Skip to contentSkip to conversation

[RFC PATCH 00/11 Allow PR and HV KVM to coexist in one kernel

RFC PATCH In review Aneesh Kumar K.V 4 participants 50 messages Sep 27, 2013 – Oct 1, 2013

Subsystems (10)
Arm (32-bit) x86 KVM PowerPC +6 more subsystems
lore
Share & tools
lore
#0 lore raw

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

-aneesh

#1 lore raw

RFC PATCH 1/11 kvm: powerpc: book3s hv: Fix vcore leak

From: Paul Mackerras <[email protected]>

add kvmppc_free_vcores() to free the kvmppc_vcore structures that we allocate for a guest, which are currently being leaked.

---
 arch/powerpc/kvm/book3s_hv.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
arch/powerpc/kvm/book3s_hv.c +10 -0
index 62a2b5a..edc7f9f 100644
@@ -1931,10 +1931,20 @@ int kvmppc_core_init_vm(struct kvm *kvm)
 	return 0;
 }
 
+static void kvmppc_free_vcores(struct kvm *kvm)
+{
+	long int i;
+
+	for (i = 0; i < KVM_MAX_VCORES; ++i)
+		kfree(kvm->arch.vcores[i]);
+	kvm->arch.online_vcores = 0;
+}
+
 void kvmppc_core_destroy_vm(struct kvm *kvm)
 {
 	uninhibit_secondary_onlining();
 
+	kvmppc_free_vcores(kvm);
 	if (kvm->arch.rma) {
 		kvm_release_rma(kvm->arch.rma);
 		kvm->arch.rma = NULL;
Signature · 1 line
1.8.1.2
#2 lore raw

Re: [RFC PATCH 01/11] kvm: powerpc: book3s hv: Fix vcore leak

Quoted 7 lines · Aneesh Kumar K.V

From: Paul Mackerras <[email protected]>

add kvmppc_free_vcores() to free the kvmppc_vcore structures that we allocate for a guest, which are currently being leaked.

This one doesn't look like an RFC to me. Applied to kvm-ppc-queue :).

Alex

Quoted 32 lines

---
arch/powerpc/kvm/book3s_hv.c | 10 ++++++++++

1 file changed, 10 insertions(+)
arch/powerpc/kvm/book3s_hv.c +10 -0
index 62a2b5a..edc7f9f 100644
@@ -1931,10 +1931,20 @@ int kvmppc_core_init_vm(struct kvm *kvm)
 	return 0;
 }
 
+static void kvmppc_free_vcores(struct kvm *kvm)
+{
+	long int i;
+
+	for (i = 0; i < KVM_MAX_VCORES; ++i)
+		kfree(kvm->arch.vcores[i]);
+	kvm->arch.online_vcores = 0;
+}
+
 void kvmppc_core_destroy_vm(struct kvm *kvm)
 {
 	uninhibit_secondary_onlining();
 
+	kvmppc_free_vcores(kvm);
 	if (kvm->arch.rma) {
 		kvm_release_rma(kvm->arch.rma);
 		kvm->arch.rma = NULL;
Signature · 1 line
1.8.1.2
#3 lore raw

RFC PATCH 2/11 kvm: powerpc: book3s: remove kvmppc_handler_highmem label

In reply to Aneesh Kumar K.V

From: Paul Mackerras <[email protected]>

This label is not used now.

---
 arch/powerpc/kvm/book3s_hv_interrupts.S | 3 ---
 arch/powerpc/kvm/book3s_interrupts.S    | 3 ---
 2 files changed, 6 deletions(-)
arch/powerpc/kvm/book3s_hv_interrupts.S +0 -3
index 37f1cc4..928142c 100644
@@ -158,9 +158,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_201)
  * Interrupts are enabled again at this point.
  */
 
-.global kvmppc_handler_highmem
-kvmppc_handler_highmem:
-
 	/*
 	 * Register usage at this point:
 	 *
arch/powerpc/kvm/book3s_interrupts.S +0 -3
index bf8f1ab..279ee3f 100644
@@ -121,9 +121,6 @@ kvm_start_lightweight:
  *
  */
 
-.global kvmppc_handler_highmem
-kvmppc_handler_highmem:
-
 	/*
 	 * Register usage at this point:
 	 *
Signature · 1 line
1.8.1.2
#4 lore raw

RFC PATCH 3/11 kvm: powerpc: book3s: move book3s_64_vio_hv.c into the main kernel binary

In reply to Aneesh Kumar K.V

From: Paul Mackerras <[email protected]>

Since the code in book3s_64_vio_hv.c is called from real mode with HV KVM, and therefore has to be built into the main kernel binary, this makes it always built-in rather than part of the KVM module. It gets called from the KVM module by PR KVM, so this adds an EXPORT_SYMBOL_GPL().

---
 arch/powerpc/kvm/Makefile           | 12 ++++++++----
 arch/powerpc/kvm/book3s_64_vio_hv.c |  2 ++
 2 files changed, 10 insertions(+), 4 deletions(-)
arch/powerpc/kvm/Makefile +8 -4
index 6646c95..104e8dc 100644
@@ -53,32 +53,36 @@ kvm-e500mc-objs := \
 	e500_emulate.o
 kvm-objs-$(CONFIG_KVM_E500MC) := $(kvm-e500mc-objs)
 
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) := \
+	book3s_64_vio_hv.o
+
 kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_PR) := \
 	$(KVM)/coalesced_mmio.o \
 	fpu.o \
 	book3s_paired_singles.o \
 	book3s_pr.o \
 	book3s_pr_papr.o \
-	book3s_64_vio_hv.o \
 	book3s_emulate.o \
 	book3s_interrupts.o \
 	book3s_mmu_hpte.o \
 	book3s_64_mmu_host.o \
 	book3s_64_mmu.o \
 	book3s_32_mmu.o
-kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_PR) := \
+
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_PR) += \
 	book3s_rmhandlers.o
 
 kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
 	book3s_hv.o \
 	book3s_hv_interrupts.o \
 	book3s_64_mmu_hv.o
+
 kvm-book3s_64-builtin-xics-objs-$(CONFIG_KVM_XICS) := \
 	book3s_hv_rm_xics.o
-kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
+
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HV) += \
 	book3s_hv_rmhandlers.o \
 	book3s_hv_rm_mmu.o \
-	book3s_64_vio_hv.o \
 	book3s_hv_ras.o \
 	book3s_hv_builtin.o \
 	book3s_hv_cma.o \
arch/powerpc/kvm/book3s_64_vio_hv.c +2 -0
index 30c2f3b..70d6594 100644
@@ -74,3 +74,5 @@ long kvmppc_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn,
 	/* Didn't find the liobn, punt it to userspace */
 	return H_TOO_HARD;
 }
+EXPORT_SYMBOL_GPL(kvmppc_h_put_tce);
+
Signature · 1 line
1.8.1.2
#5 lore raw

RFC PATCH 4/11 kvm: powerpc: book3s: Add a new config variable CONFIG_KVM_BOOK3S_HV

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help ups to select the relevant code in the kernel code when we later move HV and PR bits as seperate modules.

---
 arch/powerpc/include/asm/kvm_book3s_64.h  |  6 +++---
 arch/powerpc/include/asm/kvm_book3s_asm.h |  2 +-
 arch/powerpc/include/asm/kvm_host.h       | 10 +++++-----
 arch/powerpc/include/asm/kvm_ppc.h        |  2 +-
 arch/powerpc/kernel/asm-offsets.c         |  8 ++++----
 arch/powerpc/kernel/exceptions-64s.S      |  2 +-
 arch/powerpc/kernel/idle_power7.S         |  2 +-
 arch/powerpc/kvm/Kconfig                  |  4 ++++
 arch/powerpc/kvm/book3s_exports.c         |  5 +++--
 arch/powerpc/kvm/book3s_xics.c            |  4 ++--
 10 files changed, 25 insertions(+), 20 deletions(-)
arch/powerpc/include/asm/kvm_book3s_64.h +3 -3
index 86d638a..37af4ae 100644
@@ -35,7 +35,7 @@ static inline void svcpu_put(struct kvmppc_book3s_shadow_vcpu *svcpu)
 
 #define SPAPR_TCE_SHIFT		12
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 #define KVM_DEFAULT_HPT_ORDER	24	/* 16MB HPT by default */
 extern unsigned long kvm_rma_pages;
 #endif
@@ -278,7 +278,7 @@ static inline int is_vrma_hpte(unsigned long hpte_v)
 		(HPTE_V_1TB_SEG | (VRMA_VSID << (40 - 16)));
 }
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 /*
  * Note modification of an HPTE; set the HPTE modified bit
  * if anyone is interested.
@@ -289,6 +289,6 @@ static inline void note_hpte_modification(struct kvm *kvm,
 	if (atomic_read(&kvm->arch.hpte_mod_interest))
 		rev->guest_rpte |= HPTE_GR_MODIFIED;
 }
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
+#endif /* CONFIG_KVM_BOOK3S_HV */
 
 #endif /* __ASM_KVM_BOOK3S_64_H__ */
arch/powerpc/include/asm/kvm_book3s_asm.h +1 -1
index 360742a..1272178 100644
@@ -84,7 +84,7 @@ struct kvmppc_host_state {
 	u8 restore_hid5;
 	u8 napping;
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	u8 hwthread_req;
 	u8 hwthread_state;
 	u8 host_ipi;
arch/powerpc/include/asm/kvm_host.h +5 -5
index 78627c2..283e52e 100644
@@ -229,15 +229,15 @@ struct revmap_entry {
 #define KVMPPC_GOT_PAGE		0x80
 
 struct kvm_arch_memory_slot {
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	unsigned long *rmap;
 	unsigned long *slot_phys;
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
+#endif /* CONFIG_KVM_BOOK3S_HV */
 };
 
 struct kvm_arch {
 	unsigned int lpid;
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	unsigned long hpt_virt;
 	struct revmap_entry *revmap;
 	unsigned int host_lpid;
@@ -261,7 +261,7 @@ struct kvm_arch {
 	cpumask_t need_tlb_flush;
 	struct kvmppc_vcore *vcores[KVM_MAX_VCORES];
 	int hpt_cma_alloc;
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
+#endif /* CONFIG_KVM_BOOK3S_HV */
 #ifdef CONFIG_KVM_BOOK3S_PR
 	struct mutex hpt_mutex;
 	bool relon_disabled;
@@ -598,7 +598,7 @@ struct kvm_vcpu_arch {
 	struct kvmppc_icp *icp; /* XICS presentation controller */
 #endif
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	struct kvm_vcpu_arch_shared shregs;
 
 	unsigned long pgfault_addr;
arch/powerpc/include/asm/kvm_ppc.h +1 -1
index b15554a..9161bd1 100644
@@ -260,7 +260,7 @@ void kvmppc_set_pid(struct kvm_vcpu *vcpu, u32 pid);
 
 struct openpic;
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 extern void kvm_cma_reserve(void) __init;
 static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr)
 {
arch/powerpc/kernel/asm-offsets.c +4 -4
index 67b5d56..dd7ad4a 100644
@@ -445,7 +445,7 @@ int main(void)
 	DEFINE(VCPU_LR, offsetof(struct kvm_vcpu, arch.lr));
 	DEFINE(VCPU_CR, offsetof(struct kvm_vcpu, arch.cr));
 	DEFINE(VCPU_PC, offsetof(struct kvm_vcpu, arch.pc));
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	DEFINE(VCPU_MSR, offsetof(struct kvm_vcpu, arch.shregs.msr));
 	DEFINE(VCPU_SRR0, offsetof(struct kvm_vcpu, arch.shregs.srr0));
 	DEFINE(VCPU_SRR1, offsetof(struct kvm_vcpu, arch.shregs.srr1));
@@ -476,7 +476,7 @@ int main(void)
 	DEFINE(KVM_LPID, offsetof(struct kvm, arch.lpid));
 
 	/* book3s */
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	DEFINE(KVM_SDR1, offsetof(struct kvm, arch.sdr1));
 	DEFINE(KVM_HOST_LPID, offsetof(struct kvm, arch.host_lpid));
 	DEFINE(KVM_HOST_LPCR, offsetof(struct kvm, arch.host_lpcr));
@@ -581,7 +581,7 @@ int main(void)
 	HSTATE_FIELD(HSTATE_RESTORE_HID5, restore_hid5);
 	HSTATE_FIELD(HSTATE_NAPPING, napping);
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	HSTATE_FIELD(HSTATE_HWTHREAD_REQ, hwthread_req);
 	HSTATE_FIELD(HSTATE_HWTHREAD_STATE, hwthread_state);
 	HSTATE_FIELD(HSTATE_KVM_VCPU, kvm_vcpu);
@@ -597,7 +597,7 @@ int main(void)
 	HSTATE_FIELD(HSTATE_DABR, dabr);
 	HSTATE_FIELD(HSTATE_DECEXP, dec_expires);
 	DEFINE(IPI_PRIORITY, IPI_PRIORITY);
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
+#endif /* CONFIG_KVM_BOOK3S_HV */
 
 #ifdef CONFIG_PPC_BOOK3S_64
 	HSTATE_FIELD(HSTATE_CFAR, cfar);
arch/powerpc/kernel/exceptions-64s.S +1 -1
index 2a273be..9f2f352 100644
@@ -126,7 +126,7 @@ BEGIN_FTR_SECTION
 	bgt	cr1,.
 	GET_PACA(r13)
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	li	r0,KVM_HWTHREAD_IN_KERNEL
 	stb	r0,HSTATE_HWTHREAD_STATE(r13)
 	/* Order setting hwthread_state vs. testing hwthread_req */
arch/powerpc/kernel/idle_power7.S +1 -1
index e11863f..6aa4477 100644
@@ -84,7 +84,7 @@ _GLOBAL(power7_nap)
 	std	r9,_MSR(r1)
 	std	r1,PACAR1(r13)
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	/* Tell KVM we're napping */
 	li	r4,KVM_HWTHREAD_IN_NAP
 	stb	r4,HSTATE_HWTHREAD_STATE(r13)
arch/powerpc/kvm/Kconfig +4 -0
index ffaef2c..719c946 100644
@@ -39,6 +39,9 @@ config KVM_BOOK3S_PR
 	select KVM_MMIO
 	select MMU_NOTIFIER
 
+config KVM_BOOK3S_HV
+	bool
+
 config KVM_BOOK3S_32
 	tristate "KVM support for PowerPC book3s_32 processors"
 	depends on PPC_BOOK3S_32 && !SMP && !PTE_64BIT
@@ -71,6 +74,7 @@ config KVM_BOOK3S_64
 config KVM_BOOK3S_64_HV
 	bool "KVM support for POWER7 and PPC970 using hypervisor mode in host"
 	depends on KVM_BOOK3S_64
+	select KVM_BOOK3S_HV
 	select MMU_NOTIFIER
 	select CMA
 	---help---
arch/powerpc/kvm/book3s_exports.c +3 -2
index 7057a02..99504db 100644
@@ -20,9 +20,10 @@
 #include <linux/export.h>
 #include <asm/kvm_book3s.h>
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 EXPORT_SYMBOL_GPL(kvmppc_hv_entry_trampoline);
-#else
+#endif
+#ifdef CONFIG_KVM_BOOK3S_PR
 EXPORT_SYMBOL_GPL(kvmppc_entry_trampoline);
 EXPORT_SYMBOL_GPL(kvmppc_load_up_fpu);
 #ifdef CONFIG_ALTIVEC
arch/powerpc/kvm/book3s_xics.c +2 -2
index a3a5cb8..f0c732e 100644
@@ -1250,13 +1250,13 @@ static int kvmppc_xics_create(struct kvm_device *dev, u32 type)
 
 	xics_debugfs_init(xics);
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	if (cpu_has_feature(CPU_FTR_ARCH_206)) {
 		/* Enable real mode support */
 		xics->real_mode = ENABLE_REALMODE;
 		xics->real_mode_dbg = DEBUG_REALMODE;
 	}
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
+#endif /* CONFIG_KVM_BOOK3S_HV */
 
 	return 0;
 }
Signature · 1 line
1.8.1.2
#6 lore raw

Re: [RFC PATCH 04/11] kvm: powerpc: book3s: Add a new config variable CONFIG_KVM_BOOK3S_HV

Quoted 4 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help ups to select the relevant code in the kernel code when we later move HV and PR bits as seperate modules.

I don't think I grasp what semantically the difference between CONFIG_KVM_BOOK3S_HV and CONFIG_KVM_BOOK3S_64_HV is :).

Alex

#7 lore raw

Re: [RFC PATCH 04/11] kvm: powerpc: book3s: Add a new config variable CONFIG_KVM_BOOK3S_HV

Quoted 9 lines · Alexander Graf
Quoted 4 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help ups to select the relevant code in the kernel code when we later move HV and PR bits as seperate modules.

I don't think I grasp what semantically the difference between CONFIG_KVM_BOOK3S_HV and CONFIG_KVM_BOOK3S_64_HV is :).

When we later enable kvm-hv CONFIG_KVM_BOOK3S_64_HV becomes 'm' So we can't do #ifdef CONFIG_KVM_BOOK3S_64_HV

-aneesh

#8 lore raw

RFC PATCH 5/11 kvm: powerpc: book3s: Add kvmppc_ops callback for HV and PR specific operations

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This moves HV and PR specific functions to kvmppc_ops callback. This is needed so that we can enable HV and PR together in the same kernel. Actual changes to enable both come in the later patch.This also renames almost all of the symbols that exist in both PR and HV KVM for clarity. Symbols in the PR KVM implementation get "_pr" appended, and those in the HV KVM implementation get "_hv". Then, in book3s.c, we add a function with the name without the suffix and arrange for it to call the appropriate kvmppc_ops callback depending on which kvm type we selected during VM creation.

NOTE: we still don't enable selecting both the HV and PR together in this commit that will be done by a later commit.

---
 arch/powerpc/include/asm/kvm_book3s.h |   5 +-
 arch/powerpc/include/asm/kvm_ppc.h    |  63 ++++++++--
 arch/powerpc/kvm/Kconfig              |  15 ++-
 arch/powerpc/kvm/Makefile             |   9 +-
 arch/powerpc/kvm/book3s.c             | 145 +++++++++++++++++++++-
 arch/powerpc/kvm/book3s_32_mmu_host.c |   2 +-
 arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
 arch/powerpc/kvm/book3s_64_mmu_hv.c   |  17 ++-
 arch/powerpc/kvm/book3s_emulate.c     |   8 +-
 arch/powerpc/kvm/book3s_hv.c          | 226 +++++++++++++++++++++++++---------
 arch/powerpc/kvm/book3s_interrupts.S  |   2 +-
 arch/powerpc/kvm/book3s_pr.c          | 196 ++++++++++++++++++-----------
 arch/powerpc/kvm/emulate.c            |   6 +-
 arch/powerpc/kvm/powerpc.c            |  58 +++------
 14 files changed, 539 insertions(+), 215 deletions(-)
arch/powerpc/include/asm/kvm_book3s.h +2 -3
index 1b32f6c..3efba3c 100644
@@ -24,6 +24,8 @@
 #include <linux/kvm_host.h>
 #include <asm/kvm_book3s_asm.h>
 
+union kvmppc_one_reg;
+
 struct kvmppc_bat {
 	u64 raw;
 	u32 bepi;
@@ -124,7 +126,6 @@ extern void kvmppc_mmu_pte_flush(struct kvm_vcpu *vcpu, ulong ea, ulong ea_mask)
 extern void kvmppc_mmu_pte_vflush(struct kvm_vcpu *vcpu, u64 vp, u64 vp_mask);
 extern void kvmppc_mmu_pte_pflush(struct kvm_vcpu *vcpu, ulong pa_start, ulong pa_end);
 extern void kvmppc_set_msr(struct kvm_vcpu *vcpu, u64 new_msr);
-extern void kvmppc_set_pvr(struct kvm_vcpu *vcpu, u32 pvr);
 extern void kvmppc_mmu_book3s_64_init(struct kvm_vcpu *vcpu);
 extern void kvmppc_mmu_book3s_32_init(struct kvm_vcpu *vcpu);
 extern void kvmppc_mmu_book3s_hv_init(struct kvm_vcpu *vcpu);
@@ -193,8 +194,6 @@ static inline struct kvmppc_vcpu_book3s *to_book3s(struct kvm_vcpu *vcpu)
 	return vcpu->arch.book3s;
 }
 
-extern void kvm_return_point(void);
-
 /* Also add subarch specific defines */
 
 #ifdef CONFIG_KVM_BOOK3S_32_HANDLER
arch/powerpc/include/asm/kvm_ppc.h +53 -10

87 lines not on this page · raw message

arch/powerpc/kvm/Kconfig +14 -1
index 719c946..5c62459 100644
@@ -62,6 +62,7 @@ config KVM_BOOK3S_64
 	depends on PPC_BOOK3S_64
 	select KVM_BOOK3S_64_HANDLER
 	select KVM
+	select KVM_BOOK3S_PR if !KVM_BOOK3S_HV
 	---help---
 	  Support running unmodified book3s_64 and book3s_32 guest kernels
 	  in virtual machines on book3s_64 host processors.
@@ -93,9 +94,21 @@ config KVM_BOOK3S_64_HV
 	  If unsure, say N.
 
 config KVM_BOOK3S_64_PR
-	def_bool y
+	bool "KVM support without using hypervisor mode in host"
 	depends on KVM_BOOK3S_64 && !KVM_BOOK3S_64_HV
 	select KVM_BOOK3S_PR
+	---help---
+	  Support running guest kernels in virtual machines on processors
+	  without using hypervisor mode in the host, by running the
+	  guest in user mode (problem state) and emulating all
+	  privileged instructions and registers.
+
+	  This is not as fast as using hypervisor mode, but works on
+	  machines where hypervisor mode is not available or not usable,
+	  and can emulate processors that are different from the host
+	  processor, including emulating 32-bit processors on a 64-bit
+	  host.
+
 
 config KVM_BOOKE_HV
 	bool
arch/powerpc/kvm/Makefile +7 -2
index 104e8dc..c343793 100644
@@ -69,8 +69,13 @@ kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_PR) := \
 	book3s_64_mmu.o \
 	book3s_32_mmu.o
 
-kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_PR) += \
+ifdef CONFIG_KVM_BOOK3S_PR
+kvm-book3s_64-module-objs := \
+	$(KVM)/coalesced_mmio.o
+
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o
+endif
 
 kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
 	book3s_hv.o \
@@ -91,7 +96,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HV) += \
 kvm-book3s_64-objs-$(CONFIG_KVM_XICS) += \
 	book3s_xics.o
 
-kvm-book3s_64-module-objs := \
+kvm-book3s_64-module-objs += \
 	$(KVM)/kvm_main.o \
 	$(KVM)/eventfd.o \
 	powerpc.o \
arch/powerpc/kvm/book3s.c +141 -4

172 lines not on this page · raw message

arch/powerpc/kvm/book3s_32_mmu_host.c +1 -1
index 00e619b..c4361ef 100644
@@ -341,7 +341,7 @@ void kvmppc_mmu_flush_segments(struct kvm_vcpu *vcpu)
 	svcpu_put(svcpu);
 }
 
-void kvmppc_mmu_destroy(struct kvm_vcpu *vcpu)
+void kvmppc_mmu_destroy_pr(struct kvm_vcpu *vcpu)
 {
 	int i;
 
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 6bda504..329a978 100644
@@ -341,7 +341,7 @@ void kvmppc_mmu_flush_segments(struct kvm_vcpu *vcpu)
 	svcpu_put(svcpu);
 }
 
-void kvmppc_mmu_destroy(struct kvm_vcpu *vcpu)
+void kvmppc_mmu_destroy_pr(struct kvm_vcpu *vcpu)
 {
 	kvmppc_mmu_hpte_destroy(vcpu);
 	__destroy_context(to_book3s(vcpu)->context_id[0]);
arch/powerpc/kvm/book3s_64_mmu_hv.c +7 -10
index 043eec8..6f0485f 100644
@@ -260,10 +260,6 @@ int kvmppc_mmu_hv_init(void)
 	return 0;
 }
 
-void kvmppc_mmu_destroy(struct kvm_vcpu *vcpu)
-{
-}
-
 static void kvmppc_mmu_book3s_64_hv_reset_msr(struct kvm_vcpu *vcpu)
 {
 	kvmppc_set_msr(vcpu, MSR_SF | MSR_ME);
@@ -906,21 +902,22 @@ static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	return 0;
 }
 
-int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
+int kvm_unmap_hva_hv(struct kvm *kvm, unsigned long hva)
 {
 	if (kvm->arch.using_mmu_notifiers)
 		kvm_handle_hva(kvm, hva, kvm_unmap_rmapp);
 	return 0;
 }
 
-int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
+int kvm_unmap_hva_range_hv(struct kvm *kvm, unsigned long start, unsigned long end)
 {
 	if (kvm->arch.using_mmu_notifiers)
 		kvm_handle_hva_range(kvm, start, end, kvm_unmap_rmapp);
 	return 0;
 }
 
-void kvmppc_core_flush_memslot(struct kvm *kvm, struct kvm_memory_slot *memslot)
+void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
+				  struct kvm_memory_slot *memslot)
 {
 	unsigned long *rmapp;
 	unsigned long gfn;
@@ -994,7 +991,7 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	return ret;
 }
 
-int kvm_age_hva(struct kvm *kvm, unsigned long hva)
+int kvm_age_hva_hv(struct kvm *kvm, unsigned long hva)
 {
 	if (!kvm->arch.using_mmu_notifiers)
 		return 0;
@@ -1032,14 +1029,14 @@ static int kvm_test_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	return ret;
 }
 
-int kvm_test_age_hva(struct kvm *kvm, unsigned long hva)
+int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva)
 {
 	if (!kvm->arch.using_mmu_notifiers)
 		return 0;
 	return kvm_handle_hva(kvm, hva, kvm_test_age_rmapp);
 }
 
-void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte)
+void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte)
 {
 	if (!kvm->arch.using_mmu_notifiers)
 		return;
arch/powerpc/kvm/book3s_emulate.c +4 -4
index 34044b1..b9841ad 100644
@@ -86,8 +86,8 @@ static bool spr_allowed(struct kvm_vcpu *vcpu, enum priv_level level)
 	return true;
 }
 
-int kvmppc_core_emulate_op(struct kvm_run *run, struct kvm_vcpu *vcpu,
-                           unsigned int inst, int *advance)
+int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
+			      unsigned int inst, int *advance)
 {
 	int emulated = EMULATE_DONE;
 	int rt = get_rt(inst);
@@ -345,7 +345,7 @@ static struct kvmppc_bat *kvmppc_find_bat(struct kvm_vcpu *vcpu, int sprn)
 	return bat;
 }
 
-int kvmppc_core_emulate_mtspr(struct kvm_vcpu *vcpu, int sprn, ulong spr_val)
+int kvmppc_core_emulate_mtspr_pr(struct kvm_vcpu *vcpu, int sprn, ulong spr_val)
 {
 	int emulated = EMULATE_DONE;
 
@@ -468,7 +468,7 @@ unprivileged:
 	return emulated;
 }
 
-int kvmppc_core_emulate_mfspr(struct kvm_vcpu *vcpu, int sprn, ulong *spr_val)
+int kvmppc_core_emulate_mfspr_pr(struct kvm_vcpu *vcpu, int sprn, ulong *spr_val)
 {
 	int emulated = EMULATE_DONE;
 
arch/powerpc/kvm/book3s_hv.c +165 -61

423 lines not on this page · raw message

arch/powerpc/kvm/book3s_interrupts.S +1 -1
index 279ee3f..491d4b4 100644
@@ -181,7 +181,7 @@ kvm_start_lightweight:
 
 	/* Restore r3 (kvm_run) and r4 (vcpu) */
 	REST_2GPRS(3, r1)
-	bl	FUNC(kvmppc_handle_exit)
+	bl	FUNC(kvmppc_handle_exit_pr)
 
 	/* If RESUME_GUEST, get back in the loop */
 	cmpwi	r3, RESUME_GUEST
arch/powerpc/kvm/book3s_pr.c +126 -70

353 lines not on this page · raw message

arch/powerpc/kvm/emulate.c +3 -3
index 751cd45..81762eb 100644
@@ -130,7 +130,7 @@ static int kvmppc_emulate_mtspr(struct kvm_vcpu *vcpu, int sprn, int rs)
 	case SPRN_PIR: break;
 
 	default:
-		emulated = kvmppc_core_emulate_mtspr(vcpu, sprn,
+		emulated = kvmppc_ops->emulate_mtspr(vcpu, sprn,
 						     spr_val);
 		if (emulated == EMULATE_FAIL)
 			printk(KERN_INFO "mtspr: unknown spr "
@@ -191,7 +191,7 @@ static int kvmppc_emulate_mfspr(struct kvm_vcpu *vcpu, int sprn, int rt)
 		spr_val = kvmppc_get_dec(vcpu, get_tb());
 		break;
 	default:
-		emulated = kvmppc_core_emulate_mfspr(vcpu, sprn,
+		emulated = kvmppc_ops->emulate_mfspr(vcpu, sprn,
 						     &spr_val);
 		if (unlikely(emulated == EMULATE_FAIL)) {
 			printk(KERN_INFO "mfspr: unknown spr "
@@ -464,7 +464,7 @@ int kvmppc_emulate_instruction(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	}
 
 	if (emulated == EMULATE_FAIL) {
-		emulated = kvmppc_core_emulate_op(run, vcpu, inst, &advance);
+		emulated = kvmppc_ops->emulate_op(run, vcpu, inst, &advance);
 		if (emulated == EMULATE_AGAIN) {
 			advance = 0;
 		} else if (emulated == EMULATE_FAIL) {
arch/powerpc/kvm/powerpc.c +14 -44

92 lines not on this page · raw message

Signature · 1 line
1.8.1.2
#9 lore raw

Re: [RFC PATCH 05/11] kvm: powerpc: book3s: Add kvmppc_ops callback for HV and PR specific operation

Quoted 34 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This moves HV and PR specific functions to kvmppc_ops callback. This is needed so that we can enable HV and PR together in the same kernel. Actual changes to enable both come in the later patch.This also renames almost all of the symbols that exist in both PR and HV KVM for clarity. Symbols in the PR KVM implementation get "_pr" appended, and those in the HV KVM implementation get "_hv". Then, in book3s.c, we add a function with the name without the suffix and arrange for it to call the appropriate kvmppc_ops callback depending on which kvm type we selected during VM creation.

NOTE: we still don't enable selecting both the HV and PR together in this commit that will be done by a later commit.

---

arch/powerpc/include/asm/kvm_book3s.h |   5 +-
arch/powerpc/include/asm/kvm_ppc.h    |  63 ++++++++--
arch/powerpc/kvm/Kconfig              |  15 ++-
arch/powerpc/kvm/Makefile             |   9 +-
arch/powerpc/kvm/book3s.c             | 145 +++++++++++++++++++++-
arch/powerpc/kvm/book3s_32_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_64_mmu_hv.c   |  17 ++-
arch/powerpc/kvm/book3s_emulate.c     |   8 +-
arch/powerpc/kvm/book3s_hv.c          | 226 +++++++++++++++++++++++++---------
arch/powerpc/kvm/book3s_interrupts.S  |   2 +-
arch/powerpc/kvm/book3s_pr.c          | 196 ++++++++++++++++++-----------
arch/powerpc/kvm/emulate.c            |   6 +-
arch/powerpc/kvm/powerpc.c            |  58 +++------
14 files changed, 539 insertions(+), 215 deletions(-)

[...]

Quoted 18 lines
arch/powerpc/kvm/book3s_hv.c +2 -8
@@ -888,14 +890,8 @@ int kvmppc_set_one_reg(struct kvm_vcpu *vcpu, u64 id, union kvmppc_one_reg *val)
 	return r;
 }
 
-int kvmppc_core_check_processor_compat(void)
-{
-	if (cpu_has_feature(CPU_FTR_HVMODE))
-		return 0;
-	return -EIO;
-}
-
-struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
+static struct kvm_vcpu *kvmppc_core_vcpu_create_hv(struct kvm *kvm,
+						   unsigned int id)
 {
 	struct kvm_vcpu *vcpu;
 	int err = -EINVAL;

arch/powerpc/kvm/book3s_hv.c struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id) · line 923

Quoted 4 lines
arch/powerpc/kvm/book3s_hv.c +0 -1
@@ -920,7 +916,6 @@ struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
 	vcpu->arch.ctrl = CTRL_RUNLATCH;
 	/* default to host PVR, since we can't spoof it */
 	vcpu->arch.pvr = mfspr(SPRN_PVR);
-	kvmppc_set_pvr(vcpu, vcpu->arch.pvr);

Where is this one going?

Quoted 23 lines
arch/powerpc/kvm/book3s_hv.c +7 -1
 	spin_lock_init(&vcpu->arch.vpa_update_lock);
 	spin_lock_init(&vcpu->arch.tbacct_lock);
 	vcpu->arch.busy_preempt = TB_NIL;
@@ -972,7 +967,7 @@ static void unpin_vpa(struct kvm *kvm, struct kvmppc_vpa *vpa)
 					vpa->dirty);
 }
 
-void kvmppc_core_vcpu_free(struct kvm_vcpu *vcpu)
+static void kvmppc_core_vcpu_free_hv(struct kvm_vcpu *vcpu)
 {
 	spin_lock(&vcpu->arch.vpa_update_lock);
 	unpin_vpa(vcpu->kvm, &vcpu->arch.dtl);
@@ -983,6 +978,12 @@ void kvmppc_core_vcpu_free(struct kvm_vcpu *vcpu)
 	kmem_cache_free(kvm_vcpu_cache, vcpu);
 }
 
+static int kvmppc_core_check_requests_hv(struct kvm_vcpu *vcpu)
+{
+	/* Indicate we want to get back into the guest */
+	return 1;
+}
+
 

[...]

Quoted 17 lines
arch/powerpc/kvm/book3s_hv.c +15 -0
+	case KVM_PPC_GET_HTAB_FD: {
+		struct kvm_get_htab_fd ghf;
+
+		r = -EFAULT;
+		if (copy_from_user(&ghf, argp, sizeof(ghf)))
+			break;
+		r = kvm_vm_ioctl_get_htab_fd(kvm, &ghf);
+		break;
+	}
+
+	default:
+		r = -ENOTTY;
+	}
+
+	return r;

}

arch/powerpc/kvm/book3s_hv.c void kvmppc_core_destroy_vm(struct kvm *kvm) · line 2040

Quoted 2 lines
arch/powerpc/kvm/book3s_hv.c +1 -1
-static int kvmppc_book3s_hv_init(void)
+/* FIXME!! move to header */

Hrm :)

Quoted 45 lines
arch/powerpc/kvm/book3s_hv.c +44 -0
+extern void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
+					 struct kvm_memory_slot *memslot);
+extern int kvm_unmap_hva_hv(struct kvm *kvm, unsigned long hva);
+extern int kvm_unmap_hva_range_hv(struct kvm *kvm, unsigned long start,
+				  unsigned long end);
+extern int kvm_age_hva_hv(struct kvm *kvm, unsigned long hva);
+extern int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva);
+extern void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte);
+
+static struct kvmppc_ops kvmppc_hv_ops = {
+	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
+	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
+	.get_one_reg = kvmppc_get_one_reg_hv,
+	.set_one_reg = kvmppc_set_one_reg_hv,
+	.vcpu_load   = kvmppc_core_vcpu_load_hv,
+	.vcpu_put    = kvmppc_core_vcpu_put_hv,
+	.set_msr     = kvmppc_set_msr_hv,
+	.vcpu_run    = kvmppc_vcpu_run_hv,
+	.vcpu_create = kvmppc_core_vcpu_create_hv,
+	.vcpu_free   = kvmppc_core_vcpu_free_hv,
+	.check_requests = kvmppc_core_check_requests_hv,
+	.get_dirty_log  = kvm_vm_ioctl_get_dirty_log_hv,
+	.flush_memslot  = kvmppc_core_flush_memslot_hv,
+	.prepare_memory_region = kvmppc_core_prepare_memory_region_hv,
+	.commit_memory_region  = kvmppc_core_commit_memory_region_hv,
+	.unmap_hva = kvm_unmap_hva_hv,
+	.unmap_hva_range = kvm_unmap_hva_range_hv,
+	.age_hva  = kvm_age_hva_hv,
+	.test_age_hva = kvm_test_age_hva_hv,
+	.set_spte_hva = kvm_set_spte_hva_hv,
+	.mmu_destroy  = kvmppc_mmu_destroy_hv,
+	.free_memslot = kvmppc_core_free_memslot_hv,
+	.create_memslot = kvmppc_core_create_memslot_hv,
+	.init_vm =  kvmppc_core_init_vm_hv,
+	.destroy_vm = kvmppc_core_destroy_vm_hv,
+	.check_processor_compat = kvmppc_core_check_processor_compat_hv,
+	.get_smmu_info = kvm_vm_ioctl_get_smmu_info_hv,
+	.emulate_op = kvmppc_core_emulate_op_hv,
+	.emulate_mtspr = kvmppc_core_emulate_mtspr_hv,
+	.emulate_mfspr = kvmppc_core_emulate_mfspr_hv,
+	.fast_vcpu_kick = kvmppc_fast_vcpu_kick_hv,
+	.arch_vm_ioctl  = kvm_arch_vm_ioctl_hv,
+};
+

[...]

Quoted 31 lines
arch/powerpc/kvm/book3s_pr.c +27 -0
@@ -1390,8 +1389,42 @@ out:
 	return r;
 }
 
+static void kvmppc_core_flush_memslot_pr(struct kvm *kvm,
+					 struct kvm_memory_slot *memslot)
+{
+	return;
+}
+
+static int kvmppc_core_prepare_memory_region_pr(struct kvm *kvm,
+					struct kvm_memory_slot *memslot,
+					struct kvm_userspace_memory_region *mem)
+{
+	return 0;
+}
+
+static void kvmppc_core_commit_memory_region_pr(struct kvm *kvm,
+				struct kvm_userspace_memory_region *mem,
+				const struct kvm_memory_slot *old)
+{
+	return;
+}
+
+static void kvmppc_core_free_memslot_pr(struct kvm_memory_slot *free,
+					struct kvm_memory_slot *dont)
+{
+	return;
+}
+
+static int kvmppc_core_create_memslot_pr(struct kvm_memory_slot *slot,

arch/powerpc/kvm/book3s_pr.c out: · line 1427

Quoted 10 lines
arch/powerpc/kvm/book3s_pr.c +8 -1
+					 unsigned long npages)
+{
+	return 0;
+}
+
+
 #ifdef CONFIG_PPC64
-int kvm_vm_ioctl_get_smmu_info(struct kvm *kvm, struct kvm_ppc_smmu_info *info)
+static int kvm_vm_ioctl_get_smmu_info_pr(struct kvm *kvm,
+					 struct kvm_ppc_smmu_info *info)

You're dereferencing this function unconditionally now, probably breaking book3s_32 along the way :).

I'm not really happy with the naming scheme either, but I can't really think of anything better right now. In an ideal world all functions would still have the same names and we merely make them static and refer to them through structs :).

Alex

#10 lore raw

Re: [RFC PATCH 05/11] kvm: powerpc: book3s: Add kvmppc_ops callback for HV and PR specific operation

Quoted 55 lines · Alexander Graf
Quoted 34 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This moves HV and PR specific functions to kvmppc_ops callback. This is needed so that we can enable HV and PR together in the same kernel. Actual changes to enable both come in the later patch.This also renames almost all of the symbols that exist in both PR and HV KVM for clarity. Symbols in the PR KVM implementation get "_pr" appended, and those in the HV KVM implementation get "_hv". Then, in book3s.c, we add a function with the name without the suffix and arrange for it to call the appropriate kvmppc_ops callback depending on which kvm type we selected during VM creation.

NOTE: we still don't enable selecting both the HV and PR together in this commit that will be done by a later commit.

---

arch/powerpc/include/asm/kvm_book3s.h |   5 +-
arch/powerpc/include/asm/kvm_ppc.h    |  63 ++++++++--
arch/powerpc/kvm/Kconfig              |  15 ++-
arch/powerpc/kvm/Makefile             |   9 +-
arch/powerpc/kvm/book3s.c             | 145 +++++++++++++++++++++-
arch/powerpc/kvm/book3s_32_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_64_mmu_hv.c   |  17 ++-
arch/powerpc/kvm/book3s_emulate.c     |   8 +-
arch/powerpc/kvm/book3s_hv.c          | 226 +++++++++++++++++++++++++---------
arch/powerpc/kvm/book3s_interrupts.S  |   2 +-
arch/powerpc/kvm/book3s_pr.c          | 196 ++++++++++++++++++-----------
arch/powerpc/kvm/emulate.c            |   6 +-
arch/powerpc/kvm/powerpc.c            |  58 +++------
14 files changed, 539 insertions(+), 215 deletions(-)

[...]

Quoted 16 lines
arch/powerpc/kvm/book3s_hv.c +2 -8
@@ -888,14 +890,8 @@ int kvmppc_set_one_reg(struct kvm_vcpu *vcpu, u64 id, union kvmppc_one_reg *val)
 	return r;
 }
 
-int kvmppc_core_check_processor_compat(void)
-{
-	if (cpu_has_feature(CPU_FTR_HVMODE))
-		return 0;
-	return -EIO;
-}
-
-struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
+static struct kvm_vcpu *kvmppc_core_vcpu_create_hv(struct kvm *kvm,
+						   unsigned int id)
 {
 	struct kvm_vcpu *vcpu;
Quoted 8 lines

arch/powerpc/kvm/book3s_hv.c struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id) · line 923

Quoted 6 lines
arch/powerpc/kvm/book3s_hv.c +0 -1
 	int err = -EINVAL;
@@ -920,7 +916,6 @@ struct kvm_vcpu *kvmppc_core_vcpu_create(struct kvm *kvm, unsigned int id)
 	vcpu->arch.ctrl = CTRL_RUNLATCH;
 	/* default to host PVR, since we can't spoof it */
 	vcpu->arch.pvr = mfspr(SPRN_PVR);
-	kvmppc_set_pvr(vcpu, vcpu->arch.pvr);

Where is this one going?

That is same as the line above.

void kvmppc_set_pvr_hv(struct kvm_vcpu *vcpu, u32 pvr)

{
	vcpu->arch.pvr = pvr;
}
Quoted 44 lines
Quoted 24 lines
arch/powerpc/kvm/book3s_hv.c +7 -1
 	spin_lock_init(&vcpu->arch.vpa_update_lock);
 	spin_lock_init(&vcpu->arch.tbacct_lock);
 	vcpu->arch.busy_preempt = TB_NIL;
@@ -972,7 +967,7 @@ static void unpin_vpa(struct kvm *kvm, struct kvmppc_vpa *vpa)
 					vpa->dirty);
 }
 
-void kvmppc_core_vcpu_free(struct kvm_vcpu *vcpu)
+static void kvmppc_core_vcpu_free_hv(struct kvm_vcpu *vcpu)
 {
 	spin_lock(&vcpu->arch.vpa_update_lock);
 	unpin_vpa(vcpu->kvm, &vcpu->arch.dtl);
@@ -983,6 +978,12 @@ void kvmppc_core_vcpu_free(struct kvm_vcpu *vcpu)
 	kmem_cache_free(kvm_vcpu_cache, vcpu);
 }
 
+static int kvmppc_core_check_requests_hv(struct kvm_vcpu *vcpu)
+{
+	/* Indicate we want to get back into the guest */
+	return 1;
+}
+
 

[...]

Quoted 17 lines
arch/powerpc/kvm/book3s_hv.c +15 -0
+	case KVM_PPC_GET_HTAB_FD: {
+		struct kvm_get_htab_fd ghf;
+
+		r = -EFAULT;
+		if (copy_from_user(&ghf, argp, sizeof(ghf)))
+			break;
+		r = kvm_vm_ioctl_get_htab_fd(kvm, &ghf);
+		break;
+	}
+
+	default:
+		r = -ENOTTY;
+	}
+
+	return r;

}

Quoted 4 lines

arch/powerpc/kvm/book3s_hv.c void kvmppc_core_destroy_vm(struct kvm *kvm) · line 2040

Quoted 2 lines
arch/powerpc/kvm/book3s_hv.c +1 -1
-static int kvmppc_book3s_hv_init(void)
+/* FIXME!! move to header */

Hrm :)

yes, want to get something out for review. Will fix if we agree on the approach.

Quoted 85 lines
Quoted 46 lines
arch/powerpc/kvm/book3s_hv.c +44 -0
+extern void kvmppc_core_flush_memslot_hv(struct kvm *kvm,
+					 struct kvm_memory_slot *memslot);
+extern int kvm_unmap_hva_hv(struct kvm *kvm, unsigned long hva);
+extern int kvm_unmap_hva_range_hv(struct kvm *kvm, unsigned long start,
+				  unsigned long end);
+extern int kvm_age_hva_hv(struct kvm *kvm, unsigned long hva);
+extern int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva);
+extern void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte);
+
+static struct kvmppc_ops kvmppc_hv_ops = {
+	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
+	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
+	.get_one_reg = kvmppc_get_one_reg_hv,
+	.set_one_reg = kvmppc_set_one_reg_hv,
+	.vcpu_load   = kvmppc_core_vcpu_load_hv,
+	.vcpu_put    = kvmppc_core_vcpu_put_hv,
+	.set_msr     = kvmppc_set_msr_hv,
+	.vcpu_run    = kvmppc_vcpu_run_hv,
+	.vcpu_create = kvmppc_core_vcpu_create_hv,
+	.vcpu_free   = kvmppc_core_vcpu_free_hv,
+	.check_requests = kvmppc_core_check_requests_hv,
+	.get_dirty_log  = kvm_vm_ioctl_get_dirty_log_hv,
+	.flush_memslot  = kvmppc_core_flush_memslot_hv,
+	.prepare_memory_region = kvmppc_core_prepare_memory_region_hv,
+	.commit_memory_region  = kvmppc_core_commit_memory_region_hv,
+	.unmap_hva = kvm_unmap_hva_hv,
+	.unmap_hva_range = kvm_unmap_hva_range_hv,
+	.age_hva  = kvm_age_hva_hv,
+	.test_age_hva = kvm_test_age_hva_hv,
+	.set_spte_hva = kvm_set_spte_hva_hv,
+	.mmu_destroy  = kvmppc_mmu_destroy_hv,
+	.free_memslot = kvmppc_core_free_memslot_hv,
+	.create_memslot = kvmppc_core_create_memslot_hv,
+	.init_vm =  kvmppc_core_init_vm_hv,
+	.destroy_vm = kvmppc_core_destroy_vm_hv,
+	.check_processor_compat = kvmppc_core_check_processor_compat_hv,
+	.get_smmu_info = kvm_vm_ioctl_get_smmu_info_hv,
+	.emulate_op = kvmppc_core_emulate_op_hv,
+	.emulate_mtspr = kvmppc_core_emulate_mtspr_hv,
+	.emulate_mfspr = kvmppc_core_emulate_mfspr_hv,
+	.fast_vcpu_kick = kvmppc_fast_vcpu_kick_hv,
+	.arch_vm_ioctl  = kvm_arch_vm_ioctl_hv,
+};
+

[...]

Quoted 36 lines
arch/powerpc/kvm/book3s_pr.c +32 -0
@@ -1390,8 +1389,42 @@ out:
 	return r;
 }
 
+static void kvmppc_core_flush_memslot_pr(struct kvm *kvm,
+					 struct kvm_memory_slot *memslot)
+{
+	return;
+}
+
+static int kvmppc_core_prepare_memory_region_pr(struct kvm *kvm,
+					struct kvm_memory_slot *memslot,
+					struct kvm_userspace_memory_region *mem)
+{
+	return 0;
+}
+
+static void kvmppc_core_commit_memory_region_pr(struct kvm *kvm,
+				struct kvm_userspace_memory_region *mem,
+				const struct kvm_memory_slot *old)
+{
+	return;
+}
+
+static void kvmppc_core_free_memslot_pr(struct kvm_memory_slot *free,
+					struct kvm_memory_slot *dont)
+{
+	return;
+}
+
+static int kvmppc_core_create_memslot_pr(struct kvm_memory_slot *slot,
+					 unsigned long npages)
+{
+	return 0;
+}
+
Quoted 8 lines

arch/powerpc/kvm/book3s_pr.c out: · line 1427

Quoted 5 lines
arch/powerpc/kvm/book3s_pr.c +3 -1
+
 #ifdef CONFIG_PPC64
-int kvm_vm_ioctl_get_smmu_info(struct kvm *kvm, struct kvm_ppc_smmu_info *info)
+static int kvm_vm_ioctl_get_smmu_info_pr(struct kvm *kvm,
+					 struct kvm_ppc_smmu_info *info)

You're dereferencing this function unconditionally now, probably breaking book3s_32 along the way :).

will double check that.

Quoted 5 lines

I'm not really happy with the naming scheme either, but I can't really think of anything better right now. In an ideal world all functions would still have the same names and we merely make them static and refer to them through structs :).

I was following rest of the kernel source there. For ex: struct file_operations function pointers get pointed to by different fs specific callback, they all have fs details in their name. I also found that having _hv and _pr in the name allowed for easy grep and clarity in what different files should contain.

-aneesh

#11 lore raw

RFC PATCH 6/11 kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help us to identify whether we are running with hypervisor mode KVM enabled. The change is needed so that we can have both HV and PR kvm enabled in the same kernel.

If both HV and PR KVM are included, interrupts come in to the HV version of the kvmppc_interrupt code, which then jumps to the PR handler, renamed to kvmppc_interrupt_pr, if the guest is a PR guest.

Allowing both PR and HV in the same kernel required some changes to kvm_dev_ioctl_check_extension(), since the values returned now can't be selected with #ifdefs as much as previously. We look at is_hv_enabled to return the right value when checking for capabilities.For capabilities that are only provided by HV KVM, we return the HV value only if is_hv_enabled is true. For capabilities provided by PR KVM but not HV, we return the PR value only if is_hv_enabled is false.

---
 arch/powerpc/include/asm/kvm_book3s.h   | 53 --------------------------------
 arch/powerpc/include/asm/kvm_ppc.h      |  5 +--
 arch/powerpc/kvm/Makefile               |  2 +-
 arch/powerpc/kvm/book3s.c               | 44 +++++++++++++++++++++++++++
 arch/powerpc/kvm/book3s_hv.c            |  1 +
 arch/powerpc/kvm/book3s_hv_rmhandlers.S |  4 +++
 arch/powerpc/kvm/book3s_pr.c            |  1 +
 arch/powerpc/kvm/book3s_segment.S       |  7 ++++-
 arch/powerpc/kvm/book3s_xics.c          |  2 +-
 arch/powerpc/kvm/powerpc.c              | 54 ++++++++++++++++++---------------
 10 files changed, 90 insertions(+), 83 deletions(-)
arch/powerpc/include/asm/kvm_book3s.h +0 -53
index 3efba3c..ba33c49 100644
@@ -297,59 +297,6 @@ static inline ulong kvmppc_get_fault_dar(struct kvm_vcpu *vcpu)
 	return vcpu->arch.fault_dar;
 }
 
-#ifdef CONFIG_KVM_BOOK3S_PR
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return to_book3s(vcpu)->hior;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-	if (pending_now)
-		vcpu->arch.shared->int_pending = 1;
-	else if (old_pending)
-		vcpu->arch.shared->int_pending = 0;
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	ulong crit_raw = vcpu->arch.shared->critical;
-	ulong crit_r1 = kvmppc_get_gpr(vcpu, 1);
-	bool crit;
-
-	/* Truncate crit indicators in 32 bit mode */
-	if (!(vcpu->arch.shared->msr & MSR_SF)) {
-		crit_raw &= 0xffffffff;
-		crit_r1 &= 0xffffffff;
-	}
-
-	/* Critical section when crit == r1 */
-	crit = (crit_raw == crit_r1);
-	/* ... and we're in supervisor mode */
-	crit = crit && !(vcpu->arch.shared->msr & MSR_PR);
-
-	return crit;
-}
-#else /* CONFIG_KVM_BOOK3S_PR */
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return 0;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	return false;
-}
-#endif
-
 /* Magic register values loaded into r3 and r4 before the 'sc' assembly
  * instruction for the OSI hypercalls */
 #define OSI_SC_MAGIC_R3			0x113724FA
arch/powerpc/include/asm/kvm_ppc.h +3 -2
index 4d9641c..58e732f 100644
@@ -169,6 +169,7 @@ extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq);
 extern int kvmppc_xics_int_off(struct kvm *kvm, u32 irq);
 
 struct kvmppc_ops {
+	bool is_hv_enabled;
 	int (*get_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*set_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*get_one_reg)(struct kvm_vcpu *vcpu, u64 id,
@@ -309,10 +310,10 @@ static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr)
 
 static inline u32 kvmppc_get_xics_latch(void)
 {
-	u32 xirr = get_paca()->kvm_hstate.saved_xirr;
+	u32 xirr;
 
+	xirr = get_paca()->kvm_hstate.saved_xirr;
 	get_paca()->kvm_hstate.saved_xirr = 0;
-
 	return xirr;
 }
 
arch/powerpc/kvm/Makefile +1 -1
index c343793..a514ecd 100644
@@ -77,7 +77,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o
 endif
 
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
+kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \
 	book3s_hv.o \
 	book3s_hv_interrupts.o \
 	book3s_64_mmu_hv.o
arch/powerpc/kvm/book3s.c +44 -0

50 lines not on this page · raw message

arch/powerpc/kvm/book3s_hv.c +1 -0
index 36d7920..eec353d 100644
@@ -2048,6 +2048,7 @@ extern int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva);
 extern void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte);
 
 static struct kvmppc_ops kvmppc_hv_ops = {
+	.is_hv_enabled = true,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
 	.get_one_reg = kvmppc_get_one_reg_hv,
arch/powerpc/kvm/book3s_hv_rmhandlers.S +4 -0
index 5ede7fc..4838fdb 100644
@@ -577,6 +577,10 @@ kvmppc_interrupt:
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
+#ifdef CONFIG_KVM_BOOK3S_PR
+	cmpwi	r9, KVM_GUEST_MODE_GUEST
+	beq	kvmppc_interrupt_pr
+#endif
 	/* We're now back in the host but in guest MMU context */
 	li	r9, KVM_GUEST_MODE_HOST_HV
 	stb	r9, HSTATE_IN_GUEST(r13)
arch/powerpc/kvm/book3s_pr.c +1 -0
index ff5bd24..2a97279 100644
@@ -1509,6 +1509,7 @@ extern int kvmppc_core_emulate_mfspr_pr(struct kvm_vcpu *vcpu,
 					int sprn, ulong *spr_val);
 
 static struct kvmppc_ops kvmppc_pr_ops = {
+	.is_hv_enabled = false,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_pr,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_pr,
 	.get_one_reg = kvmppc_get_one_reg_pr,
arch/powerpc/kvm/book3s_segment.S +6 -1
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:
 .global kvmppc_handler_trampoline_exit
 kvmppc_handler_trampoline_exit:
 
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else
 .global kvmppc_interrupt
 kvmppc_interrupt:
-
+#endif
 	/* Register usage at this point:
 	 *
 	 * SPRG_SCRATCH0  = guest R13
arch/powerpc/kvm/book3s_xics.c +1 -1
index f0c732e..fa7625d 100644
@@ -818,7 +818,7 @@ int kvmppc_xics_hcall(struct kvm_vcpu *vcpu, u32 req)
 	}
 
 	/* Check for real mode returning too hard */
-	if (xics->real_mode)
+	if (xics->real_mode && kvmppc_ops->is_hv_enabled)
 		return kvmppc_xics_rm_complete(vcpu, req);
 
 	switch (req) {
arch/powerpc/kvm/powerpc.c +29 -25

119 lines not on this page · raw message

Signature · 1 line
1.8.1.2
#12 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 101 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help us to identify whether we are running with hypervisor mode KVM enabled. The change is needed so that we can have both HV and PR kvm enabled in the same kernel.

If both HV and PR KVM are included, interrupts come in to the HV version of the kvmppc_interrupt code, which then jumps to the PR handler, renamed to kvmppc_interrupt_pr, if the guest is a PR guest.

Allowing both PR and HV in the same kernel required some changes to kvm_dev_ioctl_check_extension(), since the values returned now can't be selected with #ifdefs as much as previously. We look at is_hv_enabled to return the right value when checking for capabilities.For capabilities that are only provided by HV KVM, we return the HV value only if is_hv_enabled is true. For capabilities provided by PR KVM but not HV, we return the PR value only if is_hv_enabled is false.

---

arch/powerpc/include/asm/kvm_book3s.h   | 53 --------------------------------
arch/powerpc/include/asm/kvm_ppc.h      |  5 +--
arch/powerpc/kvm/Makefile               |  2 +-
arch/powerpc/kvm/book3s.c               | 44 +++++++++++++++++++++++++++
arch/powerpc/kvm/book3s_hv.c            |  1 +
arch/powerpc/kvm/book3s_hv_rmhandlers.S |  4 +++
arch/powerpc/kvm/book3s_pr.c            |  1 +
arch/powerpc/kvm/book3s_segment.S       |  7 ++++-
arch/powerpc/kvm/book3s_xics.c          |  2 +-
arch/powerpc/kvm/powerpc.c              | 54 ++++++++++++++++++---------------
10 files changed, 90 insertions(+), 83 deletions(-)
arch/powerpc/include/asm/kvm_book3s.h +0 -53
index 3efba3c..ba33c49 100644
@@ -297,59 +297,6 @@ static inline ulong kvmppc_get_fault_dar(struct kvm_vcpu *vcpu)
 	return vcpu->arch.fault_dar;
 }
 
-#ifdef CONFIG_KVM_BOOK3S_PR
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return to_book3s(vcpu)->hior;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-	if (pending_now)
-		vcpu->arch.shared->int_pending = 1;
-	else if (old_pending)
-		vcpu->arch.shared->int_pending = 0;
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	ulong crit_raw = vcpu->arch.shared->critical;
-	ulong crit_r1 = kvmppc_get_gpr(vcpu, 1);
-	bool crit;
-
-	/* Truncate crit indicators in 32 bit mode */
-	if (!(vcpu->arch.shared->msr & MSR_SF)) {
-		crit_raw &= 0xffffffff;
-		crit_r1 &= 0xffffffff;
-	}
-
-	/* Critical section when crit = r1 */
-	crit = (crit_raw = crit_r1);
-	/* ... and we're in supervisor mode */
-	crit = crit && !(vcpu->arch.shared->msr & MSR_PR);
-
-	return crit;
-}
-#else /* CONFIG_KVM_BOOK3S_PR */
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return 0;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	return false;
-}
-#endif
-
 /* Magic register values loaded into r3 and r4 before the 'sc' assembly
 * instruction for the OSI hypercalls */
 #define OSI_SC_MAGIC_R3			0x113724FA
arch/powerpc/include/asm/kvm_ppc.h +0 -0
index 4d9641c..58e732f 100644

arch/powerpc/include/asm/kvm_ppc.h extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq); · line 172

Quoted 4 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -0
@@ -169,6 +169,7 @@ extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq);
 extern int kvmppc_xics_int_off(struct kvm *kvm, u32 irq);
 
 struct kvmppc_ops {
+	bool is_hv_enabled;

This doesn't really belong into an ops struct. Either you compare

if (kvmppc_ops = kvmppc_ops_pr)

against a global known good ops struct or you put the hint somewhere into the kvm struct.

arch/powerpc/include/asm/kvm_ppc.h static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr) · line 315

Quoted 13 lines
arch/powerpc/include/asm/kvm_ppc.h +2 -2
 	int (*get_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*set_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*get_one_reg)(struct kvm_vcpu *vcpu, u64 id,
@@ -309,10 +310,10 @@ static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr)
 
 static inline u32 kvmppc_get_xics_latch(void)
 {
-	u32 xirr = get_paca()->kvm_hstate.saved_xirr;
+	u32 xirr;
 
+	xirr = get_paca()->kvm_hstate.saved_xirr;
 	get_paca()->kvm_hstate.saved_xirr = 0;
-

I don't see any functionality change here?

arch/powerpc/kvm/Makefile kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \ · line 80

Quoted 13 lines
arch/powerpc/include/asm/kvm_ppc.h +0 -0
 	return xirr;
 }
 
arch/powerpc/kvm/Makefile +0 -0
index c343793..a514ecd 100644
@@ -77,7 +77,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o

endif

arch/powerpc/kvm/Makefile +1 -1
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
+kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \

This change looks unrelated?

arch/powerpc/kvm/book3s.c void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu) · line 72

Quoted 12 lines
	book3s_hv.o \
	book3s_hv_interrupts.o \
	book3s_64_mmu_hv.o
arch/powerpc/kvm/book3s.c +1 -0
index bdc3f95..12f94bf 100644
@@ -69,6 +69,50 @@ void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu)
 {
 }
 
+static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)

Please drop the "inline" keyword on functions in .c files :). That way the compiler tells us about unused functions.

Quoted 98 lines
arch/powerpc/kvm/book3s.c +43 -0
+{
+	if (!kvmppc_ops->is_hv_enabled)
+		return to_book3s(vcpu)->hior;
+	return 0;
+}
+
+static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
+			unsigned long pending_now, unsigned long old_pending)
+{
+	if (kvmppc_ops->is_hv_enabled)
+		return;
+	if (pending_now)
+		vcpu->arch.shared->int_pending = 1;
+	else if (old_pending)
+		vcpu->arch.shared->int_pending = 0;
+}
+
+static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
+{
+	ulong crit_raw;
+	ulong crit_r1;
+	bool crit;
+
+	if (kvmppc_ops->is_hv_enabled)
+		return false;
+
+	crit_raw = vcpu->arch.shared->critical;
+	crit_r1 = kvmppc_get_gpr(vcpu, 1);
+
+	/* Truncate crit indicators in 32 bit mode */
+	if (!(vcpu->arch.shared->msr & MSR_SF)) {
+		crit_raw &= 0xffffffff;
+		crit_r1 &= 0xffffffff;
+	}
+
+	/* Critical section when crit = r1 */
+	crit = (crit_raw = crit_r1);
+	/* ... and we're in supervisor mode */
+	crit = crit && !(vcpu->arch.shared->msr & MSR_PR);
+
+	return crit;
+}
+
 void kvmppc_inject_interrupt(struct kvm_vcpu *vcpu, int vec, u64 flags)
 {
 	vcpu->arch.shared->srr0 = kvmppc_get_pc(vcpu);
arch/powerpc/kvm/book3s_hv.c +1 -0
index 36d7920..eec353d 100644
@@ -2048,6 +2048,7 @@ extern int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva);
 extern void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte);
 
 static struct kvmppc_ops kvmppc_hv_ops = {
+	.is_hv_enabled = true,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
 	.get_one_reg = kvmppc_get_one_reg_hv,
arch/powerpc/kvm/book3s_hv_rmhandlers.S +4 -0
index 5ede7fc..4838fdb 100644
@@ -577,6 +577,10 @@ kvmppc_interrupt:
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
+#ifdef CONFIG_KVM_BOOK3S_PR
+	cmpwi	r9, KVM_GUEST_MODE_GUEST
+	beq	kvmppc_interrupt_pr
+#endif
 	/* We're now back in the host but in guest MMU context */
 	li	r9, KVM_GUEST_MODE_HOST_HV
 	stb	r9, HSTATE_IN_GUEST(r13)
arch/powerpc/kvm/book3s_pr.c +1 -0
index ff5bd24..2a97279 100644
@@ -1509,6 +1509,7 @@ extern int kvmppc_core_emulate_mfspr_pr(struct kvm_vcpu *vcpu,
 					int sprn, ulong *spr_val);
 
 static struct kvmppc_ops kvmppc_pr_ops = {
+	.is_hv_enabled = false,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_pr,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_pr,
 	.get_one_reg = kvmppc_get_one_reg_pr,
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else
Quoted 2 lines

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 58 lines
arch/powerpc/kvm/book3s_segment.S +1 -1
-
+#endif
	/* Register usage at this point:
	 *
	 * SPRG_SCRATCH0  = guest R13
arch/powerpc/kvm/book3s_xics.c +1 -1
index f0c732e..fa7625d 100644
@@ -818,7 +818,7 @@ int kvmppc_xics_hcall(struct kvm_vcpu *vcpu, u32 req)
 	}
 
 	/* Check for real mode returning too hard */
-	if (xics->real_mode)
+	if (xics->real_mode && kvmppc_ops->is_hv_enabled)
 		return kvmppc_xics_rm_complete(vcpu, req);
 
 	switch (req) {
arch/powerpc/kvm/powerpc.c +1 -6
index 69b9305..00a96fc 100644
@@ -52,7 +52,6 @@ int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 /*
 * Common checks before entering the guest world.  Call with interrupts
 * disabled.
@@ -127,7 +126,6 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
 
 	return r;
 }
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
 
 int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 {
@@ -194,11 +192,9 @@ int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
 	if ((vcpu->arch.cpu_type != KVM_CPU_3S_64) && vcpu->arch.papr_enabled)
 		goto out;
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
 	/* HV KVM can only do PAPR mode for now */
-	if (!vcpu->arch.papr_enabled)
+	if (!vcpu->arch.papr_enabled && kvmppc_ops->is_hv_enabled)
 		goto out;
-#endif
 
 #ifdef CONFIG_KVM_BOOKE_HV
 	if (!cpu_has_feature(CPU_FTR_EMB_HV))
@@ -322,22 +318,26 @@ int kvm_dev_ioctl_check_extension(long ext)
 	case KVM_CAP_DEVICE_CTRL:
 		r = 1;
 		break;
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 	case KVM_CAP_PPC_PAIRED_SINGLES:
 	case KVM_CAP_PPC_OSI:

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 328

Quoted 10 lines
arch/powerpc/kvm/powerpc.c +2 -4
 	case KVM_CAP_PPC_GET_PVINFO:
 #if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
 	case KVM_CAP_SW_TLB:
 #endif
-#ifdef CONFIG_KVM_MPIC
-	case KVM_CAP_IRQ_MPIC:
-#endif
-		r = 1;
+		/* We support this only for PR */
+		r = !kvmppc_ops->is_hv_enabled;

Reading this, have you test compiled your code against e500 configs?

Quoted 17 lines
arch/powerpc/kvm/powerpc.c +7 -0
 		break;
+#ifdef CONFIG_KVM_MMIO
 	case KVM_CAP_COALESCED_MMIO:
 		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
 		break;
 #endif
+#ifdef CONFIG_KVM_MPIC
+	case KVM_CAP_IRQ_MPIC:
+		r = 1;
+		break;
+#endif
+
 #ifdef CONFIG_PPC_BOOK3S_64
 	case KVM_CAP_SPAPR_TCE:
 	case KVM_CAP_PPC_ALLOC_HTAB:
@@ -348,32 +348,37 @@ int kvm_dev_ioctl_check_extension(long ext)
 		r = 1;

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 356

Quoted 10 lines
arch/powerpc/kvm/powerpc.c +5 -2
 		break;
 #endif /* CONFIG_PPC_BOOK3S_64 */
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
 	case KVM_CAP_PPC_SMT:
-		r = threads_per_core;
+		if (kvmppc_ops->is_hv_enabled)
+			r = threads_per_core;
+		else
+			r = 0;

0? Or 1?

Alex

#13 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 105 lines · Alexander Graf
Quoted 103 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help us to identify whether we are running with hypervisor mode KVM enabled. The change is needed so that we can have both HV and PR kvm enabled in the same kernel.

If both HV and PR KVM are included, interrupts come in to the HV version of the kvmppc_interrupt code, which then jumps to the PR handler, renamed to kvmppc_interrupt_pr, if the guest is a PR guest.

Allowing both PR and HV in the same kernel required some changes to kvm_dev_ioctl_check_extension(), since the values returned now can't be selected with #ifdefs as much as previously. We look at is_hv_enabled to return the right value when checking for capabilities.For capabilities that are only provided by HV KVM, we return the HV value only if is_hv_enabled is true. For capabilities provided by PR KVM but not HV, we return the PR value only if is_hv_enabled is false.

---

arch/powerpc/include/asm/kvm_book3s.h   | 53 --------------------------------
arch/powerpc/include/asm/kvm_ppc.h      |  5 +--
arch/powerpc/kvm/Makefile               |  2 +-
arch/powerpc/kvm/book3s.c               | 44 +++++++++++++++++++++++++++
arch/powerpc/kvm/book3s_hv.c            |  1 +
arch/powerpc/kvm/book3s_hv_rmhandlers.S |  4 +++
arch/powerpc/kvm/book3s_pr.c            |  1 +
arch/powerpc/kvm/book3s_segment.S       |  7 ++++-
arch/powerpc/kvm/book3s_xics.c          |  2 +-
arch/powerpc/kvm/powerpc.c              | 54 ++++++++++++++++++---------------
10 files changed, 90 insertions(+), 83 deletions(-)
arch/powerpc/include/asm/kvm_book3s.h +0 -53
index 3efba3c..ba33c49 100644
@@ -297,59 +297,6 @@ static inline ulong kvmppc_get_fault_dar(struct kvm_vcpu *vcpu)
 	return vcpu->arch.fault_dar;
 }
 
-#ifdef CONFIG_KVM_BOOK3S_PR
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return to_book3s(vcpu)->hior;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-	if (pending_now)
-		vcpu->arch.shared->int_pending = 1;
-	else if (old_pending)
-		vcpu->arch.shared->int_pending = 0;
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	ulong crit_raw = vcpu->arch.shared->critical;
-	ulong crit_r1 = kvmppc_get_gpr(vcpu, 1);
-	bool crit;
-
-	/* Truncate crit indicators in 32 bit mode */
-	if (!(vcpu->arch.shared->msr & MSR_SF)) {
-		crit_raw &= 0xffffffff;
-		crit_r1 &= 0xffffffff;
-	}
-
-	/* Critical section when crit == r1 */
-	crit = (crit_raw == crit_r1);
-	/* ... and we're in supervisor mode */
-	crit = crit && !(vcpu->arch.shared->msr & MSR_PR);
-
-	return crit;
-}
-#else /* CONFIG_KVM_BOOK3S_PR */
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return 0;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	return false;
-}
-#endif
-
 /* Magic register values loaded into r3 and r4 before the 'sc' assembly
 * instruction for the OSI hypercalls */
 #define OSI_SC_MAGIC_R3			0x113724FA
arch/powerpc/include/asm/kvm_ppc.h +0 -0
index 4d9641c..58e732f 100644
@@ -169,6 +169,7 @@ extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq);
 extern int kvmppc_xics_int_off(struct kvm *kvm, u32 irq);
 
Quoted 6 lines

arch/powerpc/include/asm/kvm_ppc.h extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq); · line 172

Quoted 2 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -0
 struct kvmppc_ops {
+	bool is_hv_enabled;

This doesn't really belong into an ops struct. Either you compare

if (kvmppc_ops == kvmppc_ops_pr)

will do that in the next update.

Quoted 13 lines

against a global known good ops struct or you put the hint somewhere into the kvm struct.

Quoted 10 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -1
 	int (*get_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*set_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*get_one_reg)(struct kvm_vcpu *vcpu, u64 id,
@@ -309,10 +310,10 @@ static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr)
 
 static inline u32 kvmppc_get_xics_latch(void)
 {
-	u32 xirr = get_paca()->kvm_hstate.saved_xirr;
+	u32 xirr;
 
Quoted 7 lines

arch/powerpc/include/asm/kvm_ppc.h static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr) · line 315

Quoted 3 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -1
+	xirr = get_paca()->kvm_hstate.saved_xirr;
 	get_paca()->kvm_hstate.saved_xirr = 0;
-

I don't see any functionality change here?

Quoted 1 line
arch/powerpc/include/asm/kvm_ppc.h +0 -0
 	return xirr;

Mistake on my side, I had a if condition in there before, which i later removed. But forgot to move the assignment back. Will fix.

Quoted 10 lines
Quoted 10 lines
arch/powerpc/include/asm/kvm_ppc.h +0 -0
 }
 
arch/powerpc/kvm/Makefile +0 -0
index c343793..a514ecd 100644
@@ -77,7 +77,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o

endif

Quoted 5 lines

arch/powerpc/kvm/Makefile kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \ · line 80

Quoted 2 lines
arch/powerpc/kvm/Makefile +1 -1
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
+kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \

This change looks unrelated?

yes. Will move it to the correct patch

Quoted 11 lines
Quoted 11 lines
	book3s_hv.o \
	book3s_hv_interrupts.o \
	book3s_64_mmu_hv.o
arch/powerpc/kvm/book3s.c +0 -0
index bdc3f95..12f94bf 100644
@@ -69,6 +69,50 @@ void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu)
 {
 }
 
Quoted 4 lines

arch/powerpc/kvm/book3s.c void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu) · line 72

Quoted 1 line
arch/powerpc/kvm/book3s.c +1 -0
+static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)

Please drop the "inline" keyword on functions in .c files :). That way the compiler tells us about unused functions.

ok .

Quoted 18 lines
Quoted 18 lines
arch/powerpc/kvm/book3s.c +17 -0
+{
+	if (!kvmppc_ops->is_hv_enabled)
+		return to_book3s(vcpu)->hior;
+	return 0;
+}
+
+static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
+			unsigned long pending_now, unsigned long old_pending)
+{
+	if (kvmppc_ops->is_hv_enabled)
+		return;
+	if (pending_now)
+		vcpu->arch.shared->int_pending = 1;
+	else if (old_pending)
+		vcpu->arch.shared->int_pending = 0;
+}
+

....

Quoted 8 lines
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 10 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Quoted 63 lines
Quoted 63 lines
arch/powerpc/kvm/book3s_segment.S +1 -1
-
+#endif
	/* Register usage at this point:
	 *
	 * SPRG_SCRATCH0  = guest R13
arch/powerpc/kvm/book3s_xics.c +1 -1
index f0c732e..fa7625d 100644
@@ -818,7 +818,7 @@ int kvmppc_xics_hcall(struct kvm_vcpu *vcpu, u32 req)
 	}
 
 	/* Check for real mode returning too hard */
-	if (xics->real_mode)
+	if (xics->real_mode && kvmppc_ops->is_hv_enabled)
 		return kvmppc_xics_rm_complete(vcpu, req);
 
 	switch (req) {
arch/powerpc/kvm/powerpc.c +1 -6
index 69b9305..00a96fc 100644
@@ -52,7 +52,6 @@ int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 /*
 * Common checks before entering the guest world.  Call with interrupts
 * disabled.
@@ -127,7 +126,6 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
 
 	return r;
 }
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
 
 int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 {
@@ -194,11 +192,9 @@ int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
 	if ((vcpu->arch.cpu_type != KVM_CPU_3S_64) && vcpu->arch.papr_enabled)
 		goto out;
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
 	/* HV KVM can only do PAPR mode for now */
-	if (!vcpu->arch.papr_enabled)
+	if (!vcpu->arch.papr_enabled && kvmppc_ops->is_hv_enabled)
 		goto out;
-#endif
 
 #ifdef CONFIG_KVM_BOOKE_HV
 	if (!cpu_has_feature(CPU_FTR_EMB_HV))
@@ -322,22 +318,26 @@ int kvm_dev_ioctl_check_extension(long ext)
 	case KVM_CAP_DEVICE_CTRL:
 		r = 1;
 		break;
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 	case KVM_CAP_PPC_PAIRED_SINGLES:
 	case KVM_CAP_PPC_OSI:
 	case KVM_CAP_PPC_GET_PVINFO:
 #if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
 	case KVM_CAP_SW_TLB:
 #endif
Quoted 8 lines

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 328

Quoted 6 lines
arch/powerpc/kvm/powerpc.c +2 -4
-#ifdef CONFIG_KVM_MPIC
-	case KVM_CAP_IRQ_MPIC:
-#endif
-		r = 1;
+		/* We support this only for PR */
+		r = !kvmppc_ops->is_hv_enabled;

Reading this, have you test compiled your code against e500 configs?

Not yet.

Quoted 22 lines
Quoted 22 lines
arch/powerpc/kvm/powerpc.c +8 -1
 		break;
+#ifdef CONFIG_KVM_MMIO
 	case KVM_CAP_COALESCED_MMIO:
 		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
 		break;
 #endif
+#ifdef CONFIG_KVM_MPIC
+	case KVM_CAP_IRQ_MPIC:
+		r = 1;
+		break;
+#endif
+
 #ifdef CONFIG_PPC_BOOK3S_64
 	case KVM_CAP_SPAPR_TCE:
 	case KVM_CAP_PPC_ALLOC_HTAB:
@@ -348,32 +348,37 @@ int kvm_dev_ioctl_check_extension(long ext)
 		r = 1;
 		break;
 #endif /* CONFIG_PPC_BOOK3S_64 */
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
Quoted 9 lines

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 356

Quoted 6 lines
arch/powerpc/kvm/powerpc.c +4 -1
 	case KVM_CAP_PPC_SMT:
-		r = threads_per_core;
+		if (kvmppc_ops->is_hv_enabled)
+			r = threads_per_core;
+		else
+			r = 0;

0? Or 1?

That check extension was not supported before on PR. So not sure what the value should be. May be 1 is better indicating we have one thread. Will change.

-aneesh

#14 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 217 lines · Aneesh Kumar K.V
Quoted 105 lines · Alexander Graf
Quoted 103 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This help us to identify whether we are running with hypervisor mode KVM enabled. The change is needed so that we can have both HV and PR kvm enabled in the same kernel.

If both HV and PR KVM are included, interrupts come in to the HV version of the kvmppc_interrupt code, which then jumps to the PR handler, renamed to kvmppc_interrupt_pr, if the guest is a PR guest.

Allowing both PR and HV in the same kernel required some changes to kvm_dev_ioctl_check_extension(), since the values returned now can't be selected with #ifdefs as much as previously. We look at is_hv_enabled to return the right value when checking for capabilities.For capabilities that are only provided by HV KVM, we return the HV value only if is_hv_enabled is true. For capabilities provided by PR KVM but not HV, we return the PR value only if is_hv_enabled is false.

---

arch/powerpc/include/asm/kvm_book3s.h   | 53 --------------------------------
arch/powerpc/include/asm/kvm_ppc.h      |  5 +--
arch/powerpc/kvm/Makefile               |  2 +-
arch/powerpc/kvm/book3s.c               | 44 +++++++++++++++++++++++++++
arch/powerpc/kvm/book3s_hv.c            |  1 +
arch/powerpc/kvm/book3s_hv_rmhandlers.S |  4 +++
arch/powerpc/kvm/book3s_pr.c            |  1 +
arch/powerpc/kvm/book3s_segment.S       |  7 ++++-
arch/powerpc/kvm/book3s_xics.c          |  2 +-
arch/powerpc/kvm/powerpc.c              | 54 ++++++++++++++++++---------------
10 files changed, 90 insertions(+), 83 deletions(-)
arch/powerpc/include/asm/kvm_book3s.h +0 -53
index 3efba3c..ba33c49 100644
@@ -297,59 +297,6 @@ static inline ulong kvmppc_get_fault_dar(struct kvm_vcpu *vcpu)
 	return vcpu->arch.fault_dar;
 }
 
-#ifdef CONFIG_KVM_BOOK3S_PR
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return to_book3s(vcpu)->hior;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-	if (pending_now)
-		vcpu->arch.shared->int_pending = 1;
-	else if (old_pending)
-		vcpu->arch.shared->int_pending = 0;
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	ulong crit_raw = vcpu->arch.shared->critical;
-	ulong crit_r1 = kvmppc_get_gpr(vcpu, 1);
-	bool crit;
-
-	/* Truncate crit indicators in 32 bit mode */
-	if (!(vcpu->arch.shared->msr & MSR_SF)) {
-		crit_raw &= 0xffffffff;
-		crit_r1 &= 0xffffffff;
-	}
-
-	/* Critical section when crit = r1 */
-	crit = (crit_raw = crit_r1);
-	/* ... and we're in supervisor mode */
-	crit = crit && !(vcpu->arch.shared->msr & MSR_PR);
-
-	return crit;
-}
-#else /* CONFIG_KVM_BOOK3S_PR */
-
-static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)
-{
-	return 0;
-}
-
-static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
-			unsigned long pending_now, unsigned long old_pending)
-{
-}
-
-static inline bool kvmppc_critical_section(struct kvm_vcpu *vcpu)
-{
-	return false;
-}
-#endif
-
 /* Magic register values loaded into r3 and r4 before the 'sc' assembly
 * instruction for the OSI hypercalls */
 #define OSI_SC_MAGIC_R3			0x113724FA
arch/powerpc/include/asm/kvm_ppc.h +0 -0
index 4d9641c..58e732f 100644
@@ -169,6 +169,7 @@ extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq);
 extern int kvmppc_xics_int_off(struct kvm *kvm, u32 irq);
 
Quoted 6 lines

arch/powerpc/include/asm/kvm_ppc.h extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq); · line 172

Quoted 2 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -0
 struct kvmppc_ops {
+	bool is_hv_enabled;

This doesn't really belong into an ops struct. Either you compare

if (kvmppc_ops = kvmppc_ops_pr)

will do that in the next update.

Quoted 13 lines

against a global known good ops struct or you put the hint somewhere into the kvm struct.

Quoted 10 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -1
 	int (*get_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*set_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*get_one_reg)(struct kvm_vcpu *vcpu, u64 id,
@@ -309,10 +310,10 @@ static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr)
 
 static inline u32 kvmppc_get_xics_latch(void)
 {
-	u32 xirr = get_paca()->kvm_hstate.saved_xirr;
+	u32 xirr;
 
Quoted 7 lines

arch/powerpc/include/asm/kvm_ppc.h static inline void kvmppc_set_xics_phys(int cpu, unsigned long addr) · line 315

Quoted 3 lines
arch/powerpc/include/asm/kvm_ppc.h +1 -1
+	xirr = get_paca()->kvm_hstate.saved_xirr;
 	get_paca()->kvm_hstate.saved_xirr = 0;
-

I don't see any functionality change here?

Quoted 1 line
arch/powerpc/include/asm/kvm_ppc.h +0 -0
 	return xirr;

Mistake on my side, I had a if condition in there before, which i later removed. But forgot to move the assignment back. Will fix.

Quoted 10 lines
Quoted 10 lines
arch/powerpc/include/asm/kvm_ppc.h +0 -0
 }
 
arch/powerpc/kvm/Makefile +0 -0
index c343793..a514ecd 100644
@@ -77,7 +77,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o

endif

Quoted 5 lines

arch/powerpc/kvm/Makefile kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \ · line 80

Quoted 2 lines
arch/powerpc/kvm/Makefile +1 -1
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV) := \
+kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \

This change looks unrelated?

yes. Will move it to the correct patch

Quoted 11 lines
Quoted 11 lines
	book3s_hv.o \
	book3s_hv_interrupts.o \
	book3s_64_mmu_hv.o
arch/powerpc/kvm/book3s.c +0 -0
index bdc3f95..12f94bf 100644
@@ -69,6 +69,50 @@ void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu)
 {
 }
 
Quoted 4 lines

arch/powerpc/kvm/book3s.c void kvmppc_core_load_guest_debugstate(struct kvm_vcpu *vcpu) · line 72

Quoted 1 line
arch/powerpc/kvm/book3s.c +1 -0
+static inline unsigned long kvmppc_interrupt_offset(struct kvm_vcpu *vcpu)

Please drop the "inline" keyword on functions in .c files :). That way the compiler tells us about unused functions.

ok .

Quoted 18 lines
Quoted 18 lines
arch/powerpc/kvm/book3s.c +17 -0
+{
+	if (!kvmppc_ops->is_hv_enabled)
+		return to_book3s(vcpu)->hior;
+	return 0;
+}
+
+static inline void kvmppc_update_int_pending(struct kvm_vcpu *vcpu,
+			unsigned long pending_now, unsigned long old_pending)
+{
+	if (kvmppc_ops->is_hv_enabled)
+		return;
+	if (pending_now)
+		vcpu->arch.shared->int_pending = 1;
+	else if (old_pending)
+		vcpu->arch.shared->int_pending = 0;
+}
+

....

Quoted 8 lines
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 10 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 2 lines

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Ah, because we're always jumping to kvmppc_interrupt. Can we make this slightly less magical? How about we always call kvmppc_interrupt_hv when CONFIG_KVM_BOOK3S_HV_POSSIBLE and always kvmppc_interrupt_pr when CONFIG_KVM_BOOK3S_PR_POSSIBLE and then branch to kvmppc_interrupt_pr from kvmppc_interrupt_hv?

IMHO that would make the code flow more obvious.

Quoted 74 lines
Quoted 64 lines
Quoted 64 lines
arch/powerpc/kvm/book3s_segment.S +1 -1
-
+#endif
	/* Register usage at this point:
	 *
	 * SPRG_SCRATCH0  = guest R13
arch/powerpc/kvm/book3s_xics.c +1 -1
index f0c732e..fa7625d 100644
@@ -818,7 +818,7 @@ int kvmppc_xics_hcall(struct kvm_vcpu *vcpu, u32 req)
 	}
 
 	/* Check for real mode returning too hard */
-	if (xics->real_mode)
+	if (xics->real_mode && kvmppc_ops->is_hv_enabled)
 		return kvmppc_xics_rm_complete(vcpu, req);
 
 	switch (req) {
arch/powerpc/kvm/powerpc.c +1 -6
index 69b9305..00a96fc 100644
@@ -52,7 +52,6 @@ int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 /*
 * Common checks before entering the guest world.  Call with interrupts
 * disabled.
@@ -127,7 +126,6 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
 
 	return r;
 }
-#endif /* CONFIG_KVM_BOOK3S_64_HV */
 
 int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 {
@@ -194,11 +192,9 @@ int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
 	if ((vcpu->arch.cpu_type != KVM_CPU_3S_64) && vcpu->arch.papr_enabled)
 		goto out;
 
-#ifdef CONFIG_KVM_BOOK3S_64_HV
 	/* HV KVM can only do PAPR mode for now */
-	if (!vcpu->arch.papr_enabled)
+	if (!vcpu->arch.papr_enabled && kvmppc_ops->is_hv_enabled)
 		goto out;
-#endif
 
 #ifdef CONFIG_KVM_BOOKE_HV
 	if (!cpu_has_feature(CPU_FTR_EMB_HV))
@@ -322,22 +318,26 @@ int kvm_dev_ioctl_check_extension(long ext)
 	case KVM_CAP_DEVICE_CTRL:
 		r = 1;
 		break;
-#ifndef CONFIG_KVM_BOOK3S_64_HV
 	case KVM_CAP_PPC_PAIRED_SINGLES:
 	case KVM_CAP_PPC_OSI:
 	case KVM_CAP_PPC_GET_PVINFO:
 #if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
 	case KVM_CAP_SW_TLB:
 #endif
Quoted 8 lines

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 328

Quoted 6 lines
arch/powerpc/kvm/powerpc.c +2 -4
-#ifdef CONFIG_KVM_MPIC
-	case KVM_CAP_IRQ_MPIC:
-#endif
-		r = 1;
+		/* We support this only for PR */
+		r = !kvmppc_ops->is_hv_enabled;

Reading this, have you test compiled your code against e500 configs?

Quoted 1 line

Not yet.

Please do so - for every patch in your series. If you like I can give you my example configs I usually use to test compile this.

Quoted 34 lines
Quoted 24 lines
Quoted 24 lines
arch/powerpc/kvm/powerpc.c +8 -1
 		break;
+#ifdef CONFIG_KVM_MMIO
 	case KVM_CAP_COALESCED_MMIO:
 		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
 		break;
 #endif
+#ifdef CONFIG_KVM_MPIC
+	case KVM_CAP_IRQ_MPIC:
+		r = 1;
+		break;
+#endif
+
 #ifdef CONFIG_PPC_BOOK3S_64
 	case KVM_CAP_SPAPR_TCE:
 	case KVM_CAP_PPC_ALLOC_HTAB:
@@ -348,32 +348,37 @@ int kvm_dev_ioctl_check_extension(long ext)
 		r = 1;
 		break;
 #endif /* CONFIG_PPC_BOOK3S_64 */
-#ifdef CONFIG_KVM_BOOK3S_64_HV
+#ifdef CONFIG_KVM_BOOK3S_HV
Quoted 9 lines

arch/powerpc/kvm/powerpc.c int kvm_dev_ioctl_check_extension(long ext) · line 356

Quoted 6 lines
arch/powerpc/kvm/powerpc.c +4 -1
 	case KVM_CAP_PPC_SMT:
-		r = threads_per_core;
+		if (kvmppc_ops->is_hv_enabled)
+			r = threads_per_core;
+		else
+			r = 0;

0? Or 1?

Quoted 3 lines

That check extension was not supported before on PR. So not sure what the value should be. May be 1 is better indicating we have one thread. Will change.

Ah, the default case (cap unknown) returns 0, so this is fine.

Alex

#15 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 27 lines · Alexander Graf
Quoted 22 lines · Aneesh Kumar K.V
Quoted 8 lines · Alexander Graf
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 10 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 2 lines

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Quoted 7 lines

Ah, because we're always jumping to kvmppc_interrupt. Can we make this slightly less magical? How about we always call kvmppc_interrupt_hv when CONFIG_KVM_BOOK3S_HV_POSSIBLE and always kvmppc_interrupt_pr when CONFIG_KVM_BOOK3S_PR_POSSIBLE and then branch to kvmppc_interrupt_pr from kvmppc_interrupt_hv?

IMHO that would make the code flow more obvious.

To make sure I understand you correctly, what you are suggesting is to update __KVM_HANDLER to call kvmppc_interupt_pr when HV is not enabled ?

-aneesh

#16 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 32 lines · Aneesh Kumar K.V
Quoted 30 lines · Alexander Graf
Quoted 20 lines · Aneesh Kumar K.V
Quoted 8 lines · Alexander Graf
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 9 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 2 lines

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Ah, because we're always jumping to kvmppc_interrupt. Can we make this slightly less magical? How about we always call kvmppc_interrupt_hv when CONFIG_KVM_BOOK3S_HV_POSSIBLE and always kvmppc_interrupt_pr when CONFIG_KVM_BOOK3S_PR_POSSIBLE and then branch to kvmppc_interrupt_pr from kvmppc_interrupt_hv?

Quoted 5 lines
Quoted 1 line

IMHO that would make the code flow more obvious.

To make sure I understand you correctly, what you are suggesting is to update __KVM_HANDLER to call kvmppc_interupt_pr when HV is not enabled ?

Yes, I think that makes the code flow more obvious. Every function always has the same name regardless of config options then.

Alex

#17 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 33 lines · Alexander Graf
Quoted 32 lines · Aneesh Kumar K.V
Quoted 30 lines · Alexander Graf
Quoted 20 lines · Aneesh Kumar K.V
Quoted 8 lines · Alexander Graf
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 9 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 2 lines

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Ah, because we're always jumping to kvmppc_interrupt. Can we make this slightly less magical? How about we always call kvmppc_interrupt_hv when CONFIG_KVM_BOOK3S_HV_POSSIBLE and always kvmppc_interrupt_pr when CONFIG_KVM_BOOK3S_PR_POSSIBLE and then branch to kvmppc_interrupt_pr from kvmppc_interrupt_hv?

Quoted 9 lines
Quoted 5 lines
Quoted 1 line

IMHO that would make the code flow more obvious.

To make sure I understand you correctly, what you are suggesting is to update __KVM_HANDLER to call kvmppc_interupt_pr when HV is not enabled ?

Yes, I think that makes the code flow more obvious. Every function always has the same name regardless of config options then.

Something like this ( btw non tested )

arch/powerpc/include/asm/exception-64s.h +11 -0
index cca12f0..0b798d4 100644
@@ -198,6 +198,17 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	cmpwi	r10,0;							\
 	bne	do_kvm_##n
 
+#ifdef CONFIG_KVM_BOOK3S_HV
+/*
+ * If hv is possible, interrupts come into to the hv version
+ * of the kvmppc_interrupt code, which then jumps to the PR handler,
+ * kvmppc_interrupt_pr, if the guest is a PR guest.
+ */
+#define kvmppc_interrupt kvmppc_interrupt_hv
+#else
+#define kvmppc_interrupt kvmppc_interrupt_pr
+#endif
+
 #define __KVM_HANDLER(area, h, n)					\
 do_kvm_##n:								\
 	BEGIN_FTR_SECTION_NESTED(947)					\
arch/powerpc/kvm/book3s_hv_rmhandlers.S +7 -2
index 5ede7fc..2eb6622 100644
@@ -563,8 +563,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 /*
  * We come here from the first-level interrupt handlers.
  */
-	.globl	kvmppc_interrupt
-kvmppc_interrupt:
+	.globl	kvmppc_interrupt_hv
+kvmppc_interrupt_hv:
 	/*
 	 * Register contents:
 	 * R12		= interrupt vector
@@ -577,6 +577,11 @@ kvmppc_interrupt:
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
+#ifdef CONFIG_KVM_BOOK3S_PR
+	cmpwi	r9, KVM_GUEST_MODE_GUEST
+	ld	r9, HSTATE_SCRATCH2(r13)
+	beq	kvmppc_interrupt_pr
+#endif
 	/* We're now back in the host but in guest MMU context */
 	li	r9, KVM_GUEST_MODE_HOST_HV
 	stb	r9, HSTATE_IN_GUEST(r13)
arch/powerpc/kvm/book3s_segment.S +2 -2
index 1abe478..bc50c97 100644
@@ -161,8 +161,8 @@ kvmppc_handler_trampoline_enter_end:
 .global kvmppc_handler_trampoline_exit
 kvmppc_handler_trampoline_exit:
 
-.global kvmppc_interrupt
-kvmppc_interrupt:
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
 
 	/* Register usage at this point:
 	 *
#18 lore raw

Re: [RFC PATCH 06/11] kvm: powerpc: book3s: Add is_hv_enabled to kvmppc_ops

Quoted 35 lines · Aneesh Kumar K.V
Quoted 33 lines · Alexander Graf
Quoted 32 lines · Aneesh Kumar K.V
Quoted 30 lines · Alexander Graf
Quoted 20 lines · Aneesh Kumar K.V
Quoted 8 lines · Alexander Graf
Quoted 8 lines
arch/powerpc/kvm/book3s_segment.S +0 -0
index 1abe478..e0229dd 100644
@@ -161,9 +161,14 @@ kvmppc_handler_trampoline_enter_end:

.global kvmppc_handler_trampoline_exit
kvmppc_handler_trampoline_exit:

Quoted 9 lines
Quoted 7 lines
arch/powerpc/kvm/book3s_segment.S +5 -0
+#if defined(CONFIG_KVM_BOOK3S_HV)
+.global kvmppc_interrupt_pr
+kvmppc_interrupt_pr:
+	ld	r9, HSTATE_SCRATCH2(r13)
+#else

.global kvmppc_interrupt
kvmppc_interrupt:

Just always call it kvmppc_interrupt_pr and thus share at least that part of the code :).

Quoted 2 lines

But if i don't have HV enabled, we don't compile book3s_hv_rmhandlers.S Hence don't have the kvmppc_interrupt symbol defined.

Ah, because we're always jumping to kvmppc_interrupt. Can we make this slightly less magical? How about we always call kvmppc_interrupt_hv when CONFIG_KVM_BOOK3S_HV_POSSIBLE and always kvmppc_interrupt_pr when CONFIG_KVM_BOOK3S_PR_POSSIBLE and then branch to kvmppc_interrupt_pr from kvmppc_interrupt_hv?

Quoted 8 lines
Quoted 7 lines
Quoted 4 lines
Quoted 1 line

IMHO that would make the code flow more obvious.

To make sure I understand you correctly, what you are suggesting is to update __KVM_HANDLER to call kvmppc_interupt_pr when HV is not enabled ?

Yes, I think that makes the code flow more obvious. Every function always has the same name regardless of config options then.

Something like this ( btw non tested )

Yes :).

Alex

#19 lore raw

RFC PATCH 7/11 kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---
 arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
 arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
 arch/powerpc/kvm/book3s_pr.c          |   3 +-
 arch/powerpc/kvm/trace.h              | 234 +--------------------------
 arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
 5 files changed, 308 insertions(+), 230 deletions(-)
 create mode 100644 arch/powerpc/kvm/trace_pr.h
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +7 -227

276 lines not on this page · raw message

arch/powerpc/kvm/trace_pr.h new file +297 -0

297 lines not on this page · raw message

Signature · 1 line
1.8.1.2
#20 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 61 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---

arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
arch/powerpc/kvm/book3s_pr.c          |   3 +-
arch/powerpc/kvm/trace.h              | 234 +--------------------------
arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
5 files changed, 308 insertions(+), 230 deletions(-)

create mode 100644 arch/powerpc/kvm/trace_pr.h

arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 7 lines
arch/powerpc/kvm/trace.h +4 -0
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

If your goal is to make it more obvious whether we are tracing in PR or HV land (which is a reasonable goal), then you should also split off all non-common trace points into a special hv trace header so that it's obvious whether we are looking at HV or PR.

Alex

#21 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 66 lines · Alexander Graf
Quoted 64 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---

arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
arch/powerpc/kvm/book3s_pr.c          |   3 +-
arch/powerpc/kvm/trace.h              | 234 +--------------------------
arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
5 files changed, 308 insertions(+), 230 deletions(-)

create mode 100644 arch/powerpc/kvm/trace_pr.h

arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 8 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

Quoted 5 lines

If your goal is to make it more obvious whether we are tracing in PR or HV land (which is a reasonable goal), then you should also split off all non-common trace points into a special hv trace header so that it's obvious whether we are looking at HV or PR.

-aneesh

#22 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 77 lines · Aneesh Kumar K.V
Quoted 66 lines · Alexander Graf
Quoted 64 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---

arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
arch/powerpc/kvm/book3s_pr.c          |   3 +-
arch/powerpc/kvm/trace.h              | 234 +--------------------------
arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
5 files changed, 308 insertions(+), 230 deletions(-)

create mode 100644 arch/powerpc/kvm/trace_pr.h

arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 8 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

Quoted 5 lines

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

I fail to see why we wouldn't have a trace_hv.h file then, as that can also be built as a module, no? And at that point I don't see why we would need any conditionals at all in trace.h anymore, as it would only cover generic code.

Alex

#23 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 85 lines · Alexander Graf
Quoted 77 lines · Aneesh Kumar K.V
Quoted 66 lines · Alexander Graf
Quoted 64 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---

arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
arch/powerpc/kvm/book3s_pr.c          |   3 +-
arch/powerpc/kvm/trace.h              | 234 +--------------------------
arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
5 files changed, 308 insertions(+), 230 deletions(-)

create mode 100644 arch/powerpc/kvm/trace_pr.h

arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 8 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

Quoted 5 lines

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

Quoted 4 lines

I fail to see why we wouldn't have a trace_hv.h file then, as that can also be built as a module, no? And at that point I don't see why we would need any conditionals at all in trace.h anymore, as it would only cover generic code.

Currently HV module is not using any tracepoints. Once it start using tracepoints we would have trace_hv.h

-aneesh

#24 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 84 lines · Aneesh Kumar K.V
Quoted 82 lines · Alexander Graf
Quoted 80 lines · Aneesh Kumar K.V
Quoted 66 lines · Alexander Graf
Quoted 64 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V"<[email protected]>

This patch moves PR related tracepoints to a separate header. This enables in converting PR to a kernel module which will be done in later patches

---

arch/powerpc/kvm/book3s_64_mmu_host.c |   2 +-
arch/powerpc/kvm/book3s_mmu_hpte.c    |   2 +-
arch/powerpc/kvm/book3s_pr.c          |   3 +-
arch/powerpc/kvm/trace.h              | 234 +--------------------------
arch/powerpc/kvm/trace_pr.h           | 297 ++++++++++++++++++++++++++++++++++
5 files changed, 308 insertions(+), 230 deletions(-)

create mode 100644 arch/powerpc/kvm/trace_pr.h

arch/powerpc/kvm/book3s_64_mmu_host.c +1 -1
index 329a978..fd5b393 100644
@@ -27,7 +27,7 @@
 #include<asm/machdep.h>
 #include<asm/mmu_context.h>
 #include<asm/hw_irq.h>
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE 12
 
arch/powerpc/kvm/book3s_mmu_hpte.c +1 -1
index d2d280b..4556168 100644
@@ -28,7 +28,7 @@
 #include<asm/mmu_context.h>
 #include<asm/hw_irq.h>
 
-#include "trace.h"
+#include "trace_pr.h"
 
 #define PTE_SIZE	12
 
arch/powerpc/kvm/book3s_pr.c +2 -1
index 2a97279..99d0839 100644
@@ -41,7 +41,8 @@
 #include<linux/vmalloc.h>
 #include<linux/highmem.h>
 
-#include "trace.h"
+#define CREATE_TRACE_POINTS
+#include "trace_pr.h"
 
 /* #define EXIT_DEBUG */
 /* #define DEBUG_EXT */
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 7 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

Quoted 6 lines
Quoted 4 lines

I fail to see why we wouldn't have a trace_hv.h file then, as that can also be built as a module, no? And at that point I don't see why we would need any conditionals at all in trace.h anymore, as it would only cover generic code.

Currently HV module is not using any tracepoints. Once it start using tracepoints we would have trace_hv.h

So why would there be an #ifndef in trace.h?

Alex

#25 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 22 lines · Alexander Graf
Quoted 21 lines · Aneesh Kumar K.V
Quoted 20 lines · Alexander Graf
Quoted 20 lines
Quoted 8 lines
Quoted 8 lines
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 7 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

Quoted 9 lines
Quoted 6 lines
Quoted 4 lines

I fail to see why we wouldn't have a trace_hv.h file then, as that can also be built as a module, no? And at that point I don't see why we would need any conditionals at all in trace.h anymore, as it would only cover generic code.

Currently HV module is not using any tracepoints. Once it start using tracepoints we would have trace_hv.h

So why would there be an #ifndef in trace.h?

to handle things like trace_kvm_exit in booke.c. For that we still don't have kernel module, and booke.c include trace.h.

-aneesh

#26 lore raw

Re: [RFC PATCH 07/11] kvm: powerpc: book3s: pr: move PR related tracepoints to a separate header

Quoted 28 lines · Aneesh Kumar K.V
Quoted 26 lines · Alexander Graf
Quoted 25 lines · Aneesh Kumar K.V
Quoted 24 lines · Alexander Graf
Quoted 15 lines
Quoted 8 lines
Quoted 8 lines
arch/powerpc/kvm/trace.h +0 -0
index a088e9a..7d5a136 100644
@@ -85,6 +85,12 @@ TRACE_EVENT(kvm_ppc_instr,
 	{41, "HV_PRIV"}
 #endif
 
Quoted 7 lines

arch/powerpc/kvm/trace.h TRACE_EVENT(kvm_ppc_instr, · line 91

Quoted 4 lines
arch/powerpc/kvm/trace.h +4 -0
+#ifndef CONFIG_KVM_BOOK3S_PR
+/*
+ * For pr we define this in trace_pr.h since it pr can be built as
+ * a module

Not sure I understand the need. If the config option is available, so should the struct field. Worst case that happens with HV is that we get empty shadow_srr1 values in our trace, no?

Quoted 5 lines

That is not the real reason. trace.h get built as part of kvm.ko or as part of kernel. These trace functions actually get called from kvm-pr.ko. To make they build i would either need EXPORT_SYMBOL or move the definition of them to kvm-pr.ko. I did the later and moved only pr related traces to kvm-pr.ko

I fail to see why we wouldn't have a trace_hv.h file then, as that can also be built as a module, no? And at that point I don't see why we would need any conditionals at all in trace.h anymore, as it would only cover generic code.

Quoted 6 lines
Quoted 4 lines
Quoted 2 lines

Currently HV module is not using any tracepoints. Once it start using tracepoints we would have trace_hv.h

So why would there be an #ifndef in trace.h?

to handle things like trace_kvm_exit in booke.c. For that we still don't have kernel module, and booke.c include trace.h.

Ah, so it's for booke! :)

Just move them to trace_booke.h then to clarify things.

Alex

#27 lore raw

RFC PATCH 8/11 kvm: powerpc: book3s: Support building HV and PR KVM as module

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

---
 arch/powerpc/kvm/Kconfig              |  6 +++---
 arch/powerpc/kvm/Makefile             | 12 ++++++++----
 arch/powerpc/kvm/book3s.c             | 19 ++++++++++++++++++-
 arch/powerpc/kvm/book3s_64_mmu_host.c |  1 +
 arch/powerpc/kvm/book3s_emulate.c     |  2 +-
 arch/powerpc/kvm/book3s_hv.c          |  4 ++++
 arch/powerpc/kvm/book3s_pr.c          |  5 ++++-
 arch/powerpc/kvm/book3s_rtas.c        |  1 +
 arch/powerpc/kvm/emulate.c            |  1 +
 arch/powerpc/kvm/powerpc.c            |  6 ++++++
 virt/kvm/kvm_main.c                   |  3 +++
 11 files changed, 50 insertions(+), 10 deletions(-)
arch/powerpc/kvm/Kconfig +3 -3
index 5c62459..f63fd8f 100644
@@ -73,7 +73,7 @@ config KVM_BOOK3S_64
 	  If unsure, say N.
 
 config KVM_BOOK3S_64_HV
-	bool "KVM support for POWER7 and PPC970 using hypervisor mode in host"
+	tristate "KVM support for POWER7 and PPC970 using hypervisor mode in host"
 	depends on KVM_BOOK3S_64
 	select KVM_BOOK3S_HV
 	select MMU_NOTIFIER
@@ -94,8 +94,8 @@ config KVM_BOOK3S_64_HV
 	  If unsure, say N.
 
 config KVM_BOOK3S_64_PR
-	bool "KVM support without using hypervisor mode in host"
-	depends on KVM_BOOK3S_64 && !KVM_BOOK3S_64_HV
+	tristate "KVM support without using hypervisor mode in host"
+	depends on KVM_BOOK3S_64
 	select KVM_BOOK3S_PR
 	---help---
 	  Support running guest kernels in virtual machines on processors
arch/powerpc/kvm/Makefile +8 -4
index a514ecd..861b8da 100644
@@ -56,8 +56,7 @@ kvm-objs-$(CONFIG_KVM_E500MC) := $(kvm-e500mc-objs)
 kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) := \
 	book3s_64_vio_hv.o
 
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_PR) := \
-	$(KVM)/coalesced_mmio.o \
+kvm-pr-y := \
 	fpu.o \
 	book3s_paired_singles.o \
 	book3s_pr.o \
@@ -77,7 +76,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o
 endif
 
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \
+kvm-hv-y += \
 	book3s_hv.o \
 	book3s_hv_interrupts.o \
 	book3s_64_mmu_hv.o
@@ -85,13 +84,15 @@ kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \
 kvm-book3s_64-builtin-xics-objs-$(CONFIG_KVM_XICS) := \
 	book3s_hv_rm_xics.o
 
-kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HV) += \
+ifdef CONFIG_KVM_BOOK3S_HV
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_hv_rmhandlers.o \
 	book3s_hv_rm_mmu.o \
 	book3s_hv_ras.o \
 	book3s_hv_builtin.o \
 	book3s_hv_cma.o \
 	$(kvm-book3s_64-builtin-xics-objs-y)
+endif
 
 kvm-book3s_64-objs-$(CONFIG_KVM_XICS) += \
 	book3s_xics.o
@@ -132,4 +133,7 @@ obj-$(CONFIG_KVM_E500MC) += kvm.o
 obj-$(CONFIG_KVM_BOOK3S_64) += kvm.o
 obj-$(CONFIG_KVM_BOOK3S_32) += kvm.o
 
+obj-$(CONFIG_KVM_BOOK3S_64_PR) += kvm-pr.o
+obj-$(CONFIG_KVM_BOOK3S_64_HV) += kvm-hv.o
+
 obj-y += $(kvm-book3s_64-builtin-objs-y)
arch/powerpc/kvm/book3s.c +18 -1
index 12f94bf..ca617e1 100644
@@ -170,13 +170,14 @@ void kvmppc_book3s_queue_irqprio(struct kvm_vcpu *vcpu, unsigned int vec)
 	printk(KERN_INFO "Queueing interrupt %x\n", vec);
 #endif
 }
-
+EXPORT_SYMBOL_GPL(kvmppc_book3s_queue_irqprio);
 
 void kvmppc_core_queue_program(struct kvm_vcpu *vcpu, ulong flags)
 {
 	/* might as well deliver this straight away */
 	kvmppc_inject_interrupt(vcpu, BOOK3S_INTERRUPT_PROGRAM, flags);
 }
+EXPORT_SYMBOL_GPL(kvmppc_core_queue_program);
 
 void kvmppc_core_queue_dec(struct kvm_vcpu *vcpu)
 {
@@ -351,6 +352,7 @@ pfn_t kvmppc_gfn_to_pfn(struct kvm_vcpu *vcpu, gfn_t gfn)
 
 	return gfn_to_pfn(vcpu->kvm, gfn);
 }
+EXPORT_SYMBOL_GPL(kvmppc_gfn_to_pfn);
 
 static int kvmppc_xlate(struct kvm_vcpu *vcpu, ulong eaddr, bool data,
 			 struct kvmppc_pte *pte)
@@ -418,6 +420,7 @@ int kvmppc_st(struct kvm_vcpu *vcpu, ulong *eaddr, int size, void *ptr,
 
 	return EMULATE_DONE;
 }
+EXPORT_SYMBOL_GPL(kvmppc_st);
 
 int kvmppc_ld(struct kvm_vcpu *vcpu, ulong *eaddr, int size, void *ptr,
 		      bool data)
@@ -448,6 +451,7 @@ nopte:
 mmio:
 	return EMULATE_DO_MMIO;
 }
+EXPORT_SYMBOL_GPL(kvmppc_ld);
 
 int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
 {
@@ -693,6 +697,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u64 msr)
 {
 	kvmppc_ops->set_msr(vcpu, msr);
 }
+EXPORT_SYMBOL_GPL(kvmppc_set_msr);
 
 int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
 {
@@ -774,6 +779,7 @@ int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
 {
 	return kvmppc_ops->unmap_hva(kvm, hva);
 }
+EXPORT_SYMBOL_GPL(kvm_unmap_hva);
 
 int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
 {
@@ -825,3 +831,14 @@ int kvmppc_core_check_processor_compat(void)
 {
 	return kvmppc_ops->check_processor_compat();
 }
+
+EXPORT_SYMBOL_GPL(kvm_get_dirty_log);
+EXPORT_SYMBOL_GPL(kvmppc_core_pending_dec);
+EXPORT_SYMBOL_GPL(kvmppc_init_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_core_dequeue_dec);
+EXPORT_SYMBOL_GPL(kvmppc_claim_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_alloc_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_core_prepare_to_enter);
+EXPORT_SYMBOL_GPL(kvmppc_core_queue_dec);
+EXPORT_SYMBOL_GPL(kvmppc_free_lpid);
+
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -0
index fd5b393..775d368 100644
@@ -27,6 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
+
 #include "trace_pr.h"
 
 #define PTE_SIZE 12
arch/powerpc/kvm/book3s_emulate.c +1 -1
index b9841ad..20d03c2 100644
@@ -172,7 +172,7 @@ int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			vcpu->arch.mmu.tlbie(vcpu, addr, large);
 			break;
 		}
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_KVM_BOOK3S_PR
 		case OP_31_XOP_FAKE_SC1:
 		{
 			/* SC 1 papr hypercalls */
arch/powerpc/kvm/book3s_hv.c +4 -0
index eec353d..ff57be8 100644
@@ -53,6 +53,8 @@
 #include <linux/highmem.h>
 #include <linux/hugetlb.h>
 
+#include <linux/module.h>
+
 /* #define EXIT_DEBUG */
 /* #define EXIT_DEBUG_SIMPLE */
 /* #define EXIT_DEBUG_INT */
@@ -2104,3 +2106,5 @@ static void kvmppc_book3s_exit_hv(void)
 
 module_init(kvmppc_book3s_init_hv);
 module_exit(kvmppc_book3s_exit_hv);
+
+MODULE_LICENSE("GPL");
arch/powerpc/kvm/book3s_pr.c +4 -1
index 99d0839..df48d89 100644
@@ -40,6 +40,7 @@
 #include <linux/sched.h>
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
+#include <linux/module.h>
 
 #define CREATE_TRACE_POINTS
 #include "trace_pr.h"
@@ -871,7 +872,7 @@ program_interrupt:
 			ulong cmd = kvmppc_get_gpr(vcpu, 3);
 			int i;
 
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_PPC_BOOK3S_64
 			if (kvmppc_h_pr(vcpu, cmd) == EMULATE_DONE) {
 				r = RESUME_GUEST;
 				break;
@@ -1567,3 +1568,5 @@ static void kvmppc_book3s_exit_pr(void)
 
 module_init(kvmppc_book3s_init_pr);
 module_exit(kvmppc_book3s_exit_pr);
+
+MODULE_LICENSE("GPL");
arch/powerpc/kvm/book3s_rtas.c +1 -0
index 3219ba8..cf95cde 100644
@@ -260,6 +260,7 @@ fail:
 	 */
 	return rc;
 }
+EXPORT_SYMBOL_GPL(kvmppc_rtas_hcall);
 
 void kvmppc_rtas_tokens_free(struct kvm *kvm)
 {
arch/powerpc/kvm/emulate.c +1 -0
index 81762eb..de9a340 100644
@@ -483,3 +483,4 @@ int kvmppc_emulate_instruction(struct kvm_run *run, struct kvm_vcpu *vcpu)
 
 	return emulated;
 }
+EXPORT_SYMBOL_GPL(kvmppc_emulate_instruction);
arch/powerpc/kvm/powerpc.c +6 -0
index 00a96fc..3019edc 100644
@@ -126,6 +126,7 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_prepare_to_enter);
 
 int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 {
@@ -179,6 +180,7 @@ int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_kvm_pv);
 
 int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
 {
@@ -207,6 +209,7 @@ out:
 	vcpu->arch.sane = r;
 	return r ? 0 : -EINVAL;
 }
+EXPORT_SYMBOL_GPL(kvmppc_sanity_check);
 
 int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
 {
@@ -241,6 +244,7 @@ int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_emulate_mmio);
 
 int kvm_arch_hardware_enable(void *garbage)
 {
@@ -665,6 +669,7 @@ int kvmppc_handle_load(struct kvm_run *run, struct kvm_vcpu *vcpu,
 
 	return EMULATE_DO_MMIO;
 }
+EXPORT_SYMBOL_GPL(kvmppc_handle_load);
 
 /* Same as above, but sign extends */
 int kvmppc_handle_loads(struct kvm_run *run, struct kvm_vcpu *vcpu,
@@ -726,6 +731,7 @@ int kvmppc_handle_store(struct kvm_run *run, struct kvm_vcpu *vcpu,
 
 	return EMULATE_DO_MMIO;
 }
+EXPORT_SYMBOL_GPL(kvmppc_handle_store);
 
 int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
 {
virt/kvm/kvm_main.c +3 -0
index bf040c4..66df1d2 100644
@@ -186,6 +186,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)
 		++kvm->stat.remote_tlb_flush;
 	cmpxchg(&kvm->tlbs_dirty, dirty_count, 0);
 }
+EXPORT_SYMBOL_GPL(kvm_flush_remote_tlbs);
 
 void kvm_reload_remote_mmus(struct kvm *kvm)
 {
@@ -1648,6 +1649,7 @@ void mark_page_dirty(struct kvm *kvm, gfn_t gfn)
 	memslot = gfn_to_memslot(kvm, gfn);
 	mark_page_dirty_in_slot(kvm, memslot, gfn);
 }
+EXPORT_SYMBOL_GPL(mark_page_dirty);
 
 /*
  * The vCPU has executed a HLT instruction with in-kernel mode enabled.
@@ -1673,6 +1675,7 @@ void kvm_vcpu_block(struct kvm_vcpu *vcpu)
 
 	finish_wait(&vcpu->wq, &wait);
 }
+EXPORT_SYMBOL_GPL(kvm_vcpu_block);
 
 #ifndef CONFIG_S390
 /*
Signature · 1 line
1.8.1.2
#28 lore raw

Re: [RFC PATCH 08/11] kvm: powerpc: book3s: Support building HV and PR KVM as module

Quoted 169 lines · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

---

arch/powerpc/kvm/Kconfig              |  6 +++---
arch/powerpc/kvm/Makefile             | 12 ++++++++----
arch/powerpc/kvm/book3s.c             | 19 ++++++++++++++++++-
arch/powerpc/kvm/book3s_64_mmu_host.c |  1 +
arch/powerpc/kvm/book3s_emulate.c     |  2 +-
arch/powerpc/kvm/book3s_hv.c          |  4 ++++
arch/powerpc/kvm/book3s_pr.c          |  5 ++++-
arch/powerpc/kvm/book3s_rtas.c        |  1 +
arch/powerpc/kvm/emulate.c            |  1 +
arch/powerpc/kvm/powerpc.c            |  6 ++++++
virt/kvm/kvm_main.c                   |  3 +++
11 files changed, 50 insertions(+), 10 deletions(-)
arch/powerpc/kvm/Kconfig +0 -0
index 5c62459..f63fd8f 100644
@@ -73,7 +73,7 @@ config KVM_BOOK3S_64
 	  If unsure, say N.
 

config KVM_BOOK3S_64_HV

arch/powerpc/kvm/Kconfig +1 -1
-	bool "KVM support for POWER7 and PPC970 using hypervisor mode in host"
+	tristate "KVM support for POWER7 and PPC970 using hypervisor mode in host"

depends on KVM_BOOK3S_64
select KVM_BOOK3S_HV
select MMU_NOTIFIER

arch/powerpc/kvm/Kconfig +0 -0
@@ -94,8 +94,8 @@ config KVM_BOOK3S_64_HV
 	  If unsure, say N.
 

config KVM_BOOK3S_64_PR

arch/powerpc/kvm/Kconfig +2 -2
-	bool "KVM support without using hypervisor mode in host"
-	depends on KVM_BOOK3S_64 && !KVM_BOOK3S_64_HV
+	tristate "KVM support without using hypervisor mode in host"
+	depends on KVM_BOOK3S_64
	select KVM_BOOK3S_PR
	---help---
	  Support running guest kernels in virtual machines on processors
arch/powerpc/kvm/Makefile +1 -2
index a514ecd..861b8da 100644
@@ -56,8 +56,7 @@ kvm-objs-$(CONFIG_KVM_E500MC) := $(kvm-e500mc-objs)
 kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) := \
 	book3s_64_vio_hv.o
 
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_PR) := \
-	$(KVM)/coalesced_mmio.o \
+kvm-pr-y := \
 	fpu.o \
 	book3s_paired_singles.o \
 	book3s_pr.o \
@@ -77,7 +76,7 @@ kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_rmhandlers.o

endif

arch/powerpc/kvm/Makefile +1 -1
-kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \
+kvm-hv-y += \
	book3s_hv.o \
	book3s_hv_interrupts.o \
	book3s_64_mmu_hv.o
arch/powerpc/kvm/Makefile +3 -1
@@ -85,13 +84,15 @@ kvm-book3s_64-objs-$(CONFIG_KVM_BOOK3S_64_HV)  += \
 kvm-book3s_64-builtin-xics-objs-$(CONFIG_KVM_XICS) := \
 	book3s_hv_rm_xics.o
 
-kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HV) += \
+ifdef CONFIG_KVM_BOOK3S_HV
+kvm-book3s_64-builtin-objs-$(CONFIG_KVM_BOOK3S_64_HANDLER) += \
 	book3s_hv_rmhandlers.o \
 	book3s_hv_rm_mmu.o \
 	book3s_hv_ras.o \
 	book3s_hv_builtin.o \
 	book3s_hv_cma.o \
 	$(kvm-book3s_64-builtin-xics-objs-y)
+endif
 
 kvm-book3s_64-objs-$(CONFIG_KVM_XICS) += \
 	book3s_xics.o
@@ -132,4 +133,7 @@ obj-$(CONFIG_KVM_E500MC) += kvm.o

obj-$(CONFIG_KVM_BOOK3S_64) += kvm.o
obj-$(CONFIG_KVM_BOOK3S_32) += kvm.o

arch/powerpc/kvm/Makefile +3 -0
+obj-$(CONFIG_KVM_BOOK3S_64_PR) += kvm-pr.o
+obj-$(CONFIG_KVM_BOOK3S_64_HV) += kvm-hv.o
+

obj-y += $(kvm-book3s_64-builtin-objs-y)

arch/powerpc/kvm/book3s.c +4 -1
index 12f94bf..ca617e1 100644
@@ -170,13 +170,14 @@ void kvmppc_book3s_queue_irqprio(struct kvm_vcpu *vcpu, unsigned int vec)
 	printk(KERN_INFO "Queueing interrupt %x\n", vec);
 #endif
 }
-
+EXPORT_SYMBOL_GPL(kvmppc_book3s_queue_irqprio);
 
 void kvmppc_core_queue_program(struct kvm_vcpu *vcpu, ulong flags)
 {
 	/* might as well deliver this straight away */
 	kvmppc_inject_interrupt(vcpu, BOOK3S_INTERRUPT_PROGRAM, flags);
 }
+EXPORT_SYMBOL_GPL(kvmppc_core_queue_program);
 
 void kvmppc_core_queue_dec(struct kvm_vcpu *vcpu)
 {
@@ -351,6 +352,7 @@ pfn_t kvmppc_gfn_to_pfn(struct kvm_vcpu *vcpu, gfn_t gfn)
 
 	return gfn_to_pfn(vcpu->kvm, gfn);
 }
+EXPORT_SYMBOL_GPL(kvmppc_gfn_to_pfn);
 
 static int kvmppc_xlate(struct kvm_vcpu *vcpu, ulong eaddr, bool data,
 			 struct kvmppc_pte *pte)
@@ -418,6 +420,7 @@ int kvmppc_st(struct kvm_vcpu *vcpu, ulong *eaddr, int size, void *ptr,
 
 	return EMULATE_DONE;
 }
+EXPORT_SYMBOL_GPL(kvmppc_st);
 
 int kvmppc_ld(struct kvm_vcpu *vcpu, ulong *eaddr, int size, void *ptr,
 		      bool data)
@@ -448,6 +451,7 @@ nopte:

mmio:

	return EMULATE_DO_MMIO;
}

+EXPORT_SYMBOL_GPL(kvmppc_ld);

int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
{

arch/powerpc/kvm/book3s.c +13 -0
@@ -693,6 +697,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u64 msr)
 {
 	kvmppc_ops->set_msr(vcpu, msr);
 }
+EXPORT_SYMBOL_GPL(kvmppc_set_msr);
 
 int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
 {
@@ -774,6 +779,7 @@ int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
 {
 	return kvmppc_ops->unmap_hva(kvm, hva);
 }
+EXPORT_SYMBOL_GPL(kvm_unmap_hva);
 
 int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
 {
@@ -825,3 +831,14 @@ int kvmppc_core_check_processor_compat(void)
 {
 	return kvmppc_ops->check_processor_compat();
 }
+
+EXPORT_SYMBOL_GPL(kvm_get_dirty_log);
+EXPORT_SYMBOL_GPL(kvmppc_core_pending_dec);
+EXPORT_SYMBOL_GPL(kvmppc_init_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_core_dequeue_dec);
+EXPORT_SYMBOL_GPL(kvmppc_claim_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_alloc_lpid);
+EXPORT_SYMBOL_GPL(kvmppc_core_prepare_to_enter);
+EXPORT_SYMBOL_GPL(kvmppc_core_queue_dec);
+EXPORT_SYMBOL_GPL(kvmppc_free_lpid);
+
arch/powerpc/kvm/book3s_64_mmu_host.c +0 -0
index fd5b393..775d368 100644

arch/powerpc/kvm/book3s_64_mmu_host.c line 30

Quoted 4 lines
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -0
@@ -27,6 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
+

Stray whitespace change

arch/powerpc/kvm/book3s_emulate.c int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu, · line 175

Quoted 13 lines
arch/powerpc/kvm/book3s_64_mmu_host.c +0 -0
 #include "trace_pr.h"
 
 #define PTE_SIZE 12
arch/powerpc/kvm/book3s_emulate.c +1 -1
index b9841ad..20d03c2 100644
@@ -172,7 +172,7 @@ int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			vcpu->arch.mmu.tlbie(vcpu, addr, large);
 			break;
 		}
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_KVM_BOOK3S_PR

Why?

Quoted 35 lines
arch/powerpc/kvm/book3s_emulate.c +0 -0
 		case OP_31_XOP_FAKE_SC1:
 		{
 			/* SC 1 papr hypercalls */
arch/powerpc/kvm/book3s_hv.c +4 -0
index eec353d..ff57be8 100644
@@ -53,6 +53,8 @@
 #include <linux/highmem.h>
 #include <linux/hugetlb.h>
 
+#include <linux/module.h>
+
 /* #define EXIT_DEBUG */
 /* #define EXIT_DEBUG_SIMPLE */
 /* #define EXIT_DEBUG_INT */
@@ -2104,3 +2106,5 @@ static void kvmppc_book3s_exit_hv(void)
 
 module_init(kvmppc_book3s_init_hv);
 module_exit(kvmppc_book3s_exit_hv);
+
+MODULE_LICENSE("GPL");
arch/powerpc/kvm/book3s_pr.c +1 -0
index 99d0839..df48d89 100644
@@ -40,6 +40,7 @@
 #include <linux/sched.h>
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
+#include <linux/module.h>
 
 #define CREATE_TRACE_POINTS
 #include "trace_pr.h"

arch/powerpc/kvm/book3s_pr.c program_interrupt: · line 875

Quoted 5 lines
arch/powerpc/kvm/book3s_pr.c +1 -1
@@ -871,7 +872,7 @@ program_interrupt:
 			ulong cmd = kvmppc_get_gpr(vcpu, 3);
 			int i;
 
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_PPC_BOOK3S_64

Why?

Alex

Quoted 117 lines
arch/powerpc/kvm/book3s_pr.c +2 -0
 			if (kvmppc_h_pr(vcpu, cmd) = EMULATE_DONE) {
 				r = RESUME_GUEST;
 				break;
@@ -1567,3 +1568,5 @@ static void kvmppc_book3s_exit_pr(void)
 
 module_init(kvmppc_book3s_init_pr);
 module_exit(kvmppc_book3s_exit_pr);
+
+MODULE_LICENSE("GPL");
arch/powerpc/kvm/book3s_rtas.c +1 -0
index 3219ba8..cf95cde 100644
@@ -260,6 +260,7 @@ fail:
 	 */
 	return rc;
 }
+EXPORT_SYMBOL_GPL(kvmppc_rtas_hcall);
 
 void kvmppc_rtas_tokens_free(struct kvm *kvm)
 {
arch/powerpc/kvm/emulate.c +1 -0
index 81762eb..de9a340 100644
@@ -483,3 +483,4 @@ int kvmppc_emulate_instruction(struct kvm_run *run, struct kvm_vcpu *vcpu)
 
 	return emulated;
 }
+EXPORT_SYMBOL_GPL(kvmppc_emulate_instruction);
arch/powerpc/kvm/powerpc.c +6 -0
index 00a96fc..3019edc 100644
@@ -126,6 +126,7 @@ int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_prepare_to_enter);
 
 int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 {
@@ -179,6 +180,7 @@ int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_kvm_pv);
 
 int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
 {
@@ -207,6 +209,7 @@ out:
 	vcpu->arch.sane = r;
 	return r ? 0 : -EINVAL;
 }
+EXPORT_SYMBOL_GPL(kvmppc_sanity_check);
 
 int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
 {
@@ -241,6 +244,7 @@ int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
 
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvmppc_emulate_mmio);
 
 int kvm_arch_hardware_enable(void *garbage)
 {
@@ -665,6 +669,7 @@ int kvmppc_handle_load(struct kvm_run *run, struct kvm_vcpu *vcpu,
 
 	return EMULATE_DO_MMIO;
 }
+EXPORT_SYMBOL_GPL(kvmppc_handle_load);
 
 /* Same as above, but sign extends */
 int kvmppc_handle_loads(struct kvm_run *run, struct kvm_vcpu *vcpu,
@@ -726,6 +731,7 @@ int kvmppc_handle_store(struct kvm_run *run, struct kvm_vcpu *vcpu,
 
 	return EMULATE_DO_MMIO;
 }
+EXPORT_SYMBOL_GPL(kvmppc_handle_store);
 
 int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
 {
virt/kvm/kvm_main.c +3 -0
index bf040c4..66df1d2 100644
@@ -186,6 +186,7 @@ void kvm_flush_remote_tlbs(struct kvm *kvm)
 		++kvm->stat.remote_tlb_flush;
 	cmpxchg(&kvm->tlbs_dirty, dirty_count, 0);
 }
+EXPORT_SYMBOL_GPL(kvm_flush_remote_tlbs);
 
 void kvm_reload_remote_mmus(struct kvm *kvm)
 {
@@ -1648,6 +1649,7 @@ void mark_page_dirty(struct kvm *kvm, gfn_t gfn)
 	memslot = gfn_to_memslot(kvm, gfn);
 	mark_page_dirty_in_slot(kvm, memslot, gfn);
 }
+EXPORT_SYMBOL_GPL(mark_page_dirty);
 
 /*
 * The vCPU has executed a HLT instruction with in-kernel mode enabled.
@@ -1673,6 +1675,7 @@ void kvm_vcpu_block(struct kvm_vcpu *vcpu)
 
 	finish_wait(&vcpu->wq, &wait);
 }
+EXPORT_SYMBOL_GPL(kvm_vcpu_block);
 
 #ifndef CONFIG_S390
 /*
Signature · 6 lines
1.8.1.2

--
To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#29 lore raw

Re: [RFC PATCH 08/11] kvm: powerpc: book3s: Support building HV and PR KVM as module

Quoted 14 lines · Alexander Graf

arch/powerpc/kvm/book3s_64_mmu_host.c line 30

Quoted 9 lines · Aneesh Kumar K.V
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -0
index fd5b393..775d368 100644
@@ -27,6 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
+

Stray whitespace change

will fix

Quoted 15 lines

arch/powerpc/kvm/book3s_emulate.c int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu, · line 175

Quoted 13 lines
arch/powerpc/kvm/book3s_64_mmu_host.c +0 -0
 #include "trace_pr.h"
 
 #define PTE_SIZE 12
arch/powerpc/kvm/book3s_emulate.c +1 -1
index b9841ad..20d03c2 100644
@@ -172,7 +172,7 @@ int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			vcpu->arch.mmu.tlbie(vcpu, addr, large);
 			break;
 		}
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_KVM_BOOK3S_PR

Why?

If i have CONFIG_KVM_BOOK3S_64_PR=m #ifdef CONFIG_KVM_BOOK3S_64_PR will not work. There is a runtime check I can use IS_ENABLED(). But didn't want to do those. Hence moved to the symbol which will be set as CONFIG_KVM_BOOK3S_PR = y

-aneesh

#30 lore raw

Re: [RFC PATCH 08/11] kvm: powerpc: book3s: Support building HV and PR KVM as module

Quoted 35 lines · Aneesh Kumar K.V
Quoted 14 lines · Alexander Graf

arch/powerpc/kvm/book3s_64_mmu_host.c line 30

Quoted 9 lines · Aneesh Kumar K.V
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -0
index fd5b393..775d368 100644
@@ -27,6 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
+

Stray whitespace change

will fix

Quoted 15 lines

arch/powerpc/kvm/book3s_emulate.c int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu, · line 175

Quoted 13 lines
arch/powerpc/kvm/book3s_64_mmu_host.c +0 -0
 #include "trace_pr.h"
 
 #define PTE_SIZE 12
arch/powerpc/kvm/book3s_emulate.c +1 -1
index b9841ad..20d03c2 100644
@@ -172,7 +172,7 @@ int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			vcpu->arch.mmu.tlbie(vcpu, addr, large);
 			break;
 		}
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_KVM_BOOK3S_PR

Why?

Quoted 4 lines

If i have CONFIG_KVM_BOOK3S_64_PR=m #ifdef CONFIG_KVM_BOOK3S_64_PR will not work. There is a runtime check I can use IS_ENABLED(). But didn't want to do those. Hence moved to the symbol which will be set as CONFIG_KVM_BOOK3S_PR = y

Ah, I see. The naming starts to get really confusing. Could you please give this a slightly better name, such as CONFIG_KVM_BOOK3S_PR_POSSIBLE or such, so that it's obvious that this #ifdef doesn't mean "option is enabled", but instead means "option can potentially be enabled through a module".

Alex

#31 lore raw

Re: [RFC PATCH 08/11] kvm: powerpc: book3s: Support building HV and PR KVM as module

Quoted 42 lines · Alexander Graf
Quoted 35 lines · Aneesh Kumar K.V
Quoted 14 lines · Alexander Graf

arch/powerpc/kvm/book3s_64_mmu_host.c line 30

Quoted 9 lines · Aneesh Kumar K.V
arch/powerpc/kvm/book3s_64_mmu_host.c +1 -0
index fd5b393..775d368 100644
@@ -27,6 +27,7 @@
 #include <asm/machdep.h>
 #include <asm/mmu_context.h>
 #include <asm/hw_irq.h>
+

Stray whitespace change

will fix

Quoted 15 lines

arch/powerpc/kvm/book3s_emulate.c int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu, · line 175

Quoted 13 lines
arch/powerpc/kvm/book3s_64_mmu_host.c +0 -0
 #include "trace_pr.h"
 
 #define PTE_SIZE 12
arch/powerpc/kvm/book3s_emulate.c +1 -1
index b9841ad..20d03c2 100644
@@ -172,7 +172,7 @@ int kvmppc_core_emulate_op_pr(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			vcpu->arch.mmu.tlbie(vcpu, addr, large);
 			break;
 		}
-#ifdef CONFIG_KVM_BOOK3S_64_PR
+#ifdef CONFIG_KVM_BOOK3S_PR

Why?

Quoted 4 lines

If i have CONFIG_KVM_BOOK3S_64_PR=m #ifdef CONFIG_KVM_BOOK3S_64_PR will not work. There is a runtime check I can use IS_ENABLED(). But didn't want to do those. Hence moved to the symbol which will be set as CONFIG_KVM_BOOK3S_PR = y

Quoted 5 lines

Ah, I see. The naming starts to get really confusing. Could you please give this a slightly better name, such as CONFIG_KVM_BOOK3S_PR_POSSIBLE or such, so that it's obvious that this #ifdef doesn't mean "option is enabled", but instead means "option can potentially be enabled through a module".

Will do

-aneesh

#32 lore raw

RFC PATCH 9/11 kvm: simplify processor compat check

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

---
 arch/arm/kvm/arm.c                 |  4 ++--
 arch/ia64/kvm/kvm-ia64.c           |  4 ++--
 arch/mips/kvm/kvm_mips.c           |  6 ++----
 arch/powerpc/include/asm/kvm_ppc.h |  2 +-
 arch/powerpc/kvm/44x.c             |  2 +-
 arch/powerpc/kvm/book3s.c          | 15 ++++++++++++---
 arch/powerpc/kvm/book3s_hv.c       |  9 ++++++---
 arch/powerpc/kvm/book3s_pr.c       |  5 +++--
 arch/powerpc/kvm/e500.c            |  2 +-
 arch/powerpc/kvm/e500mc.c          |  2 +-
 arch/powerpc/kvm/powerpc.c         |  5 -----
 arch/s390/kvm/kvm-s390.c           |  3 ++-
 arch/x86/kvm/x86.c                 | 13 +++++++++++--
 include/linux/kvm_host.h           |  2 +-
 virt/kvm/kvm_main.c                | 14 +++++---------
 15 files changed, 50 insertions(+), 38 deletions(-)
arch/arm/kvm/arm.c +2 -2
index 9c697db..cccb121 100644
@@ -109,9 +109,9 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	*(int *)rtn = 0;
+	return 0;
 }
 
 void kvm_arch_sync_events(struct kvm *kvm)
arch/ia64/kvm/kvm-ia64.c +2 -2
index bdfd878..065942c 100644
@@ -185,9 +185,9 @@ void kvm_arch_hardware_disable(void *garbage)
 	ia64_ptr_entry(0x3, slot);
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	*(int *)rtn = 0;
+	return 0;
 }
 
 int kvm_dev_ioctl_check_extension(long ext)
arch/mips/kvm/kvm_mips.c +2 -4
index a7b0445..4512739 100644
@@ -97,11 +97,9 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	int *r = (int *)rtn;
-	*r = 0;
-	return;
+	return 0;
 }
 
 static void kvm_mips_init_tlbs(struct kvm *kvm)
arch/powerpc/include/asm/kvm_ppc.h +1 -1
index 58e732f..592501b 100644
@@ -204,7 +204,7 @@ struct kvmppc_ops {
 			      unsigned long npages);
 	int (*init_vm)(struct kvm *kvm);
 	void (*destroy_vm)(struct kvm *kvm);
-	int (*check_processor_compat)(void);
+	void (*check_processor_compat)(void *r);
 	int (*get_smmu_info)(struct kvm *kvm, struct kvm_ppc_smmu_info *info);
 	int (*emulate_op)(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			  unsigned int inst, int *advance);
arch/powerpc/kvm/44x.c +1 -1
index 2f5c6b6..a1f4e60 100644
@@ -43,7 +43,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/book3s.c +12 -3
index ca617e1..485a6ff 100644
@@ -827,9 +827,18 @@ void kvmppc_core_destroy_vm(struct kvm *kvm)
 #endif
 }
 
-int kvmppc_core_check_processor_compat(void)
-{
-	return kvmppc_ops->check_processor_compat();
+int kvm_arch_check_processor_compat(void *opaque)
+{
+	int r,cpu;
+	struct kvmppc_ops *kvm_ops = (struct kvmppc_ops *)opaque;
+	for_each_online_cpu(cpu) {
+		smp_call_function_single(cpu,
+					 kvm_ops->check_processor_compat,
+					 &r, 1);
+		if (r < 0)
+			break;
+	}
+	return r;
 }
 
 EXPORT_SYMBOL_GPL(kvm_get_dirty_log);
arch/powerpc/kvm/book3s_hv.c +6 -3
index ff57be8..4322db4 100644
@@ -1980,11 +1980,14 @@ static int kvmppc_core_emulate_mfspr_hv(struct kvm_vcpu *vcpu, int sprn,
 	return EMULATE_FAIL;
 }
 
-static int kvmppc_core_check_processor_compat_hv(void)
+
+static void kvmppc_core_check_processor_compat_hv(void *r)
 {
 	if (!cpu_has_feature(CPU_FTR_HVMODE))
-		return -EIO;
-	return 0;
+		*(int *)r = -EIO;
+	else
+		*(int *)r = 0;
+	return;
 }
 
 static long kvm_arch_vm_ioctl_hv(struct file *filp,
arch/powerpc/kvm/book3s_pr.c +3 -2
index df48d89..127b961 100644
@@ -1490,10 +1490,11 @@ static void kvmppc_core_destroy_vm_pr(struct kvm *kvm)
 	enable_relon_interrupts(kvm);
 }
 
-static int kvmppc_core_check_processor_compat_pr(void)
+static void kvmppc_core_check_processor_compat_pr(void *r)
 {
 	/* we are always compatible */
-	return 0;
+	*(int *)r = 0;
+	return;
 }
 
 static long kvm_arch_vm_ioctl_pr(struct file *filp,
arch/powerpc/kvm/e500.c +1 -1
index ce6b73c..0681cb1 100644
@@ -323,7 +323,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/e500mc.c +1 -1
index 19c8379..48b3ba5 100644
@@ -169,7 +169,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/powerpc.c +0 -5
index 3019edc..1404f4d 100644
@@ -264,11 +264,6 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
-{
-	*(int *)rtn = kvmppc_core_check_processor_compat();
-}
-
 int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 {
 	if (type)
arch/s390/kvm/kvm-s390.c +2 -1
index 776dafe..01d4000 100644
@@ -119,8 +119,9 @@ void kvm_arch_hardware_unsetup(void)
 	gmap_unregister_ipte_notifier(&gmap_notifier);
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
+	return 0;
 }
 
 int kvm_arch_init(void *opaque)
arch/x86/kvm/x86.c +11 -2
index e5ca72a..53c8308 100644
@@ -6873,9 +6873,18 @@ void kvm_arch_hardware_unsetup(void)
 	kvm_x86_ops->hardware_unsetup();
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	kvm_x86_ops->check_processor_compatibility(rtn);
+	int r,cpu;
+
+	for_each_online_cpu(cpu) {
+		smp_call_function_single(cpu,
+			 kvm_x86_ops->check_processor_compatibility,
+			 &r, 1);
+		if (r < 0)
+			break;
+	}
+	return r;
 }
 
 bool kvm_vcpu_compatible(struct kvm_vcpu *vcpu)
include/linux/kvm_host.h +1 -1
index ca645a0..459b359 100644
@@ -649,7 +649,7 @@ int kvm_arch_hardware_enable(void *garbage);
 void kvm_arch_hardware_disable(void *garbage);
 int kvm_arch_hardware_setup(void);
 void kvm_arch_hardware_unsetup(void);
-void kvm_arch_check_processor_compat(void *rtn);
+int kvm_arch_check_processor_compat(void *opaque);
 int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu);
 int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu);
 
virt/kvm/kvm_main.c +5 -9
index 66df1d2..0594b22 100644
@@ -3166,10 +3166,9 @@ static void kvm_sched_out(struct preempt_notifier *pn,
 }
 
 int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
-		  struct module *module)
+	     struct module *module)
 {
 	int r;
-	int cpu;
 
 	r = kvm_arch_init(opaque);
 	if (r)
@@ -3195,13 +3194,10 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
 	if (r < 0)
 		goto out_free_0a;
 
-	for_each_online_cpu(cpu) {
-		smp_call_function_single(cpu,
-				kvm_arch_check_processor_compat,
-				&r, 1);
-		if (r < 0)
-			goto out_free_1;
-	}
+
+	r = kvm_arch_check_processor_compat(opaque);
+	if (r < 0)
+		goto out_free_1;
 
 	r = register_cpu_notifier(&kvm_cpu_notifier);
 	if (r)
Signature · 1 line
1.8.1.2
#33 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Which brings me to the next issue: You forgot to CC kvm@vger on your patch set. Gleb and Paolo don't read kvm-ppc@vger. And they shouldn't have to. Every kvm patch that you want review on or that should get applied needs to be sent to kvm@vger. If you want to tag it as PPC specific patch, do so by CC'ing kvm-ppc@vger.

Alex

Quoted 292 lines

---

arch/arm/kvm/arm.c                 |  4 ++--
arch/ia64/kvm/kvm-ia64.c           |  4 ++--
arch/mips/kvm/kvm_mips.c           |  6 ++----
arch/powerpc/include/asm/kvm_ppc.h |  2 +-
arch/powerpc/kvm/44x.c             |  2 +-
arch/powerpc/kvm/book3s.c          | 15 ++++++++++++---
arch/powerpc/kvm/book3s_hv.c       |  9 ++++++---
arch/powerpc/kvm/book3s_pr.c       |  5 +++--
arch/powerpc/kvm/e500.c            |  2 +-
arch/powerpc/kvm/e500mc.c          |  2 +-
arch/powerpc/kvm/powerpc.c         |  5 -----
arch/s390/kvm/kvm-s390.c           |  3 ++-
arch/x86/kvm/x86.c                 | 13 +++++++++++--
include/linux/kvm_host.h           |  2 +-
virt/kvm/kvm_main.c                | 14 +++++---------
15 files changed, 50 insertions(+), 38 deletions(-)
arch/arm/kvm/arm.c +2 -2
index 9c697db..cccb121 100644
@@ -109,9 +109,9 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	*(int *)rtn = 0;
+	return 0;
 }
 
 void kvm_arch_sync_events(struct kvm *kvm)
arch/ia64/kvm/kvm-ia64.c +2 -2
index bdfd878..065942c 100644
@@ -185,9 +185,9 @@ void kvm_arch_hardware_disable(void *garbage)
 	ia64_ptr_entry(0x3, slot);
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	*(int *)rtn = 0;
+	return 0;
 }
 
 int kvm_dev_ioctl_check_extension(long ext)
arch/mips/kvm/kvm_mips.c +2 -4
index a7b0445..4512739 100644
@@ -97,11 +97,9 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	int *r = (int *)rtn;
-	*r = 0;
-	return;
+	return 0;
 }
 
 static void kvm_mips_init_tlbs(struct kvm *kvm)
arch/powerpc/include/asm/kvm_ppc.h +1 -1
index 58e732f..592501b 100644
@@ -204,7 +204,7 @@ struct kvmppc_ops {
 			      unsigned long npages);
 	int (*init_vm)(struct kvm *kvm);
 	void (*destroy_vm)(struct kvm *kvm);
-	int (*check_processor_compat)(void);
+	void (*check_processor_compat)(void *r);
 	int (*get_smmu_info)(struct kvm *kvm, struct kvm_ppc_smmu_info *info);
 	int (*emulate_op)(struct kvm_run *run, struct kvm_vcpu *vcpu,
 			  unsigned int inst, int *advance);
arch/powerpc/kvm/44x.c +1 -1
index 2f5c6b6..a1f4e60 100644
@@ -43,7 +43,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/book3s.c +12 -3
index ca617e1..485a6ff 100644
@@ -827,9 +827,18 @@ void kvmppc_core_destroy_vm(struct kvm *kvm)
 #endif
 }
 
-int kvmppc_core_check_processor_compat(void)
-{
-	return kvmppc_ops->check_processor_compat();
+int kvm_arch_check_processor_compat(void *opaque)
+{
+	int r,cpu;
+	struct kvmppc_ops *kvm_ops = (struct kvmppc_ops *)opaque;
+	for_each_online_cpu(cpu) {
+		smp_call_function_single(cpu,
+					 kvm_ops->check_processor_compat,
+					 &r, 1);
+		if (r < 0)
+			break;
+	}
+	return r;
 }
 
 EXPORT_SYMBOL_GPL(kvm_get_dirty_log);
arch/powerpc/kvm/book3s_hv.c +6 -3
index ff57be8..4322db4 100644
@@ -1980,11 +1980,14 @@ static int kvmppc_core_emulate_mfspr_hv(struct kvm_vcpu *vcpu, int sprn,
 	return EMULATE_FAIL;
 }
 
-static int kvmppc_core_check_processor_compat_hv(void)
+
+static void kvmppc_core_check_processor_compat_hv(void *r)
 {
 	if (!cpu_has_feature(CPU_FTR_HVMODE))
-		return -EIO;
-	return 0;
+		*(int *)r = -EIO;
+	else
+		*(int *)r = 0;
+	return;
 }
 
 static long kvm_arch_vm_ioctl_hv(struct file *filp,
arch/powerpc/kvm/book3s_pr.c +3 -2
index df48d89..127b961 100644
@@ -1490,10 +1490,11 @@ static void kvmppc_core_destroy_vm_pr(struct kvm *kvm)
 	enable_relon_interrupts(kvm);
 }
 
-static int kvmppc_core_check_processor_compat_pr(void)
+static void kvmppc_core_check_processor_compat_pr(void *r)
 {
 	/* we are always compatible */
-	return 0;
+	*(int *)r = 0;
+	return;
 }
 
 static long kvm_arch_vm_ioctl_pr(struct file *filp,
arch/powerpc/kvm/e500.c +1 -1
index ce6b73c..0681cb1 100644
@@ -323,7 +323,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/e500mc.c +1 -1
index 19c8379..48b3ba5 100644
@@ -169,7 +169,7 @@ void kvmppc_core_vcpu_put(struct kvm_vcpu *vcpu)
 	kvmppc_booke_vcpu_put(vcpu);
 }
 
-int kvmppc_core_check_processor_compat(void)
+int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r;
 
arch/powerpc/kvm/powerpc.c +0 -5
index 3019edc..1404f4d 100644
@@ -264,11 +264,6 @@ void kvm_arch_hardware_unsetup(void)
 {
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
-{
-	*(int *)rtn = kvmppc_core_check_processor_compat();
-}
-
 int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 {
 	if (type)
arch/s390/kvm/kvm-s390.c +2 -1
index 776dafe..01d4000 100644
@@ -119,8 +119,9 @@ void kvm_arch_hardware_unsetup(void)
 	gmap_unregister_ipte_notifier(&gmap_notifier);
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
+	return 0;
 }
 
 int kvm_arch_init(void *opaque)
arch/x86/kvm/x86.c +11 -2
index e5ca72a..53c8308 100644
@@ -6873,9 +6873,18 @@ void kvm_arch_hardware_unsetup(void)
 	kvm_x86_ops->hardware_unsetup();
 }
 
-void kvm_arch_check_processor_compat(void *rtn)
+int kvm_arch_check_processor_compat(void *opaque)
 {
-	kvm_x86_ops->check_processor_compatibility(rtn);
+	int r,cpu;
+
+	for_each_online_cpu(cpu) {
+		smp_call_function_single(cpu,
+			 kvm_x86_ops->check_processor_compatibility,
+			 &r, 1);
+		if (r < 0)
+			break;
+	}
+	return r;
 }
 
 bool kvm_vcpu_compatible(struct kvm_vcpu *vcpu)
include/linux/kvm_host.h +1 -1
index ca645a0..459b359 100644
@@ -649,7 +649,7 @@ int kvm_arch_hardware_enable(void *garbage);
 void kvm_arch_hardware_disable(void *garbage);
 int kvm_arch_hardware_setup(void);
 void kvm_arch_hardware_unsetup(void);
-void kvm_arch_check_processor_compat(void *rtn);
+int kvm_arch_check_processor_compat(void *opaque);
 int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu);
 int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu);
 
virt/kvm/kvm_main.c +5 -9
index 66df1d2..0594b22 100644
@@ -3166,10 +3166,9 @@ static void kvm_sched_out(struct preempt_notifier *pn,
 }
 
 int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
-		  struct module *module)
+	     struct module *module)
 {
 	int r;
-	int cpu;
 
 	r = kvm_arch_init(opaque);
 	if (r)
@@ -3195,13 +3194,10 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
 	if (r < 0)
 		goto out_free_0a;
 
-	for_each_online_cpu(cpu) {
-		smp_call_function_single(cpu,
-				kvm_arch_check_processor_compat,
-				&r, 1);
-		if (r < 0)
-			goto out_free_1;
-	}
+
+	r = kvm_arch_check_processor_compat(opaque);
+	if (r < 0)
+		goto out_free_1;
 
 	r = register_cpu_notifier(&kvm_cpu_notifier);
 	if (r)
Signature · 1 line
1.8.1.2
#34 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

We also want to make the smp call function a callback of opaque. Hence this should be made arch specific.

int kvm_arch_check_processor_compat(void *opaque)

{
	int r,cpu;
	struct kvmppc_ops *kvm_ops = (struct kvmppc_ops *)opaque;
	for_each_online_cpu(cpu) {
		smp_call_function_single(cpu,
					 kvm_ops->check_processor_compat,
					 &r, 1);
		if (r < 0)
			break;
	}
	return r;
}

against

- for_each_online_cpu(cpu) {
- smp_call_function_single(cpu,
- kvm_arch_check_processor_compat,
- &r, 1);
- if (r < 0)
- goto out_free_1;
- }
+
+ r = kvm_arch_check_processor_compat(opaque);
+ if (r < 0)
+ goto out_free_1;

Quoted 6 lines

Which brings me to the next issue: You forgot to CC kvm@vger on your patch set. Gleb and Paolo don't read kvm-ppc@vger. And they shouldn't have to. Every kvm patch that you want review on or that should get applied needs to be sent to kvm@vger. If you want to tag it as PPC specific patch, do so by CC'ing kvm-ppc@vger.

Will do in the next update

-aneesh

maintainer
#35 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Il 27/09/2013 15:13, Aneesh Kumar K.V ha scritto:

Quoted 13 lines
Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Quoted 6 lines

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

I agree with Alex.

The current code is not specially awesome; having kvm_arch_check_processor_compat take an int* disguised as a void* is a bit ugly indeed.

However, the API makes sense and tells you that it is being passed as a callback (to smp_call_function_single in this case).

You are making the API more complicated to use on the arch layer (because arch maintainers now have to think "do I need to check this on all online CPUs?") and making the "leaf" POWER code less legible because it still has the weird void()(void *) calling convention.

If anything, you could change kvm_arch_check_processor_compat to return an int and accept no argument, and introduce a wrapper that kvm_init passes to smp_call_function_single.

Paolo

Quoted 44 lines

We also want to make the smp call function a callback of opaque. Hence this should be made arch specific.

int kvm_arch_check_processor_compat(void *opaque)

{
	int r,cpu;
	struct kvmppc_ops *kvm_ops = (struct kvmppc_ops *)opaque;
	for_each_online_cpu(cpu) {
		smp_call_function_single(cpu,
					 kvm_ops->check_processor_compat,
					 &r, 1);
		if (r < 0)
			break;
	}
	return r;
}

against

virt/kvm/kvm_main.c +4 -7
-	for_each_online_cpu(cpu) {
-		smp_call_function_single(cpu,
-				kvm_arch_check_processor_compat,
-				&r, 1);
-		if (r < 0)
-			goto out_free_1;
-	}
+
+	r = kvm_arch_check_processor_compat(opaque);
+	if (r < 0)
+		goto out_free_1;
Quoted 9 lines

Which brings me to the next issue: You forgot to CC kvm@vger on your patch set. Gleb and Paolo don't read kvm-ppc@vger. And they shouldn't have to. Every kvm patch that you want review on or that should get applied needs to be sent to kvm@vger. If you want to tag it as PPC specific patch, do so by CC'ing kvm-ppc@vger.

Will do in the next update

-aneesh

#36 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 21 lines · Paolo Bonzini

Il 27/09/2013 15:13, Aneesh Kumar K.V ha scritto:

Quoted 13 lines
Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Quoted 6 lines

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

Quoted 8 lines

I agree with Alex.

The current code is not specially awesome; having kvm_arch_check_processor_compat take an int* disguised as a void* is a bit ugly indeed.

However, the API makes sense and tells you that it is being passed as a callback (to smp_call_function_single in this case).

But whether to check on all cpus or not is arch dependent right?. IIUC only x86 and ppc64 need to do that. Also on ppc64 it really depends on whether HV or PR. We need to check on all cpus only if it is HV.

Quoted 6 lines

You are making the API more complicated to use on the arch layer (because arch maintainers now have to think "do I need to check this on all online CPUs?") and making the "leaf" POWER code less legible because it still has the weird void()(void *) calling convention.

IIUC what we wanted to check is to find out whether kvm can run on this system. That is really an arch specific check. So for core kvm the call should be a simple

if (kvm_arch_check_process_compat() < 0)

        error;

Now how each arch figure out whether kvm can run on this system should be arch specific. For x86 we do need to check all the cpus. On ppc64 for HV we need to. For other archs we always allow kvm.

Quoted 3 lines

If anything, you could change kvm_arch_check_processor_compat to return an int and accept no argument, and introduce a wrapper that kvm_init passes to smp_call_function_single.

What i am suggesting in the patch is to avoid calling smp_call_function_single from kvm_init and let arch decide whether to check on all cpus or not.

-aneesh

#37 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 44 lines · Aneesh Kumar K.V
Quoted 21 lines · Paolo Bonzini

Il 27/09/2013 15:13, Aneesh Kumar K.V ha scritto:

Quoted 13 lines
Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Quoted 6 lines

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

Quoted 8 lines

I agree with Alex.

The current code is not specially awesome; having kvm_arch_check_processor_compat take an int* disguised as a void* is a bit ugly indeed.

However, the API makes sense and tells you that it is being passed as a callback (to smp_call_function_single in this case).

But whether to check on all cpus or not is arch dependent right?. IIUC only x86 and ppc64 need to do that. Also on ppc64 it really depends on whether HV or PR. We need to check on all cpus only if it is HV.

Quoted 6 lines

You are making the API more complicated to use on the arch layer (because arch maintainers now have to think "do I need to check this on all online CPUs?") and making the "leaf" POWER code less legible because it still has the weird void()(void *) calling convention.

Quoted 6 lines

IIUC what we wanted to check is to find out whether kvm can run on this system. That is really an arch specific check. So for core kvm the call should be a simple

if (kvm_arch_check_process_compat() < 0)

        error;

We have that already, just return error from kvm_arch_hardware_setup. This is specific processor compatibility check and you are arguing that the processor check should be part of kvm_arch_hardware_setup().

Quoted 5 lines

Now how each arch figure out whether kvm can run on this system should be arch specific. For x86 we do need to check all the cpus. On ppc64 for HV we need to. For other archs we always allow kvm.

This is really a sanity check. Theoretically on x86 we also should not need to check all cpus since SMP configuration with different cpu models is not supported by the architecture (AFAIK), but bugs happen (BIOS bugs may cause difference in capabilities for instance). So some arches opted out from this sanity check for now and this is their choice, but the code makes it explicit what are we checking here.

Quoted 10 lines
Quoted 4 lines

If anything, you could change kvm_arch_check_processor_compat to return an int and accept no argument, and introduce a wrapper that kvm_init passes to smp_call_function_single.

What i am suggesting in the patch is to avoid calling smp_call_function_single from kvm_init and let arch decide whether to check on all cpus or not.

-aneesh

--
Gleb.

#38 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 49 lines · Gleb Natapov
Quoted 48 lines · Aneesh Kumar K.V
Quoted 21 lines · Paolo Bonzini

Il 27/09/2013 15:13, Aneesh Kumar K.V ha scritto:

Quoted 13 lines
Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Quoted 6 lines

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

Quoted 8 lines

I agree with Alex.

The current code is not specially awesome; having kvm_arch_check_processor_compat take an int* disguised as a void* is a bit ugly indeed.

However, the API makes sense and tells you that it is being passed as a callback (to smp_call_function_single in this case).

But whether to check on all cpus or not is arch dependent right?. IIUC only x86 and ppc64 need to do that. Also on ppc64 it really depends on whether HV or PR. We need to check on all cpus only if it is HV.

Quoted 6 lines

You are making the API more complicated to use on the arch layer (because arch maintainers now have to think "do I need to check this on all online CPUs?") and making the "leaf" POWER code less legible because it still has the weird void()(void *) calling convention.

IIUC what we wanted to check is to find out whether kvm can run on this system. That is really an arch specific check. So for core kvm the call should be a simple

Quoted 5 lines
Quoted 2 lines

if (kvm_arch_check_process_compat() < 0)

        error;

We have that already, just return error from kvm_arch_hardware_setup. This is specific processor compatibility check and you are arguing that the processor check should be part of kvm_arch_hardware_setup().

What about the success case ?. ie, on arch like arm we do

void kvm_arch_check_processor_compat(void *rtn)

{
	*(int *)rtn = 0;
}
for_each_online_cpu(cpu) {
	smp_call_function_single(cpu,
			kvm_arch_check_processor_compat,
			&r, 1);
	if (r < 0)
		goto out_free_1;
}

There is no need to do that for loop for arm.

The only reason I wanted this patch in the series is to make kvm_arch_check_processor_compat take additional argument opaque. I am dropping that requirement in the last patch. Considering that we have objection to this one, I will drop this patch in the next posting by rearranging the patches.

-aneesh

#39 lore raw

Re: [RFC PATCH 09/11] kvm: simplify processor compat check

Quoted 58 lines · Aneesh Kumar K.V
Quoted 49 lines · Gleb Natapov
Quoted 48 lines · Aneesh Kumar K.V
Quoted 21 lines · Paolo Bonzini

Il 27/09/2013 15:13, Aneesh Kumar K.V ha scritto:

Quoted 13 lines
Quoted 10 lines · Alexander Graf
Quoted 1 line · Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

Missing patch description.

Quoted 1 line

I fail to see how this really simplifies things, but at the end of the day it's Gleb's and Paolo's call.

Quoted 6 lines

will do. It avoid calling

	for_each_online_cpu(cpu) {
		smp_call_function_single() 

on multiple architecture.

Quoted 8 lines

I agree with Alex.

The current code is not specially awesome; having kvm_arch_check_processor_compat take an int* disguised as a void* is a bit ugly indeed.

However, the API makes sense and tells you that it is being passed as a callback (to smp_call_function_single in this case).

But whether to check on all cpus or not is arch dependent right?. IIUC only x86 and ppc64 need to do that. Also on ppc64 it really depends on whether HV or PR. We need to check on all cpus only if it is HV.

Quoted 6 lines

You are making the API more complicated to use on the arch layer (because arch maintainers now have to think "do I need to check this on all online CPUs?") and making the "leaf" POWER code less legible because it still has the weird void()(void *) calling convention.

IIUC what we wanted to check is to find out whether kvm can run on this system. That is really an arch specific check. So for core kvm the call should be a simple

Quoted 5 lines
Quoted 2 lines

if (kvm_arch_check_process_compat() < 0)

        error;

We have that already, just return error from kvm_arch_hardware_setup. This is specific processor compatibility check and you are arguing that the processor check should be part of kvm_arch_hardware_setup().

Quoted 8 lines

What about the success case ?. ie, on arch like arm we do

void kvm_arch_check_processor_compat(void *rtn)

{
	*(int *)rtn = 0;
}

for_each_online_cpu(cpu) {

As I said they opted out from doing the check. They may reconsider after first bad HW will be discovered.

Quoted 8 lines
	smp_call_function_single(cpu,
			kvm_arch_check_processor_compat,
			&r, 1);
	if (r < 0)
		goto out_free_1;
}

There is no need to do that for loop for arm.

It's done once during module initialisation. Why is this a big deal?

--
Gleb.

#40 lore raw

RFC PATCH 10/11 kvm: powerpc: book3s: Allow the HV and PR selection per virtual machine

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This moves the kvmppc_ops callbacks to be a per VM entity. This enables us to select HV and PR mode when creating a VM

---
 arch/powerpc/include/asm/kvm_host.h |  3 ++
 arch/powerpc/include/asm/kvm_ppc.h  | 11 ++++--
 arch/powerpc/kvm/book3s.c           | 60 +++++++++++++++---------------
 arch/powerpc/kvm/book3s_hv.c        | 18 +++++----
 arch/powerpc/kvm/book3s_pr.c        | 18 ++++++---
 arch/powerpc/kvm/emulate.c          | 11 +++---
 arch/powerpc/kvm/powerpc.c          | 74 ++++++++++++++++++++++++-------------
 include/linux/kvm_host.h            |  7 ++--
 include/uapi/linux/kvm.h            |  4 ++
 virt/kvm/kvm_main.c                 | 18 ++++-----
 10 files changed, 135 insertions(+), 89 deletions(-)
arch/powerpc/include/asm/kvm_host.h +3 -0
index 283e52e..61a297fc 100644
@@ -276,6 +276,9 @@ struct kvm_arch {
 #ifdef CONFIG_KVM_XICS
 	struct kvmppc_xics *xics;
 #endif
+#ifdef CONFIG_PPC_BOOK3S_64
+	struct kvmppc_ops *kvm_ops;
+#endif
 };
 
 /*
arch/powerpc/include/asm/kvm_ppc.h +7 -4
index 592501b..a4a5893 100644
@@ -134,9 +134,11 @@ extern struct page *kvm_alloc_hpt(unsigned long nr_pages);
 extern void kvm_release_hpt(struct page *page, unsigned long nr_pages);
 extern int kvmppc_core_init_vm(struct kvm *kvm);
 extern void kvmppc_core_destroy_vm(struct kvm *kvm);
-extern void kvmppc_core_free_memslot(struct kvm_memory_slot *free,
+extern void kvmppc_core_free_memslot(struct kvm *kvm,
+				     struct kvm_memory_slot *free,
 				     struct kvm_memory_slot *dont);
-extern int kvmppc_core_create_memslot(struct kvm_memory_slot *slot,
+extern int kvmppc_core_create_memslot(struct kvm *kvm,
+				      struct kvm_memory_slot *slot,
 				      unsigned long npages);
 extern int kvmppc_core_prepare_memory_region(struct kvm *kvm,
 				struct kvm_memory_slot *memslot,
@@ -216,7 +218,8 @@ struct kvmppc_ops {
 
 };
 
-extern struct kvmppc_ops *kvmppc_ops;
+extern struct kvmppc_ops *kvmppc_hv_ops;
+extern struct kvmppc_ops *kvmppc_pr_ops;
 
 /*
  * Cuts out inst bits with ordering according to spec.
@@ -324,7 +327,7 @@ static inline void kvmppc_set_host_ipi(int cpu, u8 host_ipi)
 
 static inline void kvmppc_fast_vcpu_kick(struct kvm_vcpu *vcpu)
 {
-	kvmppc_ops->fast_vcpu_kick(vcpu);
+	vcpu->kvm->arch.kvm_ops->fast_vcpu_kick(vcpu);
 }
 
 #else
arch/powerpc/kvm/book3s.c +30 -30

197 lines not on this page · raw message

arch/powerpc/kvm/book3s_hv.c +11 -7
index 4322db4..0a684a7 100644
@@ -2052,7 +2052,7 @@ extern int kvm_age_hva_hv(struct kvm *kvm, unsigned long hva);
 extern int kvm_test_age_hva_hv(struct kvm *kvm, unsigned long hva);
 extern void kvm_set_spte_hva_hv(struct kvm *kvm, unsigned long hva, pte_t pte);
 
-static struct kvmppc_ops kvmppc_hv_ops = {
+static struct kvmppc_ops kvm_ops_hv = {
 	.is_hv_enabled = true,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_hv,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_hv,
@@ -2092,11 +2092,13 @@ static int kvmppc_book3s_init_hv(void)
 {
 	int r;
 
-	r = kvm_init(&kvmppc_hv_ops, sizeof(struct kvm_vcpu), 0, THIS_MODULE);
-
-	if (r)
-		return r;
-
+	if (!kvmppc_pr_ops) {
+		r = kvm_init(&kvm_ops_hv, sizeof(struct kvm_vcpu),
+			     0, THIS_MODULE);
+		if (r)
+			return r;
+	}
+	kvmppc_hv_ops = &kvm_ops_hv;
 	r = kvmppc_mmu_hv_init();
 
 	return r;
@@ -2104,7 +2106,9 @@ static int kvmppc_book3s_init_hv(void)
 
 static void kvmppc_book3s_exit_hv(void)
 {
-	kvm_exit();
+	kvmppc_hv_ops = NULL;
+	if (!kvmppc_pr_ops)
+		kvm_exit();
 }
 
 module_init(kvmppc_book3s_init_hv);
arch/powerpc/kvm/book3s_pr.c +12 -6
index 127b961..e49e4b0 100644
@@ -1511,7 +1511,7 @@ extern int kvmppc_core_emulate_mtspr_pr(struct kvm_vcpu *vcpu,
 extern int kvmppc_core_emulate_mfspr_pr(struct kvm_vcpu *vcpu,
 					int sprn, ulong *spr_val);
 
-static struct kvmppc_ops kvmppc_pr_ops = {
+static struct kvmppc_ops kvm_ops_pr = {
 	.is_hv_enabled = false,
 	.get_sregs = kvm_arch_vcpu_ioctl_get_sregs_pr,
 	.set_sregs = kvm_arch_vcpu_ioctl_set_sregs_pr,
@@ -1551,10 +1551,14 @@ static int kvmppc_book3s_init_pr(void)
 {
 	int r;
 
-	r = kvm_init(&kvmppc_pr_ops, sizeof(struct kvm_vcpu), 0, THIS_MODULE);
-
-	if (r)
-		return r;
+	if (!kvmppc_hv_ops) {
+		r = kvm_init(&kvm_ops_pr, sizeof(struct kvm_vcpu),
+			     0, THIS_MODULE);
+		if (r)
+			return r;
+	}
+	/* Assign the global value */
+	kvmppc_pr_ops = &kvm_ops_pr;
 
 	r = kvmppc_mmu_hpte_sysinit();
 
@@ -1563,8 +1567,10 @@ static int kvmppc_book3s_init_pr(void)
 
 static void kvmppc_book3s_exit_pr(void)
 {
+	kvmppc_pr_ops = NULL;
 	kvmppc_mmu_hpte_sysexit();
-	kvm_exit();
+	if (!kvmppc_hv_ops)
+		kvm_exit();
 }
 
 module_init(kvmppc_book3s_init_pr);
arch/powerpc/kvm/emulate.c +6 -5
index de9a340..2f9a087 100644
@@ -130,8 +130,8 @@ static int kvmppc_emulate_mtspr(struct kvm_vcpu *vcpu, int sprn, int rs)
 	case SPRN_PIR: break;
 
 	default:
-		emulated = kvmppc_ops->emulate_mtspr(vcpu, sprn,
-						     spr_val);
+		emulated = vcpu->kvm->arch.kvm_ops->emulate_mtspr(vcpu, sprn,
+								  spr_val);
 		if (emulated == EMULATE_FAIL)
 			printk(KERN_INFO "mtspr: unknown spr "
 				"0x%x\n", sprn);
@@ -191,8 +191,8 @@ static int kvmppc_emulate_mfspr(struct kvm_vcpu *vcpu, int sprn, int rt)
 		spr_val = kvmppc_get_dec(vcpu, get_tb());
 		break;
 	default:
-		emulated = kvmppc_ops->emulate_mfspr(vcpu, sprn,
-						     &spr_val);
+		emulated = vcpu->kvm->arch.kvm_ops->emulate_mfspr(vcpu, sprn,
+								  &spr_val);
 		if (unlikely(emulated == EMULATE_FAIL)) {
 			printk(KERN_INFO "mfspr: unknown spr "
 				"0x%x\n", sprn);
@@ -464,7 +464,8 @@ int kvmppc_emulate_instruction(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	}
 
 	if (emulated == EMULATE_FAIL) {
-		emulated = kvmppc_ops->emulate_op(run, vcpu, inst, &advance);
+		emulated = vcpu->kvm->arch.kvm_ops->emulate_op(run, vcpu, inst,
+							       &advance);
 		if (emulated == EMULATE_AGAIN) {
 			advance = 0;
 		} else if (emulated == EMULATE_FAIL) {
arch/powerpc/kvm/powerpc.c +49 -25

171 lines not on this page · raw message

include/linux/kvm_host.h +4 -3
index 459b359..2938636 100644
@@ -508,9 +508,10 @@ int kvm_set_memory_region(struct kvm *kvm,
 			  struct kvm_userspace_memory_region *mem);
 int __kvm_set_memory_region(struct kvm *kvm,
 			    struct kvm_userspace_memory_region *mem);
-void kvm_arch_free_memslot(struct kvm_memory_slot *free,
+void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *free,
 			   struct kvm_memory_slot *dont);
-int kvm_arch_create_memslot(struct kvm_memory_slot *slot, unsigned long npages);
+int kvm_arch_create_memslot(struct kvm *kvm, struct kvm_memory_slot *slot,
+			    unsigned long npages);
 void kvm_arch_memslots_updated(struct kvm *kvm);
 int kvm_arch_prepare_memory_region(struct kvm *kvm,
 				struct kvm_memory_slot *memslot,
@@ -597,7 +598,7 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
 			 unsigned int ioctl, unsigned long arg);
 int kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf);
 
-int kvm_dev_ioctl_check_extension(long ext);
+int kvm_dev_ioctl_check_extension(struct file *filp, long ext);
 
 int kvm_get_dirty_log(struct kvm *kvm,
 			struct kvm_dirty_log *log, int *is_dirty);
include/uapi/linux/kvm.h +4 -0
index 99c2533..aebfc11 100644
@@ -518,6 +518,10 @@ struct kvm_ppc_smmu_info {
 /* machine type bits, to be used as argument to KVM_CREATE_VM */
 #define KVM_VM_S390_UCONTROL	1
 
+/* on ppc, 0 indicate default, 1 should force HV and 2 PR */
+#define KVM_VM_PPC_HV 1
+#define KVM_VM_PPC_PR 2
+
 #define KVM_S390_SIE_PAGE_OFFSET 1
 
 /*
virt/kvm/kvm_main.c +9 -9
index 0594b22..2d43702 100644
@@ -541,13 +541,13 @@ static void kvm_destroy_dirty_bitmap(struct kvm_memory_slot *memslot)
 /*
  * Free any memory in @free but not in @dont.
  */
-static void kvm_free_physmem_slot(struct kvm_memory_slot *free,
+static void kvm_free_physmem_slot(struct kvm *kvm, struct kvm_memory_slot *free,
 				  struct kvm_memory_slot *dont)
 {
 	if (!dont || free->dirty_bitmap != dont->dirty_bitmap)
 		kvm_destroy_dirty_bitmap(free);
 
-	kvm_arch_free_memslot(free, dont);
+	kvm_arch_free_memslot(kvm, free, dont);
 
 	free->npages = 0;
 }
@@ -558,7 +558,7 @@ void kvm_free_physmem(struct kvm *kvm)
 	struct kvm_memory_slot *memslot;
 
 	kvm_for_each_memslot(memslot, slots)
-		kvm_free_physmem_slot(memslot, NULL);
+		kvm_free_physmem_slot(kvm, memslot, NULL);
 
 	kfree(kvm->memslots);
 }
@@ -822,7 +822,7 @@ int __kvm_set_memory_region(struct kvm *kvm,
 	if (change == KVM_MR_CREATE) {
 		new.userspace_addr = mem->userspace_addr;
 
-		if (kvm_arch_create_memslot(&new, npages))
+		if (kvm_arch_create_memslot(kvm, &new, npages))
 			goto out_free;
 	}
 
@@ -898,7 +898,7 @@ int __kvm_set_memory_region(struct kvm *kvm,
 
 	kvm_arch_commit_memory_region(kvm, mem, &old, change);
 
-	kvm_free_physmem_slot(&old, &new);
+	kvm_free_physmem_slot(kvm, &old, &new);
 	kfree(old_memslots);
 
 	return 0;
@@ -906,7 +906,7 @@ int __kvm_set_memory_region(struct kvm *kvm,
 out_slots:
 	kfree(slots);
 out_free:
-	kvm_free_physmem_slot(&new, &old);
+	kvm_free_physmem_slot(kvm, &new, &old);
 out:
 	return r;
 }
@@ -2579,7 +2579,7 @@ static int kvm_dev_ioctl_create_vm(unsigned long type)
 	return r;
 }
 
-static long kvm_dev_ioctl_check_extension_generic(long arg)
+static long kvm_dev_ioctl_check_extension_generic(struct file *filp, long arg)
 {
 	switch (arg) {
 	case KVM_CAP_USER_MEMORY:
@@ -2603,7 +2603,7 @@ static long kvm_dev_ioctl_check_extension_generic(long arg)
 	default:
 		break;
 	}
-	return kvm_dev_ioctl_check_extension(arg);
+	return kvm_dev_ioctl_check_extension(filp, arg);
 }
 
 static long kvm_dev_ioctl(struct file *filp,
@@ -2622,7 +2622,7 @@ static long kvm_dev_ioctl(struct file *filp,
 		r = kvm_dev_ioctl_create_vm(arg);
 		break;
 	case KVM_CHECK_EXTENSION:
-		r = kvm_dev_ioctl_check_extension_generic(arg);
+		r = kvm_dev_ioctl_check_extension_generic(filp, arg);
 		break;
 	case KVM_GET_VCPU_MMAP_SIZE:
 		r = -EINVAL;
Signature · 1 line
1.8.1.2
#41 lore raw

RFC PATCH 11/11 kvm: powerpc: book3s: Fix module ownership

In reply to Aneesh Kumar K.V

From: "Aneesh Kumar K.V" <[email protected]>

This moves /dev/kvm ownership to kvm.ko module. Depending on which KVM mode we select during VM creation we take a reference count on respective module

---
 arch/powerpc/include/asm/kvm_ppc.h |  1 +
 arch/powerpc/kvm/book3s.c          | 21 +++++++++++++++++++++
 arch/powerpc/kvm/book3s_hv.c       | 15 ++++++---------
 arch/powerpc/kvm/book3s_pr.c       | 15 +++++----------
 arch/powerpc/kvm/powerpc.c         | 19 +++++++++++++++----
 5 files changed, 48 insertions(+), 23 deletions(-)
arch/powerpc/include/asm/kvm_ppc.h +1 -0
index a4a5893..2022720 100644
@@ -171,6 +171,7 @@ extern int kvmppc_xics_int_on(struct kvm *kvm, u32 irq);
 extern int kvmppc_xics_int_off(struct kvm *kvm, u32 irq);
 
 struct kvmppc_ops {
+	struct module *owner;
 	bool is_hv_enabled;
 	int (*get_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
 	int (*set_sregs)(struct kvm_vcpu *vcpu, struct kvm_sregs *sregs);
arch/powerpc/kvm/book3s.c +21 -0
index 34e189c..363df6a 100644
@@ -831,6 +831,10 @@ int kvm_arch_check_processor_compat(void *opaque)
 {
 	int r,cpu;
 	struct kvmppc_ops *kvm_ops = (struct kvmppc_ops *)opaque;
+
+	if (!kvm_ops)
+		return 0;
+
 	for_each_online_cpu(cpu) {
 		smp_call_function_single(cpu,
 					 kvm_ops->check_processor_compat,
@@ -840,6 +844,7 @@ int kvm_arch_check_processor_compat(void *opaque)
 	}
 	return r;
 }
+EXPORT_SYMBOL_GPL(kvm_arch_check_processor_compat);
 
 EXPORT_SYMBOL_GPL(kvm_get_dirty_log);
 EXPORT_SYMBOL_GPL(kvmppc_core_pending_dec);
@@ -851,3 +856,19 @@ EXPORT_SYMBOL_GPL(kvmppc_core_prepare_to_enter);
 EXPORT_SYMBOL_GPL(kvmppc_core_queue_dec);
 EXPORT_SYMBOL_GPL(kvmppc_free_lpid);
 
+static int kvmppc_book3s_init(void)
+{
+	int r;
+
+	r = kvm_init(NULL, sizeof(struct kvm_vcpu), 0, THIS_MODULE);
+	return r;
+}
+
+static void kvmppc_book3s_exit(void)
+{
+	kvm_exit();
+}
+
+module_init(kvmppc_book3s_init);
+module_exit(kvmppc_book3s_exit);
+
arch/powerpc/kvm/book3s_hv.c +6 -9
index 0a684a7..7bdc780 100644
@@ -2092,23 +2092,20 @@ static int kvmppc_book3s_init_hv(void)
 {
 	int r;
 
-	if (!kvmppc_pr_ops) {
-		r = kvm_init(&kvm_ops_hv, sizeof(struct kvm_vcpu),
-			     0, THIS_MODULE);
-		if (r)
-			return r;
-	}
+	r = kvm_arch_check_processor_compat(&kvm_ops_hv);
+	if (r < 0)
+		return r;
+
+	kvm_ops_hv.owner = THIS_MODULE;
 	kvmppc_hv_ops = &kvm_ops_hv;
-	r = kvmppc_mmu_hv_init();
 
+	r = kvmppc_mmu_hv_init();
 	return r;
 }
 
 static void kvmppc_book3s_exit_hv(void)
 {
 	kvmppc_hv_ops = NULL;
-	if (!kvmppc_pr_ops)
-		kvm_exit();
 }
 
 module_init(kvmppc_book3s_init_hv);
arch/powerpc/kvm/book3s_pr.c +5 -10
index e49e4b0..c79fada 100644
@@ -1551,17 +1551,14 @@ static int kvmppc_book3s_init_pr(void)
 {
 	int r;
 
-	if (!kvmppc_hv_ops) {
-		r = kvm_init(&kvm_ops_pr, sizeof(struct kvm_vcpu),
-			     0, THIS_MODULE);
-		if (r)
-			return r;
-	}
-	/* Assign the global value */
+	r = kvm_arch_check_processor_compat(&kvm_ops_pr);
+	if (r < 0)
+		return r;
+
+	kvm_ops_pr.owner = THIS_MODULE;
 	kvmppc_pr_ops = &kvm_ops_pr;
 
 	r = kvmppc_mmu_hpte_sysinit();
-
 	return r;
 }
 
@@ -1569,8 +1566,6 @@ static void kvmppc_book3s_exit_pr(void)
 {
 	kvmppc_pr_ops = NULL;
 	kvmppc_mmu_hpte_sysexit();
-	if (!kvmppc_hv_ops)
-		kvm_exit();
 }
 
 module_init(kvmppc_book3s_init_pr);
arch/powerpc/kvm/powerpc.c +15 -4
index 1209229..677fa7e 100644
@@ -26,6 +26,7 @@
 #include <linux/fs.h>
 #include <linux/slab.h>
 #include <linux/file.h>
+#include <linux/module.h>
 #include <asm/cputable.h>
 #include <asm/uaccess.h>
 #include <asm/kvm_ppc.h>
@@ -270,25 +271,32 @@ void kvm_arch_hardware_unsetup(void)
 
 int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 {
+	struct kvmppc_ops *kvm_ops = NULL;
 	/*
 	 * if we have both HV and PR enabled, default is HV
 	 */
 	if (type == 0) {
 		if (kvmppc_hv_ops)
-			kvm->arch.kvm_ops = kvmppc_hv_ops;
+			kvm_ops = kvmppc_hv_ops;
 		else
-			kvm->arch.kvm_ops = kvmppc_pr_ops;
+			kvm_ops = kvmppc_pr_ops;
+		if (!kvm_ops)
+			goto err_out;
 	} else	if (type == KVM_VM_PPC_HV) {
 		if (!kvmppc_hv_ops)
 			goto err_out;
-		kvm->arch.kvm_ops = kvmppc_hv_ops;
+		kvm_ops = kvmppc_hv_ops;
 	} else if (type == KVM_VM_PPC_PR) {
 		if (!kvmppc_pr_ops)
 			goto err_out;
-		kvm->arch.kvm_ops = kvmppc_pr_ops;
+		kvm_ops = kvmppc_pr_ops;
 	} else
 		goto err_out;
 
+	if (kvm_ops->owner && !try_module_get(kvm_ops->owner))
+		return -ENOENT;
+
+	kvm->arch.kvm_ops = kvm_ops;
 	return kvmppc_core_init_vm(kvm);
 err_out:
 	return -EINVAL;
@@ -311,6 +319,9 @@ void kvm_arch_destroy_vm(struct kvm *kvm)
 	kvmppc_core_destroy_vm(kvm);
 
 	mutex_unlock(&kvm->lock);
+
+	/* drop the module reference */
+	module_put(kvm->arch.kvm_ops->owner);
 }
 
 void kvm_arch_sync_events(struct kvm *kvm)
Signature · 1 line
1.8.1.2
#42 lore raw
In reply to Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V <[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

kvm-all.c +3 -2
index b87215c..a061eda 100644
@@ -1350,7 +1350,7 @@ int kvm_init(void)
     KVMState *s;
     const KVMCapabilityInfo *missing_cap;
     int ret;
-    int i;
+    int i, kvm_type;
     int max_vcpus;
 
     s = g_malloc0(sizeof(KVMState));
@@ -1407,7 +1407,8 @@ int kvm_init(void)
         goto err;
     }
 
-    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
+    kvm_type = qemu_opt_get_number(qemu_get_machine_opts(), "kvm_type", 0);
+    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, kvm_type);
     if (s->vmfd < 0) {
 #ifdef TARGET_S390X
         fprintf(stderr, "Please add the 'switch_amode' kernel parameter to "
vl.c +5 -0
index 4e709d5..4374b17 100644
@@ -427,7 +427,12 @@ static QemuOptsList qemu_machine_opts = {
             .name = "usb",
             .type = QEMU_OPT_BOOL,
             .help = "Set on/off to enable/disable usb",
+        },{
+            .name = "kvm_type",
+            .type = QEMU_OPT_NUMBER,
+            .help = "Set to kvm type to be used in create vm ioctl",
         },
+
         { /* End of list */ }
     },
 };
#43 lore raw
Quoted 41 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V <[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530
Quoted 7 lines

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

Also, users don't want to say type=0. They want to say type=PR or type=HV or type=HV,PR. In fact, can't you make this a property of -accel? Then it's truly accel specific and everything should be well.

Alex

Quoted 46 lines
kvm-all.c +3 -2
index b87215c..a061eda 100644
@@ -1350,7 +1350,7 @@ int kvm_init(void)
    KVMState *s;
    const KVMCapabilityInfo *missing_cap;
    int ret;
-    int i;
+    int i, kvm_type;
    int max_vcpus;
 
    s = g_malloc0(sizeof(KVMState));
@@ -1407,7 +1407,8 @@ int kvm_init(void)
        goto err;
    }
 
-    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
+    kvm_type = qemu_opt_get_number(qemu_get_machine_opts(), "kvm_type", 0);
+    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, kvm_type);
    if (s->vmfd < 0) {
 #ifdef TARGET_S390X
        fprintf(stderr, "Please add the 'switch_amode' kernel parameter to "
vl.c +5 -1
index 4e709d5..4374b17 100644
@@ -427,7 +427,12 @@ static QemuOptsList qemu_machine_opts = {
            .name = "usb",
            .type = QEMU_OPT_BOOL,
            .help = "Set on/off to enable/disable usb",
+        },{
+            .name = "kvm_type",
+            .type = QEMU_OPT_NUMBER,
+            .help = "Set to kvm type to be used in create vm ioctl",
        },
+
        { /* End of list */ }
    },
 };
 
--

To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#44 lore raw
Quoted 49 lines · Alexander Graf
Quoted 47 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V <[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

Quoted 6 lines
Quoted 1 line

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

With the current code CREATE_VM will fail, because we won't have kvm-hv.ko loaded and trying to create a vm with type 1 will fail. Now the challenge related to moving that to machine_init or later is, we depend on HV or PR callback early in CREATE_VM. With the changes we have

int kvmppc_core_init_vm(struct kvm *kvm)
{

#ifdef CONFIG_PPC64
	INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
	INIT_LIST_HEAD(&kvm->arch.rtas_tokens);

#endif

	return kvm->arch.kvm_ops->init_vm(kvm);
}

Also the mmu notifier callback do end up calling kvm_unmap_hva etc which are all HV/PR dependent.

Quoted 4 lines

Also, users don't want to say type=0. They want to say type=PR or type=HV or type=HV,PR. In fact, can't you make this a property of -accel? Then it's truly accel specific and everything should be well.

If we are doing this as machine property, we can't specify string, because "HV"/"PR" are all powerpc dependent, so parsing that is not possible in kvm_init in qemu. But, yes ideally it would be nice to be able to speicy the type using string. I thought accel is a machine property, hence was not sure whether I can have additional properties against that. I was using it as below.

-machine pseries,accel=kvm,kvm_type=1

will look into more details to check whether this can be accel property.

-aneesh

#45 lore raw
Quoted 66 lines · Aneesh Kumar K.V
Quoted 48 lines · Alexander Graf
Quoted 46 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V<[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

Quoted 5 lines
Quoted 1 line

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

With the current code CREATE_VM will fail, because we won't have kvm-hv.ko loaded and trying to create a vm with type 1 will fail. Now the challenge related to moving that to machine_init or later is, we depend on HV or PR callback early in CREATE_VM. With the changes we have

int kvmppc_core_init_vm(struct kvm *kvm)
{

#ifdef CONFIG_PPC64
	INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
	INIT_LIST_HEAD(&kvm->arch.rtas_tokens);
Quoted 7 lines

#endif

	return kvm->arch.kvm_ops->init_vm(kvm);
}

Also the mmu notifier callback do end up calling kvm_unmap_hva etc which are all HV/PR dependent.

Yes, so we should verify in the machine models that we're runnable with the currently selected type at least, to give the user a sensible error message.

Quoted 9 lines
Quoted 6 lines

Also, users don't want to say type=0. They want to say type=PR or type=HV or type=HV,PR. In fact, can't you make this a property of -accel? Then it's truly accel specific and everything should be well.

If we are doing this as machine property, we can't specify string, because "HV"/"PR" are all powerpc dependent, so parsing that is not possible in kvm_init in qemu. But, yes ideally it would be nice to be

Well, we could do the "name to integer" conversion in an arch specific function, no?

Quoted 7 lines

able to speicy the type using string. I thought accel is a machine property, hence was not sure whether I can have additional properties against that. I was using it as below.

-machine pseries,accel=kvm,kvm_type=1

will look into more details to check whether this can be accel property.

Great :).

Alex

#46 lore raw
Quoted 72 lines · Alexander Graf
Quoted 71 lines · Aneesh Kumar K.V
Quoted 48 lines · Alexander Graf
Quoted 46 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V<[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

Quoted 5 lines
Quoted 1 line

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

With the current code CREATE_VM will fail, because we won't have kvm-hv.ko loaded and trying to create a vm with type 1 will fail. Now the challenge related to moving that to machine_init or later is, we depend on HV or PR callback early in CREATE_VM. With the changes we have

int kvmppc_core_init_vm(struct kvm *kvm)
{

#ifdef CONFIG_PPC64
	INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
	INIT_LIST_HEAD(&kvm->arch.rtas_tokens);

#endif

	return kvm->arch.kvm_ops->init_vm(kvm);
}
Quoted 6 lines
Quoted 2 lines

Also the mmu notifier callback do end up calling kvm_unmap_hva etc which are all HV/PR dependent.

Yes, so we should verify in the machine models that we're runnable with the currently selected type at least, to give the user a sensible error message.

Something like the below

hw/ppc/spapr.c +16 -0
index 004184d..7d59ac1 100644
@@ -1337,6 +1337,21 @@ static void ppc_spapr_init(QEMUMachineInitArgs *args)
     assert(spapr->fdt_skel != NULL);
 }
 
+static int spapr_get_vm_type(const char *vm_type)
+{
+    if (!vm_type)
+        return 0;
+
+    if (!strcmp(vm_type, "HV"))
+        return 1;
+
+    if (!strcmp(vm_type, "PR"))
+        return 2;
+
+    hw_error("qemu: unknown kvm_type specified '%s'", vm_type);
+    exit(1);
+}
+
 static QEMUMachine spapr_machine = {
     .name = "pseries",
     .desc = "pSeries Logical Partition (PAPR compliant)",
@@ -1347,6 +1362,7 @@ static QEMUMachine spapr_machine = {
     .max_cpus = MAX_CPUS,
     .no_parallel = 1,
     .default_boot_order = NULL,
+    .get_vm_type = spapr_get_vm_type,
 };
 
 static void spapr_machine_init(void)
include/hw/boards.h +3 -0
index 5a7ae9f..2130488 100644
@@ -21,6 +21,8 @@ typedef void QEMUMachineResetFunc(void);
 
 typedef void QEMUMachineHotAddCPUFunc(const int64_t id, Error **errp);
 
+typedef int QEMUMachineGetVmTypeFunc(const char *arg);
+
 typedef struct QEMUMachine {
     const char *name;
     const char *alias;
@@ -28,6 +30,7 @@ typedef struct QEMUMachine {
     QEMUMachineInitFunc *init;
     QEMUMachineResetFunc *reset;
     QEMUMachineHotAddCPUFunc *hot_add_cpu;
+    QEMUMachineGetVmTypeFunc *get_vm_type;
     BlockInterfaceType block_default_type;
     int max_cpus;
     unsigned int no_serial:1,
include/hw/xen/xen.h +2 -1
index e1f88bf..acc3d74 100644
@@ -36,7 +36,8 @@ void xen_cmos_set_s3_resume(void *opaque, int irq, int level);
 
 qemu_irq *xen_interrupt_controller_init(void);
 
-int xen_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int xen_init(QEMUMachine *machine);
 int xen_hvm_init(MemoryRegion **ram_memory);
 void xenstore_store_pv_console_info(int i, struct CharDriverState *chr);
 
include/sysemu/kvm.h +2 -2
index 9bbe3db..f25caec 100644
@@ -142,8 +142,8 @@ typedef struct KVMState KVMState;
 extern KVMState *kvm_state;
 
 /* external API */
-
-int kvm_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int kvm_init(QEMUMachine *machine);
 
 int kvm_has_sync_mmu(void);
 int kvm_has_vcpu_events(void);
include/sysemu/qtest.h +3 -2
index 9a0c6b3..d71343d 100644
@@ -31,7 +31,8 @@ static inline int qtest_available(void)
     return 1;
 }
 
-int qtest_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int qtest_init(QEMUMachine *machine);
 #else
 static inline bool qtest_enabled(void)
 {
@@ -43,7 +44,7 @@ static inline int qtest_available(void)
     return 0;
 }
 
-static inline int qtest_init(void)
+static inline int qtest_init(QEMUMachine *machine)
 {
     return 0;
 }
kvm-all.c +9 -3
index b87215c..3863abd 100644
@@ -35,6 +35,8 @@
 #include "qemu/event_notifier.h"
 #include "trace.h"
 
+#include "hw/boards.h"
+
 /* This check must be after config-host.h is included */
 #ifdef CONFIG_EVENTFD
 #include <sys/eventfd.h>
@@ -1342,7 +1344,7 @@ static int kvm_max_vcpus(KVMState *s)
     return 4;
 }
 
-int kvm_init(void)
+int kvm_init(QEMUMachine *machine)
 {
     static const char upgrade_note[] =
         "Please upgrade to at least kernel 2.6.29 or recent kvm-kmod\n"
@@ -1350,7 +1352,7 @@ int kvm_init(void)
     KVMState *s;
     const KVMCapabilityInfo *missing_cap;
     int ret;
-    int i;
+    int i, kvm_type = 0;
     int max_vcpus;
 
     s = g_malloc0(sizeof(KVMState));
@@ -1407,7 +1409,11 @@ int kvm_init(void)
         goto err;
     }
 
-    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
+    if (machine->get_vm_type) {
+        kvm_type = machine->get_vm_type(qemu_opt_get(qemu_get_machine_opts(),
+                                                     "kvm_type"));
+    }
+    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, kvm_type);
     if (s->vmfd < 0) {
 #ifdef TARGET_S390X
         fprintf(stderr, "Please add the 'switch_amode' kernel parameter to "
kvm-stub.c +3 -1
index 548f471..ccb7b8c 100644
@@ -19,6 +19,8 @@
 #include "hw/pci/msi.h"
 #endif
 
+#include "hw/boards.h"
+
 KVMState *kvm_state;
 bool kvm_kernel_irqchip;
 bool kvm_async_interrupts_allowed;
@@ -33,7 +35,7 @@ int kvm_init_vcpu(CPUState *cpu)
     return -ENOSYS;
 }
 
-int kvm_init(void)
+int kvm_init(QEMUMachine *machine)
 {
     return -ENOSYS;
 }
qtest.c +1 -1
index 584c707..ef3c473 100644
@@ -502,7 +502,7 @@ static void qtest_event(void *opaque, int event)
     }
 }
 
-int qtest_init(void)
+int qtest_init(QEMUMachine *machine)
 {
     CharDriverState *chr;
 
vl.c +11 -6
index 4e709d5..7ecc581 100644
@@ -427,7 +427,12 @@ static QemuOptsList qemu_machine_opts = {
             .name = "usb",
             .type = QEMU_OPT_BOOL,
             .help = "Set on/off to enable/disable usb",
+        },{
+            .name = "kvm_type",
+            .type = QEMU_OPT_STRING,
+            .help = "Set to kvm type to be used in create vm ioctl",
         },
+
         { /* End of list */ }
     },
 };
@@ -2608,7 +2613,7 @@ static QEMUMachine *machine_parse(const char *name)
     exit(!name || !is_help_option(name));
 }
 
-static int tcg_init(void)
+static int tcg_init(QEMUMachine *machine)
 {
     tcg_exec_init(tcg_tb_size * 1024 * 1024);
     return 0;
@@ -2618,7 +2623,7 @@ static struct {
     const char *opt_name;
     const char *name;
     int (*available)(void);
-    int (*init)(void);
+    int (*init)(QEMUMachine *);
     bool *allowed;
 } accel_list[] = {
     { "tcg", "tcg", tcg_available, tcg_init, &tcg_allowed },
@@ -2627,7 +2632,7 @@ static struct {
     { "qtest", "QTest", qtest_available, qtest_init, &qtest_allowed },
 };
 
-static int configure_accelerator(void)
+static int configure_accelerator(QEMUMachine *machine)
 {
     const char *p;
     char buf[10];
@@ -2654,7 +2659,7 @@ static int configure_accelerator(void)
                     continue;
                 }
                 *(accel_list[i].allowed) = true;
-                ret = accel_list[i].init();
+                ret = accel_list[i].init(machine);
                 if (ret < 0) {
                     init_failed = true;
                     fprintf(stderr, "failed to initialize %s: %s\n",
@@ -4037,10 +4042,10 @@ int main(int argc, char **argv, char **envp)
         exit(0);
     }
 
-    configure_accelerator();
+    configure_accelerator(machine);
 
     if (!qtest_enabled() && qtest_chrdev) {
-        qtest_init();
+        qtest_init(machine);
     }
 
     machine_opts = qemu_get_machine_opts();
xen-all.c +1 -1
index 839f14f..ac3654b 100644
@@ -1000,7 +1000,7 @@ static void xen_exit_notifier(Notifier *n, void *data)
     xs_daemon_close(state->xenstore);
 }
 
-int xen_init(void)
+int xen_init(QEMUMachine *machine)
 {
     xen_xc = xen_xc_interface_open(0, 0, 0);
     if (xen_xc == XC_HANDLER_INITIAL_VALUE) {
xen-stub.c +1 -1
index ad189a6..59927cb 100644
@@ -47,7 +47,7 @@ qemu_irq *xen_interrupt_controller_init(void)
     return NULL;
 }
 
-int xen_init(void)
+int xen_init(QEMUMachine *machine)
 {
     return -ENOSYS;
 }
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

Also, users don't want to say type=0. They want to say type=PR or type=HV or type=HV,PR. In fact, can't you make this a property of -accel? Then it's truly accel specific and everything should be well.

Quoted 12 lines
Quoted 3 lines

If we are doing this as machine property, we can't specify string, because "HV"/"PR" are all powerpc dependent, so parsing that is not possible in kvm_init in qemu. But, yes ideally it would be nice to be

Well, we could do the "name to integer" conversion in an arch specific function, no?

Quoted 5 lines

able to speicy the type using string. I thought accel is a machine property, hence was not sure whether I can have additional properties against that. I was using it as below.

-machine pseries,accel=kvm,kvm_type=1

Can we really specific -accel ? I check and I am finding that as machine property.

-aneesh

#47 lore raw
Quoted 74 lines · Aneesh Kumar K.V
Quoted 72 lines · Alexander Graf
Quoted 71 lines · Aneesh Kumar K.V
Quoted 48 lines · Alexander Graf
Quoted 46 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V<[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

Quoted 5 lines
Quoted 1 line

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

With the current code CREATE_VM will fail, because we won't have kvm-hv.ko loaded and trying to create a vm with type 1 will fail. Now the challenge related to moving that to machine_init or later is, we depend on HV or PR callback early in CREATE_VM. With the changes we have

int kvmppc_core_init_vm(struct kvm *kvm)
{

#ifdef CONFIG_PPC64
	INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
	INIT_LIST_HEAD(&kvm->arch.rtas_tokens);

#endif

	return kvm->arch.kvm_ops->init_vm(kvm);
}
Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

Also the mmu notifier callback do end up calling kvm_unmap_hva etc which are all HV/PR dependent.

Yes, so we should verify in the machine models that we're runnable with the currently selected type at least, to give the user a sensible error message.

Something like the below

I like that one a lot. Andreas, Paolo, what do you think?

Alex

Quoted 301 lines
hw/ppc/spapr.c +16 -0
index 004184d..7d59ac1 100644
@@ -1337,6 +1337,21 @@ static void ppc_spapr_init(QEMUMachineInitArgs *args)
      assert(spapr->fdt_skel != NULL);
  }
 
+static int spapr_get_vm_type(const char *vm_type)
+{
+    if (!vm_type)
+        return 0;
+
+    if (!strcmp(vm_type, "HV"))
+        return 1;
+
+    if (!strcmp(vm_type, "PR"))
+        return 2;
+
+    hw_error("qemu: unknown kvm_type specified '%s'", vm_type);
+    exit(1);
+}
+
  static QEMUMachine spapr_machine = {
      .name = "pseries",
      .desc = "pSeries Logical Partition (PAPR compliant)",
@@ -1347,6 +1362,7 @@ static QEMUMachine spapr_machine = {
      .max_cpus = MAX_CPUS,
      .no_parallel = 1,
      .default_boot_order = NULL,
+    .get_vm_type = spapr_get_vm_type,
  };
 
  static void spapr_machine_init(void)
include/hw/boards.h +3 -0
index 5a7ae9f..2130488 100644
@@ -21,6 +21,8 @@ typedef void QEMUMachineResetFunc(void);
 
  typedef void QEMUMachineHotAddCPUFunc(const int64_t id, Error **errp);
 
+typedef int QEMUMachineGetVmTypeFunc(const char *arg);
+
  typedef struct QEMUMachine {
      const char *name;
      const char *alias;
@@ -28,6 +30,7 @@ typedef struct QEMUMachine {
      QEMUMachineInitFunc *init;
      QEMUMachineResetFunc *reset;
      QEMUMachineHotAddCPUFunc *hot_add_cpu;
+    QEMUMachineGetVmTypeFunc *get_vm_type;
      BlockInterfaceType block_default_type;
      int max_cpus;
      unsigned int no_serial:1,
include/hw/xen/xen.h +2 -1
index e1f88bf..acc3d74 100644
@@ -36,7 +36,8 @@ void xen_cmos_set_s3_resume(void *opaque, int irq, int level);
 
  qemu_irq *xen_interrupt_controller_init(void);
 
-int xen_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int xen_init(QEMUMachine *machine);
  int xen_hvm_init(MemoryRegion **ram_memory);
  void xenstore_store_pv_console_info(int i, struct CharDriverState *chr);
 
include/sysemu/kvm.h +2 -2
index 9bbe3db..f25caec 100644
@@ -142,8 +142,8 @@ typedef struct KVMState KVMState;
  extern KVMState *kvm_state;
 
  /* external API */
-
-int kvm_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int kvm_init(QEMUMachine *machine);
 
  int kvm_has_sync_mmu(void);
  int kvm_has_vcpu_events(void);
include/sysemu/qtest.h +3 -2
index 9a0c6b3..d71343d 100644
@@ -31,7 +31,8 @@ static inline int qtest_available(void)
      return 1;
  }
 
-int qtest_init(void);
+typedef struct QEMUMachine QEMUMachine;
+int qtest_init(QEMUMachine *machine);
  #else
  static inline bool qtest_enabled(void)
  {
@@ -43,7 +44,7 @@ static inline int qtest_available(void)
      return 0;
  }
 
-static inline int qtest_init(void)
+static inline int qtest_init(QEMUMachine *machine)
  {
      return 0;
  }
kvm-all.c +9 -3
index b87215c..3863abd 100644
@@ -35,6 +35,8 @@
  #include "qemu/event_notifier.h"
  #include "trace.h"
 
+#include "hw/boards.h"
+
  /* This check must be after config-host.h is included */
  #ifdef CONFIG_EVENTFD
  #include<sys/eventfd.h>
@@ -1342,7 +1344,7 @@ static int kvm_max_vcpus(KVMState *s)
      return 4;
  }
 
-int kvm_init(void)
+int kvm_init(QEMUMachine *machine)
  {
      static const char upgrade_note[] =
          "Please upgrade to at least kernel 2.6.29 or recent kvm-kmod\n"
@@ -1350,7 +1352,7 @@ int kvm_init(void)
      KVMState *s;
      const KVMCapabilityInfo *missing_cap;
      int ret;
-    int i;
+    int i, kvm_type = 0;
      int max_vcpus;
 
      s = g_malloc0(sizeof(KVMState));
@@ -1407,7 +1409,11 @@ int kvm_init(void)
          goto err;
      }
 
-    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, 0);
+    if (machine->get_vm_type) {
+        kvm_type = machine->get_vm_type(qemu_opt_get(qemu_get_machine_opts(),
+                                                     "kvm_type"));
+    }
+    s->vmfd = kvm_ioctl(s, KVM_CREATE_VM, kvm_type);
      if (s->vmfd<  0) {
  #ifdef TARGET_S390X
          fprintf(stderr, "Please add the 'switch_amode' kernel parameter to "
kvm-stub.c +3 -1
index 548f471..ccb7b8c 100644
@@ -19,6 +19,8 @@
  #include "hw/pci/msi.h"
  #endif
 
+#include "hw/boards.h"
+
  KVMState *kvm_state;
  bool kvm_kernel_irqchip;
  bool kvm_async_interrupts_allowed;
@@ -33,7 +35,7 @@ int kvm_init_vcpu(CPUState *cpu)
      return -ENOSYS;
  }
 
-int kvm_init(void)
+int kvm_init(QEMUMachine *machine)
  {
      return -ENOSYS;
  }
qtest.c +1 -1
index 584c707..ef3c473 100644
@@ -502,7 +502,7 @@ static void qtest_event(void *opaque, int event)
      }
  }
 
-int qtest_init(void)
+int qtest_init(QEMUMachine *machine)
  {
      CharDriverState *chr;
 
vl.c +11 -6
index 4e709d5..7ecc581 100644
@@ -427,7 +427,12 @@ static QemuOptsList qemu_machine_opts = {
              .name = "usb",
              .type = QEMU_OPT_BOOL,
              .help = "Set on/off to enable/disable usb",
+        },{
+            .name = "kvm_type",
+            .type = QEMU_OPT_STRING,
+            .help = "Set to kvm type to be used in create vm ioctl",
          },
+
          { /* End of list */ }
      },
  };
@@ -2608,7 +2613,7 @@ static QEMUMachine *machine_parse(const char *name)
      exit(!name || !is_help_option(name));
  }
 
-static int tcg_init(void)
+static int tcg_init(QEMUMachine *machine)
  {
      tcg_exec_init(tcg_tb_size * 1024 * 1024);
      return 0;
@@ -2618,7 +2623,7 @@ static struct {
      const char *opt_name;
      const char *name;
      int (*available)(void);
-    int (*init)(void);
+    int (*init)(QEMUMachine *);
      bool *allowed;
  } accel_list[] = {
      { "tcg", "tcg", tcg_available, tcg_init,&tcg_allowed },
@@ -2627,7 +2632,7 @@ static struct {
      { "qtest", "QTest", qtest_available, qtest_init,&qtest_allowed },
  };
 
-static int configure_accelerator(void)
+static int configure_accelerator(QEMUMachine *machine)
  {
      const char *p;
      char buf[10];
@@ -2654,7 +2659,7 @@ static int configure_accelerator(void)
                      continue;
                  }
                  *(accel_list[i].allowed) = true;
-                ret = accel_list[i].init();
+                ret = accel_list[i].init(machine);
                  if (ret<  0) {
                      init_failed = true;
                      fprintf(stderr, "failed to initialize %s: %s\n",
@@ -4037,10 +4042,10 @@ int main(int argc, char **argv, char **envp)
          exit(0);
      }
 
-    configure_accelerator();
+    configure_accelerator(machine);
 
      if (!qtest_enabled()&&  qtest_chrdev) {
-        qtest_init();
+        qtest_init(machine);
      }
 
      machine_opts = qemu_get_machine_opts();
xen-all.c +1 -1
index 839f14f..ac3654b 100644
@@ -1000,7 +1000,7 @@ static void xen_exit_notifier(Notifier *n, void *data)
      xs_daemon_close(state->xenstore);
  }
 
-int xen_init(void)
+int xen_init(QEMUMachine *machine)
  {
      xen_xc = xen_xc_interface_open(0, 0, 0);
      if (xen_xc == XC_HANDLER_INITIAL_VALUE) {
xen-stub.c +1 -1
index ad189a6..59927cb 100644
@@ -47,7 +47,7 @@ qemu_irq *xen_interrupt_controller_init(void)
      return NULL;
  }
 
-int xen_init(void)
+int xen_init(QEMUMachine *machine)
  {
      return -ENOSYS;
  }
Quoted 17 lines
Quoted 9 lines
Quoted 6 lines

Also, users don't want to say type=0. They want to say type=PR or type=HV or type=HV,PR. In fact, can't you make this a property of -accel? Then it's truly accel specific and everything should be well.

If we are doing this as machine property, we can't specify string, because "HV"/"PR" are all powerpc dependent, so parsing that is not possible in kvm_init in qemu. But, yes ideally it would be nice to be

Well, we could do the "name to integer" conversion in an arch specific function, no?

Quoted 5 lines

able to speicy the type using string. I thought accel is a machine property, hence was not sure whether I can have additional properties against that. I was using it as below.

-machine pseries,accel=kvm,kvm_type=1

Can we really specific -accel ? I check and I am finding that as machine property.

-aneesh

maintainer
#48 lore raw

Il 01/10/2013 13:36, Alexander Graf ha scritto:

Quoted 7 lines
Quoted 5 lines
Quoted 4 lines
Quoted 1 line

Yes, so we should verify in the machine models that we're runnable with the currently selected type at least, to give the user a sensible error message.

Something like the below

I like that one a lot. Andreas, Paolo, what do you think?

Yes, it's fine.

Paolo

#49 lore raw
In reply to Alexander Graf
Quoted 76 lines · Alexander Graf
Quoted 75 lines · Aneesh Kumar K.V
Quoted 73 lines · Alexander Graf
Quoted 72 lines · Aneesh Kumar K.V
Quoted 48 lines · Alexander Graf
Quoted 46 lines · Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Hi All,

This patch series support enabling HV and PR KVM together in the same kernel. We extend machine property with new property "kvm_type". A value of 1 will force HV KVM and 2 PR KVM. The default value is 0 which will select the fastest KVM mode. ie, HV if that is supported otherwise PR.

With Qemu command line having

-machine pseries,accel=kvm,kvm_type=1

[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]# modprobe kvm-hv
[root@llmp24l02 qemu]# bash ../qemu

now with

-machine pseries,accel=kvm,kvm_type=2

[root@llmp24l02 qemu]# rmmod kvm-pr
[root@llmp24l02 qemu]# bash ../qemu
failed to initialize KVM: Invalid argument
[root@llmp24l02 qemu]#
[root@llmp24l02 qemu]# modprobe kvm-pr
[root@llmp24l02 qemu]# bash ../qemu

Quoted 2 lines

if don't specify kvm_type machine property, it will take a default value 0, which means fastest supported.

Related qemu patch

commit 8d139053177d48a70cb710b211ea4c2843eccdfb

Author: Aneesh Kumar K.V<[email protected]>
Date:   Mon Sep 23 12:28:37 2013 +0530

kvm: Add a new machine property kvm_type

Targets like ppc64 support different type of KVM, one which use hypervisor mode and the other which doesn't. Add a new machine property kvm_type that helps in selecting the respective ones

Quoted 6 lines
Quoted 2 lines

Signed-off-by: Aneesh Kumar
K.V<[email protected]>

This really is too early, as we can't possibly run in HV mode for non-pseries machines, so the interpretation (or at least sanity checking) of what values are reasonable should occur in the machine. That's why it's a variable in the "machine opts".

With the current code CREATE_VM will fail, because we won't have kvm-hv.ko loaded and trying to create a vm with type 1 will fail. Now the challenge related to moving that to machine_init or later is, we depend on HV or PR callback early in CREATE_VM. With the changes we have

int kvmppc_core_init_vm(struct kvm *kvm)
{

#ifdef CONFIG_PPC64
    INIT_LIST_HEAD(&kvm->arch.spapr_tce_tables);
    INIT_LIST_HEAD(&kvm->arch.rtas_tokens);

#endif

    return kvm->arch.kvm_ops->init_vm(kvm);
}
Quoted 8 lines
Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

Also the mmu notifier callback do end up calling kvm_unmap_hva etc which are all HV/PR dependent.

Yes, so we should verify in the machine models that we're runnable with the currently selected type at least, to give the user a sensible error message.

Something like the below

I like that one a lot. Andreas, Paolo, what do you think?

To clarify this a bit more, this patch is missing a few critical pieces:

1) The default implementation for the kvm_type callback needs to
error out when kvm_type is set to anything
2) All non-papr-non-e500-PPC machines should force "PR only" mode and
bail out on anything else
3) We're missing sensible error messages for the user still when the
VM could not get created

but I like the overall concept :).

Alex