Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 20 May 2024 10:55:09 +0200
From:      Mario Marietto <marietto2008@gmail.com>
To:        FreeBSD virtualization <freebsd-virtualization@freebsd.org>
Subject:   How much hard is to write a version of bhyve that can accelerate qemu ?
Message-ID:  <CA%2B1FSihUy4iRjnGiAGGopyCmT3DjKtU_qegQiG8S-9GNT_T3fg@mail.gmail.com>

next in thread | raw e-mail | index | archive | help
--000000000000e9d7910618dedd68
Content-Type: text/plain; charset="UTF-8"

Hello.

some days ago I had a nice discussion with a qemu and haxm hypervisor
developer that ended with a nice question and I would like to hear your
opinions :

ME : "HAXM was created to bring Intel Virtualization Technology to Windows
and macOSusers" ; keeping in consideration that MacOS has some code of
FreeBSD inside. Is the source code of HAXM freely available ? Are you able
to port it to FreeBSD ? The idea is to accelerate qemu with HAXM on
FreeBSD. Do you know someone that can and wants to do this ?

qemu and HAXM developer : HAXM has been imported into pkgsrc/emulators/haxm
---> http://mail-index.netbsd.org/netbsd-users/2019/02/13/msg022207.html

ME : wow. So,it exists for NetBSD but not for FreeBSD ? That's so bad.

qemu and HAXM developer : Yes - https://github.com/intel/haxm. But it is
not supported. I see no point in porting an unsupported solution to a new
OS, thus dont think someone will want to port it.

ME : So,it exists for NetBSD but not for FreeBSD ? That's so bad.

qemu and HAXM developer : Yes, it supports NetBSD.

ME : If NetBSD developers have been interested to port it,so why shouldn't
FreeBSD ones be ? At least,in theory....after all,on FreeBSD qemu lacks the
acceleration of a good hypervisor.

qemu and HAXM developer : This can be solved by adding for example bhyve
support to qemu. This should be a better solution, than porting an
unsupported hypervisor. I'm not aware why FreeBSD developers didn't add it
already, you may ask for the reason.

How hard is it to write a version of bhyve that can accelerate qemu ?

-- 
Mario.

--000000000000e9d7910618dedd68
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hello.<br>
<br>
some days ago I had a nice discussion with a qemu and haxm hypervisor=20
developer that ended with a nice question and I would like to hear your=20
opinions :<br>
<br>
ME :  &quot;HAXM was created to bring Intel Virtualization Technology to=20
Windows and macOSusers&quot; ; keeping in consideration that MacOS has some=
=20
code of FreeBSD inside. Is the source code of HAXM freely available ?=20
Are you able to port it to FreeBSD ? The idea is to accelerate qemu with
 HAXM on FreeBSD. Do you know someone that can and wants to do this ?<br>
<br>
qemu and HAXM developer :  HAXM has been imported into pkgsrc/emulators/hax=
m ---&gt; <a href=3D"http://mail-index.netbsd.org/netbsd-users/2019/02/13/m=
sg022207.html" target=3D"_blank" class=3D"gmail-link gmail-link--external" =
rel=3D"nofollow ugc noopener">http://mail-index.netbsd.org/netbsd-users/201=
9/02/13/msg022207.html</a><br>
<br>
ME : wow. So,it exists for NetBSD but not for FreeBSD ? That&#39;s so bad.<=
br>
<br>
qemu and HAXM developer : Yes - <a href=3D"https://github.com/intel/haxm" t=
arget=3D"_blank" class=3D"gmail-link gmail-link--external" rel=3D"nofollow =
ugc noopener">https://github.com/intel/haxm</a>.
 But it is not supported. I see no point in porting an unsupported=20
solution to a new OS, thus dont think someone will want to port it.<br>
<br>
ME : So,it exists for NetBSD but not for FreeBSD ? That&#39;s so bad.<br>
<br>
qemu and HAXM developer : Yes, it supports NetBSD.<br>
<br>
ME : If NetBSD developers have been interested to port it,so why=20
shouldn&#39;t FreeBSD ones be ? At least,in theory....after all,on FreeBSD=
=20
qemu lacks the acceleration of a good hypervisor.<br>
<br>
qemu and HAXM developer : This can be solved by adding for example bhyve
 support to qemu. This should be a better solution, than porting an=20
unsupported hypervisor. I&#39;m not aware why FreeBSD developers didn&#39;t=
 add=20
it already, you may ask for the reason.<br>
<br>
How hard is it to write a version of bhyve that can accelerate qemu ?<br cl=
ear=3D"all"><br><span class=3D"gmail_signature_prefix">-- </span><br><div d=
ir=3D"ltr" class=3D"gmail_signature" data-smartmail=3D"gmail_signature">Mar=
io.<br></div></div>

--000000000000e9d7910618dedd68--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CA%2B1FSihUy4iRjnGiAGGopyCmT3DjKtU_qegQiG8S-9GNT_T3fg>