Re: [PATCH] Merge the changes from siimage 2.4.22-pre9 to 2.6.0-test2

[Marcelo Penna Guerra <eu@marcelopenna.org>]

On Thursday 31 July 2003 12:25, Bartlomiej Zolnierkiewicz wrote:
> What do you mean by "a lot more stable than before"?

I mean it doesn't crash the system during intensive read/write operations. It 
has something to do with limiting the max bk per request to 7.5 (hwif-
>rqsize=15).

> Can you separate your changes from forward-port?

They are really small changes. It's the one I mentioned on the first e-mail 
and this one, that was suggested by Andre Hedrick on another post:

	if(is_sata(hwif))
	{
+		drive->id->hw_config |= 0x6000;
		if(strstr(drive->id->model, "Maxtor"))
			return 3;
		return 4;
 	}

> Are you aware of side effect?
> [ Disabling DMA on ATAPI devices on SiI680. ]
>
> Please consult this change with Alan :-).

No, I'm not. Sorry. I'm look for more information. But dma is working with my 
SiI3112A, everything is stable, so maybe it should be enabled by default for 
this chipset.

Also, there's a change missing in this patch. The PCI id for SiI1210SA is 
missing in 2.6.x.

#define PCI_DEVICE_ID_SII_680      0x0680 
+#define PCI_DEVICE_ID_SII_1210SA   0x0240 
#define PCI_DEVICE_ID_SII_3112      0x3112

Anyway, thank you for taking a look at this.

Marcelo Penna Guerra