Daniela,

here's my first patch to sis5513.c : this should set the correct timings 
for ATA100 chips (and hopefully say goodbye to heavy fs corruption in 
UDMA modes).

It compiles cleanly but is **not yet tested**. I'm currently waiting for 
~30GB to move from my test drive to a safe location (gory details: 
source DMA drive, destination PIO drive, accross NFS on a 100kKiB :-P 
full-duplex Ethernet).

If there's an obvious mistake please tell me (I've a couple of hours to 
wait for the transfer before testing with and without the patch on a 
"blank" ATA-100 drive).

If others want to have a look or even (I repeat **not yet tested**) try 
this code **on backuped data** they're more than welcome.

If you want to have a quick look at the whole new sis5513.c :

http://gyver.homeip.net/sis5513.c



I'm now on to a more general sis5513.c cleanup :

1/ There is some suspicious code (sis_get_info seems broken to me from 
ATA66 on and config_art_rwp_pio is not yet ATA100-chip friendly)
2/ I'll probably move to per chip types funcs as the current generic 
ones become uglier and uglier at each new sis chipset).


LB