On 12/3/2014 7:09 AM, Grant Likely wrote:
> On Wed, Dec 3, 2014 at 3:39 AM, Frank Rowand <frowand.list@xxxxxxxxx> wrote:
>> On 11/28/2014 8:48 AM, Grant Likely wrote:
< snip >
>>> Instead of keeping it always in order, what about reordering the whole
>>> list after all the nodes at a given level are unflattened? That would
>>> avoid traversing the entire sibling list on every node addition. Like
>>> this:
>>
>> Seems pretty close to a wash to me. Either one would work. Though
>> mine would benefit from the comment you added in yours.
>>
>> -Frank
>
> Okay, if I can take that as an 'acked-by' from you then I'll merge my
> version of the patch. I prefer doing the post processing to sort the
> list because it is theoretically less costly.
Yes, that would be fine.
Acked-by: Frank Rowand <frank.rowand@xxxxxxxxxxxxxx>
>
> g.
>
>>>
>>> g.
>>>
>>> >From de46de766eb4d2240208bf83fdae955361069352 Mon Sep 17 00:00:00 2001
>>> From: Grant Likely <grant.likely@xxxxxxxxxx>
>>> Date: Fri, 28 Nov 2014 16:03:33 +0000
>>> Subject: [PATCH] of: Drop ->next pointer from struct device_node
>>>
>>> The ->next pointer in struct device_node is a hanger-on from when it was
>>> used to iterate over the whole tree by a particular device_type property
>>> value. Those days are long over, but the fdt unflattening code still
>>> uses it to put nodes in the unflattened tree into the same order as node
>>> in the flat tree. By reworking the unflattening code to reverse the list
>>> after unflattening all the children of a node, the pointer can be
>>> dropped which gives a small amount of memory savings.
>>>
>>> Signed-off-by: Grant Likely <grant.likely@xxxxxxxxxx>
>>> Cc: Gaurav Minocha <gaurav.minocha.os@xxxxxxxxx>
>>> Cc: Frank Rowand <frowand.list@xxxxxxxxx>
>>> ---
>>> drivers/of/fdt.c | 24 ++++++++++++++++++------
>>> include/linux/of.h | 1 -
>>> 2 files changed, 18 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c
>>> index 7f6ee31d5650..a41f9fdb1aa0 100644
>>> --- a/drivers/of/fdt.c
>>> +++ b/drivers/of/fdt.c
>>> @@ -226,12 +226,8 @@ static void * unflatten_dt_node(void *blob,
>>> prev_pp = &np->properties;
>>> if (dad != NULL) {
>>> np->parent = dad;
>>> - /* we temporarily use the next field as `last_child'*/
>>> - if (dad->next == NULL)
>>> - dad->child = np;
>>> - else
>>> - dad->next->sibling = np;
>>> - dad->next = np;
>>> + np->sibling = dad->child;
>>> + dad->child = np;
>>> }
>>> }
>>> /* process properties */
>>> @@ -329,6 +325,22 @@ static void * unflatten_dt_node(void *blob,
>>>
>>> if (*poffset < 0 && *poffset != -FDT_ERR_NOTFOUND)
>>> pr_err("unflatten: error %d processing FDT\n", *poffset);
>>> +
>>> + /*
>>> + * Reverse the child list. Some drivers assumes node order matches .dts
>>> + * node order
>>> + */
>>> + if (!dryrun && np->child) {
>>> + struct device_node *child = np->child;
>>> + np->child = NULL;
>>> + while (child) {
>>> + struct device_node *next = child->sibling;
>>> + child->sibling = np->child;
>>> + np->child = child;
>>> + child = next;
>>> + }
>>> + }
>>> +
>>> if (nodepp)
>>> *nodepp = np;
>>>
>>> diff --git a/include/linux/of.h b/include/linux/of.h
>>> index 8b021db3e16e..3f0f0ffbd5e5 100644
>>> --- a/include/linux/of.h
>>> +++ b/include/linux/of.h
>>> @@ -56,7 +56,6 @@ struct device_node {
>>> struct device_node *parent;
>>> struct device_node *child;
>>> struct device_node *sibling;
>>> - struct device_node *next; /* next device of same type */
>>> struct kobject kobj;
>>> unsigned long _flags;
>>> void *data;
>>>
>>
>
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
