Re: [PATCH RFC/RFT 1/4] add some block layer helpers

[Mike Christie <michaelc@cs.wisc.edu>]

Mike Christie wrote:
> add blk_rq_map_kern_iov() which takes a iovec of buffers then
> maps them into bios and a request. It can make them into multiple
> bios to support sg and st's large requests and avoid the BIO_MAX_PAGES
> limit.
> 
> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
> 
> diff --git a/drivers/block/ll_rw_blk.c b/drivers/block/ll_rw_blk.c
> --- a/drivers/block/ll_rw_blk.c
> +++ b/drivers/block/ll_rw_blk.c
> @@ -28,6 +28,7 @@
>  #include <linux/slab.h>
>  #include <linux/swap.h>
>  #include <linux/writeback.h>
> +#include <linux/uio.h>
>  #include <linux/blkdev.h>
>  
>  /*
> @@ -2278,6 +2279,70 @@ int blk_rq_map_kern(request_queue_t *q, 
>  EXPORT_SYMBOL(blk_rq_map_kern);
>  
>  /**
> + * blk_rq_map_kern_iov - map kernel data to a request, for REQ_BLOCK_PC usage
> + * @q:		request queue where request should be inserted
> + * @rq:		request to fill
> + * @vec:	pointer to kvec
> + * @vec_count:	number of elements in kvec
> + * @gfp_mask:	memory allocation flags
> + *
> + * blk_rq_map_kern_iov maps a kvec into a multiple bio request so that
> + * it can create very large requests.
> + *
> + * For now we assume that each element will fit in one bio
> + */
> +int blk_rq_map_kern_iov(request_queue_t *q, struct request *rq,
> +			struct kvec *vec, int vec_count, unsigned int gfp)
> +{

On second thought, I think this function should work on pages and not 
kernel addresses. It also has a leak in the error path.