[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

TOSFS



>Being at it, I made tos_getxattr return correkt values for block size and
>number of blocks (for files). Also, I've implemented a new scheme for
>computing something more sensible for the inode: I'm doing a 32-Bit-CRC
>over the full pathname. At least I get the same inode for multiple stat()'s
>to one file, and all inodes *seem* to be unique on my filesystem. Of course,
>that can't be guaranteed... Drawback: the CRC table takes 1K of memory.
>What do you think?

I've been thinking of a scheme like this for a while, but decided it
wasn't worth implementing unless the inode could be guaranteed unique
instead of just "usually" unique.  However what you've written would
definitely be preferable to the currently implemented method of making
up inodes.

My feeling is that the best inode computation would be to use the
starting sector number of the file as the inode.  But I don't know of
a way to compute the starting sector number without reading the FAT
directly.

cheers,
entropy

--
entropy -- it's not just a good idea, it's the second law.
Personal mail:      entropy@gnu.ai.mit.edu
MiNT library mail:  entropy@terminator.rs.itd.umich.edu
"what do you have against octal?" -jrb