From: Gavin Hu <Gavin.Hu@arm.com>
To: Marvin Liu <yong.liu@intel.com>,
"maxime.coquelin@redhat.com" <maxime.coquelin@redhat.com>,
"xiaolong.ye@intel.com" <xiaolong.ye@intel.com>,
"zhihong.wang@intel.com" <zhihong.wang@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, nd <nd@arm.com>
Subject: Re: [dpdk-dev] [PATCH v2 2/2] vhost: cache gpa to hpa translation
Date: Wed, 1 Apr 2020 10:07:15 +0000 [thread overview]
Message-ID: <VI1PR08MB5376606F77020027527404C68FC90@VI1PR08MB5376.eurprd08.prod.outlook.com> (raw)
In-Reply-To: <20200401145011.67357-2-yong.liu@intel.com>
Hi Marvin,
> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Marvin Liu
> Sent: Wednesday, April 1, 2020 10:50 PM
> To: maxime.coquelin@redhat.com; xiaolong.ye@intel.com;
> zhihong.wang@intel.com
> Cc: dev@dpdk.org; Marvin Liu <yong.liu@intel.com>
> Subject: [dpdk-dev] [PATCH v2 2/2] vhost: cache gpa to hpa translation
>
> If Tx zero copy enabled, gpa to hpa mapping table is updated one by
> one. This will harm performance when guest memory backend using 2M
> hugepages. Now add cached mapping table which will sorted by using
> sequence. Address translation will first check cached mapping table,
> then check unsorted mapping table if no match found.
>
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
>
> diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
> index 2087d1400..5cb0e83dd 100644
> --- a/lib/librte_vhost/vhost.h
> +++ b/lib/librte_vhost/vhost.h
> @@ -368,7 +368,9 @@ struct virtio_net {
> struct vhost_device_ops const *notify_ops;
>
> uint32_t nr_guest_pages;
> + uint32_t nr_cached_guest_pages;
> uint32_t max_guest_pages;
> + struct guest_page *cached_guest_pages;
> struct guest_page *guest_pages;
>
> int slave_req_fd;
> @@ -553,12 +555,25 @@ gpa_to_hpa(struct virtio_net *dev, uint64_t gpa,
> uint64_t size)
> {
> uint32_t i;
> struct guest_page *page;
> + uint32_t cached_pages = dev->nr_cached_guest_pages;
> +
> + for (i = 0; i < cached_pages; i++) {
> + page = &dev->cached_guest_pages[i];
> + if (gpa >= page->guest_phys_addr &&
> + gpa + size < page->guest_phys_addr + page->size) {
> + return gpa - page->guest_phys_addr +
> + page->host_phys_addr;
> + }
> + }
Sorry, I did not see any speedup with cached guest pages in comparison to the old code below.
Is it not a simple copy?
Is it a better idea to use hash instead to speed up the translation?
/Gavin
>
> for (i = 0; i < dev->nr_guest_pages; i++) {
> page = &dev->guest_pages[i];
>
> if (gpa >= page->guest_phys_addr &&
> gpa + size < page->guest_phys_addr + page->size) {
> + rte_memcpy(&dev-
> >cached_guest_pages[cached_pages],
> + page, sizeof(struct guest_page));
> + dev->nr_cached_guest_pages++;
> return gpa - page->guest_phys_addr +
> page->host_phys_addr;
> }
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index 79fcb9d19..1bae1fddc 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -192,7 +192,9 @@ vhost_backend_cleanup(struct virtio_net *dev)
> }
>
> rte_free(dev->guest_pages);
> + rte_free(dev->cached_guest_pages);
> dev->guest_pages = NULL;
> + dev->cached_guest_pages = NULL;
>
> if (dev->log_addr) {
> munmap((void *)(uintptr_t)dev->log_addr, dev->log_size);
> @@ -898,7 +900,7 @@ add_one_guest_page(struct virtio_net *dev,
> uint64_t guest_phys_addr,
> uint64_t host_phys_addr, uint64_t size)
> {
> struct guest_page *page, *last_page;
> - struct guest_page *old_pages;
> + struct guest_page *old_pages, *old_cached_pages;
>
> if (dev->nr_guest_pages == dev->max_guest_pages) {
> dev->max_guest_pages *= 2;
> @@ -906,9 +908,19 @@ add_one_guest_page(struct virtio_net *dev,
> uint64_t guest_phys_addr,
> dev->guest_pages = rte_realloc(dev->guest_pages,
> dev->max_guest_pages *
> sizeof(*page),
> RTE_CACHE_LINE_SIZE);
> - if (dev->guest_pages == NULL) {
> + old_cached_pages = dev->cached_guest_pages;
> + dev->cached_guest_pages = rte_realloc(dev-
> >cached_guest_pages,
> + dev->max_guest_pages *
> + sizeof(*page),
> + RTE_CACHE_LINE_SIZE);
> + dev->nr_cached_guest_pages = 0;
> + if (dev->guest_pages == NULL ||
> + dev->cached_guest_pages == NULL) {
> VHOST_LOG_CONFIG(ERR, "cannot realloc
> guest_pages\n");
> rte_free(old_pages);
> + rte_free(old_cached_pages);
> + dev->guest_pages = NULL;
> + dev->cached_guest_pages = NULL;
> return -1;
> }
> }
> @@ -1078,6 +1090,20 @@ vhost_user_set_mem_table(struct virtio_net
> **pdev, struct VhostUserMsg *msg,
> }
> }
>
> + if (dev->cached_guest_pages == NULL) {
> + dev->cached_guest_pages = rte_zmalloc(NULL,
> + dev->max_guest_pages *
> + sizeof(struct guest_page),
> + RTE_CACHE_LINE_SIZE);
> + if (dev->cached_guest_pages == NULL) {
> + VHOST_LOG_CONFIG(ERR,
> + "(%d) failed to allocate memory "
> + "for dev->cached_guest_pages\n",
> + dev->vid);
> + return RTE_VHOST_MSG_RESULT_ERR;
> + }
> + }
> +
> dev->mem = rte_zmalloc("vhost-mem-table", sizeof(struct
> rte_vhost_memory) +
> sizeof(struct rte_vhost_mem_region) * memory->nregions,
> 0);
> if (dev->mem == NULL) {
> --
> 2.17.1
next prev parent reply other threads:[~2020-04-01 10:07 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-16 15:33 [dpdk-dev] [PATCH] vhost: cache guest/vhost physical address mapping Marvin Liu
2020-03-16 13:48 ` Ye Xiaolong
2020-03-17 1:01 ` Liu, Yong
2020-04-01 14:50 ` [dpdk-dev] [PATCH v2 1/2] vhost: utilize dpdk dynamic memory allocator Marvin Liu
2020-04-01 10:08 ` Gavin Hu
2020-04-01 14:50 ` [dpdk-dev] [PATCH v2 2/2] vhost: cache gpa to hpa translation Marvin Liu
2020-04-01 10:07 ` Gavin Hu [this message]
2020-04-01 13:01 ` Liu, Yong
2020-04-02 3:04 ` Gavin Hu
2020-04-02 4:45 ` Liu, Yong
2020-04-03 8:22 ` Ma, LihongX
2020-04-02 2:57 ` Ye Xiaolong
2020-04-27 8:45 ` Maxime Coquelin
2020-04-28 0:44 ` Liu, Yong
2020-04-02 2:57 ` [dpdk-dev] [PATCH v2 1/2] vhost: utilize dpdk dynamic memory allocator Ye Xiaolong
2020-04-03 8:22 ` Ma, LihongX
2020-04-15 11:15 ` Maxime Coquelin
2020-04-28 9:13 ` [dpdk-dev] [PATCH v3 " Marvin Liu
2020-04-28 9:13 ` [dpdk-dev] [PATCH v3 2/2] vhost: binary search address mapping table Marvin Liu
2020-04-28 15:28 ` Maxime Coquelin
2020-04-28 15:38 ` Liu, Yong
2020-04-28 12:51 ` [dpdk-dev] [PATCH v3 1/2] vhost: utilize dpdk dynamic memory allocator Maxime Coquelin
2020-04-29 1:00 ` [dpdk-dev] [PATCH v4 1/2] net/virtio: add support Virtio link speed feature Marvin Liu
2020-04-29 1:00 ` [dpdk-dev] [PATCH v4 1/2] vhost: utilize dpdk dynamic memory allocator Marvin Liu
2020-04-29 1:00 ` [dpdk-dev] [PATCH v4 2/2] vhost: binary search address mapping table Marvin Liu
2020-04-29 1:01 ` [dpdk-dev] [PATCH v4 2/2] vhost: utilize dpdk dynamic memory allocator Marvin Liu
2020-04-29 1:06 ` Liu, Yong
2020-04-29 17:47 ` Maxime Coquelin
2020-04-29 1:04 ` [dpdk-dev] [PATCH v4 1/2] " Marvin Liu
2020-04-29 1:04 ` [dpdk-dev] [PATCH v4 2/2] vhost: binary search address mapping table Marvin Liu
2020-04-29 11:50 ` Maxime Coquelin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=VI1PR08MB5376606F77020027527404C68FC90@VI1PR08MB5376.eurprd08.prod.outlook.com \
--to=gavin.hu@arm.com \
--cc=dev@dpdk.org \
--cc=maxime.coquelin@redhat.com \
--cc=nd@arm.com \
--cc=xiaolong.ye@intel.com \
--cc=yong.liu@intel.com \
--cc=zhihong.wang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).