qemu QEMU instructions

P

Paul Floyd

Does anyone have concise instructions for running qemu on FreeBSD?

I'm mainly interested in running aarch64 and riscv64, both 14.2-RELEASE. That would be on an amd64 14.2-RELEASE host.

Most of the instructions that I can find are for Linux and the wiki looks very outdated, covering FreeBSD 7 and 8.
 
Paul Floyd said:
the wiki looks very outdated, covering FreeBSD 7 and 8.
Click to expand...
Try those more up to date specific arm64/QEMU (last edited 2025-03-31) and riscv/QEMU (last edited 2024-11) FreeBSD wiki pages.
Rich (BB code): 
/usr/src-main # make targets
Supported TARGET/TARGET_ARCH pairs for world and kernel targets
    amd64/amd64
    arm/armv7
    arm64/aarch64
    i386/i386
    powerpc/powerpc
    powerpc/powerpc64
    powerpc/powerpc64le
    riscv/riscv64
 
Here is my RISC-V command:
Code: 
qemu-system-riscv64 \
  -M virt \
  -smp 8 \
  -m 32G \
  -nographic \
  -bios /usr/lib/riscv64-linux-gnu/opensbi/generic/fw_jump.elf \
  -kernel u-boot.bin \
  -drive file=FreeBSD-14.0-RELEASE-riscv-riscv64.raw,format=raw,id=hd0 \
  -device virtio-blk-device,drive=hd0 \
  -nographic \
  -netdev user,id=net0,ipv6=off,hostfwd=tcp::3235-:22 \
  -device virtio-net-device,netdev=net0 \
  -device virtio-rng-pci
 
That looks better. I still get

qemu-system-aarch64: -netdev user,id=net0: network backend 'user' is not compiled into this bi
nary
 
Paul Floyd said:
I still get

qemu-system-aarch64: -netdev user,id=net0: network backend 'user' is not compiled into this bi
nary
Click to expand...
/usr/ports/emulators/qemu/Makefile shows
Rich (BB code): 
CONFIGURE_ARGS+=        --disable-attr \
...
                        --disable-user \
...
I wonder if this configuration is responsible for the message.

Try "tap" instead.

Documentation/Networking - QEMU

wiki.qemu.org wiki.qemu.org

Network backend types

In most cases, if you don't have any specific networking requirements other than to be able to access to a
web page from your guest, user networking (slirp) is a good choice. However, if you are looking to run any
kind of network service or have your guest participate in a network in any meaningful way, tap is usually
the best choice.
Click to expand...


EDIT: If you insist on "user", perhaps try (lower distversion) emulators/qemu-devel , --disable-user is not set.
Code: 
CONFIGURE_ARGS= --localstatedir=/var --mandir=share/man --extra-ldflags=-L\"/usr/lib\" \
                --extra-ldflags=-L\"${LOCALBASE}/lib\" --enable-debug \
                --prefix=${PREFIX} --cc=${CC} --enable-netmap --disable-kvm \
                --disable-linux-user --disable-linux-aio --disable-xen \
                --enable-debug-info --python=${PYTHON_CMD} --disable-guest-agent \
                --extra-cflags=-I${WRKSRC}\ -I${LOCALBASE}/include\ -DPREFIX=\\\"\"${PREFIX}\\\"\"
 
T-Daemon said:
/usr/ports/emulators/qemu/Makefile shows
Rich (BB code): 
CONFIGURE_ARGS+=        --disable-attr \
...
                        --disable-user \
...
I wonder if this configuration is responsible for the message.
Click to expand...

The text of pkg-messsage could be better

"- Needs to set net.link.tap.user_open sysctl in order to use /dev/tap*
networking as non-root. Don't forget to adjust device node
permissions in /etc/devfs.rules."

How to make the sysctl permanent? And how about actually explaining what to put in /etc/devfs.rules?

"- slirp (usermode networking) is fixed now, you still have to manually
do: echo nameserver 10.0.2.3 >/etc/resolv.conf but that is normal. And
you have to wait a bit for dhclient to do its thing; traffic to
address 10.0.2.2 is routed to 127.1 on the host."

That sounds like it conflicts with the --disable-user option.

T-Daemon said:
Try "tap" instead.
Click to expand...

With tap I can at least get both aarch64 and riscv64 images to boot. Only when running qemu as root though.

Next I need to figure out what they need to get dhcp to work.
 
Paul Floyd said:
"- Needs to set net.link.tap.user_open sysctl in order to use /dev/tap*
...
How to make the sysctl permanent?
Click to expand...
In /etc/sysctl.conf, in case the question wasn't rhetorically.

Paul Floyd said:
And how about actually explaining what to put in /etc/devfs.rules?
Click to expand...
I suppose the author of the pkg-message expected that the user should look into devfs.rules(5) manual by mentioning it, if not familiar with the configuration. Specifying the manual by name and man section number (devfs.rules(5) ) would have been more precise, though.

Considering the topic is about permissions for the "tap" device, an example rule could be:
Code: 
[qemurules=6]
add  path  'tap*'  mode  660 group qemu

/etc/rc.conf
Code: 
devfs_system_ruleset="qemurules"
Create "qemu" group, add user to that group, service devfs restart


Paul Floyd said:
"- slirp (usermode networking) is fixed now, you still have to manually
...

That sounds like it conflicts with the --disable-user option.
Click to expand...
The pkg-message you quoting from seems from emulators/qemu-devel, remember this port is not build with the "--disable-user" configuration argument (see post # 6, under EDIT, in case you haven't seen it) .

In comparison emulators/qemu pkg-message.
 
Did you get this working? All the examples I have to contribute are for Mac ARM, and I'm pretty sure they won't work on FreeBSD.
 
Jose said:
Did you get this working? All the examples I have to contribute are for Mac ARM, and I'm pretty sure they won't work on FreeBSD.
Click to expand...

Mostly. I was doing the testing in a VirtualBox VM, and I've gotten as far as being able to boot FreeBSD and get to a terminal on both aarch64 and riscv64. I haven't worked out how to get DHCP working yet. I need to decide whether to try this directly on either my old or new systems and whether I really need DHCP or not.
 
i have an old 13.0 armv7 image
it works with qemu package (vntnet usage causes a panic because of unaligned data but i think this is fixed in newer releases)
i used it to build packages for a rpi zero
sh: 
#!/bin/sh
# send image file as first arg ($1)
qemu-system-arm \
  -M virt \
  -m 2048 \
  -drive file=pflash0.img,format=raw,if=pflash,readonly=on \
  -drive file=pflash1.img,format=raw,if=pflash \
  -drive file=$1,if=virtio,cache=writethrough \
  -nographic \
  -device qemu-xhci \
  -device usb-kbd \
  -device usb-tablet \
  -device virtio-net-pci,netdev=network0 \
  -netdev tap,ifname=tap6,id=network0,script=no,downscript=no \
  -serial mon:stdio \
  -serial telnet:10.1.1.2:4321,server,nowait
tap init
Code: 
sysctl net.link.tap.up_on_open=1
sysctl net.link.tap.user_open=1
ifconfig tap6 create 2>/dev/null

ifconfig tap6 10.10.10.10/24
chmod 660 /dev/tap6
pflash images creation
Code: 
dd if=/dev/zero of=pflash0.img bs=1m count=64
dd if=/dev/zero of=pflash1.img bs=1m count=64
dd if=QEMU_EFI.fd of=pflash0.img conv=notrunc
dd if=QEMU_VARS.fd of=pflash1.img conv=notrunc
source of QEMU_* is https://retrage.github.io/edk2-nightly/
 
You must log in or register to reply here.
Back
Top