linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Xuan Zhuo <xuanzhuo@linux.alibaba.com>,
	Eric Dumazet <edumazet@google.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	Jason Wang <jasowang@redhat.com>,
	virtualization@lists.linux-foundation.org,
	"David S. Miller" <davem@davemloft.net>,
	Matthieu Baerts <matthieu.baerts@tessares.net>
Subject: [PATCH 4.14 14/38] virtio_net: Do not pull payload in skb->head
Date: Mon,  2 Aug 2021 15:44:36 +0200	[thread overview]
Message-ID: <20210802134335.285560278@linuxfoundation.org> (raw)
In-Reply-To: <20210802134334.835358048@linuxfoundation.org>

From: Eric Dumazet <edumazet@google.com>

commit 0f6925b3e8da0dbbb52447ca8a8b42b371aac7db upstream.

Xuan Zhuo reported that commit 3226b158e67c ("net: avoid 32 x truesize
under-estimation for tiny skbs") brought  a ~10% performance drop.

The reason for the performance drop was that GRO was forced
to chain sk_buff (using skb_shinfo(skb)->frag_list), which
uses more memory but also cause packet consumers to go over
a lot of overhead handling all the tiny skbs.

It turns out that virtio_net page_to_skb() has a wrong strategy :
It allocates skbs with GOOD_COPY_LEN (128) bytes in skb->head, then
copies 128 bytes from the page, before feeding the packet to GRO stack.

This was suboptimal before commit 3226b158e67c ("net: avoid 32 x truesize
under-estimation for tiny skbs") because GRO was using 2 frags per MSS,
meaning we were not packing MSS with 100% efficiency.

Fix is to pull only the ethernet header in page_to_skb()

Then, we change virtio_net_hdr_to_skb() to pull the missing
headers, instead of assuming they were already pulled by callers.

This fixes the performance regression, but could also allow virtio_net
to accept packets with more than 128bytes of headers.

Many thanks to Xuan Zhuo for his report, and his tests/help.

Fixes: 3226b158e67c ("net: avoid 32 x truesize under-estimation for tiny skbs")
Reported-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Link: https://www.spinics.net/lists/netdev/msg731397.html
Co-Developed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: virtualization@lists.linux-foundation.org
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/net/virtio_net.c   |   10 +++++++---
 include/linux/virtio_net.h |   14 +++++++++-----
 2 files changed, 16 insertions(+), 8 deletions(-)

--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -339,9 +339,13 @@ static struct sk_buff *page_to_skb(struc
 	offset += hdr_padded_len;
 	p += hdr_padded_len;
 
-	copy = len;
-	if (copy > skb_tailroom(skb))
-		copy = skb_tailroom(skb);
+	/* Copy all frame if it fits skb->head, otherwise
+	 * we let virtio_net_hdr_to_skb() and GRO pull headers as needed.
+	 */
+	if (len <= skb_tailroom(skb))
+		copy = len;
+	else
+		copy = ETH_HLEN;
 	skb_put_data(skb, p, copy);
 
 	len -= copy;
--- a/include/linux/virtio_net.h
+++ b/include/linux/virtio_net.h
@@ -65,14 +65,18 @@ static inline int virtio_net_hdr_to_skb(
 	skb_reset_mac_header(skb);
 
 	if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
-		u16 start = __virtio16_to_cpu(little_endian, hdr->csum_start);
-		u16 off = __virtio16_to_cpu(little_endian, hdr->csum_offset);
+		u32 start = __virtio16_to_cpu(little_endian, hdr->csum_start);
+		u32 off = __virtio16_to_cpu(little_endian, hdr->csum_offset);
+		u32 needed = start + max_t(u32, thlen, off + sizeof(__sum16));
+
+		if (!pskb_may_pull(skb, needed))
+			return -EINVAL;
 
 		if (!skb_partial_csum_set(skb, start, off))
 			return -EINVAL;
 
 		p_off = skb_transport_offset(skb) + thlen;
-		if (p_off > skb_headlen(skb))
+		if (!pskb_may_pull(skb, p_off))
 			return -EINVAL;
 	} else {
 		/* gso packets without NEEDS_CSUM do not set transport_offset.
@@ -100,14 +104,14 @@ retry:
 			}
 
 			p_off = keys.control.thoff + thlen;
-			if (p_off > skb_headlen(skb) ||
+			if (!pskb_may_pull(skb, p_off) ||
 			    keys.basic.ip_proto != ip_proto)
 				return -EINVAL;
 
 			skb_set_transport_header(skb, keys.control.thoff);
 		} else if (gso_type) {
 			p_off = thlen;
-			if (p_off > skb_headlen(skb))
+			if (!pskb_may_pull(skb, p_off))
 				return -EINVAL;
 		}
 	}



  parent reply	other threads:[~2021-08-02 13:52 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-02 13:44 [PATCH 4.14 00/38] 4.14.242-rc1 review Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 01/38] selftest: fix build error in tools/testing/selftests/vm/userfaultfd.c Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 02/38] KVM: x86: determine if an exception has an error code only when injecting it Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 03/38] net: split out functions related to registering inflight socket files Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 04/38] af_unix: fix garbage collect vs MSG_PEEK Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 05/38] workqueue: fix UAF in pwq_unbound_release_workfn() Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 06/38] net/802/mrp: fix memleak in mrp_request_join() Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 07/38] net/802/garp: fix memleak in garp_request_join() Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 08/38] net: annotate data race around sk_ll_usec Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 09/38] sctp: move 198 addresses from unusable to private scope Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 10/38] hfs: add missing clean-up in hfs_fill_super Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 11/38] hfs: fix high memory mapping in hfs_bnode_read Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 12/38] hfs: add lock nesting notation to hfs_find_init Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 13/38] ARM: dts: versatile: Fix up interrupt controller node names Greg Kroah-Hartman
2021-08-02 13:44 ` Greg Kroah-Hartman [this message]
2021-08-02 13:44 ` [PATCH 4.14 15/38] gro: ensure frag0 meets IP header alignment Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 16/38] x86/kvm: fix vcpu-id indexed array sizes Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 17/38] ocfs2: fix zero out valid data Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 18/38] ocfs2: issue zeroout to EOF blocks Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 19/38] can: raw: raw_setsockopt(): fix raw_rcv panic for sock UAF Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 20/38] can: mcba_usb_start(): add missing urb->transfer_dma initialization Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 21/38] can: usb_8dev: fix memory leak Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 22/38] can: ems_usb: " Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 23/38] can: esd_usb2: " Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 24/38] NIU: fix incorrect error return, missed in previous revert Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 25/38] nfc: nfcsim: fix use after free during module unload Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 26/38] x86/asm: Ensure asm/proto.h can be included stand-alone Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 27/38] cfg80211: Fix possible memory leak in function cfg80211_bss_update Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 28/38] netfilter: conntrack: adjust stop timestamp to real expiry value Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 29/38] netfilter: nft_nat: allow to specify layer 4 protocol NAT only Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 30/38] tipc: fix sleeping in tipc accept routine Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 31/38] mlx4: Fix missing error code in mlx4_load_one() Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 32/38] net: llc: fix skb_over_panic Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 33/38] net/mlx5: Fix flow table chaining Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 34/38] sctp: fix return value check in __sctp_rcv_asconf_lookup Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 35/38] tulip: windbond-840: Fix missing pci_disable_device() in probe and remove Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 36/38] sis900: " Greg Kroah-Hartman
2021-08-02 13:44 ` [PATCH 4.14 37/38] can: hi311x: fix a signedness bug in hi3110_cmd() Greg Kroah-Hartman
2021-08-02 13:45 ` [PATCH 4.14 38/38] Revert "perf map: Fix dso->nsinfo refcounting" Greg Kroah-Hartman
2021-08-03 13:17 ` [PATCH 4.14 00/38] 4.14.242-rc1 review Naresh Kamboju
2021-08-03 19:15 ` Guenter Roeck
2021-08-04  2:56 ` Samuel Zou

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=20210802134335.285560278@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=jasowang@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=matthieu.baerts@tessares.net \
    --cc=mst@redhat.com \
    --cc=stable@vger.kernel.org \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=xuanzhuo@linux.alibaba.com \
    --subject='Re: [PATCH 4.14 14/38] virtio_net: Do not pull payload in skb->head' \
    /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

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).