Re: Cell support and fix PCI PDC paths

Barely catching up, a couple of questions ;)

On 12/26/06, Matthew Wilcox <[email protected]> wrote:

> @@ -326,6 +335,10 @@ static void get_node_path(struct device *dev, struct hardware_path *path)
>  static char *print_hwpath(struct hardware_path *path, char *output)
>  {
>         int i;
> +
> +       if (is_pdc_pat())
> +               output += sprintf(output, "%u/", path->flags & HWP_CELL);
> +

I suppose this change needs to go in pdcspath_layer_read() as well?
(in pdc_stable.c)

AIUI that means that the 'write' procedure needs some rewrite to cope
with this additional field, doesn't it?

Put another way, changing the console/kbd/etc paths with the
pdc_stable driver with this patch applied will just harm the box,
right?

> diff --git a/drivers/parisc/pdc_stable.c b/drivers/parisc/pdc_stable.c
> index ea1b7a6..66afab8 100644
> --- a/drivers/parisc/pdc_stable.c
> +++ b/drivers/parisc/pdc_stable.c

> @@ -597,19 +598,28 @@ pdcs_timer_read(struct subsystem *entry, char *buf)
>  {
>         char *out = buf;
>         struct pdcspath_entry *pathentry;
> +       int timer;
>
>         if (!entry || !buf)
>                 return -EINVAL;
>
> +       /*
> +        * PDC PAT machines use the timer field for something different.
> +        * I don't know where they keep the timer field now ...
> +        */
> +       if (is_pdc_pat())
> +               return -ENODEV;

I guess the best thing would be to simply not expose the 'timer' file
in sysfs on those boxen, I'll check how to do that cleanly.

HTH, nite all ;)

T-Bone

-- 
Thibaut VARENE
http://www.parisc-linux.org/~varenet/