[Qemu-devel] HAXM is now open source

[Qemu-devel] HAXM is now open source

[Yu Ning]

Hello,

As some of you may have noticed, since QEMU 2.9.0, an accelerator known 
as “hax” has been available for Windows and macOS builds of QEMU, thanks 
to the hard work of Vincent Palatin and help from this community (Paolo 
Bonzini, Stefan Weil, et al.).

The accelerator requires a host kernel module (driver) known as Intel 
Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on 
Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator 
depends on kvm.ko on Linux.

Today, we released the source code of the HAXM kernel module under the 
BSD 3-clause license:

https://github.com/intel/haxm

We look forward to working with the community to improve HAXM (both the 
kernel module and the accelerator). The code is accompanied by some 
basic documentation (README.md and API.md), which is incomplete, but 
hopefully helps people get started. If you have any questions or 
suggestions, please create an issue or post a comment on GitHub.

Thanks,
Yu

Re: [Qemu-devel] HAXM is now open source

[Thomas Huth]

On 14.11.2017 09:54, Yu Ning wrote:
> Hello,
> 
> As some of you may have noticed, since QEMU 2.9.0, an accelerator known
> as “hax” has been available for Windows and macOS builds of QEMU, thanks
> to the hard work of Vincent Palatin and help from this community (Paolo
> Bonzini, Stefan Weil, et al.).
> 
> The accelerator requires a host kernel module (driver) known as Intel
> Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on
> Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator
> depends on kvm.ko on Linux.
> 
> Today, we released the source code of the HAXM kernel module under the
> BSD 3-clause license:
> 
> https://github.com/intel/haxm
> 
> We look forward to working with the community to improve HAXM (both the
> kernel module and the accelerator). The code is accompanied by some
> basic documentation (README.md and API.md), which is incomplete, but
> hopefully helps people get started. If you have any questions or
> suggestions, please create an issue or post a comment on GitHub.

That's great news! I hope this all will help to promote QEMU on Windows
and macOS quite a bit!

However, during the past months, I noticed a couple of times that users
ask on IRC or the qemu-discuss mailing list how they could accelerate
their QEMU on Windows - and they are running only in TCG mode when you
ask how they start QEMU. So it seems like there is not much knowledge
about "--accel hax" in the public yet. Maybe you could write a nice blog
post for the QEMU blog or something similar that explains how to use
HAXM with QEMU on Windows for the normal users? Or maybe make it more
prominent in the QEMU wiki? (e.g. the main page only mentions KVM and
Xen, but not HAXM)

 Thomas

Re: [Qemu-devel] HAXM is now open source

[John Snow]

On 11/14/2017 06:09 AM, Thomas Huth wrote:
> On 14.11.2017 09:54, Yu Ning wrote:
>> Hello,
>>
>> As some of you may have noticed, since QEMU 2.9.0, an accelerator known
>> as “hax” has been available for Windows and macOS builds of QEMU, thanks
>> to the hard work of Vincent Palatin and help from this community (Paolo
>> Bonzini, Stefan Weil, et al.).
>>
>> The accelerator requires a host kernel module (driver) known as Intel
>> Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on
>> Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator
>> depends on kvm.ko on Linux.
>>
>> Today, we released the source code of the HAXM kernel module under the
>> BSD 3-clause license:
>>
>> https://github.com/intel/haxm
>>
>> We look forward to working with the community to improve HAXM (both the
>> kernel module and the accelerator). The code is accompanied by some
>> basic documentation (README.md and API.md), which is incomplete, but
>> hopefully helps people get started. If you have any questions or
>> suggestions, please create an issue or post a comment on GitHub.
> 
> That's great news! I hope this all will help to promote QEMU on Windows
> and macOS quite a bit!
> 
> However, during the past months, I noticed a couple of times that users
> ask on IRC or the qemu-discuss mailing list how they could accelerate
> their QEMU on Windows - and they are running only in TCG mode when you
> ask how they start QEMU. So it seems like there is not much knowledge
> about "--accel hax" in the public yet. Maybe you could write a nice blog
> post for the QEMU blog or something similar that explains how to use
> HAXM with QEMU on Windows for the normal users? Or maybe make it more
> prominent in the QEMU wiki? (e.g. the main page only mentions KVM and
> Xen, but not HAXM)
> 
>  Thomas
> 

A blog post advertising this new development would be an absolute
miracle for linking to people who are just getting started with QEMU on
Windows.

(It would also be really good for idiots like me, who do not use Windows
for anything other than playing video games and sometimes forget that it
is capable of doing other things.)

--js

Re: [Qemu-devel] HAXM is now open source

[Yu Ning]

On 11/15/2017 3:13, John Snow wrote:
>
> On 11/14/2017 06:09 AM, Thomas Huth wrote:
>>
>> That's great news! I hope this all will help to promote QEMU on Windows
>> and macOS quite a bit!
>>
>> However, during the past months, I noticed a couple of times that users
>> ask on IRC or the qemu-discuss mailing list how they could accelerate
>> their QEMU on Windows - and they are running only in TCG mode when you
>> ask how they start QEMU. So it seems like there is not much knowledge
>> about "--accel hax" in the public yet. Maybe you could write a nice blog
>> post for the QEMU blog or something similar that explains how to use
>> HAXM with QEMU on Windows for the normal users? Or maybe make it more
>> prominent in the QEMU wiki? (e.g. the main page only mentions KVM and
>> Xen, but not HAXM)
> A blog post advertising this new development would be an absolute
> miracle for linking to people who are just getting started with QEMU on
> Windows.
>
> (It would also be really good for idiots like me, who do not use Windows
> for anything other than playing video games and sometimes forget that it
> is capable of doing other things.)

Thanks for the encouragement.

I think I can start by writing a small section that can be added to an 
existing document, and later expanded to a blog article. But I haven't 
found a suitable place for it on the QEMU wiki.

Does it make sense to add the proposed piece to the QEMU user manual 
(qemu-doc.texi), under the Quick Start section (2.2)? The user manual is 
published on Stefan's website and referred to by qemu.org:

https://qemu.weilnetz.de/doc/qemu-doc.html

so I think it's a popular resource among users.

BTW, I assume most Windows users begin their QEMU journey from this page 
(also credit to Stefan):

https://qemu.weilnetz.de/

and I've verified the hax accelerator module is built into the latest 
binaries there (at least the W64 ones).

-Yu

Re: [Qemu-devel] HAXM is now open source

[Paolo Bonzini]

On 15/11/2017 09:25, Yu Ning wrote:
> 
> 
> On 11/15/2017 3:13, John Snow wrote:
>>
>> On 11/14/2017 06:09 AM, Thomas Huth wrote:
>>>
>>> That's great news! I hope this all will help to promote QEMU on Windows
>>> and macOS quite a bit!
>>>
>>> However, during the past months, I noticed a couple of times that users
>>> ask on IRC or the qemu-discuss mailing list how they could accelerate
>>> their QEMU on Windows - and they are running only in TCG mode when you
>>> ask how they start QEMU. So it seems like there is not much knowledge
>>> about "--accel hax" in the public yet. Maybe you could write a nice blog
>>> post for the QEMU blog or something similar that explains how to use
>>> HAXM with QEMU on Windows for the normal users? Or maybe make it more
>>> prominent in the QEMU wiki? (e.g. the main page only mentions KVM and
>>> Xen, but not HAXM)
>> A blog post advertising this new development would be an absolute
>> miracle for linking to people who are just getting started with QEMU on
>> Windows.
>>
>> (It would also be really good for idiots like me, who do not use Windows
>> for anything other than playing video games and sometimes forget that it
>> is capable of doing other things.)
> 
> Thanks for the encouragement.
> 
> I think I can start by writing a small section that can be added to an
> existing document, and later expanded to a blog article. But I haven't
> found a suitable place for it on the QEMU wiki.
> 
> Does it make sense to add the proposed piece to the QEMU user manual
> (qemu-doc.texi), under the Quick Start section (2.2)? The user manual is
> published on Stefan's website and referred to by qemu.org:
> 
> https://qemu.weilnetz.de/doc/qemu-doc.html

Unfortunately the "Quick Start" section is quite out of date (it doesn't
even mention KVM!) and is basically a huge list of options.  We are in
the process of improving QEMU documentation, so I think a blog post is
preferrable at this stage.

Thanks,

Paolo

> so I think it's a popular resource among users.
> 
> BTW, I assume most Windows users begin their QEMU journey from this page
> (also credit to Stefan):
> 
> https://qemu.weilnetz.de/
> 
> and I've verified the hax accelerator module is built into the latest
> binaries there (at least the W64 ones).
> 
> -Yu
> 

Re: [Qemu-devel] HAXM is now open source

[Yu Ning]

On 11/15/2017 22:18, Paolo Bonzini wrote:
> On 15/11/2017 09:25, Yu Ning wrote:
>>
>> On 11/15/2017 3:13, John Snow wrote:
>>> On 11/14/2017 06:09 AM, Thomas Huth wrote:
>>>> That's great news! I hope this all will help to promote QEMU on Windows
>>>> and macOS quite a bit!
>>>>
>>>> However, during the past months, I noticed a couple of times that users
>>>> ask on IRC or the qemu-discuss mailing list how they could accelerate
>>>> their QEMU on Windows - and they are running only in TCG mode when you
>>>> ask how they start QEMU. So it seems like there is not much knowledge
>>>> about "--accel hax" in the public yet. Maybe you could write a nice blog
>>>> post for the QEMU blog or something similar that explains how to use
>>>> HAXM with QEMU on Windows for the normal users? Or maybe make it more
>>>> prominent in the QEMU wiki? (e.g. the main page only mentions KVM and
>>>> Xen, but not HAXM)
>>> A blog post advertising this new development would be an absolute
>>> miracle for linking to people who are just getting started with QEMU on
>>> Windows.
>>>
>>> (It would also be really good for idiots like me, who do not use Windows
>>> for anything other than playing video games and sometimes forget that it
>>> is capable of doing other things.)
>> Thanks for the encouragement.
>>
>> I think I can start by writing a small section that can be added to an
>> existing document, and later expanded to a blog article. But I haven't
>> found a suitable place for it on the QEMU wiki.
>>
>> Does it make sense to add the proposed piece to the QEMU user manual
>> (qemu-doc.texi), under the Quick Start section (2.2)? The user manual is
>> published on Stefan's website and referred to by qemu.org:
>>
>> https://qemu.weilnetz.de/doc/qemu-doc.html
> Unfortunately the "Quick Start" section is quite out of date (it doesn't
> even mention KVM!) and is basically a huge list of options.  We are in
> the process of improving QEMU documentation, so I think a blog post is
> preferrable at this stage.
>
>
I see. I'll take the advice.

Is there a preferred document format for the QEMU blog, e.g. Markdown? 
I'll get the content ready by early next week, and then I'll need 
someone to help me post it.

Re: [Qemu-devel] HAXM is now open source

[Thomas Huth]

On 17.11.2017 09:17, Yu Ning wrote:
[...]
> Is there a preferred document format for the QEMU blog, e.g. Markdown?

Yes, we're using markdown of Jekyll (https://jekyllrb.com/docs/posts/).
Please clone the qemu-web repository (see
https://git.qemu.org/?p=qemu-web.git), and have a look in the "_posts"
directory to see how the files should look like.

> I'll get the content ready by early next week, and then I'll need
> someone to help me post it.

Sure, Paolo and I should be able to help. The preferred way is to create
a patch against the qemu-web repository and send that patch to
qemu-devel, with CC: to Paolo and myself. We then can give feedback, do
sanity checking (in case you don't want to install Jekyll on your HD),
and push the patch to the website if everything is fine.

 Thanks!
  Thomas

Re: [Qemu-devel] HAXM is now open source

[Yu Ning]

On 11/17/2017 16:30, Thomas Huth wrote:
> On 17.11.2017 09:17, Yu Ning wrote:
> [...]
>> Is there a preferred document format for the QEMU blog, e.g. Markdown?
> Yes, we're using markdown of Jekyll (https://jekyllrb.com/docs/posts/).
> Please clone the qemu-web repository (see
> https://git.qemu.org/?p=qemu-web.git), and have a look in the "_posts"
> directory to see how the files should look like.
>
>> I'll get the content ready by early next week, and then I'll need
>> someone to help me post it.
> Sure, Paolo and I should be able to help. The preferred way is to create
> a patch against the qemu-web repository and send that patch to
> qemu-devel, with CC: to Paolo and myself. We then can give feedback, do
> sanity checking (in case you don't want to install Jekyll on your HD),
> and push the patch to the website if everything is fine.

Got it, thanks!

Re: [Qemu-devel] HAXM is now open source

[Kamil Rytarowski]

[-- Attachment #1: Type: text/plain, Size: 1335 bytes --]

On 14.11.2017 09:54, Yu Ning wrote:
> Hello,
> 
> As some of you may have noticed, since QEMU 2.9.0, an accelerator known
> as “hax” has been available for Windows and macOS builds of QEMU, thanks
> to the hard work of Vincent Palatin and help from this community (Paolo
> Bonzini, Stefan Weil, et al.).
> 
> The accelerator requires a host kernel module (driver) known as Intel
> Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on
> Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator
> depends on kvm.ko on Linux.
> 
> Today, we released the source code of the HAXM kernel module under the
> BSD 3-clause license:
> 
> https://github.com/intel/haxm
> 
> We look forward to working with the community to improve HAXM (both the
> kernel module and the accelerator). The code is accompanied by some
> basic documentation (README.md and API.md), which is incomplete, but
> hopefully helps people get started. If you have any questions or
> suggestions, please create an issue or post a comment on GitHub.
> 
> Thanks,
> Yu
> 

Please make it clear whether this module can be ported (as host) to
other OSes and whether it can support arbitrary guests OSes (for the
same CPU).

I'm researching options to add hardware assisted virtualization for
NetBSD as host.


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] HAXM is now open source

[Yu Ning]

On 11/17/2017 16:53, Kamil Rytarowski wrote:
> On 14.11.2017 09:54, Yu Ning wrote:
>> Hello,
>>
>> As some of you may have noticed, since QEMU 2.9.0, an accelerator known
>> as “hax” has been available for Windows and macOS builds of QEMU, thanks
>> to the hard work of Vincent Palatin and help from this community (Paolo
>> Bonzini, Stefan Weil, et al.).
>>
>> The accelerator requires a host kernel module (driver) known as Intel
>> Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on
>> Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator
>> depends on kvm.ko on Linux.
>>
>> Today, we released the source code of the HAXM kernel module under the
>> BSD 3-clause license:
>>
>> https://github.com/intel/haxm
>>
>> We look forward to working with the community to improve HAXM (both the
>> kernel module and the accelerator). The code is accompanied by some
>> basic documentation (README.md and API.md), which is incomplete, but
>> hopefully helps people get started. If you have any questions or
>> suggestions, please create an issue or post a comment on GitHub.
>>
>> Thanks,
>> Yu
>>
> Please make it clear whether this module can be ported (as host) to
> other OSes

It was designed with only Windows and Mac hosts in mind, and we don't 
plan to support more host OSes. But porting HAXM to another host OS is 
possible.  If you take a look at the HAXM source tree, there are pretty 
clear boundaries between host-independent code (core/, include/*.h), 
Windows-specific code (windows/, include/windows/) and Mac-specific code 
(darwin/, include/darwin/).
>   and whether it can support arbitrary guests OSes (for the
> same CPU).

Unfortunately not, but again it can be done.  Initially HAXM was 
designed to support only two guest OSes, namely Android and Tizen. When 
we upstreamed the HAXM accelerator module to QEMU, we verified it using 
various desktop OS images, mostly Linux-based (Chrome OS, Debian, 
Ubuntu, CentOS, etc.), although there were a few others (FreeDOS, etc.).

Nevertheless, you may run into issues when you boot a completely 
different guest OS (e.g. Windows, which we have never tested), or even 
with one that is similar to what I've mentioned (e.g. Fedora, which IIRC 
will boot to a kernel panic).  In such cases, the culprit is usually 
bug(s) in the HAXM kernel module, causing incorrect behavior when the 
guest executes certain x86 instructions.  Once all the bugs exposed by a 
specific guest OS are fixed, that guest OS will work correctly.

Re: [Qemu-devel] HAXM is now open source

[Kamil Rytarowski]

[-- Attachment #1: Type: text/plain, Size: 2821 bytes --]

On 17.11.2017 11:30, Yu Ning wrote:
> 
> 
> On 11/17/2017 16:53, Kamil Rytarowski wrote:
>> On 14.11.2017 09:54, Yu Ning wrote:
>>> Hello,
>>>
>>> As some of you may have noticed, since QEMU 2.9.0, an accelerator known
>>> as “hax” has been available for Windows and macOS builds of QEMU, thanks
>>> to the hard work of Vincent Palatin and help from this community (Paolo
>>> Bonzini, Stefan Weil, et al.).
>>>
>>> The accelerator requires a host kernel module (driver) known as Intel
>>> Hardware Accelerated Execution Manager (HAXM), i.e. intelhaxm.sys on
>>> Windows or intelhaxm.kext on macOS, similar to how the KVM accelerator
>>> depends on kvm.ko on Linux.
>>>
>>> Today, we released the source code of the HAXM kernel module under the
>>> BSD 3-clause license:
>>>
>>> https://github.com/intel/haxm
>>>
>>> We look forward to working with the community to improve HAXM (both the
>>> kernel module and the accelerator). The code is accompanied by some
>>> basic documentation (README.md and API.md), which is incomplete, but
>>> hopefully helps people get started. If you have any questions or
>>> suggestions, please create an issue or post a comment on GitHub.
>>>
>>> Thanks,
>>> Yu
>>>
>> Please make it clear whether this module can be ported (as host) to
>> other OSes
> 
> It was designed with only Windows and Mac hosts in mind, and we don't
> plan to support more host OSes. But porting HAXM to another host OS is
> possible.  If you take a look at the HAXM source tree, there are pretty
> clear boundaries between host-independent code (core/, include/*.h),
> Windows-specific code (windows/, include/windows/) and Mac-specific code
> (darwin/, include/darwin/).

Sounds good. I will give it a try.

>>   and whether it can support arbitrary guests OSes (for the
>> same CPU).
> 
> Unfortunately not, but again it can be done.  Initially HAXM was
> designed to support only two guest OSes, namely Android and Tizen. When
> we upstreamed the HAXM accelerator module to QEMU, we verified it using
> various desktop OS images, mostly Linux-based (Chrome OS, Debian,
> Ubuntu, CentOS, etc.), although there were a few others (FreeDOS, etc.).
> 
> Nevertheless, you may run into issues when you boot a completely
> different guest OS (e.g. Windows, which we have never tested), or even
> with one that is similar to what I've mentioned (e.g. Fedora, which IIRC
> will boot to a kernel panic).  In such cases, the culprit is usually
> bug(s) in the HAXM kernel module, causing incorrect behavior when the
> guest executes certain x86 instructions.  Once all the bugs exposed by a
> specific guest OS are fixed, that guest OS will work correctly.
> 

If it will run BSD for start it will be good enough for now.

Thank you for the answers!


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]