Re: fat_min_cluster_size for FAT32

[Andrew Clausen]

On Thu, Mar 03, 2005 at 12:07:27PM -0800, Oleg Babyak wrote:
> Hello maintainers,
>  
>  Why fat_min_cluster_size() returns 8K for FAT32? Are
>  there problems with cluster size < 8K? What will be
>  wrong if I just change it to 1024/512 for FAT32?

Some versions of Windows, in some circumstances, will ignore the cluster
size contained in the boot sector, and compute it from the partition
size, causing severe data corruption.  So you can't just pick any size
you want.

IIRC, a 4k cluster size is fine for FAT32, but it's usually a bad idea
performance-wise.

Cheers,
Andrew