Hello,
I recently bought a TX-4 which seems to be using a new chip (PCI ID:
105A:3519).
It was not supported by the kernel so i added those lines to promise_sata.c
diff sata_promise.c /usr/src/linux-2.6.11.9/drivers/scsi/sata_promise.c
170,171d169
< { PCI_VENDOR_ID_PROMISE, 0x3519, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
< board_20319 },
Is it the 'right' way?
Thanks
Laurent
On Mon, 2005-05-23 at 15:41 +0200, Laurent CARON wrote:
> Hello,
>
> I recently bought a TX-4 which seems to be using a new chip (PCI ID:
> 105A:3519).
>
> It was not supported by the kernel so i added those lines to promise_sata.c
>
> diff sata_promise.c /usr/src/linux-2.6.11.9/drivers/scsi/sata_promise.c
> 170,171d169
> < { PCI_VENDOR_ID_PROMISE, 0x3519, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
> < board_20319 },
>
>
> Is it the 'right' way?
if it works.. almost
the "right" way is to use "diff -purN" instead of just plain diff (it's
custom) and to do the files the other way around (again custom).
If you want to make it nice you add a PCI_ID_... constant for 0x3519 to
the header and use the symbolic constant in your code instead.
Greetings,
Arjan van de Ven
Arjan van de Ven a ?crit :
>if it works.. almost
>the "right" way is to use "diff -purN" instead of just plain diff (it's
>custom) and to do the files the other way around (again custom).
>
>
diff -purN /usr/src/linux-2.6.11.9/drivers/scsi/sata_promise.c
sata_promise.c
--- /usr/src/linux-2.6.11.9/drivers/scsi/sata_promise.c 2005-05-12
00:42:30.000000000 +0200
+++ sata_promise.c 2005-05-23 15:56:09.000000000 +0200
@@ -167,6 +167,8 @@ static struct pci_device_id pdc_ata_pci_
board_20319 },
{ PCI_VENDOR_ID_PROMISE, 0x3d18, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
board_20319 },
+ { PCI_VENDOR_ID_PROMISE, 0x3519, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+ board_20319 },
{ } /* terminate list */
};
>If you want to make it nice you add a PCI_ID_... constant for 0x3519 to
>the header and use the symbolic constant in your code instead.
>
>
I did it like this mainly because the file was built like this (without
pci ids in the header).
Regards,
Laurent