Skip to contentSkip to conversation

Port Xen to Hyper-V

PATCH Superseded Wei Liu 4 participants 27 messages Nov 21, 2019 – Nov 30, 2019

v4 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Wei Liu

author
#0 lore raw

v4 cover Port Xen to Hyper-V

Please see individual patches for changes.

Cc Jan Beulich <[email protected]>
Cc Andrew Cooper <[email protected]>
Cc Wei Liu <[email protected]>
Cc Roger Pau Monné <[email protected]>
Cc Paul Durrant <[email protected]>

Wei Liu (8):
x86: introduce CONFIG_GUEST and move code
x86: add missing headers in hypercall.h
x86: drop hypervisor_cpuid_base
x86: introduce hypervisor framework
x86: rename hypervisor_{alloc,free}_unused_page
x86: switch xen guest implementation to use hypervisor framework
x86: be more verbose when running on a hypervisor
x86: introduce CONFIG_HYPERV and detection code

 xen/arch/x86/Kconfig                          | 15 +++-
 xen/arch/x86/Makefile                         |  2 +-
 xen/arch/x86/guest/Makefile                   |  6 +-
 xen/arch/x86/guest/hyperv/Makefile            |  1 +
 xen/arch/x86/guest/hyperv/hyperv.c            | 54 +++++++++++++
 xen/arch/x86/guest/hypervisor.c               | 80 +++++++++++++++++++
 xen/arch/x86/guest/xen/Makefile               |  4 +
 xen/arch/x86/guest/{ => xen}/hypercall_page.S |  0
 xen/arch/x86/guest/{ => xen}/pvh-boot.c       |  2 +-
 xen/arch/x86/guest/{ => xen}/xen.c            | 60 +++++++-------
 xen/arch/x86/pv/shim.c                        |  6 +-
 xen/arch/x86/setup.c                          |  7 +-
 xen/include/asm-x86/guest.h                   |  2 +
 xen/include/asm-x86/guest/hypercall.h         |  4 +
 xen/include/asm-x86/guest/hyperv.h            | 43 ++++++++++
 xen/include/asm-x86/guest/hypervisor.h        | 62 ++++++++++++++
 xen/include/asm-x86/guest/xen.h               | 23 ++----
 17 files changed, 314 insertions(+), 57 deletions(-)
 create mode 100644 xen/arch/x86/guest/hyperv/Makefile
 create mode 100644 xen/arch/x86/guest/hyperv/hyperv.c
 create mode 100644 xen/arch/x86/guest/hypervisor.c
 create mode 100644 xen/arch/x86/guest/xen/Makefile
 rename xen/arch/x86/guest/{ => xen}/hypercall_page.S (100%)
 rename xen/arch/x86/guest/{ => xen}/pvh-boot.c (99%)
 rename xen/arch/x86/guest/{ => xen}/xen.c (94%)
 create mode 100644 xen/include/asm-x86/guest/hyperv.h
 create mode 100644 xen/include/asm-x86/guest/hypervisor.h
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author
#1 lore raw

PATCH v4 1/8 x86: introduce CONFIG_GUEST and move code

Xen is able to run as a guest on Xen. We plan to make it able to run on Hyper-V as well.

Introduce CONFIG_GUEST which is set to true if either running on Xen or Hyper-V is desired. Restructure code hierarchy for new code to come.

No functional change intended.

---
 xen/arch/x86/Kconfig                          | 4 ++++
 xen/arch/x86/Makefile                         | 2 +-
 xen/arch/x86/guest/Makefile                   | 5 +----
 xen/arch/x86/guest/xen/Makefile               | 4 ++++
 xen/arch/x86/guest/{ => xen}/hypercall_page.S | 0
 xen/arch/x86/guest/{ => xen}/pvh-boot.c       | 0
 xen/arch/x86/guest/{ => xen}/xen.c            | 0
 7 files changed, 10 insertions(+), 5 deletions(-)
 create mode 100644 xen/arch/x86/guest/xen/Makefile
 rename xen/arch/x86/guest/{ => xen}/hypercall_page.S (100%)
 rename xen/arch/x86/guest/{ => xen}/pvh-boot.c (100%)
 rename xen/arch/x86/guest/{ => xen}/xen.c (100%)
xen/arch/x86/Kconfig +4 -0
index 28b3b4692a..867de857e8 100644
@@ -161,8 +161,12 @@ config XEN_ALIGN_2M
 
 endchoice
 
+config GUEST
+	bool
+
 config XEN_GUEST
 	def_bool n
+	select GUEST
 	prompt "Xen Guest"
 	---help---
 	  Support for Xen detecting when it is running under Xen.
xen/arch/x86/Makefile +1 -1
index a0b2f4ab15..7da5a2631e 100644
@@ -1,7 +1,7 @@
 subdir-y += acpi
 subdir-y += cpu
 subdir-y += genapic
-subdir-$(CONFIG_XEN_GUEST) += guest
+subdir-$(CONFIG_GUEST) += guest
 subdir-$(CONFIG_HVM) += hvm
 subdir-y += mm
 subdir-$(CONFIG_XENOPROF) += oprofile
xen/arch/x86/guest/Makefile +1 -4
index 26fb4b1007..6806f04947 100644
@@ -1,4 +1 @@
-obj-y += hypercall_page.o
-obj-y += xen.o
-
-obj-bin-$(CONFIG_PVH_GUEST) += pvh-boot.init.o
+subdir-$(CONFIG_XEN_GUEST) += xen
xen/arch/x86/guest/xen/Makefile new file +4 -0
new file mode 100644
index 0000000000..26fb4b1007
@@ -0,0 +1,4 @@
+obj-y += hypercall_page.o
+obj-y += xen.o
+
+obj-bin-$(CONFIG_PVH_GUEST) += pvh-boot.init.o
xen/arch/x86/guest/xen/hypercall_page.S renamed from xen/arch/x86/guest/hypercall_page.S +0 -0
similarity index 100%
rename from xen/arch/x86/guest/hypercall_page.S
rename to xen/arch/x86/guest/xen/hypercall_page.S
xen/arch/x86/guest/xen/pvh-boot.c renamed from xen/arch/x86/guest/pvh-boot.c +0 -0
similarity index 100%
rename from xen/arch/x86/guest/pvh-boot.c
rename to xen/arch/x86/guest/xen/pvh-boot.c
xen/arch/x86/guest/xen/xen.c renamed from xen/arch/x86/guest/xen.c +0 -0
similarity index 100%
rename from xen/arch/x86/guest/xen.c
rename to xen/arch/x86/guest/xen/xen.c
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author
#2 lore raw

PATCH v4 2/8 x86: add missing headers in hypercall.h

In reply to Wei Liu

Include asm_defns.h because ASM_CALL_CONSTRAINT is defined there.

Include xen/lib.h because we need ASSERT_UNREACHABLE.

No functional change.

---
 xen/include/asm-x86/guest/hypercall.h | 4 ++++
 1 file changed, 4 insertions(+)
xen/include/asm-x86/guest/hypercall.h +4 -0
index d548816b30..d0d2f5022d 100644
@@ -23,6 +23,8 @@
 
 #include <xen/types.h>
 
+#include <asm/asm_defns.h>
+
 #include <public/xen.h>
 #include <public/sched.h>
 #include <public/hvm/hvm_op.h>
@@ -180,6 +182,8 @@ static inline long xen_hypercall_set_evtchn_upcall_vector(
 
 #else /* CONFIG_XEN_GUEST */
 
+#include <xen/lib.h>
+
 #include <public/sched.h>
 
 static inline void xen_hypercall_console_write(
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author
#3 lore raw

PATCH v4 3/8 x86: drop hypervisor_cpuid_base

In reply to Wei Liu

The only user is Xen specific code in PV shim. We can therefore export the variable directly.

Move __read_mostly to its standard place while at it.

---
 xen/arch/x86/guest/xen/xen.c    | 7 +------
 xen/arch/x86/pv/shim.c          | 2 +-
 xen/include/asm-x86/guest/xen.h | 2 +-
 3 files changed, 3 insertions(+), 8 deletions(-)
xen/arch/x86/guest/xen/xen.c +1 -6
index a329e7c886..1e37086518 100644
@@ -37,7 +37,7 @@
 
 bool __read_mostly xen_guest;
 
-static __read_mostly uint32_t xen_cpuid_base;
+uint32_t __read_mostly xen_cpuid_base;
 extern char hypercall_page[];
 static struct rangeset *mem;
 
@@ -297,11 +297,6 @@ int hypervisor_free_unused_page(mfn_t mfn)
     return rangeset_remove_range(mem, mfn_x(mfn), mfn_x(mfn));
 }
 
-uint32_t hypervisor_cpuid_base(void)
-{
-    return xen_cpuid_base;
-}
-
 static void ap_resume(void *unused)
 {
     map_vcpuinfo();
xen/arch/x86/pv/shim.c +1 -1
index 35bf3945ac..351da970ef 100644
@@ -995,7 +995,7 @@ domid_t get_initial_domain_id(void)
     if ( !pv_shim )
         return 0;
 
-    cpuid(hypervisor_cpuid_base() + 4, &eax, &ebx, &ecx, &edx);
+    cpuid(xen_cpuid_base + 4, &eax, &ebx, &ecx, &edx);
 
     return (eax & XEN_HVM_CPUID_DOMID_PRESENT) ? ecx : 1;
 }
xen/include/asm-x86/guest/xen.h +1 -1
index 7e04e4a7ab..b015ed1883 100644
@@ -30,13 +30,13 @@
 
 extern bool xen_guest;
 extern bool pv_console;
+extern uint32_t xen_cpuid_base;
 
 void probe_hypervisor(void);
 void hypervisor_setup(void);
 void hypervisor_ap_setup(void);
 int hypervisor_alloc_unused_page(mfn_t *mfn);
 int hypervisor_free_unused_page(mfn_t mfn);
-uint32_t hypervisor_cpuid_base(void);
 void hypervisor_resume(void);
 
 DECLARE_PER_CPU(unsigned int, vcpu_id);
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author
#4 lore raw

PATCH v4 4/8 x86: introduce hypervisor framework

In reply to Wei Liu

We will soon implement Hyper-V support for Xen. Add a framework for that.

This requires moving some of the hypervisor_* functions from xen.h to hypervisor.h.

---
Changes in v4:
1. Add ASSERT_UNREACHABLE to stubs.
2. Move __read_mostly.
3. Return hops directly.
4. Drop Paul's review tag.

---
 xen/arch/x86/guest/Makefile            |  2 +
 xen/arch/x86/guest/hypervisor.c        | 42 +++++++++++++++++
 xen/include/asm-x86/guest.h            |  1 +
 xen/include/asm-x86/guest/hypervisor.h | 62 ++++++++++++++++++++++++++
 xen/include/asm-x86/guest/xen.h        | 12 -----
 5 files changed, 107 insertions(+), 12 deletions(-)
 create mode 100644 xen/arch/x86/guest/hypervisor.c
 create mode 100644 xen/include/asm-x86/guest/hypervisor.h
xen/arch/x86/guest/Makefile +2 -0
index 6806f04947..f63d64bbee 100644
@@ -1 +1,3 @@
+obj-y += hypervisor.o
+
 subdir-$(CONFIG_XEN_GUEST) += xen
xen/arch/x86/guest/hypervisor.c new file +42 -0

42 lines not on this page · raw message

xen/include/asm-x86/guest.h +1 -0
index a38c6b5b3f..8e167165ae 100644
@@ -20,6 +20,7 @@
 #define __X86_GUEST_H__
 
 #include <asm/guest/hypercall.h>
+#include <asm/guest/hypervisor.h>
 #include <asm/guest/pvh-boot.h>
 #include <asm/guest/xen.h>
 #include <asm/pv/shim.h>
xen/include/asm-x86/guest/hypervisor.h new file +62 -0

62 lines not on this page · raw message

xen/include/asm-x86/guest/xen.h +0 -12
index b015ed1883..3145f75361 100644
@@ -33,11 +33,8 @@ extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
 void probe_hypervisor(void);
-void hypervisor_setup(void);
-void hypervisor_ap_setup(void);
 int hypervisor_alloc_unused_page(mfn_t *mfn);
 int hypervisor_free_unused_page(mfn_t mfn);
-void hypervisor_resume(void);
 
 DECLARE_PER_CPU(unsigned int, vcpu_id);
 DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
@@ -49,15 +46,6 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
 
 static inline void probe_hypervisor(void) {}
 
-static inline void hypervisor_setup(void)
-{
-    ASSERT_UNREACHABLE();
-}
-static inline void hypervisor_ap_setup(void)
-{
-    ASSERT_UNREACHABLE();
-}
-
 #endif /* CONFIG_XEN_GUEST */
 #endif /* __X86_GUEST_XEN_H__ */
 
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v4 4/8
#5 lore raw

Re: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

Quoted 10 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

Quoted 7 lines

We will soon implement Hyper-V support for Xen. Add a framework for that.

This requires moving some of the hypervisor_* functions from xen.h to hypervisor.h.

Quoted 195 lines

---
Changes in v4:
1. Add ASSERT_UNREACHABLE to stubs.
2. Move __read_mostly.
3. Return hops directly.
4. Drop Paul's review tag.

---
 xen/arch/x86/guest/Makefile            |  2 +
 xen/arch/x86/guest/hypervisor.c        | 42 +++++++++++++++++
 xen/include/asm-x86/guest.h            |  1 +
 xen/include/asm-x86/guest/hypervisor.h | 62 ++++++++++++++++++++++++++
 xen/include/asm-x86/guest/xen.h        | 12 -----
 5 files changed, 107 insertions(+), 12 deletions(-)
 create mode 100644 xen/arch/x86/guest/hypervisor.c
 create mode 100644 xen/include/asm-x86/guest/hypervisor.h
xen/arch/x86/guest/Makefile +2 -0
index 6806f04947..f63d64bbee 100644
@@ -1 +1,3 @@
+obj-y += hypervisor.o
+
 subdir-$(CONFIG_XEN_GUEST) += xen
xen/arch/x86/guest/hypervisor.c new file +1 -0
new file mode 100644
index 0000000000..103feba5d8
@@ -0,0 +1,42 @@
+/************************************************************************

******
+ * arch/x86/guest/hypervisor.c
+ *
+ * Support for detecting and running under a hypervisor.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2019 Microsoft.
+ */

xen/arch/x86/guest/hypervisor.c new file +22 -0
+
+#include <xen/types.h>
+
+#include <asm/cache.h>
+#include <asm/guest/hypervisor.h>
+
+static const struct hypervisor_ops __read_mostly *hops;
+
+const struct hypervisor_ops *hypervisor_probe(void)
+{
+    return hops;
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
xen/include/asm-x86/guest.h +1 -0
index a38c6b5b3f..8e167165ae 100644
@@ -20,6 +20,7 @@
 #define __X86_GUEST_H__
 
 #include <asm/guest/hypercall.h>
+#include <asm/guest/hypervisor.h>
 #include <asm/guest/pvh-boot.h>
 #include <asm/guest/xen.h>
 #include <asm/pv/shim.h>
xen/include/asm- renamed from xen/include/asm-x86/guest/hypervisor.h +0 -0

x86/guest/hypervisor.h
new file mode 100644
index 0000000000..2ab15a7108

xen/include/asm-x86/guest/hypervisor.h new file +1 -0
@@ -0,0 +1,62 @@
+/************************************************************************

******
+ * asm-x86/guest/hypervisor.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see
<http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2019 Microsoft.
+ */

xen/include/asm-x86/guest/hypervisor.h new file +28 -0
+
+#ifndef __X86_HYPERVISOR_H__
+#define __X86_HYPERVISOR_H__
+
+struct hypervisor_ops {
+    /* Name of the hypervisor */
+    const char *name;
+    /* Main setup routine */
+    void (*setup)(void);
+    /* AP setup */
+    void (*ap_setup)(void);
+    /* Resume from suspension */
+    void (*resume)(void);
+};
+
+#ifdef CONFIG_GUEST
+
+const struct hypervisor_ops *hypervisor_probe(void);
+void hypervisor_setup(void);
+void hypervisor_ap_setup(void);
+void hypervisor_resume(void);
+
+#else
+
+#include <xen/lib.h>
+#include <xen/types.h>
+
+static inline const struct hypervisor_ops *hypervisor_probe(void) {

return NULL; }

xen/include/asm-x86/guest/hypervisor.h new file +17 -0
+static inline void hypervisor_setup(void) { ASSERT_UNREACHABLE(); }
+static inline void hypervisor_ap_setup(void) { ASSERT_UNREACHABLE(); }
+static inline void hypervisor_resume(void) { ASSERT_UNREACHABLE(); }
+
+#endif  /* CONFIG_GUEST */
+
+#endif /* __X86_HYPERVISOR_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
xen/include/asm- renamed from xen/include/asm-x86/guest/xen.h +0 -0

x86/guest/xen.h
index b015ed1883..3145f75361 100644

xen/include/asm-x86/guest/xen.h +0 -13
@@ -33,11 +33,8 @@ extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
 void probe_hypervisor(void);
-void hypervisor_setup(void);
-void hypervisor_ap_setup(void);
 int hypervisor_alloc_unused_page(mfn_t *mfn);
 int hypervisor_free_unused_page(mfn_t mfn);
-void hypervisor_resume(void);
 
 DECLARE_PER_CPU(unsigned int, vcpu_id);
 DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
@@ -49,15 +46,6 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
 
 static inline void probe_hypervisor(void) {}
 
-static inline void hypervisor_setup(void)
-{
-    ASSERT_UNREACHABLE();
-}
-static inline void hypervisor_ap_setup(void)
-{
-    ASSERT_UNREACHABLE();
-}
-
 #endif /* CONFIG_XEN_GUEST */
 #endif /* __X86_GUEST_XEN_H__ */
 
--

2.20.1

Quoted 4 lines

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

Acked-by
#6 lore raw

Re: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

Quoted 13 lines · Paul
Quoted 13 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

We will soon implement Hyper-V support for Xen. Add a framework for that.

Quoted 6 lines
Quoted 4 lines

This requires moving some of the hypervisor_* functions from xen.h to hypervisor.h.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#7 lore raw

Re: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

In reply to Wei Liu
Quoted 22 lines · Wei Liu
xen/arch/x86/guest/hypervisor.c new file +19 -0
@@ -0,0 +1,42 @@
+/******************************************************************************
+ * arch/x86/guest/hypervisor.c
+ *
+ * Support for detecting and running under a hypervisor.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2019 Microsoft.

xen/arch/x86/guest/hypervisor.c line 29

Quoted 10 lines
xen/arch/x86/guest/hypervisor.c new file +10 -0
+ */
+
+#include <xen/types.h>
+
+#include <asm/cache.h>
+#include <asm/guest/hypervisor.h>
+
+static const struct hypervisor_ops __read_mostly *hops;
+
+const struct hypervisor_ops *hypervisor_probe(void)

Despite my earlier ack - __init?

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#8 lore raw

Re: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

In reply to Wei Liu

xen/arch/x86/guest/hypervisor.c line 27

Quoted 6 lines · Wei Liu
xen/arch/x86/guest/hypervisor.c new file +6 -0
+#include <xen/types.h>
+
+#include <asm/cache.h>
+#include <asm/guest/hypervisor.h>
+
+static const struct hypervisor_ops __read_mostly *hops;

Could I talk you into using just plain 'ops' here.  This is mostly plumbing and doesn't appear to grow significantly.  I don't think there is a risk of gaining ambiguity.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#9 lore raw

Re: [Xen-devel] [PATCH v4 4/8] x86: introduce hypervisor framework

Quoted 11 lines · Andrew Cooper

xen/arch/x86/guest/hypervisor.c line 27

Quoted 6 lines · Wei Liu
xen/arch/x86/guest/hypervisor.c new file +6 -0
+#include <xen/types.h>
+
+#include <asm/cache.h>
+#include <asm/guest/hypervisor.h>
+
+static const struct hypervisor_ops __read_mostly *hops;

Could I talk you into using just plain 'ops' here.  This is mostly plumbing and doesn't appear to grow significantly.  I don't think there is a risk of gaining ambiguity.

This is done.

Wei.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#10 lore raw

PATCH v4 5/8 x86: rename hypervisor_{alloc, free}_unused_page

In reply to Wei Liu

They are used in Xen code only.

No functional change.

---
Changes in v4:
1. Use xg_ prefix instead.
2. Drop Roger's review tag.

---
 xen/arch/x86/guest/xen/xen.c    | 6 +++---
 xen/arch/x86/pv/shim.c          | 4 ++--
 xen/include/asm-x86/guest/xen.h | 4 ++--
 3 files changed, 7 insertions(+), 7 deletions(-)
xen/arch/x86/guest/xen/xen.c +3 -3
index 1e37086518..0f5b5267c5 100644
@@ -93,7 +93,7 @@ static void map_shared_info(void)
     unsigned int i;
     unsigned long rc;
 
-    if ( hypervisor_alloc_unused_page(&mfn) )
+    if ( xg_alloc_unused_page(&mfn) )
         panic("unable to reserve shared info memory page\n");
 
     xatp.gpfn = mfn_x(mfn);
@@ -280,7 +280,7 @@ void hypervisor_ap_setup(void)
     init_evtchn();
 }
 
-int hypervisor_alloc_unused_page(mfn_t *mfn)
+int xg_alloc_unused_page(mfn_t *mfn)
 {
     unsigned long m;
     int rc;
@@ -292,7 +292,7 @@ int hypervisor_alloc_unused_page(mfn_t *mfn)
     return rc;
 }
 
-int hypervisor_free_unused_page(mfn_t mfn)
+int xg_free_unused_page(mfn_t mfn)
 {
     return rangeset_remove_range(mem, mfn_x(mfn), mfn_x(mfn));
 }
xen/arch/x86/pv/shim.c +2 -2
index 351da970ef..7a898fdbe5 100644
@@ -742,7 +742,7 @@ static long pv_shim_grant_table_op(unsigned int cmd,
                 };
                 mfn_t mfn;
 
-                rc = hypervisor_alloc_unused_page(&mfn);
+                rc = xg_alloc_unused_page(&mfn);
                 if ( rc )
                 {
                     gprintk(XENLOG_ERR,
@@ -754,7 +754,7 @@ static long pv_shim_grant_table_op(unsigned int cmd,
                 rc = xen_hypercall_memory_op(XENMEM_add_to_physmap, &xatp);
                 if ( rc )
                 {
-                    hypervisor_free_unused_page(mfn);
+                    xg_free_unused_page(mfn);
                     break;
                 }
 
xen/include/asm-x86/guest/xen.h +2 -2
index 3145f75361..01dc3ee6f6 100644
@@ -33,8 +33,8 @@ extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
 void probe_hypervisor(void);
-int hypervisor_alloc_unused_page(mfn_t *mfn);
-int hypervisor_free_unused_page(mfn_t mfn);
+int xg_alloc_unused_page(mfn_t *mfn);
+int xg_free_unused_page(mfn_t mfn);
 
 DECLARE_PER_CPU(unsigned int, vcpu_id);
 DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v4 5/8
#11 lore raw

Re: [Xen-devel] [PATCH v4 5/8] x86: rename hypervisor_{alloc, free}_unused_page

Quoted 11 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 5/8] x86: rename hypervisor_{alloc, free}_unused_page

Quoted 5 lines

They are used in Xen code only.

No functional change.

Quoted 84 lines

---
Changes in v4:
1. Use xg_ prefix instead.
2. Drop Roger's review tag.

---
 xen/arch/x86/guest/xen/xen.c    | 6 +++---
 xen/arch/x86/pv/shim.c          | 4 ++--
 xen/include/asm-x86/guest/xen.h | 4 ++--
 3 files changed, 7 insertions(+), 7 deletions(-)
xen/arch/x86/guest/xen/xen.c +3 -3
index 1e37086518..0f5b5267c5 100644
@@ -93,7 +93,7 @@ static void map_shared_info(void)
     unsigned int i;
     unsigned long rc;
 
-    if ( hypervisor_alloc_unused_page(&mfn) )
+    if ( xg_alloc_unused_page(&mfn) )
         panic("unable to reserve shared info memory page\n");
 
     xatp.gpfn = mfn_x(mfn);
@@ -280,7 +280,7 @@ void hypervisor_ap_setup(void)
     init_evtchn();
 }
 
-int hypervisor_alloc_unused_page(mfn_t *mfn)
+int xg_alloc_unused_page(mfn_t *mfn)
 {
     unsigned long m;
     int rc;
@@ -292,7 +292,7 @@ int hypervisor_alloc_unused_page(mfn_t *mfn)
     return rc;
 }
 
-int hypervisor_free_unused_page(mfn_t mfn)
+int xg_free_unused_page(mfn_t mfn)
 {
     return rangeset_remove_range(mem, mfn_x(mfn), mfn_x(mfn));
 }
xen/arch/x86/pv/shim.c +2 -2
index 351da970ef..7a898fdbe5 100644
@@ -742,7 +742,7 @@ static long pv_shim_grant_table_op(unsigned int cmd,
                 };
                 mfn_t mfn;
 
-                rc = hypervisor_alloc_unused_page(&mfn);
+                rc = xg_alloc_unused_page(&mfn);
                 if ( rc )
                 {
                     gprintk(XENLOG_ERR,
@@ -754,7 +754,7 @@ static long pv_shim_grant_table_op(unsigned int cmd,
                 rc = xen_hypercall_memory_op(XENMEM_add_to_physmap,
 &xatp);
                 if ( rc )
                 {
-                    hypervisor_free_unused_page(mfn);
+                    xg_free_unused_page(mfn);
                     break;
                 }
xen/include/asm- renamed from xen/include/asm-x86/guest/xen.h +0 -0

x86/guest/xen.h
index 3145f75361..01dc3ee6f6 100644

xen/include/asm-x86/guest/xen.h +2 -3
@@ -33,8 +33,8 @@ extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
 void probe_hypervisor(void);
-int hypervisor_alloc_unused_page(mfn_t *mfn);
-int hypervisor_free_unused_page(mfn_t mfn);
+int xg_alloc_unused_page(mfn_t *mfn);
+int xg_free_unused_page(mfn_t mfn);
 
 DECLARE_PER_CPU(unsigned int, vcpu_id);
 DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
--

2.20.1

Quoted 4 lines

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

Acked-by
#12 lore raw

Re: [Xen-devel] [PATCH v4 5/8] x86: rename hypervisor_{alloc, free}_unused_page

Quoted 11 lines · Paul
Quoted 11 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 5/8] x86: rename hypervisor_{alloc, free}_unused_page

Quoted 7 lines
Quoted 5 lines

They are used in Xen code only.

No functional change.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#13 lore raw

PATCH v4 6/8 x86: switch xen guest implementation to use hypervisor framework

In reply to Wei Liu

---
Changes in v4:
1. xen_probe returns address of ops directly.
2. Add __init to hypervisor_setup.
3. Drop Paul's review tag.

---
 xen/arch/x86/guest/hypervisor.c   | 32 ++++++++++++++++++++-
 xen/arch/x86/guest/xen/pvh-boot.c |  2 +-
 xen/arch/x86/guest/xen/xen.c      | 47 ++++++++++++++++++-------------
 xen/arch/x86/setup.c              |  2 +-
 xen/include/asm-x86/guest/xen.h   |  5 ++--
 5 files changed, 64 insertions(+), 24 deletions(-)
xen/arch/x86/guest/hypervisor.c +31 -1
index 103feba5d8..a067cacecb 100644
@@ -19,18 +19,48 @@
  * Copyright (c) 2019 Microsoft.
  */
 
+#include <xen/init.h>
 #include <xen/types.h>
 
 #include <asm/cache.h>
-#include <asm/guest/hypervisor.h>
+#include <asm/guest.h>
 
 static const struct hypervisor_ops __read_mostly *hops;
 
 const struct hypervisor_ops *hypervisor_probe(void)
 {
+    if ( hops )
+        goto out;
+
+    if ( !cpu_has_hypervisor )
+        goto out;
+
+    hops = xen_probe();
+    if ( hops )
+        goto out;
+
+ out:
     return hops;
 }
 
+void __init hypervisor_setup(void)
+{
+    if ( hops && hops->setup )
+        hops->setup();
+}
+
+void hypervisor_ap_setup(void)
+{
+    if ( hops && hops->ap_setup )
+        hops->ap_setup();
+}
+
+void hypervisor_resume(void)
+{
+    if ( hops && hops->resume )
+        hops->resume();
+}
+
 /*
  * Local variables:
  * mode: C
xen/arch/x86/guest/xen/pvh-boot.c +1 -1
index ca8e156f7d..498625eae0 100644
@@ -103,7 +103,7 @@ void __init pvh_init(multiboot_info_t **mbi, module_t **mod)
 {
     convert_pvh_info(mbi, mod);
 
-    probe_hypervisor();
+    hypervisor_probe();
     ASSERT(xen_guest);
 
     get_memory_map();
xen/arch/x86/guest/xen/xen.c +28 -19

77 lines not on this page · raw message

xen/arch/x86/setup.c +1 -1
index 00ee87bde5..19606d909b 100644
@@ -763,7 +763,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
      * allocing any xenheap structures wanted in lower memory. */
     kexec_early_calculations();
 
-    probe_hypervisor();
+    hypervisor_probe();
 
     parse_video_info();
 
xen/include/asm-x86/guest/xen.h +3 -2
index 01dc3ee6f6..db90b550a7 100644
@@ -23,6 +23,7 @@
 
 #include <asm/e820.h>
 #include <asm/fixmap.h>
+#include <asm/guest/hypervisor.h>
 
 #define XEN_shared_info ((struct shared_info *)fix_to_virt(FIX_XEN_SHARED_INFO))
 
@@ -32,7 +33,7 @@ extern bool xen_guest;
 extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
-void probe_hypervisor(void);
+const struct hypervisor_ops *xen_probe(void);
 int xg_alloc_unused_page(mfn_t *mfn);
 int xg_free_unused_page(mfn_t mfn);
 
@@ -44,7 +45,7 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
 #define xen_guest 0
 #define pv_console 0
 
-static inline void probe_hypervisor(void) {}
+static inline const struct hypervisor_ops *xen_probe(void) { return NULL; }
 
 #endif /* CONFIG_XEN_GUEST */
 #endif /* __X86_GUEST_XEN_H__ */
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v4 6/8
#14 lore raw

Re: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

Quoted 12 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

[snip]

Quoted 24 lines
xen/include/asm- renamed from xen/include/asm-x86/guest/xen.h +0 -0

x86/guest/xen.h
index 01dc3ee6f6..db90b550a7 100644

xen/include/asm-x86/guest/xen.h +2 -1
@@ -23,6 +23,7 @@
 
 #include <asm/e820.h>
 #include <asm/fixmap.h>
+#include <asm/guest/hypervisor.h>
 
 #define XEN_shared_info ((struct shared_info
 *)fix_to_virt(FIX_XEN_SHARED_INFO))
@@ -32,7 +33,7 @@ extern bool xen_guest;
 extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
-void probe_hypervisor(void);
+const struct hypervisor_ops *xen_probe(void);
 int xg_alloc_unused_page(mfn_t *mfn);
 int xg_free_unused_page(mfn_t mfn);
 

xen/include/asm-x86/guest/xen.h DECLARE_PER_CPU(struct vcpu_info *, vcpu_info); · line 46

Quoted 2 lines
xen/include/asm-x86/guest/xen.h +0 -0
@@ -44,7 +45,7 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
 #define xen_guest 0
 #define pv_console 0

Nit: These should be #defined to false rather than 0. The rest LGTM so with those fixed,

Quoted 11 lines
xen/include/asm-x86/guest/xen.h +1 -2
-static inline void probe_hypervisor(void) {}
+static inline const struct hypervisor_ops *xen_probe(void) { return NULL;
 }
 
 #endif /* CONFIG_XEN_GUEST */
 #endif /* __X86_GUEST_XEN_H__ */
--

2.20.1

Quoted 4 lines

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#15 lore raw

Re: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

Quoted 36 lines · Paul
Quoted 12 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

[snip]

Quoted 23 lines
xen/include/asm- renamed from xen/include/asm-x86/guest/xen.h +0 -0

x86/guest/xen.h
index 01dc3ee6f6..db90b550a7 100644

xen/include/asm-x86/guest/xen.h +2 -1
@@ -23,6 +23,7 @@
 
 #include <asm/e820.h>
 #include <asm/fixmap.h>
+#include <asm/guest/hypervisor.h>
 
 #define XEN_shared_info ((struct shared_info
 *)fix_to_virt(FIX_XEN_SHARED_INFO))
@@ -32,7 +33,7 @@ extern bool xen_guest;
 extern bool pv_console;
 extern uint32_t xen_cpuid_base;
 
-void probe_hypervisor(void);
+const struct hypervisor_ops *xen_probe(void);
 int xg_alloc_unused_page(mfn_t *mfn);
 int xg_free_unused_page(mfn_t mfn);
 
Quoted 5 lines

xen/include/asm-x86/guest/xen.h DECLARE_PER_CPU(struct vcpu_info *, vcpu_info); · line 46

Quoted 3 lines
xen/include/asm-x86/guest/xen.h +0 -0
@@ -44,7 +45,7 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
 #define xen_guest 0
 #define pv_console 0

Nit: These should be #defined to false rather than 0. The rest LGTM so with those fixed,

They were part of existing code so I didn't touch them.

Wei.

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#16 lore raw

Re: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

In reply to Wei Liu
Quoted 7 lines · Wei Liu
xen/arch/x86/guest/hypervisor.c +7 -0
+void __init hypervisor_setup(void)
+{
+    if ( hops && hops->setup )
+        hops->setup();
+}
+
+void hypervisor_ap_setup(void)

xen/arch/x86/guest/hypervisor.c line 62

Quoted 10 lines
xen/arch/x86/guest/hypervisor.c +10 -0
+{
+    if ( hops && hops->ap_setup )
+        hops->ap_setup();
+}
+
+void hypervisor_resume(void)
+{
+    if ( hops && hops->resume )
+        hops->resume();
+}

I assume these don't get executed frequently enough for alternatives indirect call patching to matter? Down the road, if any more frequently executed hooks should appear, we will want to switch to that model though. This has the added benefit of there then being only one runtime instance of struct hypervisor_ops; all per-hypervisor ones would become __initconstrel.

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 313

Quoted 5 lines
xen/arch/x86/guest/xen/xen.c +1 -0
@@ -326,6 +310,31 @@ void hypervisor_resume(void)
         pv_console_init();
 }
 
+static const struct hypervisor_ops xg_ops = {

With this new xg_ prefix, ...

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 317

Quoted 4 lines
xen/arch/x86/guest/xen/xen.c +4 -0
+    .name = "Xen",
+    .setup = xen_setup,
+    .ap_setup = xen_ap_setup,
+    .resume = xen_resume,

... I'd like to suggest to use it for these ones too (if they need to have a prefix in the first place, which I doubt, but which I know others view differently).

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 320

Quoted 1 line
xen/arch/x86/guest/xen/xen.c +1 -0
+const struct hypervisor_ops * __init xen_probe(void)

For this one the call is a little more difficult, but for consistency I think xg_ would be slightly better here as well.

Nit: Strictly speaking there's a stray blank after *. We normally have ones only ahead of *, but not after.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#17 lore raw

Re: [Xen-devel] [PATCH v4 6/8] x86: switch xen guest implementation to use hypervisor framework

Quoted 19 lines · Jan Beulich
Quoted 7 lines · Wei Liu
xen/arch/x86/guest/hypervisor.c +7 -0
+void __init hypervisor_setup(void)
+{
+    if ( hops && hops->setup )
+        hops->setup();
+}
+
+void hypervisor_ap_setup(void)

xen/arch/x86/guest/hypervisor.c line 62

Quoted 10 lines
xen/arch/x86/guest/hypervisor.c +10 -0
+{
+    if ( hops && hops->ap_setup )
+        hops->ap_setup();
+}
+
+void hypervisor_resume(void)
+{
+    if ( hops && hops->resume )
+        hops->resume();
+}
Quoted 8 lines

I assume these don't get executed frequently enough for alternatives indirect call patching to matter? Down the road, if any more frequently executed hooks should appear, we will want to switch to that model though. This has the added benefit of there then being only one runtime instance of struct hypervisor_ops; all per-hypervisor ones would become __initconstrel.

All the routines listed here are run very infrequently.

I agree that when more frequently executed hooks appear we will want to look into alternatives.

Quoted 8 lines

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 313

Quoted 5 lines
xen/arch/x86/guest/xen/xen.c +1 -0
@@ -326,6 +310,31 @@ void hypervisor_resume(void)
         pv_console_init();
 }
 
+static const struct hypervisor_ops xg_ops = {

With this new xg_ prefix, ...

Quoted 9 lines

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 317

Quoted 4 lines
xen/arch/x86/guest/xen/xen.c +4 -0
+    .name = "Xen",
+    .setup = xen_setup,
+    .ap_setup = xen_ap_setup,
+    .resume = xen_resume,

... I'd like to suggest to use it for these ones too (if they need to have a prefix in the first place, which I doubt, but which I know others view differently).

Since they are only visible to this file I think not having a prefix should be fine.

Quoted 4 lines

xen/arch/x86/guest/xen/xen.c void hypervisor_resume(void) · line 320

Quoted 1 line
xen/arch/x86/guest/xen/xen.c +1 -0
+const struct hypervisor_ops * __init xen_probe(void)

For this one the call is a little more difficult, but for consistency I think xg_ would be slightly better here as well.

No problem. Changed it to xg_probe.

Quoted 3 lines

Nit: Strictly speaking there's a stray blank after *. We normally have ones only ahead of *, but not after.

Fixed.

Wei.

Quoted 2 lines

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#18 lore raw

PATCH v4 7/8 x86: be more verbose when running on a hypervisor

In reply to Wei Liu

Also replace xen_guest with running_on_hypervisor boolean.

---
Changes in v4:
1. Access ->name directly.
2. Drop Roger's review tag.

---
 xen/arch/x86/setup.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
xen/arch/x86/setup.c +5 -2
index 19606d909b..123436b35a 100644
@@ -689,6 +689,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     int i, j, e820_warn = 0, bytes = 0;
     bool acpi_boot_table_init_done = false, relocated = false;
     int ret;
+    bool running_on_hypervisor;
     struct ns16550_defaults ns16550 = {
         .data_bits = 8,
         .parity    = 'n',
@@ -763,7 +764,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
      * allocing any xenheap structures wanted in lower memory. */
     kexec_early_calculations();
 
-    hypervisor_probe();
+    running_on_hypervisor = !!hypervisor_probe();
 
     parse_video_info();
 
@@ -788,6 +789,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     printk("Command line: %s\n", cmdline);
 
     printk("Xen image load base address: %#lx\n", xen_phys_start);
+    if ( running_on_hypervisor )
+        printk("Running on %s\n", hypervisor_probe()->name);
 
 #ifdef CONFIG_VIDEO
     printk("Video information:\n");
@@ -1569,7 +1572,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
             max_cpus = nr_cpu_ids;
     }
 
-    if ( xen_guest )
+    if ( running_on_hypervisor )
         hypervisor_setup();
 
     /* Low mappings were only needed for some BIOS table parsing. */
Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
#19 lore raw

Re: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

Quoted 27 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

Also replace xen_guest with running_on_hypervisor boolean.

---
Changes in v4:
1. Access ->name directly.
2. Drop Roger's review tag.

---
 xen/arch/x86/setup.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
xen/arch/x86/setup.c +0 -0
index 19606d909b..123436b35a 100644

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 692

Quoted 4 lines
xen/arch/x86/setup.c +1 -0
@@ -689,6 +689,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     int i, j, e820_warn = 0, bytes = 0;
     bool acpi_boot_table_init_done = false, relocated = false;
     int ret;
+    bool running_on_hypervisor;

Why not stash hops here? Then you can...

Quoted 13 lines
xen/arch/x86/setup.c +1 -1
     struct ns16550_defaults ns16550 = {
         .data_bits = 8,
         .parity    = 'n',
@@ -763,7 +764,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
      * allocing any xenheap structures wanted in lower memory. */
     kexec_early_calculations();
 
-    hypervisor_probe();
+    running_on_hypervisor = !!hypervisor_probe();
 
     parse_video_info();
 

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 793

Quoted 5 lines
xen/arch/x86/setup.c +2 -0
@@ -788,6 +789,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     printk("Command line: %s\n", cmdline);
 
     printk("Xen image load base address: %#lx\n", xen_phys_start);
+    if ( running_on_hypervisor )
+        printk("Running on %s\n", hypervisor_probe()->name);

...avoid calling hypervisor_probe() again here.

Paul

Quoted 13 lines
xen/arch/x86/setup.c +1 -1
 #ifdef CONFIG_VIDEO
     printk("Video information:\n");
@@ -1569,7 +1572,7 @@ void __init noreturn __start_xen(unsigned long
 mbi_p)
             max_cpus = nr_cpu_ids;
     }
 
-    if ( xen_guest )
+    if ( running_on_hypervisor )
         hypervisor_setup();
 
     /* Low mappings were only needed for some BIOS table parsing. */
Quoted 8 lines

--
2.20.1

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#20 lore raw

Re: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

In reply to Wei Liu

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 767

Quoted 6 lines · Wei Liu
xen/arch/x86/setup.c +1 -1
@@ -763,7 +764,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
      * allocing any xenheap structures wanted in lower memory. */
     kexec_early_calculations();
 
-    hypervisor_probe();
+    running_on_hypervisor = !!hypervisor_probe();

No need for !! I think?

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 793

Quoted 6 lines
xen/arch/x86/setup.c +2 -0
@@ -788,6 +789,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     printk("Command line: %s\n", cmdline);
 
     printk("Xen image load base address: %#lx\n", xen_phys_start);
+    if ( running_on_hypervisor )
+        printk("Running on %s\n", hypervisor_probe()->name);

Invoking hypervisor_probe() twice seems odd to me. I realize the function copes, but why do everything a 2nd time when not running on any hypervisor? Furthermore if this use is the only reason why struct hypervisor_ops can't be local to xen/arch/x86/guest/hypervisor.c, then I think once again that the name should be return from hypervisor_probe(). The initial if() in there then could also go away.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#21 lore raw

Re: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

Quoted 13 lines · Jan Beulich

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 767

Quoted 6 lines · Wei Liu
xen/arch/x86/setup.c +1 -1
@@ -763,7 +764,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
      * allocing any xenheap structures wanted in lower memory. */
     kexec_early_calculations();
 
-    hypervisor_probe();
+    running_on_hypervisor = !!hypervisor_probe();

No need for !! I think?

Quoted 3 lines
xen/arch/x86/setup.c +0 -0
@@ -788,6 +789,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     printk("Command line: %s\n", cmdline);
 
Quoted 10 lines

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 793

Quoted 3 lines
xen/arch/x86/setup.c +2 -0
     printk("Xen image load base address: %#lx\n", xen_phys_start);
+    if ( running_on_hypervisor )
+        printk("Running on %s\n", hypervisor_probe()->name);

Invoking hypervisor_probe() twice seems odd to me. I realize the function copes, but why do everything a 2nd time when not running on any hypervisor? Furthermore if this use is the only reason why struct hypervisor_ops can't be local to xen/arch/x86/guest/hypervisor.c, then I think once again that the name should be return from hypervisor_probe().

"Local to xen/arch/x86/guest/hypervisor.c" was rubbish of course; local to xen/arch/x86/guest/ would be more correct.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#22 lore raw

Re: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

In reply to Wei Liu
Quoted 1 line · Wei Liu

Also replace xen_guest with running_on_hypervisor boolean.

I agree with dropping xen_guest, but...

Quoted 15 lines

---
Changes in v4:
1. Access ->name directly.
2. Drop Roger's review tag.

---
 xen/arch/x86/setup.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
xen/arch/x86/setup.c +0 -0
index 19606d909b..123436b35a 100644

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 692

Quoted 4 lines
xen/arch/x86/setup.c +1 -0
@@ -689,6 +689,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     int i, j, e820_warn = 0, bytes = 0;
     bool acpi_boot_table_init_done = false, relocated = false;
     int ret;
+    bool running_on_hypervisor;

... this is semantically ambiguous with cpu_has_hypervisor.

Where they differ is whether Xen has managed to recognise the hypervisor it is running under, or not.

Given that the hypervisor_*() hooks are nops by default, I'd suggest just making blind calls.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#23 lore raw

Re: [Xen-devel] [PATCH v4 7/8] x86: be more verbose when running on a hypervisor

Quoted 27 lines · Andrew Cooper
Quoted 1 line · Wei Liu

Also replace xen_guest with running_on_hypervisor boolean.

I agree with dropping xen_guest, but...

Quoted 15 lines

---
Changes in v4:
1. Access ->name directly.
2. Drop Roger's review tag.

---
 xen/arch/x86/setup.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
xen/arch/x86/setup.c +0 -0
index 19606d909b..123436b35a 100644

xen/arch/x86/setup.c void __init noreturn __start_xen(unsigned long mbi_p) · line 692

Quoted 4 lines
xen/arch/x86/setup.c +1 -0
@@ -689,6 +689,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     int i, j, e820_warn = 0, bytes = 0;
     bool acpi_boot_table_init_done = false, relocated = false;
     int ret;
+    bool running_on_hypervisor;

... this is semantically ambiguous with cpu_has_hypervisor.

Quoted 5 lines

Where they differ is whether Xen has managed to recognise the hypervisor it is running under, or not.

Given that the hypervisor_*() hooks are nops by default, I'd suggest just making blind calls.

Well Jan asked to drop the hypervisor_name hook. I can't make blind calls here. He's unhappy with calling hypervisor_probe twice either.

I can, however, do the following:

1. Change hypervisor_probe to return NULL or a string
2. Cache and use that return value inside this function

This should make both of you happy.

Wei.

Quoted 2 lines

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

author
#24 lore raw

PATCH v4 8/8 x86: introduce CONFIG_HYPERV and detection code

In reply to Wei Liu

We use the same code structure as we did for Xen.

As starters, detect Hyper-V in probe routine. More complex functionalities will be added later.

Take the chance to fix XEN_GUEST in Kconfig.

---
Changes in V4:
1. Add comment regarding order of probe functions.
2. Adapt to changes in previous patches.

---
 xen/arch/x86/Kconfig               | 11 ++++--
 xen/arch/x86/guest/Makefile        |  1 +
 xen/arch/x86/guest/hyperv/Makefile |  1 +
 xen/arch/x86/guest/hyperv/hyperv.c | 54 ++++++++++++++++++++++++++++++
 xen/arch/x86/guest/hypervisor.c    |  8 +++++
 xen/include/asm-x86/guest.h        |  1 +
 xen/include/asm-x86/guest/hyperv.h | 43 ++++++++++++++++++++++++
 7 files changed, 117 insertions(+), 2 deletions(-)
 create mode 100644 xen/arch/x86/guest/hyperv/Makefile
 create mode 100644 xen/arch/x86/guest/hyperv/hyperv.c
 create mode 100644 xen/include/asm-x86/guest/hyperv.h
xen/arch/x86/Kconfig +9 -2
index 867de857e8..0a02b6ee3f 100644
@@ -164,10 +164,17 @@ endchoice
 config GUEST
 	bool
 
+config HYPERV_GUEST
+	bool "Hyper-V Guest"
+	select GUEST
+	---help---
+	  Support for Xen detecting when it is running under Hyper-V.
+
+	  If unsure, say N.
+
 config XEN_GUEST
-	def_bool n
+	bool "Xen Guest"
 	select GUEST
-	prompt "Xen Guest"
 	---help---
 	  Support for Xen detecting when it is running under Xen.
 
xen/arch/x86/guest/Makefile +1 -0
index f63d64bbee..f164196772 100644
@@ -1,3 +1,4 @@
 obj-y += hypervisor.o
 
+subdir-$(CONFIG_HYPERV_GUEST) += hyperv
 subdir-$(CONFIG_XEN_GUEST) += xen
xen/arch/x86/guest/hyperv/Makefile new file +1 -0
new file mode 100644
index 0000000000..68170109a9
@@ -0,0 +1 @@
+obj-y += hyperv.o
xen/arch/x86/guest/hyperv/hyperv.c new file +54 -0

54 lines not on this page · raw message

xen/arch/x86/guest/hypervisor.c +8 -0
index a067cacecb..c293e185cc 100644
@@ -39,6 +39,14 @@ const struct hypervisor_ops *hypervisor_probe(void)
     if ( hops )
         goto out;
 
+    /*
+     * Detection of Hyper-V must come after Xen to avoid false positive due
+     * to viridian support
+     */
+    hops = hyperv_probe();
+    if ( hops )
+        goto out;
+
  out:
     return hops;
 }
xen/include/asm-x86/guest.h +1 -0
index 8e167165ae..94448606d4 100644
@@ -20,6 +20,7 @@
 #define __X86_GUEST_H__
 
 #include <asm/guest/hypercall.h>
+#include <asm/guest/hyperv.h>
 #include <asm/guest/hypervisor.h>
 #include <asm/guest/pvh-boot.h>
 #include <asm/guest/xen.h>
xen/include/asm-x86/guest/hyperv.h new file +43 -0

43 lines not on this page · raw message

Signature · 7 lines
2.20.1


_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Reviewed-by on v4 8/8
#25 lore raw

Re: [Xen-devel] [PATCH v4 8/8] x86: introduce CONFIG_HYPERV and detection code

Quoted 11 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 8/8] x86: introduce CONFIG_HYPERV and detection code

Quoted 6 lines

We use the same code structure as we did for Xen.

As starters, detect Hyper-V in probe routine. More complex functionalities will be added later.

Take the chance to fix XEN_GUEST in Kconfig.

Would this fix be better in your earlier renaming patch?

Quoted 2 lines

Either way...

Quoted 210 lines

---
Changes in V4:
1. Add comment regarding order of probe functions.
2. Adapt to changes in previous patches.

---
 xen/arch/x86/Kconfig               | 11 ++++--
 xen/arch/x86/guest/Makefile        |  1 +
 xen/arch/x86/guest/hyperv/Makefile |  1 +
 xen/arch/x86/guest/hyperv/hyperv.c | 54 ++++++++++++++++++++++++++++++
 xen/arch/x86/guest/hypervisor.c    |  8 +++++
 xen/include/asm-x86/guest.h        |  1 +
 xen/include/asm-x86/guest/hyperv.h | 43 ++++++++++++++++++++++++
 7 files changed, 117 insertions(+), 2 deletions(-)
 create mode 100644 xen/arch/x86/guest/hyperv/Makefile
 create mode 100644 xen/arch/x86/guest/hyperv/hyperv.c
 create mode 100644 xen/include/asm-x86/guest/hyperv.h
xen/arch/x86/Kconfig +9 -2
index 867de857e8..0a02b6ee3f 100644
@@ -164,10 +164,17 @@ endchoice
 config GUEST
 	bool
 
+config HYPERV_GUEST
+	bool "Hyper-V Guest"
+	select GUEST
+	---help---
+	  Support for Xen detecting when it is running under Hyper-V.
+
+	  If unsure, say N.
+
 config XEN_GUEST
-	def_bool n
+	bool "Xen Guest"
 	select GUEST
-	prompt "Xen Guest"
 	---help---
 	  Support for Xen detecting when it is running under Xen.
 
xen/arch/x86/guest/Makefile +1 -0
index f63d64bbee..f164196772 100644
@@ -1,3 +1,4 @@
 obj-y += hypervisor.o
 
+subdir-$(CONFIG_HYPERV_GUEST) += hyperv
 subdir-$(CONFIG_XEN_GUEST) += xen
xen/arch/x86/guest/hyperv/Makefile new file +1 -0
new file mode 100644
index 0000000000..68170109a9
@@ -0,0 +1 @@
+obj-y += hyperv.o
xen/arch/x86/guest/hyperv/hyperv.c new file +1 -0
new file mode 100644
index 0000000000..916e08ff89
@@ -0,0 +1,54 @@
+/************************************************************************

******
+ * arch/x86/guest/hyperv/hyperv.c
+ *
+ * Support for detecting and running under Hyper-V.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2019 Microsoft.
+ */

xen/arch/x86/guest/hyperv/hyperv.c new file +34 -0
+#include <xen/init.h>
+
+#include <asm/guest.h>
+
+static const struct hypervisor_ops hyperv_ops = {
+    .name = "Hyper-V",
+};
+
+const struct hypervisor_ops * __init hyperv_probe(void)
+{
+    uint32_t eax, ebx, ecx, edx;
+
+    cpuid(0x40000000, &eax, &ebx, &ecx, &edx);
+    if ( !((ebx == 0x7263694d) &&  /* "Micr" */
+           (ecx == 0x666f736f) &&  /* "osof" */
+           (edx == 0x76482074)) )  /* "t Hv" */
+        return NULL;
+
+    cpuid(0x40000001, &eax, &ebx, &ecx, &edx);
+    if ( eax != 0x31237648 )    /* Hv#1 */
+        return NULL;
+
+    return &hyperv_ops;
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
xen/arch/x86/guest/hypervisor.c +2 -0
index a067cacecb..c293e185cc 100644
@@ -39,6 +39,14 @@ const struct hypervisor_ops *hypervisor_probe(void)
     if ( hops )
         goto out;
 
+    /*
+     * Detection of Hyper-V must come after Xen to avoid false positive

due

xen/arch/x86/guest/hypervisor.c +6 -0
+     * to viridian support
+     */
+    hops = hyperv_probe();
+    if ( hops )
+        goto out;
+
  out:
     return hops;
 }
xen/include/asm-x86/guest.h +1 -0
index 8e167165ae..94448606d4 100644
@@ -20,6 +20,7 @@
 #define __X86_GUEST_H__
 
 #include <asm/guest/hypercall.h>
+#include <asm/guest/hyperv.h>
 #include <asm/guest/hypervisor.h>
 #include <asm/guest/pvh-boot.h>
 #include <asm/guest/xen.h>
xen/include/asm- renamed from xen/include/asm-x86/guest/hyperv.h +0 -0

x86/guest/hyperv.h
new file mode 100644
index 0000000000..3f88b94c77

xen/include/asm-x86/guest/hyperv.h new file +1 -0
@@ -0,0 +1,43 @@
+/************************************************************************

******
+ * asm-x86/guest/hyperv.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see
<http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2019 Microsoft.
+ */

xen/include/asm-x86/guest/hyperv.h new file +13 -0
+
+#ifndef __X86_GUEST_HYPERV_H__
+#define __X86_GUEST_HYPERV_H__
+
+#ifdef CONFIG_HYPERV_GUEST
+
+#include <asm/guest/hypervisor.h>
+
+const struct hypervisor_ops *hyperv_probe(void);
+
+#else
+
+static inline const struct hypervisor_ops *hyperv_probe(void) { return

NULL; }

xen/include/asm-x86/guest/hyperv.h new file +13 -1
+
+#endif /* CONFIG_HYPERV_GUEST */
+#endif /* __X86_GUEST_HYPERV_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--

2.20.1

Quoted 4 lines

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

Acked-by
#26 lore raw

Re: [Xen-devel] [PATCH v4 8/8] x86: introduce CONFIG_HYPERV and detection code

Quoted 18 lines · Paul
Quoted 11 lines

-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Wei Liu
Sent: 21 November 2019 19:51
To: Xen Development List <[email protected]>
Cc: Wei Liu <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Michael Kelley <[email protected]>; Jan Beulich <[email protected]>; Roger Pau Monné <[email protected]>
Subject: [Xen-devel] [PATCH v4 8/8] x86: introduce CONFIG_HYPERV and detection code

Quoted 6 lines

We use the same code structure as we did for Xen.

As starters, detect Hyper-V in probe routine. More complex functionalities will be added later.

Take the chance to fix XEN_GUEST in Kconfig.

Quoted 8 lines

Would this fix be better in your earlier renaming patch?

Quoted 2 lines

Either way...

(either way, as Paul says)

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel