Re: [PATCH] pci: fix unterminated pci_device_id lists

[Sam Ravnborg <sam@ravnborg.org>]

On Thu, Sep 13, 2007 at 10:34:02AM +0400, Alexey Dobriyan wrote:
> [non-terminated PCI ids arrays]
> 
> Here is compile-time hack (yep, warped and whitespace damaged :))
> It's better than modpost-time hack. because it triggers earlier.
> It's worse than modpost-time hack, because of tree-wide changes.
> 
> diff --git a/drivers/net/8139too.c b/drivers/net/8139too.c
> index f4e4298..b895b5f 100644
> --- a/drivers/net/8139too.c
> +++ b/drivers/net/8139too.c
> @@ -237,7 +237,7 @@ static const struct {
>  };
> 
> 
> -static struct pci_device_id rtl8139_pci_tbl[] = {
> +PCI_MODULE_DEVICE_TABLE_BEGIN(rtl8139_pci_tbl)
>  	{0x10ec, 0x8139, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 },
>  	{0x10ec, 0x8138, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 },
>  	{0x1113, 0x1211, PCI_ANY_ID, PCI_ANY_ID, 0, 0, RTL8139 },
> @@ -273,10 +273,7 @@ static struct pci_device_id rtl8139_pci_tbl[] = {
>  	{PCI_ANY_ID, 0x8139, 0x10ec, 0x8139, 0, 0, RTL8139 },
>  	{PCI_ANY_ID, 0x8139, 0x1186, 0x1300, 0, 0, RTL8139 },
>  	{PCI_ANY_ID, 0x8139, 0x13d1, 0xab06, 0, 0, RTL8139 },
> -
> -	{0,}
> -};
> -MODULE_DEVICE_TABLE (pci, rtl8139_pci_tbl);
> +PCI_MODULE_DEVICE_TABLE_END

Looks at bit too magic to my taste.
And anyway deferring the check until modpost time should not cause troubles
anyway assuming people do build their kernel after adding a device ID.

And the above requires one to use the special macro in all places
whereas the modpost check will catch it also for new device ID users.

	Sam