Re: [PATCH] PCI Update for 2.6.3-rc1

[Marcelo Tosatti <marcelo.tosatti@cyclades.com>]

On Mon, 9 Feb 2004, Greg KH wrote:

> ChangeSet 1.1607, 2004/02/09 14:05:33-08:00, greg@kroah.com
>
> dmapool: fix up list_for_each() calls to list_for_each_entry()
>
> Now this should get that Rusty^Wmonkey off my back...
>
>
>  drivers/base/dmapool.c |   29 +++++++++++------------------
>  1 files changed, 11 insertions(+), 18 deletions(-)
>
>
> diff -Nru a/drivers/base/dmapool.c b/drivers/base/dmapool.c
> --- a/drivers/base/dmapool.c	Mon Feb  9 14:58:07 2004
> +++ b/drivers/base/dmapool.c	Mon Feb  9 14:58:07 2004
> @@ -43,9 +43,11 @@
>  static ssize_t
>  show_pools (struct device *dev, char *buf)
>  {
> -	unsigned		temp, size;
> -	char			*next;
> -	struct list_head	*i, *j;
> +	unsigned temp;
> +	unsigned size;
> +	char *next;
> +	struct dma_page *page;
> +	struct dma_pool *pool;
>
>  	next = buf;
>  	size = PAGE_SIZE;
> @@ -55,16 +57,11 @@
>  	next += temp;
>
>  	down (&pools_lock);
> -	list_for_each (i, &dev->dma_pools) {
> -		struct dma_pool	*pool;
> -		unsigned	pages = 0, blocks = 0;
> +	list_for_each_entry(pool, &dev->dma_pools, pools) {
> +		unsigned pages = 0;
> +		unsigned blocks = 0;
>
> -		pool = list_entry (i, struct dma_pool, pools);
> -
> -		list_for_each (j, &pool->page_list) {
> -			struct dma_page	*page;
> -
> -			page = list_entry (j, struct dma_page, page_list);
> +		list_for_each_entry(page, &pool->page_list, page_list) {
>  			pages++;
>  			blocks += page->in_use;

Hi Rusty,

I have you ask riel to use list_for_each_entry() instead list_for_each()
in the 2.4 inode highmem patches.

Is this a "beauty" only modification or is there any potential problem
with "list_for_each()" ?

Thank you.