From patchwork Fri May 27 10:34:09 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Liang Li X-Patchwork-Id: 9138095 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 3EA2160759 for ; Fri, 27 May 2016 10:42:21 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 0E4F828135 for ; Fri, 27 May 2016 10:42:21 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 0321628159; Fri, 27 May 2016 10:42:21 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 19C0728135 for ; Fri, 27 May 2016 10:42:20 +0000 (UTC) Received: from localhost ([::1]:45146 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b6FDf-0002ud-5e for patchwork-qemu-devel@patchwork.kernel.org; Fri, 27 May 2016 06:42:19 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40162) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b6FAt-0000ZM-SK for qemu-devel@nongnu.org; Fri, 27 May 2016 06:39:29 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1b6FAp-0006m9-LS for qemu-devel@nongnu.org; Fri, 27 May 2016 06:39:26 -0400 Received: from mga11.intel.com ([192.55.52.93]:37070) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b6FAp-0006lz-8v for qemu-devel@nongnu.org; Fri, 27 May 2016 06:39:23 -0400 Received: from orsmga001.jf.intel.com ([10.7.209.18]) by fmsmga102.fm.intel.com with ESMTP; 27 May 2016 03:39:21 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.26,373,1459839600"; d="scan'208";a="963504435" Received: from ll.sh.intel.com (HELO localhost) ([10.239.13.123]) by orsmga001.jf.intel.com with ESMTP; 27 May 2016 03:39:19 -0700 From: Liang Li To: linux-kernel@vger.kernel.org Date: Fri, 27 May 2016 18:34:09 +0800 Message-Id: <1464345250-6388-1-git-send-email-liang.z.li@intel.com> X-Mailer: git-send-email 1.9.1 X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 192.55.52.93 Subject: [Qemu-devel] [PATCH RFC v2 kernel] balloon: speed up inflating/deflating process X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, "Michael S. Tsirkin" , Liang Li , qemu-devel@nongnu.org, Amit Shah , Cornelia Huck , Paolo Bonzini Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" X-Virus-Scanned: ClamAV using ClamSMTP The implementation of the current virtio-balloon is not very efficient, Bellow is test result of time spends on inflating the balloon to 3GB of a 4GB idle guest: a. allocating pages (6.5%, 103ms) b. sending PFNs to host (68.3%, 787ms) c. address translation (6.1%, 96ms) d. madvise (19%, 300ms) It takes about 1577ms for the whole inflating process to complete. The test shows that the bottle neck is the stage b and stage d. If using a bitmap to send the page info instead of the PFNs, we can reduce the overhead spends on stage b quite a lot. Furthermore, it's possible to do the address translation and do the madvise with a bulk of pages, instead of the current page per page way, so the overhead of stage c and stage d can also be reduced a lot. This patch is the kernel side implementation which is intended to speed up the inflating & deflating process by adding a new feature to the virtio-balloon device. And now, inflating the balloon to 3GB of a 4GB idle guest only takes 200ms, it's about 8 times as fast as before. TODO: optimize stage a by allocating/freeing a chunk of pages instead of a single page at a time. v2 change: 1. Use a small page bitmap instead of a large one. 2. Address some of comments of v1. Signed-off-by: Liang Li Suggested-by: Michael S. Tsirkin Cc: Michael S. Tsirkin Cc: Paolo Bonzini Cc: Cornelia Huck Cc: Amit Shah --- drivers/virtio/virtio_balloon.c | 207 ++++++++++++++++++++++++++++++++++-- include/uapi/linux/virtio_balloon.h | 1 + 2 files changed, 200 insertions(+), 8 deletions(-) diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c index 476c0e3..823b4e4 100644 --- a/drivers/virtio/virtio_balloon.c +++ b/drivers/virtio/virtio_balloon.c @@ -40,11 +40,19 @@ #define VIRTIO_BALLOON_ARRAY_PFNS_MAX 256 #define OOM_VBALLOON_DEFAULT_PAGES 256 #define VIRTBALLOON_OOM_NOTIFY_PRIORITY 80 +#define VIRTIO_BALLOON_PFNS_LIMIT ((2 * (1ULL << 30)) >> PAGE_SHIFT) /* 2GB */ 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"); +struct balloon_bmap_hdr { + __virtio32 type; + __virtio32 page_shift; + __virtio64 start_pfn; + __virtio64 bmap_len; +}; + struct virtio_balloon { struct virtio_device *vdev; struct virtqueue *inflate_vq, *deflate_vq, *stats_vq; @@ -62,6 +70,13 @@ struct virtio_balloon { /* Number of balloon pages we've told the Host we're not using. */ unsigned int num_pages; + /* Bitmap and length used to tell the host the pages */ + unsigned long *page_bitmap; + unsigned long bmap_len; + /* Used to record the processed pfn range */ + unsigned long min_pfn, max_pfn, start_pfn, end_pfn; + /* Used for sending page bitmap and header */ + struct scatterlist sg[2]; /* * The pages we've told the Host we're not using are enqueued * at vb_dev_info->pages list. @@ -111,15 +126,39 @@ static void balloon_ack(struct virtqueue *vq) wake_up(&vb->acked); } +static inline void init_pfn_range(struct virtio_balloon *vb) +{ + vb->min_pfn = (1UL << 48); + vb->max_pfn = 0; +} + static void tell_host(struct virtio_balloon *vb, struct virtqueue *vq) { - struct scatterlist sg; unsigned int len; - sg_init_one(&sg, vb->pfns, sizeof(vb->pfns[0]) * vb->num_pfns); + if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_PAGE_BITMAP)) { + struct balloon_bmap_hdr hdr; + unsigned long bmap_len; + + hdr.type = cpu_to_virtio32(vb->vdev, 0); + hdr.page_shift = cpu_to_virtio32(vb->vdev, PAGE_SHIFT); + hdr.start_pfn = cpu_to_virtio64(vb->vdev, vb->start_pfn); + bmap_len = min(vb->bmap_len, + (vb->end_pfn - vb->start_pfn) / BITS_PER_BYTE); + hdr.bmap_len = cpu_to_virtio64(vb-vdev, bmap_len); + sg_set_buf(&vb->sg[0], &hdr, sizeof(hdr)); + sg_set_buf(&vb->sg[1], vb->page_bitmap, bmap_len); + virtqueue_add_outbuf(vq, vb->sg, 2, vb, GFP_KERNEL); + } else { + struct scatterlist sg; + + sg_init_one(&sg, vb->pfns, sizeof(vb->pfns[0]) * vb->num_pfns); + /* We should always be able to add one buffer to an + * empty queue. + */ + virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL); + } - /* We should always be able to add one buffer to an empty queue. */ - virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL); virtqueue_kick(vq); /* When host has read buffer, this completes via balloon_ack */ @@ -139,7 +178,18 @@ static void set_page_pfns(struct virtio_balloon *vb, page_to_balloon_pfn(page) + i); } -static unsigned fill_balloon(struct virtio_balloon *vb, size_t num) +static inline void update_pfn_range(struct virtio_balloon *vb, + struct page *page) +{ + unsigned long balloon_pfn = page_to_balloon_pfn(page); + + if (balloon_pfn < vb->min_pfn) + vb->min_pfn = balloon_pfn; + if (balloon_pfn > vb->max_pfn) + vb->max_pfn = balloon_pfn; +} + +static unsigned int fill_balloon_pfns(struct virtio_balloon *vb, size_t num) { struct balloon_dev_info *vb_dev_info = &vb->vb_dev_info; unsigned num_allocated_pages; @@ -176,7 +226,126 @@ static unsigned fill_balloon(struct virtio_balloon *vb, size_t num) return num_allocated_pages; } -static void release_pages_balloon(struct virtio_balloon *vb) +static void tell_host_bitmap(struct virtio_balloon *vb, + struct list_head *pages, struct virtqueue *vq) +{ + unsigned long pfn; + struct page *page, *next; + bool find; + + vb->min_pfn = rounddown(vb->min_pfn, BITS_PER_LONG); + vb->max_pfn = roundup(vb->max_pfn, BITS_PER_LONG); + for (pfn = vb->min_pfn; pfn < vb->max_pfn; + pfn += VIRTIO_BALLOON_PFNS_LIMIT) { + vb->start_pfn = pfn; + vb->end_pfn = pfn; + memset(vb->page_bitmap, 0, vb->bmap_len); + find = false; + list_for_each_entry_safe(page, next, pages, lru) { + unsigned long balloon_pfn = page_to_balloon_pfn(page); + + if (balloon_pfn < pfn || + balloon_pfn >= pfn + VIRTIO_BALLOON_PFNS_LIMIT) + continue; + set_bit(balloon_pfn - pfn, vb->page_bitmap); + if (balloon_pfn > vb->end_pfn) + vb->end_pfn = balloon_pfn; + find = true; + } + if (find) { + vb->end_pfn = roundup(vb->end_pfn, BITS_PER_LONG); + tell_host(vb, vq); + } + } +} + +static long fill_balloon_bitmap(struct virtio_balloon *vb, size_t num) +{ + struct balloon_dev_info *vb_dev_info = &vb->vb_dev_info; + long num_allocated_pages; + + init_pfn_range(vb); + mutex_lock(&vb->balloon_lock); + for (vb->num_pfns = 0; vb->num_pfns < num; + vb->num_pfns += VIRTIO_BALLOON_PAGES_PER_PAGE) { + struct page *page = balloon_page_enqueue(vb_dev_info); + + if (!page) { + dev_info_ratelimited(&vb->vdev->dev, + "Out of puff! Can't get %u pages\n", + VIRTIO_BALLOON_PAGES_PER_PAGE); + /* Sleep for at least 1/5 of a second before retry. */ + msleep(200); + break; + } + update_pfn_range(vb, page); + vb->num_pages += VIRTIO_BALLOON_PAGES_PER_PAGE; + if (!virtio_has_feature(vb->vdev, + VIRTIO_BALLOON_F_DEFLATE_ON_OOM)) + adjust_managed_page_count(page, -1); + } + + num_allocated_pages = vb->num_pfns; + tell_host_bitmap(vb, &vb_dev_info->pages, vb->inflate_vq); + mutex_unlock(&vb->balloon_lock); + + return num_allocated_pages; +} + +static long fill_balloon(struct virtio_balloon *vb, size_t num) +{ + long num_allocated_pages; + + if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_PAGE_BITMAP)) + num_allocated_pages = fill_balloon_bitmap(vb, num); + else + num_allocated_pages = fill_balloon_pfns(vb, num); + + return num_allocated_pages; +} + +static void release_pages_balloon_bitmap(struct virtio_balloon *vb, + struct list_head *pages) +{ + struct page *page, *next; + + list_for_each_entry_safe(page, next, pages, lru) { + if (!virtio_has_feature(vb->vdev, + VIRTIO_BALLOON_F_DEFLATE_ON_OOM)) + adjust_managed_page_count(page, 1); + list_del(&page->lru); + put_page(page); + } +} + +static unsigned long leak_balloon_bitmap(struct virtio_balloon *vb, size_t num) +{ + unsigned long num_freed_pages; + struct page *page; + struct balloon_dev_info *vb_dev_info = &vb->vb_dev_info; + LIST_HEAD(pages); + + init_pfn_range(vb); + mutex_lock(&vb->balloon_lock); + for (vb->num_pfns = 0; vb->num_pfns < num; + vb->num_pfns += VIRTIO_BALLOON_PAGES_PER_PAGE) { + page = balloon_page_dequeue(vb_dev_info); + if (!page) + break; + update_pfn_range(vb, page); + list_add(&page->lru, &pages); + vb->num_pages -= VIRTIO_BALLOON_PAGES_PER_PAGE; + } + + num_freed_pages = vb->num_pfns; + tell_host_bitmap(vb, &pages, vb->deflate_vq); + release_pages_balloon_bitmap(vb, &pages); + mutex_unlock(&vb->balloon_lock); + + return num_freed_pages; +} + +static void release_pages_balloon_pfns(struct virtio_balloon *vb) { unsigned int i; struct page *page; @@ -192,7 +361,7 @@ static void release_pages_balloon(struct virtio_balloon *vb) } } -static unsigned leak_balloon(struct virtio_balloon *vb, size_t num) +static unsigned int leak_balloon_pfns(struct virtio_balloon *vb, size_t num) { unsigned num_freed_pages; struct page *page; @@ -219,11 +388,23 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num) */ if (vb->num_pfns != 0) tell_host(vb, vb->deflate_vq); - release_pages_balloon(vb); + release_pages_balloon_pfns(vb); mutex_unlock(&vb->balloon_lock); return num_freed_pages; } +static long leak_balloon(struct virtio_balloon *vb, size_t num) +{ + long num_freed_pages; + + if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_PAGE_BITMAP)) + num_freed_pages = leak_balloon_bitmap(vb, num); + else + num_freed_pages = leak_balloon_pfns(vb, num); + + return num_freed_pages; +} + static inline void update_stat(struct virtio_balloon *vb, int idx, u16 tag, u64 val) { @@ -514,6 +695,14 @@ static int virtballoon_probe(struct virtio_device *vdev) spin_lock_init(&vb->stop_update_lock); vb->stop_update = false; vb->num_pages = 0; + vb->bmap_len = ALIGN(VIRTIO_BALLOON_PFNS_LIMIT, BITS_PER_LONG) / + BITS_PER_BYTE + 2 * sizeof(unsigned long); + vb->page_bitmap = kzalloc(vb->bmap_len, GFP_KERNEL); + if (!vb->page_bitmap) { + err = -ENOMEM; + goto out; + } + sg_init_table(vb->sg, 2); mutex_init(&vb->balloon_lock); init_waitqueue_head(&vb->acked); vb->vdev = vdev; @@ -571,6 +760,7 @@ static void virtballoon_remove(struct virtio_device *vdev) cancel_work_sync(&vb->update_balloon_stats_work); remove_common(vb); + kfree(vb->page_bitmap); kfree(vb); } @@ -609,6 +799,7 @@ static unsigned int features[] = { VIRTIO_BALLOON_F_MUST_TELL_HOST, VIRTIO_BALLOON_F_STATS_VQ, VIRTIO_BALLOON_F_DEFLATE_ON_OOM, + VIRTIO_BALLOON_F_PAGE_BITMAP, }; static struct virtio_driver virtio_balloon_driver = { diff --git a/include/uapi/linux/virtio_balloon.h b/include/uapi/linux/virtio_balloon.h index 343d7dd..f78fa47 100644 --- a/include/uapi/linux/virtio_balloon.h +++ b/include/uapi/linux/virtio_balloon.h @@ -34,6 +34,7 @@ #define VIRTIO_BALLOON_F_MUST_TELL_HOST 0 /* Tell before reclaiming pages */ #define VIRTIO_BALLOON_F_STATS_VQ 1 /* Memory Stats virtqueue */ #define VIRTIO_BALLOON_F_DEFLATE_ON_OOM 2 /* Deflate balloon on OOM */ +#define VIRTIO_BALLOON_F_PAGE_BITMAP 3 /* Send page info with bitmap */ /* Size of a PFN in the balloon interface. */ #define VIRTIO_BALLOON_PFN_SHIFT 12