答复: S2885 failing on 1.1.6

[YhLu]

Eric,

The problem caused by

ht_setup_chain(). in northbridge/amd/amdk8/incoherent_ht.c

The S2885 CPU link0 is connected to 8151 and link 2 is connected to 8131.

The ht_setup_chain() seems not to be complete as the ht_scan_chain in
devices/hypertransport.c

Regards

YH

-----邮件原件-----
发件人: ron minnich [mailto:[e-mail]] 
发送时间: 2004年3月12日 13:07
收件人: Eric W. Biederman
抄送: Li-Ta Lo; YhLu; LinuxBIOS
主题: Re: S2885 failing on 1.1.6

On 12 Mar 2004, Eric W. Biederman wrote:
> Why you flashed the entire chip is something I don't want to
> contemplate, instead of leaving fallback alone is something I don't
> want to contemplate :)

it's a test node, we don't worry about fallback on this one.
ron

[Li-Ta Lo]

On Fri, 2004-03-12 at 14:54, YhLu wrote:
> Eric,
>
> The problem caused by
>
> ht_setup_chain(). in northbridge/amd/amdk8/incoherent_ht.c
>
> The S2885 CPU link0 is connected to 8151 and link 2 is connected to 8131.
>
> The ht_setup_chain() seems not to be complete as the ht_scan_chain in
> devices/hypertransport.c

I change the reset code to

static void hard_reset(void)
{
        set_bios_reset();

/* enable cf9 */
        pci_write_config8(PCI_DEV(1, 0x04, 3), 0x41, 0xf1);
        /* reset */
        outb(0x0e, 0x0cf9);
}

static void soft_reset(void)
{
        set_bios_reset();
        pci_write_config8(PCI_DEV(1, 0x04, 0), 0x47, 1);
}

Now it works.

Is that PCI(1, ...) mean another link ?

Ollie