Date:   Fri, 29 Sep 2017 14:55:18 +0800
From:   Wei Wang <wei.w.wang@...el.com>
To:     "Michael S. Tsirkin" <mst@...hat.com>
CC:     virtio-dev@...ts.oasis-open.org, linux-kernel@...r.kernel.org,
        qemu-devel@...gnu.org, virtualization@...ts.linux-foundation.org,
        kvm@...r.kernel.org, linux-mm@...ck.org, mhocko@...nel.org,
        akpm@...ux-foundation.org, mawilcox@...rosoft.com,
        david@...hat.com, cornelia.huck@...ibm.com,
        mgorman@...hsingularity.net, aarcange@...hat.com,
        amit.shah@...hat.com, pbonzini@...hat.com, willy@...radead.org,
        liliang.opensource@...il.com, yang.zhang.wz@...il.com,
        quan.xu@...yun.com
Subject: Re: [virtio-dev] Re: [PATCH v15 3/5] virtio-balloon: VIRTIO_BALLOON_F_SG

On 09/29/2017 12:01 PM, Michael S. Tsirkin wrote:
> On Fri, Sep 08, 2017 at 07:09:24PM +0800, Wei Wang wrote:
>> On 09/08/2017 11:36 AM, Michael S. Tsirkin wrote:
>>> On Tue, Aug 29, 2017 at 11:09:18AM +0800, Wei Wang wrote:
>>>> On 08/29/2017 02:03 AM, Michael S. Tsirkin wrote:
>>>>> On Mon, Aug 28, 2017 at 06:08:31PM +0800, Wei Wang wrote:
>>>>>> Add a new feature, VIRTIO_BALLOON_F_SG, which enables the transfer
>>>>>> of balloon (i.e. inflated/deflated) pages using scatter-gather lists
>>>>>> to the host.
>>>>>>
>>>>>> The implementation of the previous virtio-balloon is not very
>>>>>> efficient, because the balloon 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
>>>>>> sgs. An sg describes a chunk of guest physically continuous pages.
>>>>>> With this mechanism, step 4) can also be optimized by doing address
>>>>>> translation and madvise() in chunks rather than page by page.
>>>>>>
>>>>>> With this new feature, the above ballooning process takes ~597ms
>>>>>> resulting in an improvement of ~86%.
>>>>>>
>>>>>> TODO: optimize stage 1) by allocating/freeing a chunk of pages
>>>>>> instead of a single page each time.
>>>>>>
>>>>>> Signed-off-by: Wei Wang <wei.w.wang@...el.com>
>>>>>> Signed-off-by: Liang Li <liang.z.li@...el.com>
>>>>>> Suggested-by: Michael S. Tsirkin <mst@...hat.com>
>>>>>> ---
>>>>>>     drivers/virtio/virtio_balloon.c     | 171 ++++++++++++++++++++++++++++++++----
>>>>>>     include/uapi/linux/virtio_balloon.h |   1 +
>>>>>>     2 files changed, 155 insertions(+), 17 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
>>>>>> index f0b3a0b..8ecc1d4 100644
>>>>>> --- a/drivers/virtio/virtio_balloon.c
>>>>>> +++ b/drivers/virtio/virtio_balloon.c
>>>>>> @@ -32,6 +32,8 @@
>>>>>>     #include <linux/mm.h>
>>>>>>     #include <linux/mount.h>
>>>>>>     #include <linux/magic.h>
>>>>>> +#include <linux/xbitmap.h>
>>>>>> +#include <asm/page.h>
>>>>>>     /*
>>>>>>      * Balloon device works in 4K page units.  So each page is pointed to by
>>>>>> @@ -79,6 +81,9 @@ struct virtio_balloon {
>>>>>>     	/* Synchronize access/update to this struct virtio_balloon elements */
>>>>>>     	struct mutex balloon_lock;
>>>>>> +	/* The xbitmap used to record balloon pages */
>>>>>> +	struct xb page_xb;
>>>>>> +
>>>>>>     	/* The array of pfns we tell the Host about. */
>>>>>>     	unsigned int num_pfns;
>>>>>>     	__virtio32 pfns[VIRTIO_BALLOON_ARRAY_PFNS_MAX];
>>>>>> @@ -141,13 +146,111 @@ static void set_page_pfns(struct virtio_balloon *vb,
>>>>>>     					  page_to_balloon_pfn(page) + i);
>>>>>>     }
>>>>>> +static int add_one_sg(struct virtqueue *vq, void *addr, uint32_t size)
>>>>>> +{
>>>>>> +	struct scatterlist sg;
>>>>>> +
>>>>>> +	sg_init_one(&sg, addr, size);
>>>>>> +	return virtqueue_add_inbuf(vq, &sg, 1, vq, GFP_KERNEL);
>>>>>> +}
>>>>>> +
>>>>>> +static void send_balloon_page_sg(struct virtio_balloon *vb,
>>>>>> +				 struct virtqueue *vq,
>>>>>> +				 void *addr,
>>>>>> +				 uint32_t size,
>>>>>> +				 bool batch)
>>>>>> +{
>>>>>> +	unsigned int len;
>>>>>> +	int err;
>>>>>> +
>>>>>> +	err = add_one_sg(vq, addr, size);
>>>>>> +	/* Sanity check: this can't really happen */
>>>>>> +	WARN_ON(err);
>>>>> It might be cleaner to detect that add failed due to
>>>>> ring full and kick then. Just an idea, up to you
>>>>> whether to do it.
>>>>>
>>>>>> +
>>>>>> +	/* If batching is in use, we batch the sgs till the vq is full. */
>>>>>> +	if (!batch || !vq->num_free) {
>>>>>> +		virtqueue_kick(vq);
>>>>>> +		wait_event(vb->acked, virtqueue_get_buf(vq, &len));
>>>>>> +		/* Release all the entries if there are */
>>>>> Meaning
>>>>> 	Account for all used entries if any
>>>>> ?
>>>>>
>>>>>> +		while (virtqueue_get_buf(vq, &len))
>>>>>> +			;
>>>>> Above code is reused below. Add a function?
>>>>>
>>>>>> +	}
>>>>>> +}
>>>>>> +
>>>>>> +/*
>>>>>> + * Send balloon pages in sgs to host. The balloon pages are recorded in the
>>>>>> + * page xbitmap. Each bit in the bitmap corresponds to a page of PAGE_SIZE.
>>>>>> + * The page xbitmap is searched for continuous "1" bits, which correspond
>>>>>> + * to continuous pages, to chunk into sgs.
>>>>>> + *
>>>>>> + * @page_xb_start and @page_xb_end form the range of bits in the xbitmap that
>>>>>> + * need to be searched.
>>>>>> + */
>>>>>> +static void tell_host_sgs(struct virtio_balloon *vb,
>>>>>> +			  struct virtqueue *vq,
>>>>>> +			  unsigned long page_xb_start,
>>>>>> +			  unsigned long page_xb_end)
>>>>>> +{
>>>>>> +	unsigned long sg_pfn_start, sg_pfn_end;
>>>>>> +	void *sg_addr;
>>>>>> +	uint32_t sg_len, sg_max_len = round_down(UINT_MAX, PAGE_SIZE);
>>>>>> +
>>>>>> +	sg_pfn_start = page_xb_start;
>>>>>> +	while (sg_pfn_start < page_xb_end) {
>>>>>> +		sg_pfn_start = xb_find_next_bit(&vb->page_xb, sg_pfn_start,
>>>>>> +						page_xb_end, 1);
>>>>>> +		if (sg_pfn_start == page_xb_end + 1)
>>>>>> +			break;
>>>>>> +		sg_pfn_end = xb_find_next_bit(&vb->page_xb, sg_pfn_start + 1,
>>>>>> +					      page_xb_end, 0);
>>>>>> +		sg_addr = (void *)pfn_to_kaddr(sg_pfn_start);
>>>>>> +		sg_len = (sg_pfn_end - sg_pfn_start) << PAGE_SHIFT;
>>>>>> +		while (sg_len > sg_max_len) {
>>>>>> +			send_balloon_page_sg(vb, vq, sg_addr, sg_max_len, 1);
>>>>> Last argument should be true, not 1.
>>>>>
>>>>>> +			sg_addr += sg_max_len;
>>>>>> +			sg_len -= sg_max_len;
>>>>>> +		}
>>>>>> +		send_balloon_page_sg(vb, vq, sg_addr, sg_len, 1);
>>>>>> +		xb_zero(&vb->page_xb, sg_pfn_start, sg_pfn_end);
>>>>>> +		sg_pfn_start = sg_pfn_end + 1;
>>>>>> +	}
>>>>>> +
>>>>>> +	/*
>>>>>> +	 * The last few sgs may not reach the batch size, but need a kick to
>>>>>> +	 * notify the device to handle them.
>>>>>> +	 */
>>>>>> +	if (vq->num_free != virtqueue_get_vring_size(vq)) {
>>>>>> +		virtqueue_kick(vq);
>>>>>> +		wait_event(vb->acked, virtqueue_get_buf(vq, &sg_len));
>>>>>> +		while (virtqueue_get_buf(vq, &sg_len))
>>>>>> +			;
>>>>> Some entries can get used after a pause. Looks like they will leak then?
>>>>> One fix would be to convert above if to a while loop.
>>>>> I don't know whether to do it like this in send_balloon_page_sg too.
>>>>>
>>>> Thanks for the above comments. I've re-written this part of code.
>>>> Please have a check below if there is anything more we could improve:
>>>>
>>>> static void kick_and_wait(struct virtqueue *vq, wait_queue_head_t wq_head)
>>>> {
>>>>           unsigned int len;
>>>>
>>>>           virtqueue_kick(vq);
>>>>           wait_event(wq_head, virtqueue_get_buf(vq, &len));
>>>>           /* Detach all the used buffers from the vq */
>>>>           while (virtqueue_get_buf(vq, &len))
>>>>                   ;
>>> I would move this last part to before add_buf. Increases chances
>>> it succeeds even in case of a bug.
>>>> }
>>>>
>>>> static int add_one_sg(struct virtqueue *vq, void *addr, uint32_t size)
>>>> {
>>>>           struct scatterlist sg;
>>>>           int ret;
>>>>
>>>>           sg_init_one(&sg, addr, size);
>>>>           ret = virtqueue_add_inbuf(vq, &sg, 1, vq, GFP_KERNEL);
>>>>           if (unlikely(ret == -ENOSPC))
>>>>                   dev_warn(&vq->vdev->dev, "%s: failed due to ring full\n",
>>>>                                    __func__);
>>> So if this ever triggers then kick and wait might fail, right?
>>> I think you should not special-case this one then.
>> OK, I will remove the check above, and take other suggestions as well.
>> Thanks.
>>
>> Best,
>> Wei
> Any updates here? It's been a while.
>

Yes. with some major optimization on xbitmap, we can improve the 
ballooning time to ~492ms.
I will send out the patches soon.

Best,
Wei

Powered by blists - more mailing lists