[Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Daniel P. Berrange]

This enables report of a signature in CPUID for the TCG
interpretor.

Changed in v4:

 - Report 0x40000001 in EAX for 0x40000000 index (Eduardo)
 - Report all zeros for 0x40000001 index (Eduardo)
 - Make code style consistent when checking limits (Eduardo)

Changed in v3:

 - Simplify CPU limit code still further (Eduardo)

Changed in v2:

 - Rewrite the way we bounds check / cap the CPUID index
   to use a flat switch, instead of nested ifs (Eduardo)
 - Add a 'tcg-cpuid' property to allow it to be hidden
   (Eduardo)
 - Hide the TCG signature for old machine types
 - Force code to a no-op if tcg_enabled() is false (Eduardo)


NB, I did not introduce a general 'hypervisor-cpuid' property
to obsolete the existing 'kvm=off|on' -cpu property, since it
appears impossible to get the back compat semantics right,
as described in a previous reply.

Daniel P. Berrange (2):
  i386: rewrite way CPUID index is validated
  i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf

 include/hw/i386/pc.h |  5 ++++
 target/i386/cpu.c    | 65 +++++++++++++++++++++++++++++++++++-----------------
 target/i386/cpu.h    |  1 +
 3 files changed, 50 insertions(+), 21 deletions(-)

-- 
2.9.3

[Qemu-devel] [PATCH v4 1/2] i386: rewrite way CPUID index is validated

[Daniel P. Berrange]

Change the nested if statements into a flat format, to make
it clearer what validation / capping is being performed on
different CPUID index values.

NB this changes behaviour when "index > env->cpuid_xlevel2".
This won't have any guest-visible effect because no there is
no CPUID[0xC0000001] feature supported by TCG, and KVM code
will never call cpu_x86_cpuid() with such an index value.

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
---
 target/i386/cpu.c | 35 +++++++++++++++--------------------
 1 file changed, 15 insertions(+), 20 deletions(-)

diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 13c0985..8cb4af4 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -2626,28 +2626,23 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
     X86CPU *cpu = x86_env_get_cpu(env);
     CPUState *cs = CPU(cpu);
     uint32_t pkg_offset;
+    uint32_t limit;
 
-    /* test if maximum index reached */
-    if (index & 0x80000000) {
-        if (index > env->cpuid_xlevel) {
-            if (env->cpuid_xlevel2 > 0) {
-                /* Handle the Centaur's CPUID instruction. */
-                if (index > env->cpuid_xlevel2) {
-                    index = env->cpuid_xlevel2;
-                } else if (index < 0xC0000000) {
-                    index = env->cpuid_xlevel;
-                }
-            } else {
-                /* Intel documentation states that invalid EAX input will
-                 * return the same information as EAX=cpuid_level
-                 * (Intel SDM Vol. 2A - Instruction Set Reference - CPUID)
-                 */
-                index =  env->cpuid_level;
-            }
-        }
+    /* Calculate & apply limits for different index ranges */
+    if (index >= 0xC0000000) {
+        limit = env->cpuid_xlevel2;
+    } else if (index >= 0x80000000) {
+        limit = env->cpuid_xlevel;
     } else {
-        if (index > env->cpuid_level)
-            index = env->cpuid_level;
+        limit = env->cpuid_level;
+    }
+
+    if (index > limit) {
+        /* Intel documentation states that invalid EAX input will
+         * return the same information as EAX=cpuid_level
+         * (Intel SDM Vol. 2A - Instruction Set Reference - CPUID)
+         */
+        index = env->cpuid_level;
     }
 
     switch(index) {
-- 
2.9.3

Re: [Qemu-devel] [PATCH v4 1/2] i386: rewrite way CPUID index is validated

[Kashyap Chamarthy]

On Tue, May 09, 2017 at 02:27:35PM +0100, Daniel P. Berrange wrote:
> Change the nested if statements into a flat format, to make
> it clearer what validation / capping is being performed on
> different CPUID index values.
> 
> NB this changes behaviour when "index > env->cpuid_xlevel2".
> This won't have any guest-visible effect because no there is
> no CPUID[0xC0000001]

Nit: When applying, maybe the maintainer could fix the typo:

"because no there is no" -> "because there is no"

> feature supported by TCG, and KVM code
> will never call cpu_x86_cpuid() with such an index value.
> 
> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
> ---
>  target/i386/cpu.c | 35 +++++++++++++++--------------------
>  1 file changed, 15 insertions(+), 20 deletions(-)
> 

[...]

-- 
/kashyap

Re: [Qemu-devel] [PATCH v4 1/2] i386: rewrite way CPUID index is validated

[Eduardo Habkost]

On Fri, May 19, 2017 at 04:58:23PM +0200, Kashyap Chamarthy wrote:
> On Tue, May 09, 2017 at 02:27:35PM +0100, Daniel P. Berrange wrote:
> > Change the nested if statements into a flat format, to make
> > it clearer what validation / capping is being performed on
> > different CPUID index values.
> > 
> > NB this changes behaviour when "index > env->cpuid_xlevel2".
> > This won't have any guest-visible effect because no there is
> > no CPUID[0xC0000001]
> 
> Nit: When applying, maybe the maintainer could fix the typo:
> 
> "because no there is no" -> "because there is no"

This was already merged, unfortunately.  Thanks for noting that,
anyway.

-- 
Eduardo

[Qemu-devel] [PATCH v4 2/2] i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf

[Daniel P. Berrange]

Currently when running KVM, we expose "KVMKVMKVM\0\0\0" in
the 0x40000000 CPUID leaf. Other hypervisors (VMWare,
HyperV, Xen, BHyve) all do the same thing, which leaves
TCG as the odd one out.

The CPUID signature is used by software to detect which
virtual environment they are running in and (potentially)
change behaviour in certain ways. For example, systemd
supports a ConditionVirtualization= setting in unit files.
The virt-what command can also report the virt type it is
running on

Currently both these apps have to resort to custom hacks
like looking for 'fw-cfg' entry in the /proc/device-tree
file to identify TCG.

This change thus proposes a signature "TCGTCGTCGTCG" to be
reported when running under TCG.

To hide this, the -cpu option tcg-cpuid=off can be used.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
---
 include/hw/i386/pc.h |  5 +++++
 target/i386/cpu.c    | 28 ++++++++++++++++++++++++++++
 target/i386/cpu.h    |  1 +
 3 files changed, 34 insertions(+)

diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
index f278b3a..3aec60f 100644
--- a/include/hw/i386/pc.h
+++ b/include/hw/i386/pc.h
@@ -376,6 +376,11 @@ bool e820_get_entry(int, uint32_t, uint64_t *, uint64_t *);
 #define PC_COMPAT_2_8 \
     HW_COMPAT_2_8 \
     {\
+        .driver   = TYPE_X86_CPU,\
+        .property = "tcg-cpuid",\
+        .value    = "off",\
+    },\
+    {\
         .driver   = "kvmclock",\
         .property = "x-mach-use-reliable-get-clock",\
         .value    = "off",\
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 8cb4af4..edf451c 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -2627,12 +2627,15 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
     CPUState *cs = CPU(cpu);
     uint32_t pkg_offset;
     uint32_t limit;
+    uint32_t signature[3];
 
     /* Calculate & apply limits for different index ranges */
     if (index >= 0xC0000000) {
         limit = env->cpuid_xlevel2;
     } else if (index >= 0x80000000) {
         limit = env->cpuid_xlevel;
+    } else if (index >= 0x40000000) {
+        limit = 0x40000001;
     } else {
         limit = env->cpuid_level;
     }
@@ -2867,6 +2870,30 @@ void cpu_x86_cpuid(CPUX86State *env, uint32_t index, uint32_t count,
         }
         break;
     }
+    case 0x40000000:
+        /*
+         * CPUID code in kvm_arch_init_vcpu() ignores stuff
+         * set here, but we restrict to TCG none the less.
+         */
+        if (tcg_enabled() && cpu->expose_tcg) {
+            memcpy(signature, "TCGTCGTCGTCG", 12);
+            *eax = 0x40000001;
+            *ebx = signature[0];
+            *ecx = signature[1];
+            *edx = signature[2];
+        } else {
+            *eax = 0;
+            *ebx = 0;
+            *ecx = 0;
+            *edx = 0;
+        }
+        break;
+    case 0x40000001:
+        *eax = 0;
+        *ebx = 0;
+        *ecx = 0;
+        *edx = 0;
+        break;
     case 0x80000000:
         *eax = env->cpuid_xlevel;
         *ebx = env->cpuid_vendor1;
@@ -4008,6 +4035,7 @@ static Property x86_cpu_properties[] = {
     DEFINE_PROP_BOOL("kvm-no-smi-migration", X86CPU, kvm_no_smi_migration,
                      false),
     DEFINE_PROP_BOOL("vmware-cpuid-freq", X86CPU, vmware_cpuid_freq, true),
+    DEFINE_PROP_BOOL("tcg-cpuid", X86CPU, expose_tcg, true),
     DEFINE_PROP_END_OF_LIST()
 };
 
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index c4602ca..c25f0ce 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -1216,6 +1216,7 @@ struct X86CPU {
     bool check_cpuid;
     bool enforce_cpuid;
     bool expose_kvm;
+    bool expose_tcg;
     bool migratable;
     bool max_features; /* Enable all supported features automatically */
     uint32_t apic_id;
-- 
2.9.3

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Richard Henderson]

On 05/09/2017 06:27 AM, Daniel P. Berrange wrote:
> This enables report of a signature in CPUID for the TCG
> interpretor.
> 
> Changed in v4:
> 
>   - Report 0x40000001 in EAX for 0x40000000 index (Eduardo)
>   - Report all zeros for 0x40000001 index (Eduardo)
>   - Make code style consistent when checking limits (Eduardo)
> 
> Changed in v3:
> 
>   - Simplify CPU limit code still further (Eduardo)
> 
> Changed in v2:
> 
>   - Rewrite the way we bounds check / cap the CPUID index
>     to use a flat switch, instead of nested ifs (Eduardo)
>   - Add a 'tcg-cpuid' property to allow it to be hidden
>     (Eduardo)
>   - Hide the TCG signature for old machine types
>   - Force code to a no-op if tcg_enabled() is false (Eduardo)
> 
> 
> NB, I did not introduce a general 'hypervisor-cpuid' property
> to obsolete the existing 'kvm=off|on' -cpu property, since it
> appears impossible to get the back compat semantics right,
> as described in a previous reply.
> 
> Daniel P. Berrange (2):
>    i386: rewrite way CPUID index is validated
>    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf

I probably should have commented earlier but...  what's the point?

If you want the guest os to actually do anything with this, what do you gain 
for advertising TCG over KVM?


r~

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Daniel P. Berrange]

On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> On 05/09/2017 06:27 AM, Daniel P. Berrange wrote:
> > This enables report of a signature in CPUID for the TCG
> > interpretor.
> > 
> > Changed in v4:
> > 
> >   - Report 0x40000001 in EAX for 0x40000000 index (Eduardo)
> >   - Report all zeros for 0x40000001 index (Eduardo)
> >   - Make code style consistent when checking limits (Eduardo)
> > 
> > Changed in v3:
> > 
> >   - Simplify CPU limit code still further (Eduardo)
> > 
> > Changed in v2:
> > 
> >   - Rewrite the way we bounds check / cap the CPUID index
> >     to use a flat switch, instead of nested ifs (Eduardo)
> >   - Add a 'tcg-cpuid' property to allow it to be hidden
> >     (Eduardo)
> >   - Hide the TCG signature for old machine types
> >   - Force code to a no-op if tcg_enabled() is false (Eduardo)
> > 
> > 
> > NB, I did not introduce a general 'hypervisor-cpuid' property
> > to obsolete the existing 'kvm=off|on' -cpu property, since it
> > appears impossible to get the back compat semantics right,
> > as described in a previous reply.
> > 
> > Daniel P. Berrange (2):
> >    i386: rewrite way CPUID index is validated
> >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> 
> I probably should have commented earlier but...  what's the point?
> 
> If you want the guest os to actually do anything with this, what do
> you gain for advertising TCG over KVM?

TCG doesn't expose any of the paravirt features that KVM does, so from that
POV reporting "KVMKVMKVM" when running TCG would be wrong. As mentioned in
the 2nd patch too, every hypervisor out there reports a distinct CPUID
string, so again I think TCG should have its own identity, as it is
different CPU emulation other users of QEMU (Xen, KVM, $whatever Windows
and OS-X CPU accelerators are).

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 :|

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Richard W.M. Jones]

On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> On 05/09/2017 06:27 AM, Daniel P. Berrange wrote:
> >This enables report of a signature in CPUID for the TCG
> >interpretor.
> >
> >Changed in v4:
> >
> >  - Report 0x40000001 in EAX for 0x40000000 index (Eduardo)
> >  - Report all zeros for 0x40000001 index (Eduardo)
> >  - Make code style consistent when checking limits (Eduardo)
> >
> >Changed in v3:
> >
> >  - Simplify CPU limit code still further (Eduardo)
> >
> >Changed in v2:
> >
> >  - Rewrite the way we bounds check / cap the CPUID index
> >    to use a flat switch, instead of nested ifs (Eduardo)
> >  - Add a 'tcg-cpuid' property to allow it to be hidden
> >    (Eduardo)
> >  - Hide the TCG signature for old machine types
> >  - Force code to a no-op if tcg_enabled() is false (Eduardo)
> >
> >
> >NB, I did not introduce a general 'hypervisor-cpuid' property
> >to obsolete the existing 'kvm=off|on' -cpu property, since it
> >appears impossible to get the back compat semantics right,
> >as described in a previous reply.
> >
> >Daniel P. Berrange (2):
> >   i386: rewrite way CPUID index is validated
> >   i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> 
> I probably should have commented earlier but...  what's the point?
> 
> If you want the guest os to actually do anything with this, what do
> you gain for advertising TCG over KVM?

I can see this being useful from virt-what, since it would allow
vendors to diagnose problems being caused by TCG (since as Dan
mentions in the other reply and as you are already well acquainted
with, TCG and KVM don't present or emulate quite the same thing).

That said, I don't think it's anything more than a nice to have.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine.  Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Richard Henderson]

On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
>>> Daniel P. Berrange (2):
>>>    i386: rewrite way CPUID index is validated
>>>    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
>>
>> I probably should have commented earlier but...  what's the point?
>>
>> If you want the guest os to actually do anything with this, what do
>> you gain for advertising TCG over KVM?
> 
> I can see this being useful from virt-what, since it would allow
> vendors to diagnose problems being caused by TCG ...

Thanks, that's an excellent point.


r~

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Eduardo Habkost]

On Tue, May 09, 2017 at 07:18:07AM -0700, Richard Henderson wrote:
> On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> > On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > > > Daniel P. Berrange (2):
> > > >    i386: rewrite way CPUID index is validated
> > > >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > > 
> > > I probably should have commented earlier but...  what's the point?
> > > 
> > > If you want the guest os to actually do anything with this, what do
> > > you gain for advertising TCG over KVM?
> > 
> > I can see this being useful from virt-what, since it would allow
> > vendors to diagnose problems being caused by TCG ...
> 
> Thanks, that's an excellent point.

Richard, Paolo: I would like to merge this. Do you agree?

-- 
Eduardo

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Richard W.M. Jones]

On Fri, May 19, 2017 at 04:41:46PM -0300, Eduardo Habkost wrote:
> On Tue, May 09, 2017 at 07:18:07AM -0700, Richard Henderson wrote:
> > On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> > > On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > > > > Daniel P. Berrange (2):
> > > > >    i386: rewrite way CPUID index is validated
> > > > >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > > > 
> > > > I probably should have commented earlier but...  what's the point?
> > > > 
> > > > If you want the guest os to actually do anything with this, what do
> > > > you gain for advertising TCG over KVM?
> > > 
> > > I can see this being useful from virt-what, since it would allow
> > > vendors to diagnose problems being caused by TCG ...
> > 
> > Thanks, that's an excellent point.
> 
> Richard, Paolo: I would like to merge this. Do you agree?

It's fine by me but note that I didn't review the patch.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines.  Supports shell scripting,
bindings from many languages.  http://libguestfs.org

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Richard W.M. Jones]

On Fri, May 19, 2017 at 04:41:46PM -0300, Eduardo Habkost wrote:
> On Tue, May 09, 2017 at 07:18:07AM -0700, Richard Henderson wrote:
> > On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> > > On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > > > > Daniel P. Berrange (2):
> > > > >    i386: rewrite way CPUID index is validated
> > > > >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > > > 
> > > > I probably should have commented earlier but...  what's the point?
> > > > 
> > > > If you want the guest os to actually do anything with this, what do
> > > > you gain for advertising TCG over KVM?
> > > 
> > > I can see this being useful from virt-what, since it would allow
> > > vendors to diagnose problems being caused by TCG ...
> > 
> > Thanks, that's an excellent point.
> 
> Richard, Paolo: I would like to merge this. Do you agree?

Ah, other Richard, OK!

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Daniel P. Berrange]

On Fri, May 19, 2017 at 04:41:46PM -0300, Eduardo Habkost wrote:
> On Tue, May 09, 2017 at 07:18:07AM -0700, Richard Henderson wrote:
> > On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> > > On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > > > > Daniel P. Berrange (2):
> > > > >    i386: rewrite way CPUID index is validated
> > > > >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > > > 
> > > > I probably should have commented earlier but...  what's the point?
> > > > 
> > > > If you want the guest os to actually do anything with this, what do
> > > > you gain for advertising TCG over KVM?
> > > 
> > > I can see this being useful from virt-what, since it would allow
> > > vendors to diagnose problems being caused by TCG ...
> > 
> > Thanks, that's an excellent point.
> 
> Richard, Paolo: I would like to merge this. Do you agree?

ping, since there's no objections can we merge this now ?


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 :|

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Eduardo Habkost]

On Wed, Jun 14, 2017 at 11:18:46AM +0100, Daniel P. Berrange wrote:
> On Fri, May 19, 2017 at 04:41:46PM -0300, Eduardo Habkost wrote:
> > On Tue, May 09, 2017 at 07:18:07AM -0700, Richard Henderson wrote:
> > > On 05/09/2017 07:13 AM, Richard W.M. Jones wrote:
> > > > On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > > > > > Daniel P. Berrange (2):
> > > > > >    i386: rewrite way CPUID index is validated
> > > > > >    i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > > > > 
> > > > > I probably should have commented earlier but...  what's the point?
> > > > > 
> > > > > If you want the guest os to actually do anything with this, what do
> > > > > you gain for advertising TCG over KVM?
> > > > 
> > > > I can see this being useful from virt-what, since it would allow
> > > > vendors to diagnose problems being caused by TCG ...
> > > 
> > > Thanks, that's an excellent point.
> > 
> > Richard, Paolo: I would like to merge this. Do you agree?
> 
> ping, since there's no objections can we merge this now ?

Queued on x86-next.  Thanks!

-- 
Eduardo

Re: [Qemu-devel] [PATCH v4 0/2] Support CPUID signature for TCG

[Eduardo Habkost]

On Tue, May 09, 2017 at 03:13:37PM +0100, Richard W.M. Jones wrote:
> On Tue, May 09, 2017 at 07:05:51AM -0700, Richard Henderson wrote:
> > On 05/09/2017 06:27 AM, Daniel P. Berrange wrote:
> > >This enables report of a signature in CPUID for the TCG
> > >interpretor.
> > >
> > >Changed in v4:
> > >
> > >  - Report 0x40000001 in EAX for 0x40000000 index (Eduardo)
> > >  - Report all zeros for 0x40000001 index (Eduardo)
> > >  - Make code style consistent when checking limits (Eduardo)
> > >
> > >Changed in v3:
> > >
> > >  - Simplify CPU limit code still further (Eduardo)
> > >
> > >Changed in v2:
> > >
> > >  - Rewrite the way we bounds check / cap the CPUID index
> > >    to use a flat switch, instead of nested ifs (Eduardo)
> > >  - Add a 'tcg-cpuid' property to allow it to be hidden
> > >    (Eduardo)
> > >  - Hide the TCG signature for old machine types
> > >  - Force code to a no-op if tcg_enabled() is false (Eduardo)
> > >
> > >
> > >NB, I did not introduce a general 'hypervisor-cpuid' property
> > >to obsolete the existing 'kvm=off|on' -cpu property, since it
> > >appears impossible to get the back compat semantics right,
> > >as described in a previous reply.
> > >
> > >Daniel P. Berrange (2):
> > >   i386: rewrite way CPUID index is validated
> > >   i386: expose "TCGTCGTCGTCG" in the 0x40000000 CPUID leaf
> > 
> > I probably should have commented earlier but...  what's the point?
> > 
> > If you want the guest os to actually do anything with this, what do
> > you gain for advertising TCG over KVM?
> 
> I can see this being useful from virt-what, since it would allow
> vendors to diagnose problems being caused by TCG (since as Dan
> mentions in the other reply and as you are already well acquainted
> with, TCG and KVM don't present or emulate quite the same thing).
> 
> That said, I don't think it's anything more than a nice to have.

We also have systemd-detect-virt, and the ConditionVirtualization
option in systemd.

-- 
Eduardo