Re: [PATCH 8/12] x86-64: update iommu/dma mapping functions to sg helpers

[Jens Axboe <jens.axboe@oracle.com>]

On Thu, May 10 2007, Benny Halevy wrote:
> @@ -411,12 +406,13 @@ int gart_map_sg(struct device *dev, struct scatterlist *sg, int nents, int dir)
>  			   boundary and the new one doesn't have an offset. */
>  			if (!iommu_merge || !nextneed || !need || s->offset ||
>  			    (ps->offset + ps->length) % PAGE_SIZE) { 
> -				if (dma_map_cont(sg, start, i, sg+out, pages,
> -						 need) < 0)
> +				if (dma_map_cont(start_sg, i - start, sg+out,
> +						  pages, need) < 0)
>  					goto error;
>  				out++;
>  				pages = 0;
> -				start = i;	
> +				start = i;
> +				start_sg = s;
>  			}
>  		}
>  
> @@ -424,7 +420,7 @@ int gart_map_sg(struct device *dev, struct scatterlist *sg, int nents, int dir)
>  		pages += to_pages(s->offset, s->length);
>  		ps = s;
>  	}
> -	if (dma_map_cont(sg, start, i, sg+out, pages, need) < 0)
> +	if (dma_map_cont(start_sg, i - start, sg+out, pages, need) < 0)
>  		goto error;
>  	out++;
>  	flush_gart();

Your patch is (very) buggy, the whole premise of doing chained sg
entries makes sg + int illegal!

-- 
Jens Axboe