[PATCH 1/2] mtdpart: memory accessor interface for MTD layer

[Sudhakar Rajashekhara]

Hi David,

On Sun, Aug 08, 2010 at 17:53:18, David Woodhouse wrote:
> On Fri, 2010-08-06 at 12:18 +0530, Sudhakar Rajashekhara wrote:
> > 
> > Thanks for the feedback. I'll be re-working on this patch and will
> > re-post
> > the updated patch soon. 
> 
> Start with this, perhaps...
> 
> Subject: mtd/partitions: Add add_mtd_partitions_ret() function
> 
> Some callers want access to the MTD devices which get registered for
> them when they call add_mtd_partitions(). Add a variant on the function
> which does that.
> 
> Signed-off-by: David Woodhouse <David.Woodhouse at intel.com>
> 
> diff --git a/drivers/mtd/mtdpart.c b/drivers/mtd/mtdpart.c
> index 4c539de..b9ee79b 100644
> --- a/drivers/mtd/mtdpart.c
> +++ b/drivers/mtd/mtdpart.c
> @@ -522,26 +522,36 @@ out_register:
>   * for reasons of data integrity.
>   */
>  
> -int add_mtd_partitions(struct mtd_info *master,
> -		       const struct mtd_partition *parts,
> -		       int nbparts)
> +int add_mtd_partitions_ret(struct mtd_info *master,
> +			   const struct mtd_partition *parts,
> +			   int nbparts, struct mtd_info ***mtds_ret)
>  {
>  	struct mtd_part *slave;
>  	uint64_t cur_offset = 0;
> +	struct mtd_info **mtds = NULL;
>  	int i;
>  
>  	printk(KERN_NOTICE "Creating %d MTD partitions on \"%s\":\n", nbparts, master->name);
>  
> +	if (mtds_ret) {
> +		mtds = kmalloc(sizeof(*mtds) * nbparts, GFP_KERNEL);
> +		if (!mtds)
> +			return -ENOMEM;
> +	}
>  	for (i = 0; i < nbparts; i++) {
>  		slave = add_one_partition(master, parts + i, i, cur_offset);
> -		if (!slave)
> +		if (!slave) {
> +			kfree(mtds);
>  			return -ENOMEM;
> +		}
>  		cur_offset = slave->offset + slave->mtd.size;
>  	}
>  
> +	if (mtds_ret)
> +		*mtds_ret = mtds;
>  	return 0;
>  }
> -EXPORT_SYMBOL(add_mtd_partitions);
> +EXPORT_SYMBOL_GPL(add_mtd_partitions_ret);
>  
>  static DEFINE_SPINLOCK(part_parser_lock);
>  static LIST_HEAD(part_parsers);
> diff --git a/include/linux/mtd/partitions.h b/include/linux/mtd/partitions.h
> index 274b619..f7935fa 100644
> --- a/include/linux/mtd/partitions.h
> +++ b/include/linux/mtd/partitions.h
> @@ -49,9 +49,12 @@ struct mtd_partition {
>  
>  struct mtd_info;
>  
> -int add_mtd_partitions(struct mtd_info *, const struct mtd_partition *, int);
> +int add_mtd_partitions_ret(struct mtd_info *, const struct mtd_partition *, int,
> +			   struct mtd_info ***);
>  int del_mtd_partitions(struct mtd_info *);
>  
> +#define add_mtd_partitions(m, p, n) add_mtd_partitions_ret(m, p, n, NULL)
> +
>  /*
>   * Functions dealing with the various ways of partitioning the space
>   */
> 
> -- 

Thanks very much for this piece of code, it reduced my work. But while working
on this, I found out that, for m25p80 device, the MTD device is available even
for the add_mtd_partitions() case. So the method to read the MAC address remains
same both for un-partitioned and partitioned MTD device. I'll post the modified
patch tomorrow. I'll be looking forward to hearing your comments on that patch
as well.

Thanks,
Sudhakar