V2 EPIA FIXED!

[ron minnich]

Here's what I had to do. First, though, let me say: the new linuxbios 
architecture, while hard to get the hang of (learning curve) is very 
powerful.

You can have a mainboard-specific pci scan function, and that's what I 
needed. I may put this as a generic thing into the pci device code.

See below for how it works. Basically, for your mainboard, you provide a 
function that scans the root of the "mainboard PCI" tree.

ron

Here goes

================

#include <console/console.h>
#include <device/device.h>
#include <device/pci.h>
#include <device/pci_ids.h>
#include <device/pci_ops.h>

#include <arch/io.h>
#include <device/chip.h>
#include "chip.h"

static int
mainboard_scan_bus(device_t root, int maxbus) 
{
    int retval;
    printk_spew("%s: root %p maxbus %d\n", __FUNCTION__, root, maxbus);
    retval = pci_scan_bus(root->bus, 0, 0xff, maxbus);
    printk_spew("DONE %s: return %d\n", __FUNCTION__, maxbus);
    return maxbus;
}

static struct device_operations mainboard_operations = {
    .read_resources   = root_dev_read_resources,
    .set_resources    = root_dev_set_resources,
    .enable_resources = enable_childrens_resources,
    .init             = 0,
    .scan_bus         = mainboard_scan_bus,
    .enable           = 0,
};

static void enumerate(struct chip *chip)
{
    struct chip *child;
    dev_root.ops = &mainboard_operations;
    chip->dev = &dev_root;
    chip->bus = 0;
    for(child = chip->children; child; child = child->next) {
    	child->bus = &dev_root.link[0];
    }
}
struct chip_control mainboard_via_epia_control = {
    .enumerate = enumerate, 
    .name      = "VIA EPIA mainboard ",
};

ron minnich  writes:
> Here's what I had to do. First, though, let me say: the new linuxbios 
> architecture, while hard to get the hang of (learning curve) is very 
> powerful.
>
> You can have a mainboard-specific pci scan function, and that's what I 
> needed. I may put this as a generic thing into the pci device code.
>
> See below for how it works. Basically, for your mainboard, you provide a 
> function that scans the root of the "mainboard PCI" tree.

Hmm.  I am curious why the default did not take.
See: devices/root_device.c

All you should need is included below.  And potentially we can have
a generic enumeration function in root_device.c that does the rest
of it as well.

The Opteron has rather special needs so it needs to override the defaults,
but the VIA EPIA should not.

Eric
> ================
>
> #include <console/console.h>
> #include <device/device.h>
> #include <device/pci.h>
> #include <device/pci_ids.h>
> #include <device/pci_ops.h>
>
> #include <arch/io.h>
> #include <device/chip.h>
> #include "chip.h"
>
> static void enumerate(struct chip *chip)
> {
>    struct chip *child;
>    chip->dev = &dev_root;
>    chip->bus = 0;
>    for(child = chip->children; child; child = child->next) {
>    	child->bus = &dev_root.link[0];
>    }
> }
> struct chip_control mainboard_via_epia_control = {
>    .enumerate = enumerate, 
>    .name      = "VIA EPIA mainboard ",
> };

[ron minnich]

On 29 Sep 2003, Eric W. Biederman wrote:
> All you should need is included below.  And potentially we can have
> a generic enumeration function in root_device.c that does the rest
> of it as well.

yes, we should.

The default did not take, but I will take a look.

ron

Linux Boot In Under 200ms

[Adam Agnew]

This shouldn't slip under anyone's radar. Not that it would, being on
slashdot and all.

Slashdot Headline: "Software Tweak Makes Linux Boot In Under 200ms"
http://developers.slashdot.org/developers/03/09/30/003209.shtml?tid=106&...

- Adam A.

Linux Boot In Under 200ms

Adam Agnew  writes:
> This shouldn't slip under anyone's radar. Not that it would, being on
> slashdot and all.
>
> Slashdot Headline: "Software Tweak Makes Linux Boot In Under 200ms"
> http://developers.slashdot.org/developers/03/09/30/003209.shtml?tid=106&...

Have you found the patches yet.  This sounds like the old technique
of making everything modular so that init will start faster.

Eric