[En-Nut-Discussion] NutRegisterDevice().....

PragmaLab info at pragmalab.nl
Fri Apr 11 13:44:45 CEST 2008


Hello Harald,

thanks for your comments. I can imagine the hard work and time it took you
to design and code the phat-file system. Support for FAT32, directories and
long filenames is quite complicated (as I found out myself ;-( ). I think
you did a nice job so far, especially the layers in the code (FAT12/16/32,
volume, util etc) make it easy to understand. Also the use of sectorbuffers
that cache sectors is a step forward, allthough my perception of FAT is that
it is hard to design a good predective caching scheme. 

So far I did optimize the 'PhatCountFreeClusters()' for FAT16 and FAT32,
just by parsing the FAT-table per sector iso loading a sector per
clusterentry. Mounting time decreased from 42 seconds to 10 seconds for a 1
GB SD card (still not acceptable). Combined with the fact that we use the
Mega256 iso the AVR-arm makes mounting and writing too slow. We use a HW SPI
already. But I'm going to look for more optimisations and keep you informed.


> > By the way, SD-HC cards use the SD specifications 2.0. I'm 
> having to 
> > re-design the mmcard.c for that as well. Only SD-HC cards have a 
> > capacity >2 GB
> 
> Not sure about this, but there may be patent problems when 
> using SD Card mode instead of MMC/SPI.
> http://en.wikipedia.org/wiki/Sd_card
> 

Thanks, I will have a look at it. 

> P.S.: Inline quoting is more common in mailing lists (and 
> Usenet) for several reasons:
> 1. It's easier for readers to follow discussion 2. The 
> articles are much shorter

I'm sorry. Since I'm always replying to the digest, I must not forget to cut
the tail as well....
Inline quoting indeed makes the messages much easier to read, especially the
digest.




More information about the En-Nut-Discussion mailing list