After commit e2b3b35eb989 ("vhost_net: batch used ring update in rx"),
we tend to batch updating used heads. But it doesn't flush batched
heads before trying to do busy polling, this will cause vhost to wait
for guest TX which waits for the used RX. Fixing by flush batched
heads before busy loop.
1 byte TCP_RR performance recovers from 13107.83 to 50402.65.
Fixes: e2b3b35eb989 ("vhost_net: batch used ring update in rx")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
DaeRyong Jeong reports a race between vhost_dev_cleanup() and
vhost_process_iotlb_msg():
Thread interleaving:
CPU0 (vhost_process_iotlb_msg) CPU1 (vhost_dev_cleanup)
(In the case of both VHOST_IOTLB_UPDATE and
VHOST_IOTLB_INVALIDATE)
===== =====
vhost_umem_clean(dev->iotlb);
if (!dev->iotlb) {
ret = -EFAULT;
break;
}
dev->iotlb = NULL;
The reason is we don't synchronize between them, fixing by protecting
vhost_process_iotlb_msg() with dev mutex.
Reported-by: DaeRyong Jeong <threeearcat@gmail.com>
Fixes: 6b1e6cc7855b0 ("vhost: new device IOTLB API")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently vhost *_access_ok() functions return int. This is error-prone because there are two popular conventions: 1. 0 means failure, 1 means success 2. -errno means failure, 0 means success Although vhost mostly uses #1, it does not do so consistently. umem_access_ok() uses #2. This patch changes the return type from int to bool so that false means failure and true means success. This eliminates a potential source of errors. Suggested-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Commit d65026c6c62e7d9616c8ceb5a53b68bcdc050525 ("vhost: validate log
when IOTLB is enabled") introduced a regression. The logic was
originally:
if (vq->iotlb)
return 1;
return A && B;
After the patch the short-circuit logic for A was inverted:
if (A || vq->iotlb)
return A;
return B;
This patch fixes the regression by rewriting the checks in the obvious
way, no longer returning A when vq->iotlb is non-NULL (which is hard to
understand).
Reported-by: syzbot+65a84dde0214b0387ccd@syzkaller.appspotmail.com
Cc: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
vhost_copy_to_user is used to copy vring used elements to userspace.
We should use VHOST_ADDR_USED instead of VHOST_ADDR_DESC.
Fixes: f88949138058 ("vhost: introduce O(1) vq metadata cache")
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
1) The sockmap code has to free socket memory on close if there is
corked data, from John Fastabend.
2) Tunnel names coming from userspace need to be length validated. From
Eric Dumazet.
3) arp_filter() has to take VRFs properly into account, from Miguel
Fadon Perlines.
4) Fix oops in error path of tcf_bpf_init(), from Davide Caratti.
5) Missing idr_remove() in u32_delete_key(), from Cong Wang.
6) More syzbot stuff. Several use of uninitialized value fixes all
over, from Eric Dumazet.
7) Do not leak kernel memory to userspace in sctp, also from Eric
Dumazet.
8) Discard frames from unused ports in DSA, from Andrew Lunn.
9) Fix DMA mapping and reset/failover problems in ibmvnic, from Thomas
Falcon.
10) Do not access dp83640 PHY registers prematurely after reset, from
Esben Haabendal.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (46 commits)
vhost-net: set packet weight of tx polling to 2 * vq size
net: thunderx: rework mac addresses list to u64 array
inetpeer: fix uninit-value in inet_getpeer
dp83640: Ensure against premature access to PHY registers after reset
devlink: convert occ_get op to separate registration
ARM: dts: ls1021a: Specify TBIPA register address
net/fsl_pq_mdio: Allow explicit speficition of TBIPA address
ibmvnic: Do not reset CRQ for Mobility driver resets
ibmvnic: Fix failover case for non-redundant configuration
ibmvnic: Fix reset scheduler error handling
ibmvnic: Zero used TX descriptor counter on reset
ibmvnic: Fix DMA mapping mistakes
tipc: use the right skb in tipc_sk_fill_sock_diag()
sctp: sctp_sockaddr_af must check minimal addr length for AF_INET6
net: dsa: Discard frames from unused ports
sctp: do not leak kernel memory to user space
soreuseport: initialise timewait reuseport field
ipv4: fix uninit-value in ip_route_output_key_hash_rcu()
dccp: initialize ireq->ir_mark
net: fix uninit-value in __hw_addr_add_ex()
...
handle_tx will delay rx for tens or even hundreds of milliseconds when tx busy
polling udp packets with small length(e.g. 1byte udp payload), because setting
VHOST_NET_WEIGHT takes into account only sent-bytes but no single packet length.
Ping-Latencies shown below were tested between two Virtual Machines using
netperf (UDP_STREAM, len=1), and then another machine pinged the client:
vq size=256
Packet-Weight Ping-Latencies(millisecond)
min avg max
Origin 3.319 18.489 57.303
64 1.643 2.021 2.552
128 1.825 2.600 3.224
256 1.997 2.710 4.295
512 1.860 3.171 4.631
1024 2.002 4.173 9.056
2048 2.257 5.650 9.688
4096 2.093 8.508 15.943
vq size=512
Packet-Weight Ping-Latencies(millisecond)
min avg max
Origin 6.537 29.177 66.245
64 2.798 3.614 4.403
128 2.861 3.820 4.775
256 3.008 4.018 4.807
512 3.254 4.523 5.824
1024 3.079 5.335 7.747
2048 3.944 8.201 12.762
4096 4.158 11.057 19.985
Seems pretty consistent, a small dip at 2 VQ sizes.
Ring size is a hint from device about a burst size it can tolerate. Based on
benchmarks, set the weight to 2 * vq size.
To evaluate this change, another tests were done using netperf(RR, TX) between
two machines with Intel(R) Xeon(R) Gold 6133 CPU @ 2.50GHz, and vq size was
tweaked through qemu. Results shown below does not show obvious changes.
vq size=256 TCP_RR vq size=512 TCP_RR
size/sessions/+thu%/+normalize% size/sessions/+thu%/+normalize%
1/ 1/ -7%/ -2% 1/ 1/ 0%/ -2%
1/ 4/ +1%/ 0% 1/ 4/ +1%/ 0%
1/ 8/ +1%/ -2% 1/ 8/ 0%/ +1%
64/ 1/ -6%/ 0% 64/ 1/ +7%/ +3%
64/ 4/ 0%/ +2% 64/ 4/ -1%/ +1%
64/ 8/ 0%/ 0% 64/ 8/ -1%/ -2%
256/ 1/ -3%/ -4% 256/ 1/ -4%/ -2%
256/ 4/ +3%/ +4% 256/ 4/ +1%/ +2%
256/ 8/ +2%/ 0% 256/ 8/ +1%/ -1%
vq size=256 UDP_RR vq size=512 UDP_RR
size/sessions/+thu%/+normalize% size/sessions/+thu%/+normalize%
1/ 1/ -5%/ +1% 1/ 1/ -3%/ -2%
1/ 4/ +4%/ +1% 1/ 4/ -2%/ +2%
1/ 8/ -1%/ -1% 1/ 8/ -1%/ 0%
64/ 1/ -2%/ -3% 64/ 1/ +1%/ +1%
64/ 4/ -5%/ -1% 64/ 4/ +2%/ 0%
64/ 8/ 0%/ -1% 64/ 8/ -2%/ +1%
256/ 1/ +7%/ +1% 256/ 1/ -7%/ 0%
256/ 4/ +1%/ +1% 256/ 4/ -3%/ -4%
256/ 8/ +2%/ +2% 256/ 8/ +1%/ +1%
vq size=256 TCP_STREAM vq size=512 TCP_STREAM
size/sessions/+thu%/+normalize% size/sessions/+thu%/+normalize%
64/ 1/ 0%/ -3% 64/ 1/ 0%/ 0%
64/ 4/ +3%/ -1% 64/ 4/ -2%/ +4%
64/ 8/ +9%/ -4% 64/ 8/ -1%/ +2%
256/ 1/ +1%/ -4% 256/ 1/ +1%/ +1%
256/ 4/ -1%/ -1% 256/ 4/ -3%/ 0%
256/ 8/ +7%/ +5% 256/ 8/ -3%/ 0%
512/ 1/ +1%/ 0% 512/ 1/ -1%/ -1%
512/ 4/ +1%/ -1% 512/ 4/ 0%/ 0%
512/ 8/ +7%/ -5% 512/ 8/ +6%/ -1%
1024/ 1/ 0%/ -1% 1024/ 1/ 0%/ +1%
1024/ 4/ +3%/ 0% 1024/ 4/ +1%/ 0%
1024/ 8/ +8%/ +5% 1024/ 8/ -1%/ 0%
2048/ 1/ +2%/ +2% 2048/ 1/ -1%/ 0%
2048/ 4/ +1%/ 0% 2048/ 4/ 0%/ -1%
2048/ 8/ -2%/ 0% 2048/ 8/ 5%/ -1%
4096/ 1/ -2%/ 0% 4096/ 1/ -2%/ 0%
4096/ 4/ +2%/ 0% 4096/ 4/ 0%/ 0%
4096/ 8/ +9%/ -2% 4096/ 8/ -5%/ -1%
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Haibin Zhang <haibinzhang@tencent.com>
Signed-off-by: Yunfang Tai <yunfangtai@tencent.com>
Signed-off-by: Lidong Chen <lidongchen@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull fw_cfg, vhost updates from Michael Tsirkin: "This cleans up the qemu fw cfg device driver. On top of this, vmcore is dumped there on crash to help debugging with kASLR enabled. Also included are some fixes in vhost" * tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: vhost: add vsock compat ioctl vhost: fix vhost ioctl signature to build with clang fw_cfg: write vmcoreinfo details crash: export paddr_vmcoreinfo_note() fw_cfg: add DMA register fw_cfg: add a public uapi header fw_cfg: handle fw_cfg_read_blob() error fw_cfg: remove inline from fw_cfg_read_blob() fw_cfg: fix sparse warnings around FW_CFG_FILE_DIR read fw_cfg: fix sparse warning reading FW_CFG_ID fw_cfg: fix sparse warnings with fw_cfg_file fw_cfg: fix sparse warnings in fw_cfg_sel_endianness() ptr_ring: fix build
Minor conflicts in drivers/net/ethernet/mellanox/mlx5/core/en_rep.c,
we had some overlapping changes:
1) In 'net' MLX5E_PARAMS_LOG_{SQ,RQ}_SIZE -->
MLX5E_REP_PARAMS_LOG_{SQ,RQ}_SIZE
2) In 'net-next' params->log_rq_size is renamed to be
params->log_rq_mtu_frames.
3) In 'net-next' params->hard_mtu is added.
Signed-off-by: David S. Miller <davem@davemloft.net>
Vq log_base is the userspace address of bitmap which has nothing to do
with IOTLB. So it needs to be validated unconditionally otherwise we
may try use 0 as log_base which may lead to pin pages that will lead
unexpected result (e.g trigger BUG_ON() in set_bit_to_user()).
Fixes: 6b1e6cc7855b0 ("vhost: new device IOTLB API")
Reported-by: syzbot+6304bf97ef436580fede@syzkaller.appspotmail.com
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>