Michael Schmitz wrote:
take out one POSTPONED of the linkThe connection has timed out. I am not lucky..I looked at the patch and it appears that we will still need this.
I just had a look as well. IMHO it just forces the FAT to be 16 bit for HDD and 12 bit for floppy which sounds correct (and Atari users need that for very small harddisks or very large floppy disks?).
One thing I don't understand is this: /* If more clusters than fat entries in 16-bit fat, we assume * it's a real MSDOS partition with 12-bit fat. */ if (sbi->fat_bits != 32 && total_clusters+2 > sbi-> fat_length*SECTOR_SIZE*8/sbi->fat_bits) sbi->fat_bits = 12;Please explain - if there are more than 2^16 clusters how can they fit in 2^12 number?
I am wondering if the patch isn't over-complicated - for example, can floppy have more clusters than MAX_FAT12 and still have FAT12? It yes then I don't understand the MAX_FAT12 constant - seems like it's not real MAX_ - or the Atari GEMDOS has different MAX than MS-DOS?
there's a related problem with FAT support: in 2.6, the logical sector size is not allowed to be larger than the MMU pagesize. That causes large partitions to be unmountable from Linux.
oh! What is the MMU pagesize? 4 or 8 kB? It makes the FAT almost unusable (64 or 128 MB partition size if I guess right).
Petr