Re: [PATCH v4 2/3] KVM: PPC: epapr: Add idle hcall support for host

On 17.02.2012, at 03:13, Liu Yu-B13201 wrote:

=20
=20

quoted

-----Original Message-----
From: Alexander Graf [mailto:agraf@suse.de]
Sent: Thursday, February 16, 2012 6:20 PM
To: Liu Yu-B13201
Cc: <redacted>; <redacted>; <linuxppc-
dev@ozlabs.org>; Wood Scott-B07421; Liu Yu-B13201
Subject: Re: [PATCH v4 2/3] KVM: PPC: epapr: Add idle hcall support =

for

quoted

host
=20
=20
=20
On 16.02.2012, at 10:26, Liu Yu [off-list ref] wrote:
=20

quoted

And add a new flag definition in kvm_ppc_pvinfo to indicate whether
host support EV_IDLE hcall.
=20
Signed-off-by: Liu Yu <redacted>
---
v4:
no change
=20
arch/powerpc/include/asm/kvm_para.h |   14 ++++++++++++--
arch/powerpc/kvm/powerpc.c          |    8 ++++++++
include/linux/kvm.h                 |    2 ++
3 files changed, 22 insertions(+), 2 deletions(-)
=20

diff --git a/arch/powerpc/include/asm/kvm_para.h

b/arch/powerpc/include/asm/kvm_para.h
index 7b754e7..81a34c9 100644

--- a/arch/powerpc/include/asm/kvm_para.h
+++ b/arch/powerpc/include/asm/kvm_para.h

@@ -75,9 +75,19 @@ struct kvm_vcpu_arch_shared { };

=20
#define KVM_SC_MAGIC_R0        0x4b564d21 /* "KVM!" */
-#define HC_VENDOR_KVM        (42 << 16)
+
+#include <asm/epapr_hcalls.h>
+
+/* ePAPR Hypercall Vendor ID */
+#define HC_VENDOR_EPAPR        (EV_EPAPR_VENDOR_ID << 16)
+#define HC_VENDOR_KVM        (EV_KVM_VENDOR_ID << 16)
+
+/* ePAPR Hypercall Token */
+#define HC_EV_IDLE        EV_IDLE
+
+/* ePAPR Hypercall Return Codes */
#define HC_EV_SUCCESS        0
-#define HC_EV_UNIMPLEMENTED    12
+#define HC_EV_UNIMPLEMENTED    EV_UNIMPLEMENTED
=20
#define KVM_FEATURE_MAGIC_PAGE    1
=20

diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c
index 0e21d15..03ebd5d 100644
--- a/arch/powerpc/kvm/powerpc.c
+++ b/arch/powerpc/kvm/powerpc.c

@@ -81,6 +81,10 @@ int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)

=20
      /* Second return value is in r4 */
      break;
+    case HC_VENDOR_EPAPR | HC_EV_IDLE:
+        r =3D HC_EV_SUCCESS;
+        kvm_vcpu_block(vcpu);
+        break;
  default:
      r =3D HC_EV_UNIMPLEMENTED;
      break;

@@ -746,6 +750,10 @@ static int kvm_vm_ioctl_get_pvinfo(struct

kvm_ppc_pvinfo *pvinfo)

quoted

  pvinfo->hcall[2] =3D inst_sc;
  pvinfo->hcall[3] =3D inst_nop;
=20
+#ifdef CONFIG_BOOKE
+    pvinfo->flags |=3D KVM_PPC_PVINFO_FLAGS_EV_IDLE; #endif
+
  return 0;
}>=20

=20
Why limit it to booke? The less ifdefs our code has, the better :)

=20
The code here tells userspace that kvm support ev_idle.
But I'm not sure if the ev_idle code works for other platforms.
=20
So I think we should keep the ifdef until other platform test the code =

:)

But the implementation is in generic code and is not #ifdef'ed, so a =
guest could still call it just fine. It looks simple enough to work =
without major testing on different platforms, so I'd say just expose it =
and be done :)


Alex