[dpdk-dev] [Bug 794] [dpdk-19.11.10] loopback_multi_queues: core dumped when quit vhost testpmd
bugzilla at dpdk.org
bugzilla at dpdk.org
Tue Aug 24 10:58:53 CEST 2021
https://bugs.dpdk.org/show_bug.cgi?id=794
Bug ID: 794
Summary: [dpdk-19.11.10] loopback_multi_queues: core dumped
when quit vhost testpmd
Product: DPDK
Version: 19.11
Hardware: x86
OS: Linux
Status: UNCONFIRMED
Severity: normal
Priority: Normal
Component: vhost/virtio
Assignee: dev at dpdk.org
Reporter: linglix.chen at intel.com
Target Milestone: ---
Environment
DPDK version:
19.11.10-rc1:179f2a3a8a56600a66e4e0830bbe993be67f9d7e
Other software versions: N/A
OS: Ubuntu 20.04.2 LTS/Linux 5.11.16-051116-generic
Compiler: gcc version 9.3.0 (Ubuntu 9.3.0-17ubuntu1~20.04)
Hardware platform: Intel(R) Xeon(R) Platinum 8280M CPU @ 2.70GHz
NIC hardware: FVL-40g
NIC firmware & driver:
driver: i40e
version: 5.11.16-051116-generic
firmware-version: 8.30 0x8000a4ae 1.2926.0
Steps to reproduce
1. Launch vhost testpmd
x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 31,32 -n 4 --file-prefix=vhost
--no-pci --vdev 'net_vhost0,iface=vhost-net,queues=1' -- -i --nb-cores=1
--rxq=1 --txq=1 --txd=1024 --rxd=1024
2. Launch virtio-user
x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29,30 -n 4 --file-prefix=virtio
--no-pci --vdev
'net_virtio_user0,mac=00:01:02:03:04:05,path=./vhost-net,queues=1,packed_vq=0,in_order=1,mrg_rxbuf=0'
--single-file-segments -- -i --nb-cores=1 --rxq=1 --txq=1 --txd=1024 --rxd=1024
3.quit vhost testpmd
testpmd> quit
Show the output from the previous commands.
Stopping port 0...
Stopping ports...
Done
Shutting down port 0...
Closing ports...
VHOST_CONFIG: free connfd = 24 for device 'vhost-net'
Port 0: link state change event
free(): invalid pointer
Aborted (core dumped)
Expected Result:
quit normal
Regression
Is this issue a regression: (Y/N) Y
First bad commit id: commit eb35473b23045d32d4abd4d31ca36483e37953df
Author: Maxime Coquelin <maxime.coquelin at redhat.com>
Date: Tue Jun 29 18:11:28 2021 +0200
vhost: fix missing guest pages table NUMA realloc
[ upstream commit 57589cdfd784e6704ded490d80038584d6ade9df ]
When the guest allocates virtqueues on a different NUMA node
than the one the Vhost metadata are allocated, both the Vhost
device struct and the virtqueues struct are reallocated.
However, reallocating the guest pages table was missing, which
likely causes at least one cross-NUMA accesses for every burst
of packets.
This patch reallocates this table on the same NUMA node as the
other metadata.
Fixes: e246896178e6 ("vhost: get guest/host physical address mappings")
Signed-off-by: Maxime Coquelin <maxime.coquelin at redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia at intel.com>
--
You are receiving this mail because:
You are the assignee for the bug.
More information about the dev
mailing list