[PATCH] 2.5 scsi changes : qlogicfas.c fixed

[PATCH] 2.5 scsi changes : qlogicfas.c fixed

[Brett]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 364 bytes --]


Hey,

Once again, thanks to people messing with the scsi layer, I'm forced to 
attempt to fix this driver, which it seems I'm the only one using :)

Like last time, I cannot confirm that this _works_ as my card is dodgy, 
but I can confirm that it compiles, and performs exactly as is did before 
the changes.

The changes seems simple enough.

thanks,

	/ Brett

[-- Attachment #2: Type: TEXT/PLAIN, Size: 649 bytes --]

--- drivers/scsi/qlogicfas.c.bak	Wed Feb 20 21:23:11 2002
+++ drivers/scsi/qlogicfas.c	Wed Feb 20 21:18:37 2002
@@ -344,6 +344,7 @@
 unsigned int	reqlen; 		/* total length of transfer */
 struct scatterlist	*sglist;	/* scatter-gather list pointer */
 unsigned int	sgcount;		/* sg counter */
+char *buf;
 
 rtrc(1)
 	j = inb(qbase + 6);
@@ -391,7 +392,8 @@
 					REG0;
 					return ((qabort == 1 ? DID_ABORT : DID_RESET) << 16);
 				}
-				if (ql_pdma(phase, sglist->address, sglist->length))
+				buf = page_address(sglist->page) + sglist->offset;
+				if (ql_pdma(phase, buf, sglist->length))
 					break;
 				sglist++;
 			}

Re: [PATCH] 2.5 scsi changes : qlogicfas.c fixed

[David S. Miller]

   From: Brett <brett@bad-sports.com>
   Date: Wed, 20 Feb 2002 22:00:37 +1100 (EST)
   
   Once again, thanks to people messing with the scsi layer, I'm forced to 
   attempt to fix this driver, which it seems I'm the only one using :)

The esp.c sources would be a better basis for writing a driver to
drive this chip, BTW.

qlogicfas.c is rarely if ever used (as you contend) and doesn't work
at all in the rare case we know it is used (as you also contend :-)