Re: [ppc patch] support SCSI disks

[Hollis Blanchard]

On Oct 4, 2004, at 6:21 AM, Marco Gerards wrote:
> As far as I am concerned this patch can be committed.  So I will do
> that when I have the time for that.

Great!

> > +      else if ((! grub_strcmp (alias->type, "scsi"))
> > +              || (! grub_strcmp (alias->type, "ide"))
> > +              || (! grub_strcmp (alias->type, "ata")))
>
> What I meant with my previous email was finding these aliases that
> function as controller dynamically if possible.  But this is ok for
> now.

To handle something like USB devices, we would need to fully recurse 
below each node. The USB bus topology is represented in device tree 
nodes, so a USB block device could appear N levels deep below a "usb" 
alias. Firewire may be the same way. While I do have a USB flash 
device, my B&W G3's firmware doesn't recognize it as a block device, so 
I can't really test. However, this code supports SCSI for now, so I 
agree we can address the more general problem later.

> > +      /* XXX: Don't use hardcoded path lengths.  */
> > +      char childtype[64];
> > +      char childpath[64];
> > +      char childname[64];
> > +      char fullname[64];
>
> I assume you tested this length is sufficient.

"Works for me", but as we discussed once I would prefer a get_property 
function that returns malloc'ed data. There may indeed be systems whose 
bus topology is so deep that we exhaust these buffers.

-Hollis