Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Wei Liu]

On Tue, Mar 13, 2018 at 02:23:15PM +, Roger Pau Monné wrote:
> On Mon, Mar 05, 2018 at 11:36:58AM +, Ian Jackson wrote:
> > Roger Pau Monne writes ("[PATCH] xl: remove apic option for PVH guests"):
> > > XSA-256 forces the local APIC to always be enabled for PVH guests, so
> > > ignore any apic option for PVH guests. Update the documentation
> > > accordingly.
> > ...
> > > diff --git a/tools/xl/xl_parse.c b/tools/xl/xl_parse.c
> > > index f6842540ca..8b999825d2 100644
> > > --- a/tools/xl/xl_parse.c
> > > +++ b/tools/xl/xl_parse.c
> > > @@ -1208,7 +1208,6 @@ void parse_config_data(const char *config_source,
> > >  }
> > >  
> > >  xlu_cfg_get_defbool(config, "nestedhvm", _info->nested_hvm, 0);
> > > -xlu_cfg_get_defbool(config, "apic", _info->apic, 0);
> > 
> > Is this hunk not in a path also used by HVM ?
> 
> Yes, this hunk is moved to a HVM-specific section a little bit below:
> 
> @@ -1243,6 +1242,7 @@ void parse_config_data(const char *config_source,
>  xlu_cfg_get_defbool(config, "nx", _info->u.hvm.nx, 0);
>  xlu_cfg_get_defbool(config, "hpet", _info->u.hvm.hpet, 0);
>  xlu_cfg_get_defbool(config, "vpt_align", _info->u.hvm.vpt_align, 
> 0);
> +xlu_cfg_get_defbool(config, "apic", _info->apic, 0);
> 
>  switch (xlu_cfg_get_list(config, "viridian",
>   , _viridian, 1))
> 
> So it's only set for HVM, which AFAICT should be fine.
> 

I tried to apply this patch to staging but it failed.

Wei.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Roger Pau Monné]

On Mon, Mar 05, 2018 at 11:36:58AM +, Ian Jackson wrote:
> Roger Pau Monne writes ("[PATCH] xl: remove apic option for PVH guests"):
> > XSA-256 forces the local APIC to always be enabled for PVH guests, so
> > ignore any apic option for PVH guests. Update the documentation
> > accordingly.
> ...
> > diff --git a/tools/xl/xl_parse.c b/tools/xl/xl_parse.c
> > index f6842540ca..8b999825d2 100644
> > --- a/tools/xl/xl_parse.c
> > +++ b/tools/xl/xl_parse.c
> > @@ -1208,7 +1208,6 @@ void parse_config_data(const char *config_source,
> >  }
> >  
> >  xlu_cfg_get_defbool(config, "nestedhvm", _info->nested_hvm, 0);
> > -xlu_cfg_get_defbool(config, "apic", _info->apic, 0);
> 
> Is this hunk not in a path also used by HVM ?

Yes, this hunk is moved to a HVM-specific section a little bit below:

@@ -1243,6 +1242,7 @@ void parse_config_data(const char *config_source,
 xlu_cfg_get_defbool(config, "nx", _info->u.hvm.nx, 0);
 xlu_cfg_get_defbool(config, "hpet", _info->u.hvm.hpet, 0);
 xlu_cfg_get_defbool(config, "vpt_align", _info->u.hvm.vpt_align, 0);
+xlu_cfg_get_defbool(config, "apic", _info->apic, 0);

 switch (xlu_cfg_get_list(config, "viridian",
  , _viridian, 1))

So it's only set for HVM, which AFAICT should be fine.

Roger.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Wei Liu]

On Tue, Mar 13, 2018 at 10:31:20AM +, Roger Pau Monné wrote:
> On Mon, Mar 05, 2018 at 08:43:48AM -0600, Doug Goldstein wrote:
> > On 3/2/18 5:29 AM, Jan Beulich wrote:
> >  On 02.03.18 at 12:09,  wrote:
> > >> On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
> > >>> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
> >  On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> > > On 01/03/18 12:22, Wei Liu wrote:
> > >> On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> > >>> XSA-256 forces the local APIC to always be enabled for PVH guests, 
> > >>> so
> > >>> ignore any apic option for PVH guests. Update the documentation
> > >>> accordingly.
> > >> I think how I will approach this is to dictate that PVH always has 
> > >> LAPIC
> > >> in our in-tree document, then use that as the justification for this
> > >> change. That's the consensus from 2 years ago, right?
> > >>
> > >> Or we're just working around the limitation in our code base, and 
> > >> users
> > >> may demand a no-LAPIC PVH guest just because...
> > >
> > > Currently, Xen enforces that HVM guests have an LAPIC.  This is 
> > > because
> > > making the non-LAPIC case function correctly/safely devolved into a
> > > massive rats nest and I stopped trying to fix it after 2 days of 
> > > trying.
> > >
> > > At the moment, it would be wise to discuss whether the non-LAPIC case 
> > > is
> > > actually sensible.  I personally see no value in keeping it.
> > >
> > 
> >  +1
> > 
> > > If someone can come up with a convincing usecase for keeping it, then
> > > ok, but the barrier for this is increasing all the time, especially 
> > > now
> > > that hardware acceleration and posted interrupts means that a
> > > pipeline-virtualised APIC is faster and more efficient than any of our
> > > event channel mechanisms.
> > 
> >  +1
> > >>>
> > >>> I've looked at the in-tree pvh document and it just refers to the local
> > >>> APIC in this sentence:
> > >>>
> > >>> "AP startup can be performed using hypercalls or the local APIC if 
> > >>> present."
> > >>>
> > >>> I guess the trailing "if present" could be removed, but it's not
> > >>> colliding with this patch.
> > >>>
> > >>> I'm happy with rebasing this patch and applying the above change, is
> > >>> there any other document that should be changed?
> > >>
> > >> Can we make it more explicit. Like
> > >>
> > >>   VCPUs for PVH must have local APIC and it can't be disabled.
> > >>
> > >> ?
> > > 
> > > To be honest I liker Roger's suggestion better. And yet better
> > > would imo be if we left that sentence alone, unless we really mean
> > > to close that road for anyone wanting to take on making APIC-
> > > less guests work securely.
> > > 
> > > Jan
> > 
> > I believe that's exactly what Andrew proposed in
> > https://lists.xenproject.org/archives/html/xen-devel/2018-03/msg00089.html
> > removing the wording doesn't exclude someone from adding it later but it
> > does make it clear that its not available today.
> 
> I'm kind of lost regarding whether we reached consensus or not. Is the
> current patch suitable, or should I change some of the wording?

I'm fine with the wording of the doc for now.

Ian had a question on HVM path that is yet to be answered.

Wei.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Roger Pau Monné]

On Mon, Mar 05, 2018 at 08:43:48AM -0600, Doug Goldstein wrote:
> On 3/2/18 5:29 AM, Jan Beulich wrote:
>  On 02.03.18 at 12:09,  wrote:
> >> On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
> >>> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
>  On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> > On 01/03/18 12:22, Wei Liu wrote:
> >> On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> >>> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> >>> ignore any apic option for PVH guests. Update the documentation
> >>> accordingly.
> >> I think how I will approach this is to dictate that PVH always has 
> >> LAPIC
> >> in our in-tree document, then use that as the justification for this
> >> change. That's the consensus from 2 years ago, right?
> >>
> >> Or we're just working around the limitation in our code base, and users
> >> may demand a no-LAPIC PVH guest just because...
> >
> > Currently, Xen enforces that HVM guests have an LAPIC.  This is because
> > making the non-LAPIC case function correctly/safely devolved into a
> > massive rats nest and I stopped trying to fix it after 2 days of trying.
> >
> > At the moment, it would be wise to discuss whether the non-LAPIC case is
> > actually sensible.  I personally see no value in keeping it.
> >
> 
>  +1
> 
> > If someone can come up with a convincing usecase for keeping it, then
> > ok, but the barrier for this is increasing all the time, especially now
> > that hardware acceleration and posted interrupts means that a
> > pipeline-virtualised APIC is faster and more efficient than any of our
> > event channel mechanisms.
> 
>  +1
> >>>
> >>> I've looked at the in-tree pvh document and it just refers to the local
> >>> APIC in this sentence:
> >>>
> >>> "AP startup can be performed using hypercalls or the local APIC if 
> >>> present."
> >>>
> >>> I guess the trailing "if present" could be removed, but it's not
> >>> colliding with this patch.
> >>>
> >>> I'm happy with rebasing this patch and applying the above change, is
> >>> there any other document that should be changed?
> >>
> >> Can we make it more explicit. Like
> >>
> >>   VCPUs for PVH must have local APIC and it can't be disabled.
> >>
> >> ?
> > 
> > To be honest I liker Roger's suggestion better. And yet better
> > would imo be if we left that sentence alone, unless we really mean
> > to close that road for anyone wanting to take on making APIC-
> > less guests work securely.
> > 
> > Jan
> 
> I believe that's exactly what Andrew proposed in
> https://lists.xenproject.org/archives/html/xen-devel/2018-03/msg00089.html
> removing the wording doesn't exclude someone from adding it later but it
> does make it clear that its not available today.

I'm kind of lost regarding whether we reached consensus or not. Is the
current patch suitable, or should I change some of the wording?

Thanks, Roger.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Doug Goldstein]

On 3/2/18 5:29 AM, Jan Beulich wrote:
 On 02.03.18 at 12:09,  wrote:
>> On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
>>> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
 On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> On 01/03/18 12:22, Wei Liu wrote:
>> On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
>>> XSA-256 forces the local APIC to always be enabled for PVH guests, so
>>> ignore any apic option for PVH guests. Update the documentation
>>> accordingly.
>> I think how I will approach this is to dictate that PVH always has LAPIC
>> in our in-tree document, then use that as the justification for this
>> change. That's the consensus from 2 years ago, right?
>>
>> Or we're just working around the limitation in our code base, and users
>> may demand a no-LAPIC PVH guest just because...
>
> Currently, Xen enforces that HVM guests have an LAPIC.  This is because
> making the non-LAPIC case function correctly/safely devolved into a
> massive rats nest and I stopped trying to fix it after 2 days of trying.
>
> At the moment, it would be wise to discuss whether the non-LAPIC case is
> actually sensible.  I personally see no value in keeping it.
>

 +1

> If someone can come up with a convincing usecase for keeping it, then
> ok, but the barrier for this is increasing all the time, especially now
> that hardware acceleration and posted interrupts means that a
> pipeline-virtualised APIC is faster and more efficient than any of our
> event channel mechanisms.

 +1
>>>
>>> I've looked at the in-tree pvh document and it just refers to the local
>>> APIC in this sentence:
>>>
>>> "AP startup can be performed using hypercalls or the local APIC if present."
>>>
>>> I guess the trailing "if present" could be removed, but it's not
>>> colliding with this patch.
>>>
>>> I'm happy with rebasing this patch and applying the above change, is
>>> there any other document that should be changed?
>>
>> Can we make it more explicit. Like
>>
>>   VCPUs for PVH must have local APIC and it can't be disabled.
>>
>> ?
> 
> To be honest I liker Roger's suggestion better. And yet better
> would imo be if we left that sentence alone, unless we really mean
> to close that road for anyone wanting to take on making APIC-
> less guests work securely.
> 
> Jan

I believe that's exactly what Andrew proposed in
https://lists.xenproject.org/archives/html/xen-devel/2018-03/msg00089.html
removing the wording doesn't exclude someone from adding it later but it
does make it clear that its not available today.

-- 
Doug Goldstein



signature.asc
Description: OpenPGP digital signature
___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Ian Jackson]

Roger Pau Monne writes ("[PATCH] xl: remove apic option for PVH guests"):
> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> ignore any apic option for PVH guests. Update the documentation
> accordingly.
...
> diff --git a/tools/xl/xl_parse.c b/tools/xl/xl_parse.c
> index f6842540ca..8b999825d2 100644
> --- a/tools/xl/xl_parse.c
> +++ b/tools/xl/xl_parse.c
> @@ -1208,7 +1208,6 @@ void parse_config_data(const char *config_source,
>  }
>  
>  xlu_cfg_get_defbool(config, "nestedhvm", _info->nested_hvm, 0);
> -xlu_cfg_get_defbool(config, "apic", _info->apic, 0);

Is this hunk not in a path also used by HVM ?

Ian.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Andrew Cooper]

On 02/03/18 11:09, Wei Liu wrote:
> On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
>> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
>>> On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
 On 01/03/18 12:22, Wei Liu wrote:
> On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
>> XSA-256 forces the local APIC to always be enabled for PVH guests, so
>> ignore any apic option for PVH guests. Update the documentation
>> accordingly.
> I think how I will approach this is to dictate that PVH always has LAPIC
> in our in-tree document, then use that as the justification for this
> change. That's the consensus from 2 years ago, right?
>
> Or we're just working around the limitation in our code base, and users
> may demand a no-LAPIC PVH guest just because...
 Currently, Xen enforces that HVM guests have an LAPIC.  This is because
 making the non-LAPIC case function correctly/safely devolved into a
 massive rats nest and I stopped trying to fix it after 2 days of trying.

 At the moment, it would be wise to discuss whether the non-LAPIC case is
 actually sensible.  I personally see no value in keeping it.

>>> +1
>>>
 If someone can come up with a convincing usecase for keeping it, then
 ok, but the barrier for this is increasing all the time, especially now
 that hardware acceleration and posted interrupts means that a
 pipeline-virtualised APIC is faster and more efficient than any of our
 event channel mechanisms.
>>> +1
>> I've looked at the in-tree pvh document and it just refers to the local
>> APIC in this sentence:
>>
>> "AP startup can be performed using hypercalls or the local APIC if present."
>>
>> I guess the trailing "if present" could be removed, but it's not
>> colliding with this patch.
>>
>> I'm happy with rebasing this patch and applying the above change, is
>> there any other document that should be changed?
> Can we make it more explicit. Like
>
>   VCPUs for PVH must have local APIC and it can't be disabled.

-1 to this.  When an APIC is available to the guest, there is soft
disable and hard disable as part of the state model.  Saying this will
only confuse the matter.

~Andrew

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Jan Beulich]

>>> On 02.03.18 at 12:09,  wrote:
> On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
>> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
>> > On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
>> > > On 01/03/18 12:22, Wei Liu wrote:
>> > > > On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
>> > > >> XSA-256 forces the local APIC to always be enabled for PVH guests, so
>> > > >> ignore any apic option for PVH guests. Update the documentation
>> > > >> accordingly.
>> > > > I think how I will approach this is to dictate that PVH always has 
>> > > > LAPIC
>> > > > in our in-tree document, then use that as the justification for this
>> > > > change. That's the consensus from 2 years ago, right?
>> > > >
>> > > > Or we're just working around the limitation in our code base, and users
>> > > > may demand a no-LAPIC PVH guest just because...
>> > > 
>> > > Currently, Xen enforces that HVM guests have an LAPIC.  This is because
>> > > making the non-LAPIC case function correctly/safely devolved into a
>> > > massive rats nest and I stopped trying to fix it after 2 days of trying.
>> > > 
>> > > At the moment, it would be wise to discuss whether the non-LAPIC case is
>> > > actually sensible.  I personally see no value in keeping it.
>> > > 
>> > 
>> > +1
>> > 
>> > > If someone can come up with a convincing usecase for keeping it, then
>> > > ok, but the barrier for this is increasing all the time, especially now
>> > > that hardware acceleration and posted interrupts means that a
>> > > pipeline-virtualised APIC is faster and more efficient than any of our
>> > > event channel mechanisms.
>> > 
>> > +1
>> 
>> I've looked at the in-tree pvh document and it just refers to the local
>> APIC in this sentence:
>> 
>> "AP startup can be performed using hypercalls or the local APIC if present."
>> 
>> I guess the trailing "if present" could be removed, but it's not
>> colliding with this patch.
>> 
>> I'm happy with rebasing this patch and applying the above change, is
>> there any other document that should be changed?
> 
> Can we make it more explicit. Like
> 
>   VCPUs for PVH must have local APIC and it can't be disabled.
> 
> ?

To be honest I liker Roger's suggestion better. And yet better
would imo be if we left that sentence alone, unless we really mean
to close that road for anyone wanting to take on making APIC-
less guests work securely.

Jan

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Wei Liu]

On Thu, Mar 01, 2018 at 05:01:55PM +, Roger Pau Monné wrote:
> On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
> > On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> > > On 01/03/18 12:22, Wei Liu wrote:
> > > > On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> > > >> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> > > >> ignore any apic option for PVH guests. Update the documentation
> > > >> accordingly.
> > > > I think how I will approach this is to dictate that PVH always has LAPIC
> > > > in our in-tree document, then use that as the justification for this
> > > > change. That's the consensus from 2 years ago, right?
> > > >
> > > > Or we're just working around the limitation in our code base, and users
> > > > may demand a no-LAPIC PVH guest just because...
> > > 
> > > Currently, Xen enforces that HVM guests have an LAPIC.  This is because
> > > making the non-LAPIC case function correctly/safely devolved into a
> > > massive rats nest and I stopped trying to fix it after 2 days of trying.
> > > 
> > > At the moment, it would be wise to discuss whether the non-LAPIC case is
> > > actually sensible.  I personally see no value in keeping it.
> > > 
> > 
> > +1
> > 
> > > If someone can come up with a convincing usecase for keeping it, then
> > > ok, but the barrier for this is increasing all the time, especially now
> > > that hardware acceleration and posted interrupts means that a
> > > pipeline-virtualised APIC is faster and more efficient than any of our
> > > event channel mechanisms.
> > 
> > +1
> 
> I've looked at the in-tree pvh document and it just refers to the local
> APIC in this sentence:
> 
> "AP startup can be performed using hypercalls or the local APIC if present."
> 
> I guess the trailing "if present" could be removed, but it's not
> colliding with this patch.
> 
> I'm happy with rebasing this patch and applying the above change, is
> there any other document that should be changed?

Can we make it more explicit. Like

  VCPUs for PVH must have local APIC and it can't be disabled.

?

(CC Jan as well)

Wei.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Roger Pau Monné]

On Thu, Mar 01, 2018 at 04:01:23PM +, Wei Liu wrote:
> On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> > On 01/03/18 12:22, Wei Liu wrote:
> > > On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> > >> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> > >> ignore any apic option for PVH guests. Update the documentation
> > >> accordingly.
> > > I think how I will approach this is to dictate that PVH always has LAPIC
> > > in our in-tree document, then use that as the justification for this
> > > change. That's the consensus from 2 years ago, right?
> > >
> > > Or we're just working around the limitation in our code base, and users
> > > may demand a no-LAPIC PVH guest just because...
> > 
> > Currently, Xen enforces that HVM guests have an LAPIC.  This is because
> > making the non-LAPIC case function correctly/safely devolved into a
> > massive rats nest and I stopped trying to fix it after 2 days of trying.
> > 
> > At the moment, it would be wise to discuss whether the non-LAPIC case is
> > actually sensible.  I personally see no value in keeping it.
> > 
> 
> +1
> 
> > If someone can come up with a convincing usecase for keeping it, then
> > ok, but the barrier for this is increasing all the time, especially now
> > that hardware acceleration and posted interrupts means that a
> > pipeline-virtualised APIC is faster and more efficient than any of our
> > event channel mechanisms.
> 
> +1

I've looked at the in-tree pvh document and it just refers to the local
APIC in this sentence:

"AP startup can be performed using hypercalls or the local APIC if present."

I guess the trailing "if present" could be removed, but it's not
colliding with this patch.

I'm happy with rebasing this patch and applying the above change, is
there any other document that should be changed?

Thanks, Roger.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Wei Liu]

On Thu, Mar 01, 2018 at 03:57:18PM +, Andrew Cooper wrote:
> On 01/03/18 12:22, Wei Liu wrote:
> > On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> >> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> >> ignore any apic option for PVH guests. Update the documentation
> >> accordingly.
> > I think how I will approach this is to dictate that PVH always has LAPIC
> > in our in-tree document, then use that as the justification for this
> > change. That's the consensus from 2 years ago, right?
> >
> > Or we're just working around the limitation in our code base, and users
> > may demand a no-LAPIC PVH guest just because...
> 
> Currently, Xen enforces that HVM guests have an LAPIC.  This is because
> making the non-LAPIC case function correctly/safely devolved into a
> massive rats nest and I stopped trying to fix it after 2 days of trying.
> 
> At the moment, it would be wise to discuss whether the non-LAPIC case is
> actually sensible.  I personally see no value in keeping it.
> 

+1

> If someone can come up with a convincing usecase for keeping it, then
> ok, but the barrier for this is increasing all the time, especially now
> that hardware acceleration and posted interrupts means that a
> pipeline-virtualised APIC is faster and more efficient than any of our
> event channel mechanisms.

+1

Wei.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Andrew Cooper]

On 01/03/18 12:22, Wei Liu wrote:
> On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
>> XSA-256 forces the local APIC to always be enabled for PVH guests, so
>> ignore any apic option for PVH guests. Update the documentation
>> accordingly.
> I think how I will approach this is to dictate that PVH always has LAPIC
> in our in-tree document, then use that as the justification for this
> change. That's the consensus from 2 years ago, right?
>
> Or we're just working around the limitation in our code base, and users
> may demand a no-LAPIC PVH guest just because...

Currently, Xen enforces that HVM guests have an LAPIC.  This is because
making the non-LAPIC case function correctly/safely devolved into a
massive rats nest and I stopped trying to fix it after 2 days of trying.

At the moment, it would be wise to discuss whether the non-LAPIC case is
actually sensible.  I personally see no value in keeping it.

If someone can come up with a convincing usecase for keeping it, then
ok, but the barrier for this is increasing all the time, especially now
that hardware acceleration and posted interrupts means that a
pipeline-virtualised APIC is faster and more efficient than any of our
event channel mechanisms.

~Andrew

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Wei Liu]

On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> ignore any apic option for PVH guests. Update the documentation
> accordingly.

I think how I will approach this is to dictate that PVH always has LAPIC
in our in-tree document, then use that as the justification for this
change. That's the consensus from 2 years ago, right?

Or we're just working around the limitation in our code base, and users
may demand a no-LAPIC PVH guest just because...

Wei.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Re: [Xen-devel] [PATCH] xl: remove apic option for PVH guests

[Roger Pau Monné]

On Wed, Feb 28, 2018 at 10:20:53AM +, Roger Pau Monne wrote:
> XSA-256 forces the local APIC to always be enabled for PVH guests, so
> ignore any apic option for PVH guests. Update the documentation
> accordingly.
> 
> Signed-off-by: Roger Pau Monné 
> ---
> Cc: Ian Jackson 
> Cc: Wei Liu 

I've wrongly based this on top of my vpci series, so it's likely not
going to apply correctly on top of staging. I will wait for review
before resending, so that I can also apply any comments.

Thanks, Roger.

___
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel