diff mbox series

[v4,6/7] x86/hyperv: retrieve vp_index from Hyper-V

Message ID 20200122202343.5703-7-liuwe@microsoft.com (mailing list archive)
State New, archived
Headers show
Series More Hyper-V infrastructure | expand

Commit Message

Wei Liu Jan. 22, 2020, 8:23 p.m. UTC 
This will be useful when invoking hypercall that targets specific
vcpu(s).

Signed-off-by: Wei Liu <liuwe@microsoft.com>
Reviewed-by: Paul Durrant <paul@xen.org>
---
v4:
1. Use private.h
2. Add Paul's review tag

v2:
1. Fold into setup_pcpu_arg function
---
 xen/arch/x86/guest/hyperv/hyperv.c  | 5 +++++
 xen/arch/x86/guest/hyperv/private.h | 1 +
 2 files changed, 6 insertions(+)

Comments

Jan Beulich Jan. 23, 2020, 3:48 p.m. UTC | #1 
On 22.01.2020 21:23, Wei Liu wrote:
> This will be useful when invoking hypercall that targets specific
> vcpu(s).
> 
> Signed-off-by: Wei Liu <liuwe@microsoft.com>
> Reviewed-by: Paul Durrant <paul@xen.org>

For formal reasons
Acked-by: Jan Beulich <jbeulich@suse.com>

However I wonder whether the Viridian entry in MAINTAINERS shouldn't
be extended by

F:	xen/arch/x86/guest/hyperv/

(and possibly have its title adjusted). Thoughts?

Jan
Wei Liu Jan. 28, 2020, 3:55 p.m. UTC | #2 
On Thu, Jan 23, 2020 at 04:48:38PM +0100, Jan Beulich wrote:
> On 22.01.2020 21:23, Wei Liu wrote:
> > This will be useful when invoking hypercall that targets specific
> > vcpu(s).
> > 
> > Signed-off-by: Wei Liu <liuwe@microsoft.com>
> > Reviewed-by: Paul Durrant <paul@xen.org>
> 
> For formal reasons
> Acked-by: Jan Beulich <jbeulich@suse.com>
> 
> However I wonder whether the Viridian entry in MAINTAINERS shouldn't
> be extended by
> 
> F:	xen/arch/x86/guest/hyperv/
> 
> (and possibly have its title adjusted). Thoughts?

This isn't about emulating Hyper-V inside Xen, so I don't think that's
the right approach here.

That said, if Paul wants to take this under his purview, it's fine by me
-- that would make me easier to upstream my patch. ;-)  I also wouldn't
mind adding myself as maintainer for this path.

Wei.

> 
> Jan
Jan Beulich Jan. 28, 2020, 4:18 p.m. UTC | #3 
On 28.01.2020 16:55, Wei Liu wrote:
> On Thu, Jan 23, 2020 at 04:48:38PM +0100, Jan Beulich wrote:
>> On 22.01.2020 21:23, Wei Liu wrote:
>>> This will be useful when invoking hypercall that targets specific
>>> vcpu(s).
>>>
>>> Signed-off-by: Wei Liu <liuwe@microsoft.com>
>>> Reviewed-by: Paul Durrant <paul@xen.org>
>>
>> For formal reasons
>> Acked-by: Jan Beulich <jbeulich@suse.com>
>>
>> However I wonder whether the Viridian entry in MAINTAINERS shouldn't
>> be extended by
>>
>> F:	xen/arch/x86/guest/hyperv/
>>
>> (and possibly have its title adjusted). Thoughts?
> 
> This isn't about emulating Hyper-V inside Xen, so I don't think that's
> the right approach here.

Well, it's the code producing the interface in one case, and
consuming it here. So there is some overlap at least.

> That said, if Paul wants to take this under his purview, it's fine by me
> -- that would make me easier to upstream my patch. ;-)  I also wouldn't
> mind adding myself as maintainer for this path.

Perhaps best both of you? Paul, Andrew, what do you think?

Jan
Durrant, Paul Jan. 28, 2020, 4:33 p.m. UTC | #4 
> -----Original Message-----
> From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of Jan
> Beulich
> Sent: 28 January 2020 16:19
> To: Wei Liu <wl@xen.org>; Paul Durrant <paul@xen.org>; Andrew Cooper
> <andrew.cooper3@citrix.com>
> Cc: Xen Development List <xen-devel@lists.xenproject.org>; Roger Pau Monné
> <roger.pau@citrix.com>; Wei Liu <liuwe@microsoft.com>; Michael Kelley
> <mikelley@microsoft.com>
> Subject: Re: [Xen-devel] [PATCH v4 6/7] x86/hyperv: retrieve vp_index from
> Hyper-V
> 
> On 28.01.2020 16:55, Wei Liu wrote:
> > On Thu, Jan 23, 2020 at 04:48:38PM +0100, Jan Beulich wrote:
> >> On 22.01.2020 21:23, Wei Liu wrote:
> >>> This will be useful when invoking hypercall that targets specific
> >>> vcpu(s).
> >>>
> >>> Signed-off-by: Wei Liu <liuwe@microsoft.com>
> >>> Reviewed-by: Paul Durrant <paul@xen.org>
> >>
> >> For formal reasons
> >> Acked-by: Jan Beulich <jbeulich@suse.com>
> >>
> >> However I wonder whether the Viridian entry in MAINTAINERS shouldn't
> >> be extended by
> >>
> >> F:	xen/arch/x86/guest/hyperv/
> >>
> >> (and possibly have its title adjusted). Thoughts?
> >
> > This isn't about emulating Hyper-V inside Xen, so I don't think that's
> > the right approach here.
> 
> Well, it's the code producing the interface in one case, and
> consuming it here. So there is some overlap at least.
> 
> > That said, if Paul wants to take this under his purview, it's fine by me
> > -- that would make me easier to upstream my patch. ;-)  I also wouldn't
> > mind adding myself as maintainer for this path.
> 
> Perhaps best both of you? Paul, Andrew, what do you think?
> 

IMO it's probably best to the put the Hyper-V stuff under 'Viridian' and add yourself as a maintainer there. There really is likely to be significant overlap and it'd make it easier (for me at least) to keep track of the bigger picture (i.e. Xen using enlightenments as well as implementing them).

  Cheers,

    Paul
Wei Liu Jan. 28, 2020, 4:53 p.m. UTC | #5 
On Tue, Jan 28, 2020 at 04:33:00PM +0000, Durrant, Paul wrote:
> > -----Original Message-----
> > From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of Jan
> > Beulich
> > Sent: 28 January 2020 16:19
> > To: Wei Liu <wl@xen.org>; Paul Durrant <paul@xen.org>; Andrew Cooper
> > <andrew.cooper3@citrix.com>
> > Cc: Xen Development List <xen-devel@lists.xenproject.org>; Roger Pau Monné
> > <roger.pau@citrix.com>; Wei Liu <liuwe@microsoft.com>; Michael Kelley
> > <mikelley@microsoft.com>
> > Subject: Re: [Xen-devel] [PATCH v4 6/7] x86/hyperv: retrieve vp_index from
> > Hyper-V
> > 
> > On 28.01.2020 16:55, Wei Liu wrote:
> > > On Thu, Jan 23, 2020 at 04:48:38PM +0100, Jan Beulich wrote:
> > >> On 22.01.2020 21:23, Wei Liu wrote:
> > >>> This will be useful when invoking hypercall that targets specific
> > >>> vcpu(s).
> > >>>
> > >>> Signed-off-by: Wei Liu <liuwe@microsoft.com>
> > >>> Reviewed-by: Paul Durrant <paul@xen.org>
> > >>
> > >> For formal reasons
> > >> Acked-by: Jan Beulich <jbeulich@suse.com>
> > >>
> > >> However I wonder whether the Viridian entry in MAINTAINERS shouldn't
> > >> be extended by
> > >>
> > >> F:	xen/arch/x86/guest/hyperv/
> > >>
> > >> (and possibly have its title adjusted). Thoughts?
> > >
> > > This isn't about emulating Hyper-V inside Xen, so I don't think that's
> > > the right approach here.
> > 
> > Well, it's the code producing the interface in one case, and
> > consuming it here. So there is some overlap at least.
> > 
> > > That said, if Paul wants to take this under his purview, it's fine by me
> > > -- that would make me easier to upstream my patch. ;-)  I also wouldn't
> > > mind adding myself as maintainer for this path.
> > 
> > Perhaps best both of you? Paul, Andrew, what do you think?
> > 
> 
> IMO it's probably best to the put the Hyper-V stuff under 'Viridian'
> and add yourself as a maintainer there. There really is likely to be
> significant overlap and it'd make it easier (for me at least) to keep
> track of the bigger picture (i.e. Xen using enlightenments as well as
> implementing them).

When you said "yourself", did you mean me or Jan?

Wei.

> 
>   Cheers,
> 
>     Paul
Durrant, Paul Jan. 28, 2020, 5:01 p.m. UTC | #6 
> -----Original Message-----
> From: Wei Liu <wl@xen.org>
> Sent: 28 January 2020 16:54
> To: Durrant, Paul <pdurrant@amazon.co.uk>
> Cc: Jan Beulich <jbeulich@suse.com>; Wei Liu <wl@xen.org>; Paul Durrant
> <paul@xen.org>; Andrew Cooper <andrew.cooper3@citrix.com>; Xen Development
> List <xen-devel@lists.xenproject.org>; Roger Pau Monné
> <roger.pau@citrix.com>; Wei Liu <liuwe@microsoft.com>; Michael Kelley
> <mikelley@microsoft.com>
> Subject: Re: [Xen-devel] [PATCH v4 6/7] x86/hyperv: retrieve vp_index from
> Hyper-V
> 
> On Tue, Jan 28, 2020 at 04:33:00PM +0000, Durrant, Paul wrote:
> > > -----Original Message-----
> > > From: Xen-devel <xen-devel-bounces@lists.xenproject.org> On Behalf Of
> Jan
> > > Beulich
> > > Sent: 28 January 2020 16:19
> > > To: Wei Liu <wl@xen.org>; Paul Durrant <paul@xen.org>; Andrew Cooper
> > > <andrew.cooper3@citrix.com>
> > > Cc: Xen Development List <xen-devel@lists.xenproject.org>; Roger Pau
> Monné
> > > <roger.pau@citrix.com>; Wei Liu <liuwe@microsoft.com>; Michael Kelley
> > > <mikelley@microsoft.com>
> > > Subject: Re: [Xen-devel] [PATCH v4 6/7] x86/hyperv: retrieve vp_index
> from
> > > Hyper-V
> > >
> > > On 28.01.2020 16:55, Wei Liu wrote:
> > > > On Thu, Jan 23, 2020 at 04:48:38PM +0100, Jan Beulich wrote:
> > > >> On 22.01.2020 21:23, Wei Liu wrote:
> > > >>> This will be useful when invoking hypercall that targets specific
> > > >>> vcpu(s).
> > > >>>
> > > >>> Signed-off-by: Wei Liu <liuwe@microsoft.com>
> > > >>> Reviewed-by: Paul Durrant <paul@xen.org>
> > > >>
> > > >> For formal reasons
> > > >> Acked-by: Jan Beulich <jbeulich@suse.com>
> > > >>
> > > >> However I wonder whether the Viridian entry in MAINTAINERS
> shouldn't
> > > >> be extended by
> > > >>
> > > >> F:	xen/arch/x86/guest/hyperv/
> > > >>
> > > >> (and possibly have its title adjusted). Thoughts?
> > > >
> > > > This isn't about emulating Hyper-V inside Xen, so I don't think
> that's
> > > > the right approach here.
> > >
> > > Well, it's the code producing the interface in one case, and
> > > consuming it here. So there is some overlap at least.
> > >
> > > > That said, if Paul wants to take this under his purview, it's fine
> by me
> > > > -- that would make me easier to upstream my patch. ;-)  I also
> wouldn't
> > > > mind adding myself as maintainer for this path.
> > >
> > > Perhaps best both of you? Paul, Andrew, what do you think?
> > >
> >
> > IMO it's probably best to the put the Hyper-V stuff under 'Viridian'
> > and add yourself as a maintainer there. There really is likely to be
> > significant overlap and it'd make it easier (for me at least) to keep
> > track of the bigger picture (i.e. Xen using enlightenments as well as
> > implementing them).
> 
> When you said "yourself", did you mean me or Jan?
>

You. I thought was replying to your question. Apologies for any confusion.

  Paul
 
> Wei.
> 
> >
> >   Cheers,
> >
> >     Paul
diff mbox series

Patch

diff --git a/xen/arch/x86/guest/hyperv/hyperv.c b/xen/arch/x86/guest/hyperv/hyperv.c
index c5195af948..085e646dc6 100644
--- a/xen/arch/x86/guest/hyperv/hyperv.c
+++ b/xen/arch/x86/guest/hyperv/hyperv.c
@@ -31,6 +31,7 @@ 
 
 struct ms_hyperv_info __read_mostly ms_hyperv;
 DEFINE_PER_CPU_READ_MOSTLY(void *, hv_pcpu_input_arg);
+DEFINE_PER_CPU_READ_MOSTLY(unsigned int, hv_vp_index);
 
 static uint64_t generate_guest_id(void)
 {
@@ -125,6 +126,7 @@  static void __init setup_hypercall_page(void)
 static void setup_hypercall_pcpu_arg(void)
 {
     void *mapping;
+    uint64_t vp_index_msr;
 
     if ( this_cpu(hv_pcpu_input_arg) )
         return;
@@ -135,6 +137,9 @@  static void setup_hypercall_pcpu_arg(void)
               smp_processor_id());
 
     this_cpu(hv_pcpu_input_arg) = mapping;
+
+    rdmsrl(HV_X64_MSR_VP_INDEX, vp_index_msr);
+    this_cpu(hv_vp_index) = vp_index_msr;
 }
 
 static void __init setup(void)
diff --git a/xen/arch/x86/guest/hyperv/private.h b/xen/arch/x86/guest/hyperv/private.h
index b6902b5639..da70990401 100644
--- a/xen/arch/x86/guest/hyperv/private.h
+++ b/xen/arch/x86/guest/hyperv/private.h
@@ -25,5 +25,6 @@ 
 #include <xen/percpu.h>
 
 DECLARE_PER_CPU(void *, hv_pcpu_input_arg);
+DECLARE_PER_CPU(unsigned int, hv_vp_index);
 
 #endif /* __XEN_HYPERV_PRIVIATE_H__  */