From: Bobby Eshleman <bobbyeshleman@gmail.com>
To: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Stefano Garzarella <sgarzare@redhat.com>,
"Michael S. Tsirkin" <mst@redhat.com>,
Bobby Eshleman <bobby.eshleman@gmail.com>,
Bobby Eshleman <bobby.eshleman@bytedance.com>,
Cong Wang <cong.wang@bytedance.com>,
Jiang Wang <jiang.wang@bytedance.com>,
Stefan Hajnoczi <stefanha@redhat.com>,
Jason Wang <jasowang@redhat.com>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
kvm@vger.kernel.org, virtualization@lists.linux-foundation.org,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] vsock: replace virtio_vsock_pkt with sk_buff
Date: Tue, 4 Oct 2022 22:50:28 +0000 [thread overview]
Message-ID: <Yzy4tEGXSZcLYkLv@bullseye> (raw)
In-Reply-To: <Yzyr60cn468ph8Io@bullseye>
On Tue, Oct 04, 2022 at 09:55:55PM +0000, Bobby Eshleman wrote:
> On Sat, Oct 15, 2022 at 12:49:59PM -0700, Cong Wang wrote:
> > On Mon, Oct 03, 2022 at 12:11:39AM +0000, Bobby Eshleman wrote:
> > > On Thu, Oct 06, 2022 at 09:34:10AM +0200, Stefano Garzarella wrote:
> > > > On Thu, Oct 06, 2022 at 03:08:12AM -0400, Michael S. Tsirkin wrote:
> > > > > On Wed, Oct 05, 2022 at 06:19:44PM -0700, Bobby Eshleman wrote:
> > > > > > This patch replaces the struct virtio_vsock_pkt with struct sk_buff.
> > > > > >
> > > > > > Using sk_buff in vsock benefits it by a) allowing vsock to be extended
> > > > > > for socket-related features like sockmap, b) vsock may in the future
> > > > > > use other sk_buff-dependent kernel capabilities, and c) vsock shares
> > > > > > commonality with other socket types.
> > > > > >
> > > > > > This patch is taken from the original series found here:
> > > > > > https://lore.kernel.org/all/cover.1660362668.git.bobby.eshleman@bytedance.com/
> > > > > >
> > > > > > Small-sized packet throughput improved by ~5% (from 18.53 Mb/s to 19.51
> > > > > > Mb/s). Tested using uperf, 16B payloads, 64 threads, 100s, averaged from
> > > > > > 10 test runs (n=10). This improvement is likely due to packet merging.
> > > > > >
> > > > > > Large-sized packet throughput decreases ~9% (from 27.25 Gb/s to 25.04
> > > > > > Gb/s). Tested using uperf, 64KB payloads, 64 threads, 100s, averaged
> > > > > > from 10 test runs (n=10).
> > > > > >
> > > > > > Medium-sized packet throughput decreases ~5% (from 4.0 Gb/s to 3.81
> > > > > > Gb/s). Tested using uperf, 4k to 8k payload sizes picked randomly
> > > > > > according to normal distribution, 64 threads, 100s, averaged from 10
> > > > > > test runs (n=10).
> > > > >
> > > > > It is surprizing to me that the original vsock code managed to outperform
> > > > > the new one, given that to my knowledge we did not focus on optimizing it.
> > > >
> > > > Yeah mee to.
> > > >
> > >
> > > Indeed.
> > >
> > > > From this numbers maybe the allocation cost has been reduced as it performs
> > > > better with small packets. But with medium to large packets we perform
> > > > worse, perhaps because previously we were allocating a contiguous buffer up
> > > > to 64k?
> > > > Instead alloc_skb() could allocate non-contiguous pages ? (which would solve
> > > > the problems we saw a few days ago)
> > > >
> > >
> > > I think this would be the case with alloc_skb_with_frags(), but
> > > internally alloc_skb() uses kmalloc() for the payload and sk_buff_head
> > > slab allocations for the sk_buff itself (all the more confusing to me,
> > > as the prior allocator also uses two separate allocations per packet).
> >
> > I think it is related to your implementation of
> > virtio_transport_add_to_queue(), where you introduced much more
> > complicated logic than before:
> >
> > - spin_lock_bh(&vsock->send_pkt_list_lock);
> > - list_add_tail(&pkt->list, &vsock->send_pkt_list);
> > - spin_unlock_bh(&vsock->send_pkt_list_lock);
> > -
> > + virtio_transport_add_to_queue(&vsock->send_pkt_queue, skb);
> >
>
> I wish it were that easy, but I included this change because it actually
> boosts performance.
>
> For 16B payloads, this change improves throughput from 16 Mb/s to 20Mb/s
> in my test harness, and reduces the memory usage of the kmalloc-512 and
> skbuff_head_cache slab caches by ~50MB at cache size peak (total slab
> cache size from ~540MB to ~390MB), but typically (not at peak) the slab
Edit: from ~590MB to ~540MB. Mixed up numbers in editing the
paragraph.
> cache size when this merging is used keeps the memory slab caches closer
> to ~150MB smaller. Tests done using uperf.
>
> For payloads greater than GOOD_COPY_LEN I don't see any any notable
> difference between the skb code with merging and the skb code without
> merging in terms of throughput. I assume this is because the skb->len
> comparison with GOOD_COPY_LEN should short circuit the expression and
> the other memory operations should not occur.
>
> > A simple list_add_tail() is definitely faster than your
> > virtio_transport_skbs_can_merge() check. So, why do you have to merge
> > skb while we don't merge virtio_vsock_pkt?
> >
>
> sk_buff is over twice the size of virtio_vsock_pkt (96B vs 232B). It
> seems wise to reduce the footprint in other ways to try and keep it
> comparable.
>
> > _If_ you are trying to mimic TCP, I think you are doing it wrong, it can
> > be much more efficient if you could do the merge in sendmsg() before skb
> > is even allocated, see tcp_sendmsg_locked().
>
> I'll definitely give it a read, merging before allocating an skb sounds
> better.
>
> Best,
> Bobby
next prev parent reply other threads:[~2022-10-18 3:42 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-06 1:19 [PATCH v2] vsock: replace virtio_vsock_pkt with sk_buff Bobby Eshleman
2022-10-06 3:54 ` Jakub Kicinski
2022-10-06 7:08 ` Michael S. Tsirkin
2022-10-06 7:34 ` Stefano Garzarella
2022-10-03 0:11 ` Bobby Eshleman
2022-10-15 19:49 ` Cong Wang
2022-10-04 21:55 ` Bobby Eshleman
2022-10-04 22:50 ` Bobby Eshleman [this message]
2022-10-14 13:31 ` Stefano Garzarella
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=Yzy4tEGXSZcLYkLv@bullseye \
--to=bobbyeshleman@gmail.com \
--cc=bobby.eshleman@bytedance.com \
--cc=bobby.eshleman@gmail.com \
--cc=cong.wang@bytedance.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=jasowang@redhat.com \
--cc=jiang.wang@bytedance.com \
--cc=kuba@kernel.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mst@redhat.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sgarzare@redhat.com \
--cc=stefanha@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
--cc=xiyou.wangcong@gmail.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).