[PATCH] Linux Secondary Slave IDE timings

[PATCH] Linux Secondary Slave IDE timings

[Daniel Quinlan]

This fix is from Dave Bogdanoff <bog@transmeta.com>.

Linux incorrectly sets up IDE timings for secondary slave drives on PC
systems that use Intel PIIX Southbridges.

This will correctly shift IDE slave PCI timings for register 44h so
that:

 - secondary slave (drive1) uses bits 4-7
 - primary slave   (drive1) uses bits 0-3

(The addition of the parentheses is needed so the shift will take
place after the bitwise-or.  Without the parentheses, the shift will
incorrectly always take place before the bitwise-or.)

--- linux/drivers/ide/piix.c.orig	Tue Jan 29 23:26:55 2002
+++ linux/drivers/ide/piix.c	Tue Jan 29 23:27:33 2002
@@ -258,8 +258,8 @@
 			master_data = master_data | 0x0070;
 		pci_read_config_byte(HWIF(drive)->pci_dev, slave_port, &slave_data);
 		slave_data = slave_data & (HWIF(drive)->index ? 0x0f : 0xf0);
-		slave_data = slave_data | ((timings[pio][0] << 2) | (timings[pio][1]
-					   << (HWIF(drive)->index ? 4 : 0)));
+		slave_data = slave_data | (((timings[pio][0] << 2) | timings[pio][1])
+					   << (HWIF(drive)->index ? 4 : 0));
 	} else {
 		master_data = master_data & 0xccf8;
 		if (pio > 1)

Re: [PATCH] Linux Secondary Slave IDE timings

[Andre Hedrick]

It is a valid and correct fix.
There must have been a bug or something in an odd compiler, but the reason
for the <> | order problem escapes me now.
The verification came from a commumication I had with the author of the
Intel programmer docs.

I hope my comments do not kill the patch.

Regards,

Andre Hedrick
Linux Disk Certification Project                Linux ATA Development

On Tue, 26 Feb 2002, Daniel Quinlan wrote:

> This fix is from Dave Bogdanoff <bog@transmeta.com>.
> 
> Linux incorrectly sets up IDE timings for secondary slave drives on PC
> systems that use Intel PIIX Southbridges.
> 
> This will correctly shift IDE slave PCI timings for register 44h so
> that:
> 
>  - secondary slave (drive1) uses bits 4-7
>  - primary slave   (drive1) uses bits 0-3
> 
> (The addition of the parentheses is needed so the shift will take
> place after the bitwise-or.  Without the parentheses, the shift will
> incorrectly always take place before the bitwise-or.)
> 
> --- linux/drivers/ide/piix.c.orig	Tue Jan 29 23:26:55 2002
> +++ linux/drivers/ide/piix.c	Tue Jan 29 23:27:33 2002
> @@ -258,8 +258,8 @@
>  			master_data = master_data | 0x0070;
>  		pci_read_config_byte(HWIF(drive)->pci_dev, slave_port, &slave_data);
>  		slave_data = slave_data & (HWIF(drive)->index ? 0x0f : 0xf0);
> -		slave_data = slave_data | ((timings[pio][0] << 2) | (timings[pio][1]
> -					   << (HWIF(drive)->index ? 4 : 0)));
> +		slave_data = slave_data | (((timings[pio][0] << 2) | timings[pio][1])
> +					   << (HWIF(drive)->index ? 4 : 0));
>  	} else {
>  		master_data = master_data & 0xccf8;
>  		if (pio > 1)
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: [PATCH] Linux Secondary Slave IDE timings

[Martin Dalecki]

Andre Hedrick wrote:
> It is a valid and correct fix.
> There must have been a bug or something in an odd compiler, but the reason
> for the <> | order problem escapes me now.
> The verification came from a commumication I had with the author of the
> Intel programmer docs.
> 
> I hope my comments do not kill the patch.

Certainly NOT. Thank's for confirmation.