3

I have recently encountered an embedded system with IDE drives that are FAT32 but have >4gb files. It appears to do this by setting the file size in the 32byte directory entry to how many bytes the final cluster uses - instead of being the actual file size. This allows the the files to have arbitrary FAT chains. The downside is the only way to know the file sizes are to stream through the chains, which are enormous.

I have never seen this before. Is this a known standard or variation and does anything on windows or linux support it?

Are there any 'quirky' FAT32 modifications that are more standard that allow files greater than 4GB?

Bibhas Debnath
  • 14,559
  • 17
  • 68
  • 96
  • Does Windows7 read that IDE drive properly? If so it's exFAT, if not, it's some propietary extension. – Dominik Weber Aug 31 '10 at 17:20
  • I found this question to be on topic, and I'd love an answer if anybody can post what operating system does this. I'd also like a hexdump of the directory entry if possible. (Yes I know OP is long gone. If somebody can answer, they can provide.) – Joshua Apr 04 '22 at 03:52

2 Answers2

5

Probably exFAT

gbn
  • 422,506
  • 82
  • 585
  • 676
  • 1
    It would be immediately obvious to somebody looking at an exFAT directory entry that it's not FAT32. – Joshua Apr 04 '22 at 03:56
4

This I think is a complete hack.

I've implemented FAT32 and I've never seen or heard of anything like this. I will be very surprised if there is any support anywhere for this.

  • Because your answer is incorrect and was written after the correct answer was given. – Vinko Vrsalovic Sep 29 '09 at 05:52
  • indeed it is so - exFAT is the answer –  Oct 03 '09 at 08:53
  • It still seems like a complete hack. And its a patented hack to boot! – RBerteig May 28 '10 at 03:15
  • 1
    +1 because it is a HACK. exFAT has 64 bit file sizes and not the "bytes in last cluster" thing. This is NOT compatible and a propietary format. and exFAT is not an extension; it's a new file format, since the core structures are different. – Dominik Weber Aug 31 '10 at 17:22