs just fine
from the attached USB floppy drive.
Re: The 1.44MB FAT32 floppy disk
Posted: Tue Jun 07, 2016 1:45 pm
by kzinti
Very exciting stuff. Makes me want to go order a USB foppy drive too!
Does it work on 5.25 inch disks too?
Re: The 1.44MB FAT32 floppy disk - Win7 update
Posted: Tue Jun 07, 2016 2:05 pm
by Kevin
mikegonta wrote:The
1.44MB FAT32 floppy disk works just fine on Win7 (chkdsk shows no errors) and of course it also boots and runs just fine from the attached USB floppy drive.
I wouldn't call it "works fine" when it shows the wrong disk size.
Did you try what happens when you try to store more than 1.44 MB on the disk?
Re: The 1.44MB FAT32 floppy disk - Win7 update
Posted: Tue Jun 07, 2016 2:57 pm
by onlyonemac
Kevin wrote:I wouldn't call it "works fine" when it shows the wrong disk size.
Did you try what happens when you try to store more than 1.44 MB on the disk?
If I remember correctly, he marked the non-existent space as "bad sectors". And that's probably what would happen if you tried to use them - the storage device would return an I/O error just like if you try to access a bad sector on any other disk. It would still be interesting to know though - I imagine that it would be pretty system-dependent, with some systems possibly freezing and other returning errors.
On a slightly unrelated note, I can't remember the exact details but I once did something along the lines of formatting a floppy disk with extra sectors per track and storing more data on it, which worked fine. I then, if I remember correctly, formatted it with 18 sectors per track but somehow made Linux believe that it had 20 sectors per track. Linux was able to read and write the disk correctly, even though there were two "missing" sectors per track (it didn't care what physical track the sectors were in, because that is abstracted at the block-device level), but the reported capacity was larger than the disk actually had sectors for and once I got past the end of the real sectors the storage device gave I/O errors for any sectors that didn't exist.
Re: The 1.44MB FAT32 floppy disk
Posted: Tue Jun 07, 2016 5:44 pm
by Kazinsal
Formatting more sectors per track on a floppy was actually more than just a valid method if you did it at the disk level and not just the filesystem level! In the 90s, Microsoft formatted 3.5 inch installation floppies at 21 sectors per track to get 1.68 MB on a disk, thereby reducing the number of floppies needed for larger software considerably.
With even more trickery (like lowering the amount of space between tracks on the disk to get 82 tracks in, and changing the sizes of the sectors on each track) you can get over 1.7 MB on a standard 3.5 inch high density floppy disk. IBM introduced a completely non-compatible format called XDF that pushed the formatted capacity of a floppy to 1840 KiB*. They technically have the space for 2,000,000 bytes of storage on them, but if you just used them raw like that you'd have no way to seek to specific sectors and tracks since those 2,000,000 raw bytes include the space for synchronization information and inter-sector gaps.
* IBM's proprietary XDF distribution format took advantage of the fact that you can technically format different sectors on a single track with arbitrary power-of-two sizes. With 12,500 raw bytes available on a single track, you can fit the equivalent of 23 512-byte sectors onto the track by formatting it as four sectors: the first is 8192 bytes, the second 2048, the third 1024, and the fourth 512. You can do this in the sector ID gaps, but you can't do this in the track header, so the track header was programmed to think there were 39 128 byte sectors instead. This gives you 11776 bytes per track, and with two sides of 80 tracks, you get exactly 1840 KiB. More technical details can be found
here.
Re: The 1.44MB FAT32 floppy disk - Linux update
Posted: Wed Jun 08, 2016 8:46 am
by mikegonta
The 1.44MB FAT32 floppy disk works just fine on Linux.
Of course, this is running on a Win64 VM with the physical USB floppy drive mounted.
Re: The 1.44MB FAT32 floppy disk
Posted: Wed Jun 08, 2016 9:27 am
by Techel
I somehow still don't get the point of fat32 on fd
Re: The 1.44MB FAT32 floppy disk
Posted: Wed Jun 08, 2016 12:08 pm
by SpyderTL
Because.... why not?
What's the point of OSDev?
For me, I like the idea of my computer only running code that I wrote, and nothing else. And I've also learned a lot about hardware, Linux, assembler, and even Windows that I didn't know before.
I also prefer flexibility over performance. So having the option to format a floppy drive with any file system that I choose appeals to me. Windows gives you no options, whatsoever.
Not sure about Linux. Never tried to format a floppy with FAT32, or EXT4. Anyone know if this is possible?
EDIT: Apparently, Debian has support for formatting floppies as EXT2.
https://wiki.debian.org/FormatFloppy
Re: The 1.44MB FAT32 floppy disk
Posted: Wed Jun 08, 2016 12:34 pm
by Kazinsal
ext2 is scalable. FAT32 has lower bounds of 32 MB.
This is still a bad hack. A clever hack would be adding support for DMF or XDF. Or something actually useful, like giving MikeOS the ability to use subdirectories.
Re: The 1.44MB FAT32 floppy disk
Posted: Wed Jun 08, 2016 3:10 pm
by onlyonemac
Kazinsal wrote:Formatting more sectors per track on a floppy was actually more than just a valid method if you did it at the disk level and not just the filesystem level! In the 90s, Microsoft formatted 3.5 inch installation floppies at 21 sectors per track to get 1.68 MB on a disk, thereby reducing the number of floppies needed for larger software considerably.
Yes, I know it's a valid method. That's why I was doing it (except I did only 20 sectors per track because one of the floppy drives that I was using them in couldn't read 21 sectors per track (old/worn hardware) and I didn't want to risk damaging my drive with 81 or 82 tracks so I left it at 80 tracks - 82 tracks doesn't change the distance between the tracks; that distance is fixed by the floppy drive's stepper motor, so putting 82 tracks on the disk relies on the fact that most floppy drives can seek beyond track 80 before they reach their physical limits - a lot of old software used to put copy-protection data in tracks 81 and 82 as well because most disk copying utilities would only copy the first 80 tracks). But then i started messing around to see what would happen if the system thought there were 20 sectors per track but there were only 18.