On Jun 10, 2009, at 4:46 PM, Nate Case wrote:
Hi Kumar,
Thanks for the feedback. Response below:
+
+ root = of_find_node_by_path("/");
+ if (root == NULL)
+ return;
+
+ model = of_get_property(root, "model", NULL);
+ if (strncasecmp(model, "xes,", strlen("xes,")) != 0)
+ return;
What is this check for?
Historical accident. A while back we somehow ended up with a case of
probe() succeeding when it shouldn't have, and this was the
workaround.
I'll remove it since it's useless now.
+
+ printk(KERN_INFO "X-ES MPC85xx-based single-board computer: %s\n",
+ model + strlen("xes,"));
Why not print the name from machine_id.name
Because machine_id->name will give us "X-ES MPC8572" from the machine
description rather than the actual board name (e.g., xpedite5370).
In our situation, we have several boards that share largely the same
hardware design. I decided to just put 1 machine definition per
processor rather than having a machine definition for every board
(which
would get pretty crazy for our case). For example, we have at least 6
MPC8572-based boards with more on the way, but all of their changes
can
be described in the device tree (e.g., changes in number of ethernet
ports or PCI express controllers), so there's no need for a machine
definition for all of them.
+machine_device_initcall(xes_mpc8572, xes_mpc85xx_publish_devices);
+machine_device_initcall(xes_mpc8548, xes_mpc85xx_publish_devices);
Do you not need this for xes_mpc8540?
Yes, thanks. I'll fix this and the other things you mentioned and
re-submit. Will this be able to make it into your next tree destined
for 2.6.31?
If it get submitted pretty soon I'd be willing to take it for .31.
Also can you split the patch into 3 pieces (defconfig, .dts, and
actual platform code)
- k
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
