The FreeBSD project has a number of enhancements to bsd-user. Add myself
as maintainer and Kyle Evans as a reviewer. Also add our github repo.

Signed-off-by: Warner Losh <imp@bsdimp.com>
Signed-off-by: Kyle Evans <kevans@freebsd.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
---
 MAINTAINERS | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index 26c9454823..ec0e935038 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2896,9 +2896,12 @@ F: thunk.c
 F: accel/tcg/user-exec*.c
 
 BSD user
-S: Orphan
+M: Warner Losh <imp@bsdimp.com>
+R: Kyle Evans <kevans@freebsd.org>
+S: Maintained
 F: bsd-user/
 F: default-configs/targets/*-bsd-user.mak
+T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
 
 Linux user
 M: Laurent Vivier <laurent@vivier.eu>
-- 
2.30.0

On 07/03/2021 16.56, Warner Losh wrote:
> The FreeBSD project has a number of enhancements to bsd-user. Add myself
> as maintainer and Kyle Evans as a reviewer. Also add our github repo.
> 
> Signed-off-by: Warner Losh <imp@bsdimp.com>
> Signed-off-by: Kyle Evans <kevans@freebsd.org>
> Reviewed-by: Thomas Huth <thuth@redhat.com>
> ---
>   MAINTAINERS | 5 ++++-
>   1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 26c9454823..ec0e935038 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -2896,9 +2896,12 @@ F: thunk.c
>   F: accel/tcg/user-exec*.c
>   
>   BSD user
> -S: Orphan
> +M: Warner Losh <imp@bsdimp.com>
> +R: Kyle Evans <kevans@freebsd.org>
> +S: Maintained
>   F: bsd-user/
>   F: default-configs/targets/*-bsd-user.mak
> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1

BSD is not really my home turf, but since nobody else picked this up and I 
plan to send a pull request for a bunch of patches anyway this week, I can 
also put it into my queue.

  Thomas

On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
>
> On 07/03/2021 16.56, Warner Losh wrote:
> > The FreeBSD project has a number of enhancements to bsd-user. Add myself
> > as maintainer and Kyle Evans as a reviewer. Also add our github repo.
> >
> > Signed-off-by: Warner Losh <imp@bsdimp.com>
> > Signed-off-by: Kyle Evans <kevans@freebsd.org>
> > Reviewed-by: Thomas Huth <thuth@redhat.com>
> > ---
> >   MAINTAINERS | 5 ++++-
> >   1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/MAINTAINERS b/MAINTAINERS
> > index 26c9454823..ec0e935038 100644
> > --- a/MAINTAINERS
> > +++ b/MAINTAINERS
> > @@ -2896,9 +2896,12 @@ F: thunk.c
> >   F: accel/tcg/user-exec*.c
> >
> >   BSD user
> > -S: Orphan
> > +M: Warner Losh <imp@bsdimp.com>
> > +R: Kyle Evans <kevans@freebsd.org>
> > +S: Maintained
> >   F: bsd-user/
> >   F: default-configs/targets/*-bsd-user.mak
> > +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
>
> BSD is not really my home turf, but since nobody else picked this up and I
> plan to send a pull request for a bunch of patches anyway this week, I can
> also put it into my queue.

Fine with me. (The v1 was in my to-review queue, but I'm currently
running somewhat behind on processing patches.)

-- PMM

Hi Peter,

+Markus/Paolo/Laurent/Richard

On 3/8/21 11:24 AM, Peter Maydell wrote:
> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
>>
>> On 07/03/2021 16.56, Warner Losh wrote:
>>> The FreeBSD project has a number of enhancements to bsd-user. Add myself
>>> as maintainer and Kyle Evans as a reviewer. Also add our github repo.
>>>
>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
>>> ---
>>>   MAINTAINERS | 5 ++++-
>>>   1 file changed, 4 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>> index 26c9454823..ec0e935038 100644
>>> --- a/MAINTAINERS
>>> +++ b/MAINTAINERS
>>> @@ -2896,9 +2896,12 @@ F: thunk.c
>>>   F: accel/tcg/user-exec*.c
>>>
>>>   BSD user
>>> -S: Orphan
>>> +M: Warner Losh <imp@bsdimp.com>
>>> +R: Kyle Evans <kevans@freebsd.org>
>>> +S: Maintained
>>>   F: bsd-user/
>>>   F: default-configs/targets/*-bsd-user.mak
>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
>>
>> BSD is not really my home turf, but since nobody else picked this up and I
>> plan to send a pull request for a bunch of patches anyway this week, I can
>> also put it into my queue.
> 
> Fine with me. (The v1 was in my to-review queue, but I'm currently
> running somewhat behind on processing patches.)

This is a patch for mainstream QEMU, I'm having hard time
understanding the point of it. This is some official way
to say that BSD-user is not maintained in mainstream but
has to be used in the referred fork which is way different
that mainstream...

I'd rather wait for more mainstream contributions from Warner
and Kyle, or blow the current orphan/dead code and import
bsd-user-rebase-3.1 adding the maintainer entries along, but
certainly not mark this dead code as maintained.

Please convince me why I'm wrong, because I'd prefer NAck this
patch...

Regards,

Phil.

On Mon, 8 Mar 2021 at 11:16, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>
> Hi Peter,
>
> +Markus/Paolo/Laurent/Richard
>
> On 3/8/21 11:24 AM, Peter Maydell wrote:
> > On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
> >>
> >> On 07/03/2021 16.56, Warner Losh wrote:
> >>> The FreeBSD project has a number of enhancements to bsd-user. Add myself
> >>> as maintainer and Kyle Evans as a reviewer. Also add our github repo.
> >>>
> >>> Signed-off-by: Warner Losh <imp@bsdimp.com>
> >>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
> >>> Reviewed-by: Thomas Huth <thuth@redhat.com>
> >>> ---
> >>>   MAINTAINERS | 5 ++++-
> >>>   1 file changed, 4 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/MAINTAINERS b/MAINTAINERS
> >>> index 26c9454823..ec0e935038 100644
> >>> --- a/MAINTAINERS
> >>> +++ b/MAINTAINERS
> >>> @@ -2896,9 +2896,12 @@ F: thunk.c
> >>>   F: accel/tcg/user-exec*.c
> >>>
> >>>   BSD user
> >>> -S: Orphan
> >>> +M: Warner Losh <imp@bsdimp.com>
> >>> +R: Kyle Evans <kevans@freebsd.org>
> >>> +S: Maintained
> >>>   F: bsd-user/
> >>>   F: default-configs/targets/*-bsd-user.mak
> >>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
> >>
> >> BSD is not really my home turf, but since nobody else picked this up and I
> >> plan to send a pull request for a bunch of patches anyway this week, I can
> >> also put it into my queue.
> >
> > Fine with me. (The v1 was in my to-review queue, but I'm currently
> > running somewhat behind on processing patches.)
>
> This is a patch for mainstream QEMU, I'm having hard time
> understanding the point of it. This is some official way
> to say that BSD-user is not maintained in mainstream but
> has to be used in the referred fork which is way different
> that mainstream...
>
> I'd rather wait for more mainstream contributions from Warner
> and Kyle, or blow the current orphan/dead code and import
> bsd-user-rebase-3.1 adding the maintainer entries along, but
> certainly not mark this dead code as maintained.

I guess it depends what the intention is here. If the idea is
that now Warner and Kyle will look at patches that come in to
the existing bsd-user/ code in mainline, then that counts as
maintained. If they don't want to do that, then maybe we should
wait until we actually have something in mainline that they want
to maintain.

I think I agree with you about not referring to the out-of-tree
3.1-based fork; that's not relevant to upstream.

thanks
-- PMM

On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
> Hi Peter,
> 
> +Markus/Paolo/Laurent/Richard
> 
> On 3/8/21 11:24 AM, Peter Maydell wrote:
>> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
>>>
>>> On 07/03/2021 16.56, Warner Losh wrote:
>>>> The FreeBSD project has a number of enhancements to bsd-user. Add myself
>>>> as maintainer and Kyle Evans as a reviewer. Also add our github repo.
>>>>
>>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
>>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
>>>> ---
>>>>    MAINTAINERS | 5 ++++-
>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>>> index 26c9454823..ec0e935038 100644
>>>> --- a/MAINTAINERS
>>>> +++ b/MAINTAINERS
>>>> @@ -2896,9 +2896,12 @@ F: thunk.c
>>>>    F: accel/tcg/user-exec*.c
>>>>
>>>>    BSD user
>>>> -S: Orphan
>>>> +M: Warner Losh <imp@bsdimp.com>
>>>> +R: Kyle Evans <kevans@freebsd.org>
>>>> +S: Maintained
>>>>    F: bsd-user/
>>>>    F: default-configs/targets/*-bsd-user.mak
>>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
>>>
>>> BSD is not really my home turf, but since nobody else picked this up and I
>>> plan to send a pull request for a bunch of patches anyway this week, I can
>>> also put it into my queue.
>>
>> Fine with me. (The v1 was in my to-review queue, but I'm currently
>> running somewhat behind on processing patches.)
> 
> This is a patch for mainstream QEMU, I'm having hard time
> understanding the point of it. This is some official way
> to say that BSD-user is not maintained in mainstream but
> has to be used in the referred fork which is way different
> that mainstream...
> 
> I'd rather wait for more mainstream contributions from Warner
> and Kyle, or blow the current orphan/dead code and import
> bsd-user-rebase-3.1 adding the maintainer entries along, but
> certainly not mark this dead code as maintained.
> 
> Please convince me why I'm wrong, because I'd prefer NAck this
> patch...

The idea has been discussed here:

https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html

So this is not about declaring that bsd-user is maintained in a different 
repository, but about giving Warner et al. a chance to finally upstream 
their work.

  Thomas

On Mon, Mar 08, 2021 at 12:55:10PM +0100, Thomas Huth wrote:
> On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
> > Hi Peter,
> > 
> > +Markus/Paolo/Laurent/Richard
> > 
> > On 3/8/21 11:24 AM, Peter Maydell wrote:
> > > On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
> > > > 
> > > > On 07/03/2021 16.56, Warner Losh wrote:
> > > > > The FreeBSD project has a number of enhancements to bsd-user. Add myself
> > > > > as maintainer and Kyle Evans as a reviewer. Also add our github repo.
> > > > > 
> > > > > Signed-off-by: Warner Losh <imp@bsdimp.com>
> > > > > Signed-off-by: Kyle Evans <kevans@freebsd.org>
> > > > > Reviewed-by: Thomas Huth <thuth@redhat.com>
> > > > > ---
> > > > >    MAINTAINERS | 5 ++++-
> > > > >    1 file changed, 4 insertions(+), 1 deletion(-)
> > > > > 
> > > > > diff --git a/MAINTAINERS b/MAINTAINERS
> > > > > index 26c9454823..ec0e935038 100644
> > > > > --- a/MAINTAINERS
> > > > > +++ b/MAINTAINERS
> > > > > @@ -2896,9 +2896,12 @@ F: thunk.c
> > > > >    F: accel/tcg/user-exec*.c
> > > > > 
> > > > >    BSD user
> > > > > -S: Orphan
> > > > > +M: Warner Losh <imp@bsdimp.com>
> > > > > +R: Kyle Evans <kevans@freebsd.org>
> > > > > +S: Maintained
> > > > >    F: bsd-user/
> > > > >    F: default-configs/targets/*-bsd-user.mak
> > > > > +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
> > > > 
> > > > BSD is not really my home turf, but since nobody else picked this up and I
> > > > plan to send a pull request for a bunch of patches anyway this week, I can
> > > > also put it into my queue.
> > > 
> > > Fine with me. (The v1 was in my to-review queue, but I'm currently
> > > running somewhat behind on processing patches.)
> > 
> > This is a patch for mainstream QEMU, I'm having hard time
> > understanding the point of it. This is some official way
> > to say that BSD-user is not maintained in mainstream but
> > has to be used in the referred fork which is way different
> > that mainstream...
> > 
> > I'd rather wait for more mainstream contributions from Warner
> > and Kyle, or blow the current orphan/dead code and import
> > bsd-user-rebase-3.1 adding the maintainer entries along, but
> > certainly not mark this dead code as maintained.
> > 
> > Please convince me why I'm wrong, because I'd prefer NAck this
> > patch...
> 
> The idea has been discussed here:
> 
> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
> 
> So this is not about declaring that bsd-user is maintained in a different
> repository, but about giving Warner et al. a chance to finally upstream
> their work.

Yep, I think this change in MAINTAINERS file is primarily about signalling
intent for future.

Marking the subsystem as maintained isn't saying the current code is great,
just that there is someone committed to improving it hereafter.

If we want to warn people that the current impl isn't great, that's goes
back to the topic of having a way to classify QEMU features into quality
levels Tier 1/2/3. 

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

On 3/8/21 1:14 PM, Daniel P. Berrangé wrote:
> On Mon, Mar 08, 2021 at 12:55:10PM +0100, Thomas Huth wrote:
>> On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
>>> Hi Peter,
>>>
>>> +Markus/Paolo/Laurent/Richard
>>>
>>> On 3/8/21 11:24 AM, Peter Maydell wrote:
>>>> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
>>>>>
>>>>> On 07/03/2021 16.56, Warner Losh wrote:
>>>>>> The FreeBSD project has a number of enhancements to bsd-user. Add myself
>>>>>> as maintainer and Kyle Evans as a reviewer. Also add our github repo.
>>>>>>
>>>>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
>>>>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
>>>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
>>>>>> ---
>>>>>>    MAINTAINERS | 5 ++++-
>>>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>>>>> index 26c9454823..ec0e935038 100644
>>>>>> --- a/MAINTAINERS
>>>>>> +++ b/MAINTAINERS
>>>>>> @@ -2896,9 +2896,12 @@ F: thunk.c
>>>>>>    F: accel/tcg/user-exec*.c
>>>>>>
>>>>>>    BSD user
>>>>>> -S: Orphan
>>>>>> +M: Warner Losh <imp@bsdimp.com>
>>>>>> +R: Kyle Evans <kevans@freebsd.org>
>>>>>> +S: Maintained
>>>>>>    F: bsd-user/
>>>>>>    F: default-configs/targets/*-bsd-user.mak
>>>>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user bsd-user-rebase-3.1
>>>>>
>>>>> BSD is not really my home turf, but since nobody else picked this up and I
>>>>> plan to send a pull request for a bunch of patches anyway this week, I can
>>>>> also put it into my queue.
>>>>
>>>> Fine with me. (The v1 was in my to-review queue, but I'm currently
>>>> running somewhat behind on processing patches.)
>>>
>>> This is a patch for mainstream QEMU, I'm having hard time
>>> understanding the point of it. This is some official way
>>> to say that BSD-user is not maintained in mainstream but
>>> has to be used in the referred fork which is way different
>>> that mainstream...
>>>
>>> I'd rather wait for more mainstream contributions from Warner
>>> and Kyle, or blow the current orphan/dead code and import
>>> bsd-user-rebase-3.1 adding the maintainer entries along, but
>>> certainly not mark this dead code as maintained.
>>>
>>> Please convince me why I'm wrong, because I'd prefer NAck this
>>> patch...
>>
>> The idea has been discussed here:
>>
>> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
>>
>> So this is not about declaring that bsd-user is maintained in a different
>> repository, but about giving Warner et al. a chance to finally upstream
>> their work.
> 
> Yep, I think this change in MAINTAINERS file is primarily about signalling
> intent for future.
> 
> Marking the subsystem as maintained isn't saying the current code is great,
> just that there is someone committed to improving it hereafter.

OK, thank Thomas / Daniel for explaining and referring to the "BSD-user
plans" (which I didn't notice earlier).

Warner, what about mentioning your plans here in this patch?

Resumed ideally, else a simple link to the thread.

> If we want to warn people that the current impl isn't great, that's goes
> back to the topic of having a way to classify QEMU features into quality
> levels Tier 1/2/3.

That indeed sounds good w.r.t. contributors / users expectations.

I suppose 1=hw_accel/security, 2=tested, 3=rest?

Not a single clue how to do that although.

On Mon, Mar 8, 2021 at 6:14 AM Philippe Mathieu-Daudé <philmd@redhat.com>
wrote:

> On 3/8/21 1:14 PM, Daniel P. Berrangé wrote:
> > On Mon, Mar 08, 2021 at 12:55:10PM +0100, Thomas Huth wrote:
> >> On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
> >>> Hi Peter,
> >>>
> >>> +Markus/Paolo/Laurent/Richard
> >>>
> >>> On 3/8/21 11:24 AM, Peter Maydell wrote:
> >>>> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
> >>>>>
> >>>>> On 07/03/2021 16.56, Warner Losh wrote:
> >>>>>> The FreeBSD project has a number of enhancements to bsd-user. Add
> myself
> >>>>>> as maintainer and Kyle Evans as a reviewer. Also add our github
> repo.
> >>>>>>
> >>>>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
> >>>>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
> >>>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
> >>>>>> ---
> >>>>>>    MAINTAINERS | 5 ++++-
> >>>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
> >>>>>>
> >>>>>> diff --git a/MAINTAINERS b/MAINTAINERS
> >>>>>> index 26c9454823..ec0e935038 100644
> >>>>>> --- a/MAINTAINERS
> >>>>>> +++ b/MAINTAINERS
> >>>>>> @@ -2896,9 +2896,12 @@ F: thunk.c
> >>>>>>    F: accel/tcg/user-exec*.c
> >>>>>>
> >>>>>>    BSD user
> >>>>>> -S: Orphan
> >>>>>> +M: Warner Losh <imp@bsdimp.com>
> >>>>>> +R: Kyle Evans <kevans@freebsd.org>
> >>>>>> +S: Maintained
> >>>>>>    F: bsd-user/
> >>>>>>    F: default-configs/targets/*-bsd-user.mak
> >>>>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user
> bsd-user-rebase-3.1
> >>>>>
> >>>>> BSD is not really my home turf, but since nobody else picked this up
> and I
> >>>>> plan to send a pull request for a bunch of patches anyway this week,
> I can
> >>>>> also put it into my queue.
> >>>>
> >>>> Fine with me. (The v1 was in my to-review queue, but I'm currently
> >>>> running somewhat behind on processing patches.)
> >>>
> >>> This is a patch for mainstream QEMU, I'm having hard time
> >>> understanding the point of it. This is some official way
> >>> to say that BSD-user is not maintained in mainstream but
> >>> has to be used in the referred fork which is way different
> >>> that mainstream...
> >>>
> >>> I'd rather wait for more mainstream contributions from Warner
> >>> and Kyle, or blow the current orphan/dead code and import
> >>> bsd-user-rebase-3.1 adding the maintainer entries along, but
> >>> certainly not mark this dead code as maintained.
> >>>
> >>> Please convince me why I'm wrong, because I'd prefer NAck this
> >>> patch...
> >>
> >> The idea has been discussed here:
> >>
> >> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
> >>
> >> So this is not about declaring that bsd-user is maintained in a
> different
> >> repository, but about giving Warner et al. a chance to finally upstream
> >> their work.
> >
> > Yep, I think this change in MAINTAINERS file is primarily about
> signalling
> > intent for future
>

 Yes. Just so. We have the older fork that we're trying to port forward. If
people have changes as we do that, it sure would be nice to have them go
through us rather than create more conflict with the main tree. I don't
need to have a new set of conflicts with main because someone thought that
it would be a good idea to move the space before or after the '(' or after
in a bunch of files. It's just extra work if I have to do it as part of the
merging. Better that goes into the top of our queue so it's managed and
easy and a click on github than another half hour I have to spend sorting
that out when I'd rather be sorting out the substantial change that go on
upstream in other areas that legitimately do make things much better.

We talked about all this in the above thread, I thought, and I thought it
was all settled, so I was rather surprised to wake up to this thread this
morning.


> > Marking the subsystem as maintained isn't saying the current code is
> great,
> > just that there is someone committed to improving it hereafter.
>
> OK, thank Thomas / Daniel for explaining and referring to the "BSD-user
> plans" (which I didn't notice earlier).
>
> Warner, what about mentioning your plans here in this patch?
>

Where is there room in the MAINTAINERS file for that? How would you like me
to do that?


> Resumed ideally, else a simple link to the thread.
>

I'm not sure what you are asking here.


> > If we want to warn people that the current impl isn't great, that's goes
> > back to the topic of having a way to classify QEMU features into quality
> > levels Tier 1/2/3.
>
> That indeed sounds good w.r.t. contributors / users expectations.
>
> I suppose 1=hw_accel/security, 2=tested, 3=rest?
>
> Not a single clue how to do that although.
>

Yup. Why invent something new just to make it harder for me to get things
into the tree? There's already the tiered maintainer stuff, and I'm trying
to get our stuff that turn the current bsd-user that's crap into something
that's quite solid.

The plans are to get our changes upstream so there's no daylight between
what we do and what's upstream, except for the newest system calls that are
being implemented just after they enter FreeBSD. We have a bunch of changes
that make bsd-user able to build tens of thousands of packages more or less
natively (the setup is a hybrid environment where the compilers are cross
compilers, but everything built is native and run in emulation).

I tried getting one big omnibus patch together, but qemu's head is moving
too fast, so those plans failed to ever reach the tip of the tree. In
discussions referenced in this thread, people suggested I submit changes to
the MAINTAINER stuff as a signal moving forward. I've submitted other
patches as well to get things going (those are already in). I plan on
submitting more, but wanted to get the simple stuff ironed out first.
QEMU's workflow is utterly alien to me, to be honest, but I wanted to be a
good citizen and learn on the basics before submitting things that were
harder to do and/or explain.

Anyway, I'm trying to learn the local customs here. I'm unsure what the
next step here is. Can someone explicitly tell me that? I thought all these
preliminaries were already sorted out.

Warner

On Mon, Mar 08, 2021 at 08:21:34AM -0700, Warner Losh wrote:
> On Mon, Mar 8, 2021 at 6:14 AM Philippe Mathieu-Daudé <philmd@redhat.com>
> wrote:
> 
> > On 3/8/21 1:14 PM, Daniel P. Berrangé wrote:
> > > On Mon, Mar 08, 2021 at 12:55:10PM +0100, Thomas Huth wrote:
> > >> On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
> > >>> Hi Peter,
> > >>>
> > >>> +Markus/Paolo/Laurent/Richard
> > >>>
> > >>> On 3/8/21 11:24 AM, Peter Maydell wrote:
> > >>>> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com> wrote:
> > >>>>>
> > >>>>> On 07/03/2021 16.56, Warner Losh wrote:
> > >>>>>> The FreeBSD project has a number of enhancements to bsd-user. Add
> > myself
> > >>>>>> as maintainer and Kyle Evans as a reviewer. Also add our github
> > repo.
> > >>>>>>
> > >>>>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
> > >>>>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
> > >>>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
> > >>>>>> ---
> > >>>>>>    MAINTAINERS | 5 ++++-
> > >>>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
> > >>>>>>
> > >>>>>> diff --git a/MAINTAINERS b/MAINTAINERS
> > >>>>>> index 26c9454823..ec0e935038 100644
> > >>>>>> --- a/MAINTAINERS
> > >>>>>> +++ b/MAINTAINERS
> > >>>>>> @@ -2896,9 +2896,12 @@ F: thunk.c
> > >>>>>>    F: accel/tcg/user-exec*.c
> > >>>>>>
> > >>>>>>    BSD user
> > >>>>>> -S: Orphan
> > >>>>>> +M: Warner Losh <imp@bsdimp.com>
> > >>>>>> +R: Kyle Evans <kevans@freebsd.org>
> > >>>>>> +S: Maintained
> > >>>>>>    F: bsd-user/
> > >>>>>>    F: default-configs/targets/*-bsd-user.mak
> > >>>>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user
> > bsd-user-rebase-3.1
> > >>>>>
> > >>>>> BSD is not really my home turf, but since nobody else picked this up
> > and I
> > >>>>> plan to send a pull request for a bunch of patches anyway this week,
> > I can
> > >>>>> also put it into my queue.
> > >>>>
> > >>>> Fine with me. (The v1 was in my to-review queue, but I'm currently
> > >>>> running somewhat behind on processing patches.)
> > >>>
> > >>> This is a patch for mainstream QEMU, I'm having hard time
> > >>> understanding the point of it. This is some official way
> > >>> to say that BSD-user is not maintained in mainstream but
> > >>> has to be used in the referred fork which is way different
> > >>> that mainstream...
> > >>>
> > >>> I'd rather wait for more mainstream contributions from Warner
> > >>> and Kyle, or blow the current orphan/dead code and import
> > >>> bsd-user-rebase-3.1 adding the maintainer entries along, but
> > >>> certainly not mark this dead code as maintained.
> > >>>
> > >>> Please convince me why I'm wrong, because I'd prefer NAck this
> > >>> patch...
> > >>
> > >> The idea has been discussed here:
> > >>
> > >> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
> > >>
> > >> So this is not about declaring that bsd-user is maintained in a
> > different
> > >> repository, but about giving Warner et al. a chance to finally upstream
> > >> their work.
> > >
> > > Yep, I think this change in MAINTAINERS file is primarily about
> > signalling
> > > intent for future
> >
> 
>  Yes. Just so. We have the older fork that we're trying to port forward. If
> people have changes as we do that, it sure would be nice to have them go
> through us rather than create more conflict with the main tree. I don't
> need to have a new set of conflicts with main because someone thought that
> it would be a good idea to move the space before or after the '(' or after
> in a bunch of files. It's just extra work if I have to do it as part of the
> merging. Better that goes into the top of our queue so it's managed and
> easy and a click on github than another half hour I have to spend sorting
> that out when I'd rather be sorting out the substantial change that go on
> upstream in other areas that legitimately do make things much better.
> 
> We talked about all this in the above thread, I thought, and I thought it
> was all settled, so I was rather surprised to wake up to this thread this
> morning.

From my POV, nothing has changed - I think the MAINTAINERS change is
reasonable.


> > > Marking the subsystem as maintained isn't saying the current code is
> > great,
> > > just that there is someone committed to improving it hereafter.
> >
> > OK, thank Thomas / Daniel for explaining and referring to the "BSD-user
> > plans" (which I didn't notice earlier).
> >
> > Warner, what about mentioning your plans here in this patch?
> >
> 
> Where is there room in the MAINTAINERS file for that? How would you like me
> to do that?

I think he was just referring to mentioning the intention in the
commit message.  The MAINTAINERS file isn't a suitable place for
describing plans.

> > > If we want to warn people that the current impl isn't great, that's goes
> > > back to the topic of having a way to classify QEMU features into quality
> > > levels Tier 1/2/3.
> >
> > That indeed sounds good w.r.t. contributors / users expectations.
> >
> > I suppose 1=hw_accel/security, 2=tested, 3=rest?
> >
> > Not a single clue how to do that although.
> >
> 
> Yup. Why invent something new just to make it harder for me to get things
> into the tree? There's already the tiered maintainer stuff, and I'm trying
> to get our stuff that turn the current bsd-user that's crap into something
> that's quite solid.


...snip...

You don't have to worry about this  - it is tangential to your patch
here - a problem facing QEMU as a whole, not BSD specific, and there's
no immediate answer here either.


Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

On Mon, Mar 8, 2021 at 8:28 AM Daniel P. Berrangé <berrange@redhat.com>
wrote:

> On Mon, Mar 08, 2021 at 08:21:34AM -0700, Warner Losh wrote:
> > On Mon, Mar 8, 2021 at 6:14 AM Philippe Mathieu-Daudé <philmd@redhat.com
> >
> > wrote:
> >
> > > On 3/8/21 1:14 PM, Daniel P. Berrangé wrote:
> > > > On Mon, Mar 08, 2021 at 12:55:10PM +0100, Thomas Huth wrote:
> > > >> On 08/03/2021 12.16, Philippe Mathieu-Daudé wrote:
> > > >>> Hi Peter,
> > > >>>
> > > >>> +Markus/Paolo/Laurent/Richard
> > > >>>
> > > >>> On 3/8/21 11:24 AM, Peter Maydell wrote:
> > > >>>> On Mon, 8 Mar 2021 at 10:09, Thomas Huth <thuth@redhat.com>
> wrote:
> > > >>>>>
> > > >>>>> On 07/03/2021 16.56, Warner Losh wrote:
> > > >>>>>> The FreeBSD project has a number of enhancements to bsd-user.
> Add
> > > myself
> > > >>>>>> as maintainer and Kyle Evans as a reviewer. Also add our github
> > > repo.
> > > >>>>>>
> > > >>>>>> Signed-off-by: Warner Losh <imp@bsdimp.com>
> > > >>>>>> Signed-off-by: Kyle Evans <kevans@freebsd.org>
> > > >>>>>> Reviewed-by: Thomas Huth <thuth@redhat.com>
> > > >>>>>> ---
> > > >>>>>>    MAINTAINERS | 5 ++++-
> > > >>>>>>    1 file changed, 4 insertions(+), 1 deletion(-)
> > > >>>>>>
> > > >>>>>> diff --git a/MAINTAINERS b/MAINTAINERS
> > > >>>>>> index 26c9454823..ec0e935038 100644
> > > >>>>>> --- a/MAINTAINERS
> > > >>>>>> +++ b/MAINTAINERS
> > > >>>>>> @@ -2896,9 +2896,12 @@ F: thunk.c
> > > >>>>>>    F: accel/tcg/user-exec*.c
> > > >>>>>>
> > > >>>>>>    BSD user
> > > >>>>>> -S: Orphan
> > > >>>>>> +M: Warner Losh <imp@bsdimp.com>
> > > >>>>>> +R: Kyle Evans <kevans@freebsd.org>
> > > >>>>>> +S: Maintained
> > > >>>>>>    F: bsd-user/
> > > >>>>>>    F: default-configs/targets/*-bsd-user.mak
> > > >>>>>> +T: git https://github.com/qemu-bsd-user/qemu-bsd-user
> > > bsd-user-rebase-3.1
> > > >>>>>
> > > >>>>> BSD is not really my home turf, but since nobody else picked
> this up
> > > and I
> > > >>>>> plan to send a pull request for a bunch of patches anyway this
> week,
> > > I can
> > > >>>>> also put it into my queue.
> > > >>>>
> > > >>>> Fine with me. (The v1 was in my to-review queue, but I'm currently
> > > >>>> running somewhat behind on processing patches.)
> > > >>>
> > > >>> This is a patch for mainstream QEMU, I'm having hard time
> > > >>> understanding the point of it. This is some official way
> > > >>> to say that BSD-user is not maintained in mainstream but
> > > >>> has to be used in the referred fork which is way different
> > > >>> that mainstream...
> > > >>>
> > > >>> I'd rather wait for more mainstream contributions from Warner
> > > >>> and Kyle, or blow the current orphan/dead code and import
> > > >>> bsd-user-rebase-3.1 adding the maintainer entries along, but
> > > >>> certainly not mark this dead code as maintained.
> > > >>>
> > > >>> Please convince me why I'm wrong, because I'd prefer NAck this
> > > >>> patch...
> > > >>
> > > >> The idea has been discussed here:
> > > >>
> > > >> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
> > > >>
> > > >> So this is not about declaring that bsd-user is maintained in a
> > > different
> > > >> repository, but about giving Warner et al. a chance to finally
> upstream
> > > >> their work.
> > > >
> > > > Yep, I think this change in MAINTAINERS file is primarily about
> > > signalling
> > > > intent for future
> > >
> >
> >  Yes. Just so. We have the older fork that we're trying to port forward.
> If
> > people have changes as we do that, it sure would be nice to have them go
> > through us rather than create more conflict with the main tree. I don't
> > need to have a new set of conflicts with main because someone thought
> that
> > it would be a good idea to move the space before or after the '(' or
> after
> > in a bunch of files. It's just extra work if I have to do it as part of
> the
> > merging. Better that goes into the top of our queue so it's managed and
> > easy and a click on github than another half hour I have to spend sorting
> > that out when I'd rather be sorting out the substantial change that go on
> > upstream in other areas that legitimately do make things much better.
> >
> > We talked about all this in the above thread, I thought, and I thought it
> > was all settled, so I was rather surprised to wake up to this thread this
> > morning.
>
> From my POV, nothing has changed - I think the MAINTAINERS change is
> reasonable.
>
>
> > > > Marking the subsystem as maintained isn't saying the current code is
> > > great,
> > > > just that there is someone committed to improving it hereafter.
> > >
> > > OK, thank Thomas / Daniel for explaining and referring to the "BSD-user
> > > plans" (which I didn't notice earlier).
> > >
> > > Warner, what about mentioning your plans here in this patch?
> > >
> >
> > Where is there room in the MAINTAINERS file for that? How would you like
> me
> > to do that?
>
> I think he was just referring to mentioning the intention in the
> commit message.  The MAINTAINERS file isn't a suitable place for
> describing plans.
>

Ah, that makes sense... I can do that and resend.


> > > > If we want to warn people that the current impl isn't great, that's
> goes
> > > > back to the topic of having a way to classify QEMU features into
> quality
> > > > levels Tier 1/2/3.
> > >
> > > That indeed sounds good w.r.t. contributors / users expectations.
> > >
> > > I suppose 1=hw_accel/security, 2=tested, 3=rest?
> > >
> > > Not a single clue how to do that although.
> > >
> >
> > Yup. Why invent something new just to make it harder for me to get things
> > into the tree? There's already the tiered maintainer stuff, and I'm
> trying
> > to get our stuff that turn the current bsd-user that's crap into
> something
> > that's quite solid.
>
>
> ...snip...
>
> You don't have to worry about this  - it is tangential to your patch
> here - a problem facing QEMU as a whole, not BSD specific, and there's
> no immediate answer here either.
>

Gotcha. I wasn't sure if there was something else I needed to do here or
not.

Warner

On Mon, 8 Mar 2021 at 11:55, Thomas Huth <thuth@redhat.com> wrote:
> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01399.html
>
> So this is not about declaring that bsd-user is maintained in a different
> repository, but about giving Warner et al. a chance to finally upstream
> their work.

Thanks; I'd forgotten (or had not read that email carefully enough
the first time around to realize) that the current plan for bsd-user is
improve-in-place rather than delete-and-restart. (I'm supportive of
whatever Warner and Kyle think is the best approach, in any case.)

thanks
-- PMM

FreeBSD 12.1 has reached end of life. Use 12.2 instead so that FreeBSD's
project's packages will work.  Update which timezone to pick. Work around a QEMU
bug that incorrectly raises an exception on a CRC32 instruction with the FPU
disabled.  The qemu bug is described here:
https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html

Signed-off-by: Warner Losh <imp@bsdimp.com>

---
 tests/vm/freebsd | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/tests/vm/freebsd b/tests/vm/freebsd
index 09f3ee6cb8..81389d527e 100755
--- a/tests/vm/freebsd
+++ b/tests/vm/freebsd
@@ -20,12 +20,16 @@ import socket
 import subprocess
 import basevm
 
+FREEBSD_CONFIG = {
+    'cpu'	: "max,sse4.2=off",
+}
+
 class FreeBSDVM(basevm.BaseVM):
     name = "freebsd"
     arch = "x86_64"
 
-    link = "https://download.freebsd.org/ftp/releases/ISO-IMAGES/12.1/FreeBSD-12.1-RELEASE-amd64-disc1.iso.xz"
-    csum = "7394c3f60a1e236e7bd3a05809cf43ae39a3b8e5d42d782004cf2f26b1cfcd88"
+    link = "https://download.freebsd.org/ftp/releases/ISO-IMAGES/12.2/FreeBSD-12.2-RELEASE-amd64-disc1.iso.xz"
+    csum = "a4530246cafbf1dd42a9bd3ea441ca9a78a6a0cd070278cbdf63f3a6f803ecae"
     size = "20G"
     pkgs = [
         # build tools
@@ -125,7 +129,7 @@ class FreeBSDVM(basevm.BaseVM):
         self.console_wait_send("IPv6",                  "n")
         self.console_wait_send("Resolver",              "\n")
 
-        self.console_wait_send("Time Zone Selector",    "a\n")
+        self.console_wait_send("Time Zone Selector",    "0\n")
         self.console_wait_send("Confirmation",          "y")
         self.console_wait_send("Time & Date",           "\n")
         self.console_wait_send("Time & Date",           "\n")
@@ -206,4 +210,4 @@ class FreeBSDVM(basevm.BaseVM):
         self.print_step("All done")
 
 if __name__ == "__main__":
-    sys.exit(basevm.main(FreeBSDVM))
+    sys.exit(basevm.main(FreeBSDVM, config=FREEBSD_CONFIG))
-- 
2.30.0

On 07/03/2021 16.56, Warner Losh wrote:
> FreeBSD 12.1 has reached end of life. Use 12.2 instead so that FreeBSD's
> project's packages will work.  Update which timezone to pick. Work around a QEMU
> bug that incorrectly raises an exception on a CRC32 instruction with the FPU
> disabled.  The qemu bug is described here:
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
> 
> Signed-off-by: Warner Losh <imp@bsdimp.com>
> 
> ---
>   tests/vm/freebsd | 14 +++++++++-----
>   1 file changed, 9 insertions(+), 5 deletions(-)

I gave this a try, but it's currently failing in the unit tests:

Running test test-crypto-tlscredsx509

** (tests/test-crypto-tlscredsx509:9999): CRITICAL **: 12:56:35.157: Failed 
to sign certificate ASN1 parser: Value is not valid.
ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed to sign 
certificate ASN1 parser: Value is not valid.
gmake: *** [Makefile.mtest:576: run-test-70] Error 1
gmake: *** Waiting for unfinished jobs....
Running test test-crypto-tlssession

** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288: Failed 
to sign certificate ASN1 parser: Value is not valid.
ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to sign 
certificate ASN1 parser: Value is not valid.

I guess it's the same problem as:

https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html

... so this would require a bug fix in the libtasn of FreeBSD first? See:

https://gitlab.com/gnutls/libtasn1/-/merge_requests/71

  Thomas

On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com> wrote:

> On 07/03/2021 16.56, Warner Losh wrote:
> > FreeBSD 12.1 has reached end of life. Use 12.2 instead so that FreeBSD's
> > project's packages will work.  Update which timezone to pick. Work
> around a QEMU
> > bug that incorrectly raises an exception on a CRC32 instruction with the
> FPU
> > disabled.  The qemu bug is described here:
> > https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
> >
> > Signed-off-by: Warner Losh <imp@bsdimp.com>
> >
> > ---
> >   tests/vm/freebsd | 14 +++++++++-----
> >   1 file changed, 9 insertions(+), 5 deletions(-)
>
> I gave this a try, but it's currently failing in the unit tests:
>
> Running test test-crypto-tlscredsx509
>
> ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **: 12:56:35.157:
> Failed
> to sign certificate ASN1 parser: Value is not valid.
> ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed to sign
> certificate ASN1 parser: Value is not valid.
> gmake: *** [Makefile.mtest:576: run-test-70] Error 1
> gmake: *** Waiting for unfinished jobs....
> Running test test-crypto-tlssession
>
> ** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288: Failed
> to sign certificate ASN1 parser: Value is not valid.
> ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to sign
> certificate ASN1 parser: Value is not valid.
>

That's totally unrelated to my change. Was it failing before? What
environment was it failing in because it all seemed to work for me...


> I guess it's the same problem as:
>
> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>
> ... so this would require a bug fix in the libtasn of FreeBSD first? See:
>
> https://gitlab.com/gnutls/libtasn1/-/merge_requests/71


 Is this on the host that built qemu, or inside the VM or where exactly?

Warner

On 08/03/2021 16.26, Warner Losh wrote:
> 
> 
> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com 
> <mailto:thuth@redhat.com>> wrote:
> 
>     On 07/03/2021 16.56, Warner Losh wrote:
>      > FreeBSD 12.1 has reached end of life. Use 12.2 instead so that FreeBSD's
>      > project's packages will work.  Update which timezone to pick. Work
>     around a QEMU
>      > bug that incorrectly raises an exception on a CRC32 instruction with
>     the FPU
>      > disabled.  The qemu bug is described here:
>      > https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>     <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
>      >
>      > Signed-off-by: Warner Losh <imp@bsdimp.com <mailto:imp@bsdimp.com>>
>      >
>      > ---
>      >   tests/vm/freebsd | 14 +++++++++-----
>      >   1 file changed, 9 insertions(+), 5 deletions(-)
> 
>     I gave this a try, but it's currently failing in the unit tests:
> 
>     Running test test-crypto-tlscredsx509
> 
>     ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **: 12:56:35.157: Failed
>     to sign certificate ASN1 parser: Value is not valid.
>     ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed to sign
>     certificate ASN1 parser: Value is not valid.
>     gmake: *** [Makefile.mtest:576: run-test-70] Error 1
>     gmake: *** Waiting for unfinished jobs....
>     Running test test-crypto-tlssession
> 
>     ** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288: Failed
>     to sign certificate ASN1 parser: Value is not valid.
>     ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to sign
>     certificate ASN1 parser: Value is not valid.
> 
> 
> That's totally unrelated to my change. Was it failing before? What 
> environment was it failing in because it all seemed to work for me...

It's been a while since I last ran "make vm-build-freebsd", so I can't 
really tell whether the problem was already there before ... when I now try 
to run it without your patch, it fails for me, too, but rather due to 
FreeBSD 12.1 being out of service instead.

>     I guess it's the same problem as:
> 
>     https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>     <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>
> 
>     ... so this would require a bug fix in the libtasn of FreeBSD first? See:
> 
>     https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>     <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
> 
>   Is this on the host that built qemu, or inside the VM or where exactly?

It's inside the VM ... I assume the libtasn there has the same bug as the 
one on macOS?

  Thomas

> 8 марта 2021 г., в 18:41, Thomas Huth <thuth@redhat.com> написал(а):
> 
> On 08/03/2021 16.26, Warner Losh wrote:
>> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com <mailto:thuth@redhat.com>> wrote:
>>    On 07/03/2021 16.56, Warner Losh wrote:
>>     > FreeBSD 12.1 has reached end of life. Use 12.2 instead so that FreeBSD's
>>     > project's packages will work.  Update which timezone to pick. Work
>>    around a QEMU
>>     > bug that incorrectly raises an exception on a CRC32 instruction with
>>    the FPU
>>     > disabled.  The qemu bug is described here:
>>     > https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>>    <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
>>     >
>>     > Signed-off-by: Warner Losh <imp@bsdimp.com <mailto:imp@bsdimp.com>>
>>     >
>>     > ---
>>     >   tests/vm/freebsd | 14 +++++++++-----
>>     >   1 file changed, 9 insertions(+), 5 deletions(-)
>>    I gave this a try, but it's currently failing in the unit tests:
>>    Running test test-crypto-tlscredsx509
>>    ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **: 12:56:35.157: Failed
>>    to sign certificate ASN1 parser: Value is not valid.
>>    ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed to sign
>>    certificate ASN1 parser: Value is not valid.
>>    gmake: *** [Makefile.mtest:576: run-test-70] Error 1
>>    gmake: *** Waiting for unfinished jobs....
>>    Running test test-crypto-tlssession
>>    ** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288: Failed
>>    to sign certificate ASN1 parser: Value is not valid.
>>    ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to sign
>>    certificate ASN1 parser: Value is not valid.
>> That's totally unrelated to my change. Was it failing before? What environment was it failing in because it all seemed to work for me...
> 
> It's been a while since I last ran "make vm-build-freebsd", so I can't really tell whether the problem was already there before ... when I now try to run it without your patch, it fails for me, too, but rather due to FreeBSD 12.1 being out of service instead.
> 
>>    I guess it's the same problem as:
>>    https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>>    <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>
>>    ... so this would require a bug fix in the libtasn of FreeBSD first? See:
>>    https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>>    <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
>>  Is this on the host that built qemu, or inside the VM or where exactly?
> 
> It's inside the VM ... I assume the libtasn there has the same bug as the one on macOS?
> 

The gnutls failures on macOS and FreeBSD (with clang as main compiler) won’t happen only if libtasn1 from master is used. Otherwise libtasn1 has to be compiled with -O1/-O0.

Отправлено с iPhone

> Thomas
> 

On Mon, Mar 8, 2021 at 8:46 AM Roman Bolshakov <r.bolshakov@yadro.com>
wrote:

>
> > 8 марта 2021 г., в 18:41, Thomas Huth <thuth@redhat.com> написал(а):
> >
> > On 08/03/2021 16.26, Warner Losh wrote:
> >> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com <mailto:
> thuth@redhat.com>> wrote:
> >>    On 07/03/2021 16.56, Warner Losh wrote:
> >>     > FreeBSD 12.1 has reached end of life. Use 12.2 instead so that
> FreeBSD's
> >>     > project's packages will work.  Update which timezone to pick. Work
> >>    around a QEMU
> >>     > bug that incorrectly raises an exception on a CRC32 instruction
> with
> >>    the FPU
> >>     > disabled.  The qemu bug is described here:
> >>     > https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
> >>    <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
> >>     >
> >>     > Signed-off-by: Warner Losh <imp@bsdimp.com <mailto:imp@bsdimp.com
> >>
> >>     >
> >>     > ---
> >>     >   tests/vm/freebsd | 14 +++++++++-----
> >>     >   1 file changed, 9 insertions(+), 5 deletions(-)
> >>    I gave this a try, but it's currently failing in the unit tests:
> >>    Running test test-crypto-tlscredsx509
> >>    ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **: 12:56:35.157:
> Failed
> >>    to sign certificate ASN1 parser: Value is not valid.
> >>    ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed to
> sign
> >>    certificate ASN1 parser: Value is not valid.
> >>    gmake: *** [Makefile.mtest:576: run-test-70] Error 1
> >>    gmake: *** Waiting for unfinished jobs....
> >>    Running test test-crypto-tlssession
> >>    ** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288:
> Failed
> >>    to sign certificate ASN1 parser: Value is not valid.
> >>    ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to
> sign
> >>    certificate ASN1 parser: Value is not valid.
> >> That's totally unrelated to my change. Was it failing before? What
> environment was it failing in because it all seemed to work for me...
> >
> > It's been a while since I last ran "make vm-build-freebsd", so I can't
> really tell whether the problem was already there before ... when I now try
> to run it without your patch, it fails for me, too, but rather due to
> FreeBSD 12.1 being out of service instead.
> >
> >>    I guess it's the same problem as:
> >>    https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
> >>    <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
> >
> >>    ... so this would require a bug fix in the libtasn of FreeBSD first?
> See:
> >>    https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
> >>    <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
> >>  Is this on the host that built qemu, or inside the VM or where exactly?
> >
> > It's inside the VM ... I assume the libtasn there has the same bug as
> the one on macOS?
> >
>
> The gnutls failures on macOS and FreeBSD (with clang as main compiler)
> won’t happen only if libtasn1 from master is used. Otherwise libtasn1 has
> to be compiled with -O1/-O0.
>

Forgive my ignorance...

Does this come from a submodule? From something I've built and installed on
my FreeBSD host? Or something that's installed into the FreeBSD guest?

Warner


> Отправлено с iPhone
>
> > Thomas
> >
>

On Mon, Mar 8, 2021 at 12:51 PM Warner Losh <imp@bsdimp.com> wrote:

>
>
> On Mon, Mar 8, 2021 at 8:46 AM Roman Bolshakov <r.bolshakov@yadro.com>
> wrote:
>
>>
>> > 8 марта 2021 г., в 18:41, Thomas Huth <thuth@redhat.com> написал(а):
>> >
>> > On 08/03/2021 16.26, Warner Losh wrote:
>> >> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com <mailto:
>> thuth@redhat.com>> wrote:
>> >>    On 07/03/2021 16.56, Warner Losh wrote:
>> >>     > FreeBSD 12.1 has reached end of life. Use 12.2 instead so that
>> FreeBSD's
>> >>     > project's packages will work.  Update which timezone to pick.
>> Work
>> >>    around a QEMU
>> >>     > bug that incorrectly raises an exception on a CRC32 instruction
>> with
>> >>    the FPU
>> >>     > disabled.  The qemu bug is described here:
>> >>     >
>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>> >>    <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
>> >>     >
>> >>     > Signed-off-by: Warner Losh <imp@bsdimp.com <mailto:
>> imp@bsdimp.com>>
>> >>     >
>> >>     > ---
>> >>     >   tests/vm/freebsd | 14 +++++++++-----
>> >>     >   1 file changed, 9 insertions(+), 5 deletions(-)
>> >>    I gave this a try, but it's currently failing in the unit tests:
>> >>    Running test test-crypto-tlscredsx509
>> >>    ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **:
>> 12:56:35.157: Failed
>> >>    to sign certificate ASN1 parser: Value is not valid.
>> >>    ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL: Failed
>> to sign
>> >>    certificate ASN1 parser: Value is not valid.
>> >>    gmake: *** [Makefile.mtest:576: run-test-70] Error 1
>> >>    gmake: *** Waiting for unfinished jobs....
>> >>    Running test test-crypto-tlssession
>> >>    ** (tests/test-crypto-tlssession:10002): CRITICAL **: 12:56:35.288:
>> Failed
>> >>    to sign certificate ASN1 parser: Value is not valid.
>> >>    ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL: Failed to
>> sign
>> >>    certificate ASN1 parser: Value is not valid.
>> >> That's totally unrelated to my change. Was it failing before? What
>> environment was it failing in because it all seemed to work for me...
>> >
>> > It's been a while since I last ran "make vm-build-freebsd", so I can't
>> really tell whether the problem was already there before ... when I now try
>> to run it without your patch, it fails for me, too, but rather due to
>> FreeBSD 12.1 being out of service instead.
>> >
>> >>    I guess it's the same problem as:
>> >>    https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>> >>    <
>> https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>
>> >>    ... so this would require a bug fix in the libtasn of FreeBSD
>> first? See:
>> >>    https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>> >>    <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
>> >>  Is this on the host that built qemu, or inside the VM or where
>> exactly?
>> >
>> > It's inside the VM ... I assume the libtasn there has the same bug as
>> the one on macOS?
>> >
>>
>> The gnutls failures on macOS and FreeBSD (with clang as main compiler)
>> won’t happen only if libtasn1 from master is used. Otherwise libtasn1 has
>> to be compiled with -O1/-O0.
>>
>
> Forgive my ignorance...
>
> Does this come from a submodule? From something I've built and installed
> on my FreeBSD host? Or something that's installed into the FreeBSD guest
>

ah, it's inside the guest.

Warner

On 08/03/2021 20.57, Warner Losh wrote:
> 
> 
> On Mon, Mar 8, 2021 at 12:51 PM Warner Losh <imp@bsdimp.com 
> <mailto:imp@bsdimp.com>> wrote:
> 
> 
> 
>     On Mon, Mar 8, 2021 at 8:46 AM Roman Bolshakov <r.bolshakov@yadro.com
>     <mailto:r.bolshakov@yadro.com>> wrote:
> 
> 
>          > 8 марта 2021 г., в 18:41, Thomas Huth <thuth@redhat.com
>         <mailto:thuth@redhat.com>> написал(а):
>          >
>          > On 08/03/2021 16.26, Warner Losh wrote:
>          >> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com
>         <mailto:thuth@redhat.com> <mailto:thuth@redhat.com
>         <mailto:thuth@redhat.com>>> wrote:
>          >>    On 07/03/2021 16.56, Warner Losh wrote:
>          >>     > FreeBSD 12.1 has reached end of life. Use 12.2 instead so
>         that FreeBSD's
>          >>     > project's packages will work.  Update which timezone to
>         pick. Work
>          >>    around a QEMU
>          >>     > bug that incorrectly raises an exception on a CRC32
>         instruction with
>          >>    the FPU
>          >>     > disabled.  The qemu bug is described here:
>          >>     >
>         https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>         <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
>          >>   
>         <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>         <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>>
>          >>     >
>          >>     > Signed-off-by: Warner Losh <imp@bsdimp.com
>         <mailto:imp@bsdimp.com> <mailto:imp@bsdimp.com <mailto:imp@bsdimp.com>>>
>          >>     >
>          >>     > ---
>          >>     >   tests/vm/freebsd | 14 +++++++++-----
>          >>     >   1 file changed, 9 insertions(+), 5 deletions(-)
>          >>    I gave this a try, but it's currently failing in the unit tests:
>          >>    Running test test-crypto-tlscredsx509
>          >>    ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **:
>         12:56:35.157: Failed
>          >>    to sign certificate ASN1 parser: Value is not valid.
>          >>    ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL:
>         Failed to sign
>          >>    certificate ASN1 parser: Value is not valid.
>          >>    gmake: *** [Makefile.mtest:576: run-test-70] Error 1
>          >>    gmake: *** Waiting for unfinished jobs....
>          >>    Running test test-crypto-tlssession
>          >>    ** (tests/test-crypto-tlssession:10002): CRITICAL **:
>         12:56:35.288: Failed
>          >>    to sign certificate ASN1 parser: Value is not valid.
>          >>    ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL:
>         Failed to sign
>          >>    certificate ASN1 parser: Value is not valid.
>          >> That's totally unrelated to my change. Was it failing before?
>         What environment was it failing in because it all seemed to work for
>         me...
>          >
>          > It's been a while since I last ran "make vm-build-freebsd", so I
>         can't really tell whether the problem was already there before ...
>         when I now try to run it without your patch, it fails for me, too,
>         but rather due to FreeBSD 12.1 being out of service instead.
>          >
>          >>    I guess it's the same problem as:
>          >>
>         https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>         <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>
>          >>   
>         <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>         <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>>
>          >>    ... so this would require a bug fix in the libtasn of FreeBSD
>         first? See:
>          >> https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>         <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
>          >>    <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>         <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>>
>          >>  Is this on the host that built qemu, or inside the VM or where
>         exactly?
>          >
>          > It's inside the VM ... I assume the libtasn there has the same
>         bug as the one on macOS?
>          >
> 
>         The gnutls failures on macOS and FreeBSD (with clang as main
>         compiler) won’t happen only if libtasn1 from master is used.
>         Otherwise libtasn1 has to be compiled with -O1/-O0.
> 
> 
>     Forgive my ignorance...
> 
>     Does this come from a submodule? From something I've built and installed
>     on my FreeBSD host? Or something that's installed into the FreeBSD guest
> 
> 
> ah, it's inside the guest.

Yes, IIUC it's the pre-installed libtasn1 in the guest, and the problem 
likely occurs just now since the clang that has been used to compile the 
library has been updated, too?

Anyway, can you reproduce the issue? If so, could you please file a bug 
ticket against libtasn1 there, so that it gets fixed? (sorry, I don't have 
any clue about the FreeBSD process here)

  Thomas

On 09/03/2021 06.08, Thomas Huth wrote:
> On 08/03/2021 20.57, Warner Losh wrote:
>>
>>
>> On Mon, Mar 8, 2021 at 12:51 PM Warner Losh <imp@bsdimp.com 
>> <mailto:imp@bsdimp.com>> wrote:
>>
>>
>>
>>     On Mon, Mar 8, 2021 at 8:46 AM Roman Bolshakov <r.bolshakov@yadro.com
>>     <mailto:r.bolshakov@yadro.com>> wrote:
>>
>>
>>          > 8 марта 2021 г., в 18:41, Thomas Huth <thuth@redhat.com
>>         <mailto:thuth@redhat.com>> написал(а):
>>          >
>>          > On 08/03/2021 16.26, Warner Losh wrote:
>>          >> On Mon, Mar 8, 2021 at 6:30 AM Thomas Huth <thuth@redhat.com
>>         <mailto:thuth@redhat.com> <mailto:thuth@redhat.com
>>         <mailto:thuth@redhat.com>>> wrote:
>>          >>    On 07/03/2021 16.56, Warner Losh wrote:
>>          >>     > FreeBSD 12.1 has reached end of life. Use 12.2 instead so
>>         that FreeBSD's
>>          >>     > project's packages will work.  Update which timezone to
>>         pick. Work
>>          >>    around a QEMU
>>          >>     > bug that incorrectly raises an exception on a CRC32
>>         instruction with
>>          >>    the FPU
>>          >>     > disabled.  The qemu bug is described here:
>>          >>     >
>>         https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>>         <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>
>>          >>         
>> <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html
>>         <https://www.mail-archive.com/qemu-devel@nongnu.org/msg784158.html>>
>>          >>     >
>>          >>     > Signed-off-by: Warner Losh <imp@bsdimp.com
>>         <mailto:imp@bsdimp.com> <mailto:imp@bsdimp.com 
>> <mailto:imp@bsdimp.com>>>
>>          >>     >
>>          >>     > ---
>>          >>     >   tests/vm/freebsd | 14 +++++++++-----
>>          >>     >   1 file changed, 9 insertions(+), 5 deletions(-)
>>          >>    I gave this a try, but it's currently failing in the unit 
>> tests:
>>          >>    Running test test-crypto-tlscredsx509
>>          >>    ** (tests/test-crypto-tlscredsx509:9999): CRITICAL **:
>>         12:56:35.157: Failed
>>          >>    to sign certificate ASN1 parser: Value is not valid.
>>          >>    ERROR test-crypto-tlscredsx509 - Bail out! FATAL-CRITICAL:
>>         Failed to sign
>>          >>    certificate ASN1 parser: Value is not valid.
>>          >>    gmake: *** [Makefile.mtest:576: run-test-70] Error 1
>>          >>    gmake: *** Waiting for unfinished jobs....
>>          >>    Running test test-crypto-tlssession
>>          >>    ** (tests/test-crypto-tlssession:10002): CRITICAL **:
>>         12:56:35.288: Failed
>>          >>    to sign certificate ASN1 parser: Value is not valid.
>>          >>    ERROR test-crypto-tlssession - Bail out! FATAL-CRITICAL:
>>         Failed to sign
>>          >>    certificate ASN1 parser: Value is not valid.
>>          >> That's totally unrelated to my change. Was it failing before?
>>         What environment was it failing in because it all seemed to work for
>>         me...
>>          >
>>          > It's been a while since I last ran "make vm-build-freebsd", so I
>>         can't really tell whether the problem was already there before ...
>>         when I now try to run it without your patch, it fails for me, too,
>>         but rather due to FreeBSD 12.1 being out of service instead.
>>          >
>>          >>    I guess it's the same problem as:
>>          >>
>>         https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>>         <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>
>>          >>         
>> <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html
>>         
>> <https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg06750.html>>
>>          >>    ... so this would require a bug fix in the libtasn of FreeBSD
>>         first? See:
>>          >> https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>>         <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>
>>          >>    <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
>>         <https://gitlab.com/gnutls/libtasn1/-/merge_requests/71>>
>>          >>  Is this on the host that built qemu, or inside the VM or where
>>         exactly?
>>          >
>>          > It's inside the VM ... I assume the libtasn there has the same
>>         bug as the one on macOS?
>>          >
>>
>>         The gnutls failures on macOS and FreeBSD (with clang as main
>>         compiler) won’t happen only if libtasn1 from master is used.
>>         Otherwise libtasn1 has to be compiled with -O1/-O0.
>>
>>
>>     Forgive my ignorance...
>>
>>     Does this come from a submodule? From something I've built and installed
>>     on my FreeBSD host? Or something that's installed into the FreeBSD guest
>>
>>
>> ah, it's inside the guest.
> 
> Yes, IIUC it's the pre-installed libtasn1 in the guest, and the problem 
> likely occurs just now since the clang that has been used to compile the 
> library has been updated, too?
> 
> Anyway, can you reproduce the issue? If so, could you please file a bug 
> ticket against libtasn1 there, so that it gets fixed? (sorry, I don't have 
> any clue about the FreeBSD process here)

For the time being, I suggest we add this hunk to your patch here to get
the VM-build going again:

diff --git a/tests/vm/freebsd b/tests/vm/freebsd
--- a/tests/vm/freebsd
+++ b/tests/vm/freebsd
@@ -46,7 +46,9 @@ class FreeBSDVM(basevm.BaseVM):
          "gettext",
  
          # libs: crypto
-        "gnutls",
+        # TODO: Enable gnutls again once FreeBSD's libtasn1 got fixed
+        # See: https://gitlab.com/gnutls/libtasn1/-/merge_requests/71
+        # "gnutls",
  
          # libs: images
          "jpeg-turbo",

  Thomas

Am 08.03.21 um 16:46 schrieb Roman Bolshakov:

> gnutls failures on macOS and FreeBSD (with clang as main compiler) won’t happen only if libtasn1 from master is used. Otherwise libtasn1 has to be compiled with -O1/-O0.


Of course it is also possible to patch older versions of libtasn1 to
make them work with clang -O2.

Stefan