Re: [PATCH v6 4/5] dax: for truncate/hole-punch, do zeroing through the driver if possible

["Verma, Vishal L" <vishal.l.verma@intel.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8", Size: 834 bytes --]

On Tue, 2016-05-10 at 12:49 -0600, Vishal Verma wrote:
> 
...

> @@ -1240,11 +1254,16 @@ int dax_zero_page_range(struct inode *inode,
> loff_t from, unsigned length,
>  			.size = PAGE_SIZE,
>  		};
>  
> -		if (dax_map_atomic(bdev, &dax) < 0)
> -			return PTR_ERR(dax.addr);
> -		clear_pmem(dax.addr + offset, length);
> -		wmb_pmem();
> -		dax_unmap_atomic(bdev, &dax);
> +		if (dax_range_is_aligned(bdev, &dax, offset, length))
> +			return blkdev_issue_zeroout(bdev, dax.sector,
> +					length >> 9, GFP_NOFS, true);

Found another bug here while testing. The zeroout needs to be done for
sector + (offset >> 9). The above just zeroed out the first sector of
the page irrespective of offset, which is wrong.N‹§²æìr¸›zǧu©ž²Æ {\b­†éì¹»\x1c®&Þ–)îÆi¢žØ^n‡r¶‰šŽŠÝ¢j$½§$¢¸\x05¢¹¨­è§~Š'.)îÄÃ,yèm¶ŸÿÃ\f%Š{±šj+ƒðèž×¦j)Z†·Ÿ