linux-kernel - RE: [PATCH kernel v8 2/4] virtio-balloon: VIRTIO_BALLOON_F_CHUNK

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 5 Apr 2017 03:31:36 +0000
From:   "Wang, Wei W" <wei.w.wang@...el.com>
To:     "virtio-dev@...ts.oasis-open.org" <virtio-dev@...ts.oasis-open.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "qemu-devel@...gnu.org" <qemu-devel@...gnu.org>,
        "virtualization@...ts.linux-foundation.org" 
        <virtualization@...ts.linux-foundation.org>,
        "kvm@...r.kernel.org" <kvm@...r.kernel.org>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        "mst@...hat.com" <mst@...hat.com>,
        "david@...hat.com" <david@...hat.com>,
        "Hansen, Dave" <dave.hansen@...el.com>,
        "cornelia.huck@...ibm.com" <cornelia.huck@...ibm.com>,
        "akpm@...ux-foundation.org" <akpm@...ux-foundation.org>,
        "mgorman@...hsingularity.net" <mgorman@...hsingularity.net>,
        "aarcange@...hat.com" <aarcange@...hat.com>,
        "amit.shah@...hat.com" <amit.shah@...hat.com>,
        "pbonzini@...hat.com" <pbonzini@...hat.com>,
        "liliang.opensource@...il.com" <liliang.opensource@...il.com>
Subject: RE: [PATCH kernel v8 2/4] virtio-balloon:
 VIRTIO_BALLOON_F_CHUNK_TRANSFER

On Thursday, March 16, 2017 3:09 PM Wei Wang wrote:
> The implementation of the current virtio-balloon is not very efficient, because
> the ballooned pages are transferred to the host one by one. Here is the
> breakdown of the time in percentage spent on each step of the balloon inflating
> process (inflating 7GB of an 8GB idle guest).
> 
> 1) allocating pages (6.5%)
> 2) sending PFNs to host (68.3%)
> 3) address translation (6.1%)
> 4) madvise (19%)
> 
> It takes about 4126ms for the inflating process to complete.
> The above profiling shows that the bottlenecks are stage 2) and stage 4).
> 
> This patch optimizes step 2) by transferring pages to the host in chunks. A chunk
> consists of guest physically continuous pages, and it is offered to the host via a
> base PFN (i.e. the start PFN of those physically continuous pages) and the size
> (i.e. the total number of the pages). A chunk is formated as below:
> 
> --------------------------------------------------------
> |                 Base (52 bit)        | Rsvd (12 bit) |
> --------------------------------------------------------
> --------------------------------------------------------
> |                 Size (52 bit)        | Rsvd (12 bit) |
> --------------------------------------------------------
> 
> By doing so, step 4) can also be optimized by doing address translation and
> madvise() in chunks rather than page by page.
> 
> This optimization requires the negotiation of a new feature bit,
> VIRTIO_BALLOON_F_CHUNK_TRANSFER.
> 
> With this new feature, the above ballooning process takes ~590ms resulting in
> an improvement of ~85%.
> 
> TODO: optimize stage 1) by allocating/freeing a chunk of pages instead of a
> single page each time.
> 
> Signed-off-by: Liang Li <liang.z.li@...el.com>
> Signed-off-by: Wei Wang <wei.w.wang@...el.com>
> Suggested-by: Michael S. Tsirkin <mst@...hat.com>
> ---
>  drivers/virtio/virtio_balloon.c     | 371 +++++++++++++++++++++++++++++++++-
> --
>  include/uapi/linux/virtio_balloon.h |   9 +
>  2 files changed, 353 insertions(+), 27 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c index
> f59cb4f..3f4a161 100644
> --- a/drivers/virtio/virtio_balloon.c
> +++ b/drivers/virtio/virtio_balloon.c
> @@ -42,6 +42,10 @@
>  #define OOM_VBALLOON_DEFAULT_PAGES 256
>  #define VIRTBALLOON_OOM_NOTIFY_PRIORITY 80
> 
> +#define PAGE_BMAP_SIZE	(8 * PAGE_SIZE)
> +#define PFNS_PER_PAGE_BMAP	(PAGE_BMAP_SIZE * BITS_PER_BYTE)
> +#define PAGE_BMAP_COUNT_MAX	32
> +
>  static int oom_pages = OOM_VBALLOON_DEFAULT_PAGES;
> module_param(oom_pages, int, S_IRUSR | S_IWUSR);
> MODULE_PARM_DESC(oom_pages, "pages to free on OOM"); @@ -50,6 +54,14
> @@ MODULE_PARM_DESC(oom_pages, "pages to free on OOM");  static struct
> vfsmount *balloon_mnt;  #endif
> 
> +#define BALLOON_CHUNK_BASE_SHIFT 12
> +#define BALLOON_CHUNK_SIZE_SHIFT 12
> +struct balloon_page_chunk {
> +	__le64 base;
> +	__le64 size;
> +};
> +
> +typedef __le64 resp_data_t;
>  struct virtio_balloon {
>  	struct virtio_device *vdev;
>  	struct virtqueue *inflate_vq, *deflate_vq, *stats_vq; @@ -67,6 +79,31
> @@ struct virtio_balloon {
> 
>  	/* Number of balloon pages we've told the Host we're not using. */
>  	unsigned int num_pages;
> +	/* Pointer to the response header. */
> +	struct virtio_balloon_resp_hdr *resp_hdr;
> +	/* Pointer to the start address of response data. */
> +	resp_data_t *resp_data;

I think the implementation has an issue here - both the balloon pages and the unused pages use the same buffer ("resp_data" above) to store chunks. It would cause a race in this case: live migration starts while ballooning is also in progress. I plan to use separate buffers for CHUNKS_OF_BALLOON_PAGES and CHUNKS_OF_UNUSED_PAGES. Please let me know if you have a different suggestion. Thanks.

Best,
Wei