From: Alan Cox To: Alessandro Zummo Cc: Jeff Garzik , Alan Cox Subject: Re: drivers/ata/pata_artop.c Date: Sun, 15 Oct 2006 14:25:16 -0400 User-Agent: Mutt/1.4.1i On Sun, Oct 15, 2006 at 07:18:31PM +0200, Alessandro Zummo wrote: > In the discovery phase there's a lot of time spent in the detection > of the second port. The error recovery is a bit determined right now - Tejun's been doing some work on SRST behaviour and also for the worst cases polled detect so it should come out ok > What's the correct way to inform the driver > to avoid checking the second port? Set the number of ports to 1 in your own tree for now. The real fix is not to go poking at pata ports if the ret is 0xFF --- drivers/ata/pata_artop.c | 6 ++++++ 1 file changed, 6 insertions(+) Index: linux-2.6.21.7/drivers/ata/pata_artop.c =================================================================== --- linux-2.6.21.7.orig/drivers/ata/pata_artop.c +++ linux-2.6.21.7/drivers/ata/pata_artop.c @@ -26,6 +26,7 @@ #include #include #include +#include #define DRV_NAME "pata_artop" #define DRV_VERSION "0.4.2" @@ -469,6 +470,11 @@ static int artop_init_one (struct pci_de pci_read_config_byte(pdev, 0x4a, ®); pci_write_config_byte(pdev, 0x4a, (reg & ~0x01) | 0x80); + /* NAS100D workaround */ +#ifdef CONFIG_MACH_NAS100D + if (machine_is_nas100d()) + ports = 1; +#endif } BUG_ON(info == NULL);