Warning : previous patch was oviously bad for anything but ATA100 SIS 
controller. Thanks to John Fremlin for the proof-reading.
New patch attached.

John Fremlin wrote:

> Lionel Bouton <Lionel.Bouton@free.fr> writes:
> 
> [...]
> 
> 
>>+static unsigned int capabilities = 0x00000000;
>>
> 
> Can't you make this per dev in some way?


Is there a SIS standalone IDE controller? I'm not aware of such a beast, 
only know integrated chipsets.
Anyhow here's a line of the original sis5513.c :

static struct pci_dev *bmide_dev;

If I'm not mistaken, the driver was not designed for multiple devices. I 
don't want to rewrite sis5513.c (yet), if there's need for such support, 
see end of mail.


> 
> [...]
> 
> 
>>+	if (capabilities && SIS5513_FLAG_ATA_100) {
>>
> 
> Typo?
> 
> [...]
> 
> 
>>+	if (capabilities && SIS5513_FLAG_ATA_100) {
>>
> 
> Ditto?
> 
> [...]
> 
> 
>>+	if (capabilities && SIS5513_FLAG_ATA_100) {
>>
> 
> Aarrgh
> 
> 


Beginners like me should not code stuffed with chocolate!
Can someone pass me a spare brown paper bag please?


>>+		switch(timing) { /*   active  recovery
>>+					  v     v */
>>+		case 4:		test1 = 0x30|0x01; break;
>>+		case 3:		test1 = 0x30|0x03; break;
>>+		case 2:		test1 = 0x40|0x04; break;
>>+		case 1:		test1 = 0x60|0x07; break;
>>+		default:	break;
>>+		}
>>+		pci_write_config_byte(dev, drive_pci, test1);
>>+		return; /* temporary hack for tests */
>>
> 
> Is this the new magic you have added?
> 
> [...]
> 


It's a part of it yes. Other parts should fix the cycle time and the 
compatibility mode too. Less importantly the /proc/sis/ide should be 
correct now.


> 
>>+	if (capabilities & SIS5513_FLAG_ATA_100) {
>>
> 
> Aaah much better
> 


I'll punch one hole in my paper bag.


> [...]
> 
> I think that this driver needs to be tested on something that is not
> ATA 100 :-)
> 


It'll surely fail. How 3 simple chars can ruin your day...
New patch attached should correct the problem.

I've a rewritten sis5513.c in the works. I'll address code "prettiness" 
in this future one.
If there's a need for multiple SIS IDE controller devices I'll put the 
support in this new one too.

LB.