|
From: | Do Vincenzo |
Subject: | [Qemu-ppc] High ping round trip time between guests |
Date: | Thu, 14 Jan 2016 12:07:33 +0000 |
Hello,
I'm doing some network tests on a Freescale T1040 board by starting two guest
with both a tap interface bridged on the hypervisor.
Once the two guest are up I can ping between them but the ping round trip time
is very high if compared to the same measures on x86-64 or ARMv8.
I get an average of 2ms per ping on the PowerPC board, 0.3 ms on x86-64 and
pretty much the same on ARMv8.
I'm using the following command to start the guests:
qemu-system-ppc64 -enable-kvm -nographic -m 512 -M ppce500 -kernel uImage_4.2 \
-initrd rootfs_guest8.ext2.gz \
-append "root=/dev/ram rw loglevel=7 console=ttyS0,115200" \
-netdev tap,id=tap0,script=qemu-ifup,downscript=qemu-ifdown,vhost=on \
-device virtio-net-pci,netdev=tap0,mac="52:54:00:12:34:01"
I've tried with and without vhost, but the ping round trip time stays the same.
Would it be correct to say that the high latency comes from the interrupts that
the network driver sends to KVM (who sends it to the correct guest) for each
received packet?
Also, is there a way to improve it?
Thank you for your time and best regards,
Vincenzo Do
|
[Prev in Thread] | Current Thread | [Next in Thread] |