On 2018/12/25 下午8:50, Michael S. Tsirkin wrote:
On Tue, Dec 25, 2018 at 06:05:25PM +0800, Jason Wang wrote:
On 2018/12/25 上午2:10, Michael S. Tsirkin wrote:
On Mon, Dec 24, 2018 at 03:53:16PM +0800, Jason Wang wrote:
On 2018/12/14 下午8:36, Michael S. Tsirkin wrote:
On Fri, Dec 14, 2018 at 11:57:35AM +0800, Jason Wang wrote:
On 2018/12/13 下午11:44, Michael S. Tsirkin wrote:
On Thu, Dec 13, 2018 at 06:10:22PM +0800, Jason Wang wrote:
It was noticed that the copy_user() friends that was used to access
virtqueue metdata tends to be very expensive for dataplane
implementation like vhost since it involves lots of software check,
speculation barrier, hardware feature toggling (e.g SMAP). The
extra cost will be more obvious when transferring small packets.
This patch tries to eliminate those overhead by pin vq metadata pages
and access them through vmap(). During SET_VRING_ADDR, we will setup
those mappings and memory accessors are modified to use pointers to
access the metadata directly.
Note, this was only done when device IOTLB is not enabled. We could
use similar method to optimize it in the future.
Tests shows about ~24% improvement on TX PPS when using virtio-user +
vhost_net + xdp1 on TAP (CONFIG_HARDENED_USERCOPY is not enabled):
Before: ~5.0Mpps
After: ~6.1Mpps
Signed-off-by: Jason Wang<jasowang@xxxxxxxxxx>
---
drivers/vhost/vhost.c | 178 ++++++++++++++++++++++++++++++++++++++++++
drivers/vhost/vhost.h | 11 +++
2 files changed, 189 insertions(+)
diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
index bafe39d2e637..1bd24203afb6 100644
--- a/drivers/vhost/vhost.c
+++ b/drivers/vhost/vhost.c
@@ -443,6 +443,9 @@ void vhost_dev_init(struct vhost_dev *dev,
vq->indirect = NULL;
vq->heads = NULL;
vq->dev = dev;
+ memset(&vq->avail_ring, 0, sizeof(vq->avail_ring));
+ memset(&vq->used_ring, 0, sizeof(vq->used_ring));
+ memset(&vq->desc_ring, 0, sizeof(vq->desc_ring));
mutex_init(&vq->mutex);
vhost_vq_reset(dev, vq);
if (vq->handle_kick)
@@ -614,6 +617,102 @@ static void vhost_clear_msg(struct vhost_dev *dev)
spin_unlock(&dev->iotlb_lock);
}
+static int vhost_init_vmap(struct vhost_vmap *map, unsigned long uaddr,
+ size_t size, int write)
+{
+ struct page **pages;
+ int npages = DIV_ROUND_UP(size, PAGE_SIZE);
+ int npinned;
+ void *vaddr;
+
+ pages = kmalloc_array(npages, sizeof(struct page *), GFP_KERNEL);
+ if (!pages)
+ return -ENOMEM;
+
+ npinned = get_user_pages_fast(uaddr, npages, write, pages);
+ if (npinned != npages)
+ goto err;
+
As I said I have doubts about the whole approach, but this
implementation in particular isn't a good idea
as it keeps the page around forever.
The pages wil be released during set features.
So no THP, no NUMA rebalancing,
For THP, we will probably miss 2 or 4 pages, but does this really matter
consider the gain we have?
We as in vhost? networking isn't the only thing guest does.
We don't even know if this guest does a lot of networking.
You don't
know what else is in this huge page. Can be something very important
that guest touches all the time.
Well, the probability should be very small consider we usually give several
gigabytes to guest. The rest of the pages that doesn't sit in the same
hugepage with metadata can still be merged by THP. Anyway, I can test the
differences.
Thanks!
For NUMA rebalancing, I'm even not quite sure if
it can helps for the case of IPC (vhost). It looks to me the worst case it
may cause page to be thrash between nodes if vhost and userspace are running
in two nodes.
So again it's a gain for vhost but has a completely unpredictable effect on
other functionality of the guest.
That's what bothers me with this approach.
So:
- The rest of the pages could still be balanced to other nodes, no?
- try to balance metadata pages (belongs to co-operate processes) itself is
still questionable
I am not sure why. It should be easy enough to force the VCPU and vhost
to move (e.g. start them pinned to 1 cpu, then pin them to another one).
Clearly sometimes this would be necessary for load balancing reasons.
Yes, but it looks to me the part of motivation of auto NUMA is to avoid
manual pinning.
With autonuma after a while (could take seconds but it will happen) the
memory will migrate.
Yes. As you mentioned during the discuss, I wonder we could do it
similarly through mmu notifier like APIC access page in commit
c24ae0dcd3e ("kvm: x86: Unpin and remove kvm_arch->apic_access_page")
This is the price of all GUP users not only vhost itself.
Yes. GUP is just not a great interface for vhost to use.
Zerocopy codes (enabled by defualt) use them for years.
But only for TX and temporarily. We pin, read, unpin.
Probably not. For several reasons that the page will be not be released soon
or held for a very long period of time or even forever.
With zero copy? Well it's pinned until transmit. Takes a while
but could be enough for autocopy to work esp since
its the packet memory so not reused immediately.
Your patch is different
- it writes into memory and GUP has known issues with file
backed memory
The ordinary user for vhost is anonymous pages I think?
It's not the most common scenario and not the fastest one
(e.g. THP does not work) but file backed is useful sometimes.
It would not be nice at all to corrupt guest memory in that case.
Ok.
- it keeps pages pinned forever
What's more
important, the goal is not to be left too much behind for other backends
like DPDK or AF_XDP (all of which are using GUP).
So these guys assume userspace knows what it's doing.
We can't assume that.
What kind of assumption do you they have?
userspace-controlled
amount of memory locked up and not accounted for.
It's pretty easy to add this since the slow path was still kept. If we
exceeds the limitation, we can switch back to slow path.
Don't get me wrong it's a great patch in an ideal world.
But then in an ideal world no barriers smap etc are necessary at all.
Again, this is only for metadata accessing not the data which has been used
for years for real use cases.
For SMAP, it makes senses for the address that kernel can not forcast. But
it's not the case for the vhost metadata since we know the address will be
accessed very frequently. For speculation barrier, it helps nothing for the
data path of vhost which is a kthread.
I don't see how a kthread makes any difference. We do have a validation
step which makes some difference.
The problem is not kthread but the address of userspace address. The
addresses of vq metadata tends to be consistent for a while, and vhost knows
they will be frequently. SMAP doesn't help too much in this case.
Thanks.
It's true for a real life applications but a malicious one
can call the setup ioctls any number of times. And SMAP is
all about malcious applications.
We don't do this in the path of ioctl, there's no context switch between
userspace and kernel in the worker thread. SMAP is used to prevent kernel
from accessing userspace pages unexpectedly which is not the case for
metadata access.
Thanks
OK let's forget smap for now.
Some numbers I measured:
On an old Sandy bridge machine without SMAP support. Remove speculation
barrier boost the performance from 4.6Mpps to 5.1Mpps
On a newer Broadwell machine with SMAP support. Remove speculation
barrier only gives 2%-5% improvement, disable SMAP completely through
Kconfig boost 57% performance from 4.8Mpps to 7.5Mpps. (Vmap gives 6Mpps
- 6.1Mpps, it only bypass SMAP for metadata).
So it looks like for recent machine, SMAP becomes pain point when the
copy is short (e.g 64B) for high PPS.
Thanks
Packet or AF_XDP benefit from
accessing metadata directly, we should do it as well.
Thanks
