diff mbox series

[v4,44/46] KVM: PPC: Book3S HV P9: implement hash guest support

Message ID 20210323010305.1045293-45-npiggin@gmail.com
State New
Headers show
Series KVM: PPC: Book3S: C-ify the P9 entry/exit code | expand

Commit Message

Nicholas Piggin March 23, 2021, 1:03 a.m. UTC 
Guest entry/exit has to restore and save/clear the SLB, plus several
other bits to accommodate hash guests in the P9 path.

Radix host, hash guest support is removed from the P7/8 path.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kvm/book3s_hv.c            |  20 ++-
 arch/powerpc/kvm/book3s_hv_interrupt.c  | 160 ++++++++++++++++--------
 arch/powerpc/kvm/book3s_hv_rm_mmu.c     |   4 +
 arch/powerpc/kvm/book3s_hv_rmhandlers.S |  14 +--
 4 files changed, 131 insertions(+), 67 deletions(-)

Comments

Fabiano Rosas March 23, 2021, 3:53 p.m. UTC | #1 
Nicholas Piggin <npiggin@gmail.com> writes:

> Guest entry/exit has to restore and save/clear the SLB, plus several
> other bits to accommodate hash guests in the P9 path.
>
> Radix host, hash guest support is removed from the P7/8 path.
>
> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> ---

<snip>

> diff --git a/arch/powerpc/kvm/book3s_hv_interrupt.c b/arch/powerpc/kvm/book3s_hv_interrupt.c
> index cd84d2c37632..03fbfef708a8 100644
> --- a/arch/powerpc/kvm/book3s_hv_interrupt.c
> +++ b/arch/powerpc/kvm/book3s_hv_interrupt.c
> @@ -55,6 +55,50 @@ static void __accumulate_time(struct kvm_vcpu *vcpu, struct kvmhv_tb_accumulator
>  #define accumulate_time(vcpu, next) do {} while (0)
>  #endif
>
> +static inline void mfslb(unsigned int idx, u64 *slbee, u64 *slbev)
> +{
> +	asm volatile("slbmfev  %0,%1" : "=r" (*slbev) : "r" (idx));
> +	asm volatile("slbmfee  %0,%1" : "=r" (*slbee) : "r" (idx));
> +}
> +
> +static inline void __mtslb(u64 slbee, u64 slbev)
> +{
> +	asm volatile("slbmte %0,%1" :: "r" (slbev), "r" (slbee));
> +}
> +
> +static inline void mtslb(unsigned int idx, u64 slbee, u64 slbev)
> +{
> +	BUG_ON((slbee & 0xfff) != idx);
> +
> +	__mtslb(slbee, slbev);
> +}
> +
> +static inline void slb_invalidate(unsigned int ih)
> +{
> +	asm volatile("slbia %0" :: "i"(ih));
> +}

Fyi, in my environment the assembler complains:

{standard input}: Assembler messages:                                    
{standard input}:1293: Error: junk at end of line: `6'                             
{standard input}:2138: Error: junk at end of line: `6'                    
make[3]: *** [../scripts/Makefile.build:271:
arch/powerpc/kvm/book3s_hv_interrupt.o] Error 1

This works:

-       asm volatile("slbia %0" :: "i"(ih));
+       asm volatile(PPC_SLBIA(%0) :: "i"(ih));

But I don't know what is going on.
Nicholas Piggin March 24, 2021, 1:34 a.m. UTC | #2 
Excerpts from Fabiano Rosas's message of March 24, 2021 1:53 am:
> Nicholas Piggin <npiggin@gmail.com> writes:
> 
>> Guest entry/exit has to restore and save/clear the SLB, plus several
>> other bits to accommodate hash guests in the P9 path.
>>
>> Radix host, hash guest support is removed from the P7/8 path.
>>
>> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
>> ---
> 
> <snip>
> 
>> diff --git a/arch/powerpc/kvm/book3s_hv_interrupt.c b/arch/powerpc/kvm/book3s_hv_interrupt.c
>> index cd84d2c37632..03fbfef708a8 100644
>> --- a/arch/powerpc/kvm/book3s_hv_interrupt.c
>> +++ b/arch/powerpc/kvm/book3s_hv_interrupt.c
>> @@ -55,6 +55,50 @@ static void __accumulate_time(struct kvm_vcpu *vcpu, struct kvmhv_tb_accumulator
>>  #define accumulate_time(vcpu, next) do {} while (0)
>>  #endif
>>
>> +static inline void mfslb(unsigned int idx, u64 *slbee, u64 *slbev)
>> +{
>> +	asm volatile("slbmfev  %0,%1" : "=r" (*slbev) : "r" (idx));
>> +	asm volatile("slbmfee  %0,%1" : "=r" (*slbee) : "r" (idx));
>> +}
>> +
>> +static inline void __mtslb(u64 slbee, u64 slbev)
>> +{
>> +	asm volatile("slbmte %0,%1" :: "r" (slbev), "r" (slbee));
>> +}
>> +
>> +static inline void mtslb(unsigned int idx, u64 slbee, u64 slbev)
>> +{
>> +	BUG_ON((slbee & 0xfff) != idx);
>> +
>> +	__mtslb(slbee, slbev);
>> +}
>> +
>> +static inline void slb_invalidate(unsigned int ih)
>> +{
>> +	asm volatile("slbia %0" :: "i"(ih));
>> +}
> 
> Fyi, in my environment the assembler complains:
> 
> {standard input}: Assembler messages:                                    
> {standard input}:1293: Error: junk at end of line: `6'                             
> {standard input}:2138: Error: junk at end of line: `6'                    
> make[3]: *** [../scripts/Makefile.build:271:
> arch/powerpc/kvm/book3s_hv_interrupt.o] Error 1
> 
> This works:
> 
> -       asm volatile("slbia %0" :: "i"(ih));
> +       asm volatile(PPC_SLBIA(%0) :: "i"(ih));
> 
> But I don't know what is going on.

Ah yes, we still need to use PPC_SLBIA. IH parameter to slbia was only 
added in binutils 2.27 and we support down to 2.23.

Thanks for the fix I'll add it.

Thanks,
Nick
diff mbox series

Patch

diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
index 5b3b6842d6b9..c151a60c0daa 100644
--- a/arch/powerpc/kvm/book3s_hv.c
+++ b/arch/powerpc/kvm/book3s_hv.c
@@ -3828,7 +3828,8 @@  static int kvmhv_p9_guest_entry(struct kvm_vcpu *vcpu, u64 time_limit,
 		}
 		kvmppc_xive_pull_vcpu(vcpu);
 
-		vcpu->arch.slb_max = 0;
+		if (kvm_is_radix(vcpu->kvm))
+			vcpu->arch.slb_max = 0;
 	}
 
 	dec = mfspr(SPRN_DEC);
@@ -4061,7 +4062,6 @@  static void kvmppc_vcore_blocked(struct kvmppc_vcore *vc)
 /*
  * This never fails for a radix guest, as none of the operations it does
  * for a radix guest can fail or have a way to report failure.
- * kvmhv_run_single_vcpu() relies on this fact.
  */
 static int kvmhv_setup_mmu(struct kvm_vcpu *vcpu)
 {
@@ -4240,8 +4240,15 @@  int kvmhv_run_single_vcpu(struct kvm_vcpu *vcpu, u64 time_limit,
 	vc->runner = vcpu;
 
 	/* See if the MMU is ready to go */
-	if (!kvm->arch.mmu_ready)
-		kvmhv_setup_mmu(vcpu);
+	if (!kvm->arch.mmu_ready) {
+		r = kvmhv_setup_mmu(vcpu);
+		if (r) {
+			run->exit_reason = KVM_EXIT_FAIL_ENTRY;
+			run->fail_entry.hardware_entry_failure_reason = 0;
+			vcpu->arch.ret = r;
+			return r;
+		}
+	}
 
 	if (need_resched())
 		cond_resched();
@@ -4254,7 +4261,8 @@  int kvmhv_run_single_vcpu(struct kvm_vcpu *vcpu, u64 time_limit,
 	preempt_disable();
 	pcpu = smp_processor_id();
 	vc->pcpu = pcpu;
-	kvmppc_prepare_radix_vcpu(vcpu, pcpu);
+	if (kvm_is_radix(kvm))
+		kvmppc_prepare_radix_vcpu(vcpu, pcpu);
 
 	local_irq_disable();
 	hard_irq_disable();
@@ -4454,7 +4462,7 @@  static int kvmppc_vcpu_run_hv(struct kvm_vcpu *vcpu)
 	vcpu->arch.state = KVMPPC_VCPU_BUSY_IN_HOST;
 
 	do {
-		if (kvm_is_radix(kvm))
+		if (radix_enabled())
 			r = kvmhv_run_single_vcpu(vcpu, ~(u64)0,
 						  vcpu->arch.vcore->lpcr);
 		else
diff --git a/arch/powerpc/kvm/book3s_hv_interrupt.c b/arch/powerpc/kvm/book3s_hv_interrupt.c
index cd84d2c37632..03fbfef708a8 100644
--- a/arch/powerpc/kvm/book3s_hv_interrupt.c
+++ b/arch/powerpc/kvm/book3s_hv_interrupt.c
@@ -55,6 +55,50 @@  static void __accumulate_time(struct kvm_vcpu *vcpu, struct kvmhv_tb_accumulator
 #define accumulate_time(vcpu, next) do {} while (0)
 #endif
 
+static inline void mfslb(unsigned int idx, u64 *slbee, u64 *slbev)
+{
+	asm volatile("slbmfev  %0,%1" : "=r" (*slbev) : "r" (idx));
+	asm volatile("slbmfee  %0,%1" : "=r" (*slbee) : "r" (idx));
+}
+
+static inline void __mtslb(u64 slbee, u64 slbev)
+{
+	asm volatile("slbmte %0,%1" :: "r" (slbev), "r" (slbee));
+}
+
+static inline void mtslb(unsigned int idx, u64 slbee, u64 slbev)
+{
+	BUG_ON((slbee & 0xfff) != idx);
+
+	__mtslb(slbee, slbev);
+}
+
+static inline void slb_invalidate(unsigned int ih)
+{
+	asm volatile("slbia %0" :: "i"(ih));
+}
+
+/*
+ * Malicious or buggy radix guests may have inserted SLB entries
+ * (only 0..3 because radix always runs with UPRT=1), so these must
+ * be cleared here to avoid side-channels. slbmte is used rather
+ * than slbia, as it won't clear cached translations.
+ */
+static void radix_clear_slb(void)
+{
+	u64 slbee, slbev;
+	int i;
+
+	for (i = 0; i < 4; i++) {
+		mfslb(i, &slbee, &slbev);
+		if (unlikely(slbee || slbev)) {
+			slbee = i;
+			slbev = 0;
+			mtslb(i, slbee, slbev);
+		}
+	}
+}
+
 static void switch_mmu_to_guest_radix(struct kvm *kvm, struct kvm_vcpu *vcpu, u64 lpcr)
 {
 	struct kvmppc_vcore *vc = vcpu->arch.vcore;
@@ -80,6 +124,31 @@  static void switch_mmu_to_guest_radix(struct kvm *kvm, struct kvm_vcpu *vcpu, u6
 	kvmppc_check_need_tlb_flush(kvm, vc->pcpu, nested);
 }
 
+static void switch_mmu_to_guest_hpt(struct kvm *kvm, struct kvm_vcpu *vcpu, u64 lpcr)
+{
+	struct kvm_nested_guest *nested = vcpu->arch.nested;
+	u32 lpid;
+	int i;
+
+	BUG_ON(nested);
+
+	lpid = kvm->arch.lpid;
+
+	mtspr(SPRN_LPID, lpid);
+	mtspr(SPRN_LPCR, lpcr);
+	mtspr(SPRN_PID, vcpu->arch.pid);
+
+	for (i = 0; i < vcpu->arch.slb_max; i++)
+		__mtslb(vcpu->arch.slb[i].orige, vcpu->arch.slb[i].origv);
+
+	isync();
+
+	/*
+	 * TLBIEL is not virtualised for HPT guests, so check_need_tlb_flush
+	 * is not required here.
+	 */
+}
+
 static void switch_mmu_to_host_radix(struct kvm *kvm, u32 pid)
 {
 	isync();
@@ -91,42 +160,30 @@  static void switch_mmu_to_host_radix(struct kvm *kvm, u32 pid)
 	isync();
 }
 
-static inline void mfslb(unsigned int idx, u64 *slbee, u64 *slbev)
-{
-	asm volatile("slbmfev  %0,%1" : "=r" (*slbev) : "r" (idx));
-	asm volatile("slbmfee  %0,%1" : "=r" (*slbee) : "r" (idx));
-}
-
-static inline void mtslb(unsigned int idx, u64 slbee, u64 slbev)
-{
-	BUG_ON((slbee & 0xfff) != idx);
-
-	asm volatile("slbmte %0,%1" :: "r" (slbev), "r" (slbee));
-}
-
-static inline void slb_invalidate(unsigned int ih)
-{
-	asm volatile("slbia %0" :: "i"(ih));
-}
-
-/*
- * Malicious or buggy radix guests may have inserted SLB entries
- * (only 0..3 because radix always runs with UPRT=1), so these must
- * be cleared here to avoid side-channels. slbmte is used rather
- * than slbia, as it won't clear cached translations.
- */
-static void radix_clear_slb(void)
+static void save_clear_guest_mmu(struct kvm *kvm, struct kvm_vcpu *vcpu)
 {
-	u64 slbee, slbev;
-	int i;
+	if (kvm_is_radix(kvm)) {
+		radix_clear_slb();
+	} else {
+		int i;
+		int nr = 0;
 
-	for (i = 0; i < 4; i++) {
-		mfslb(i, &slbee, &slbev);
-		if (unlikely(slbee || slbev)) {
-			slbee = i;
-			slbev = 0;
-			mtslb(i, slbee, slbev);
+		/*
+		 * This must run before switching to host (radix host can't
+		 * access all SLBs).
+		 */
+		for (i = 0; i < vcpu->arch.slb_nr; i++) {
+			u64 slbee, slbev;
+			mfslb(i, &slbee, &slbev);
+			if (slbee & SLB_ESID_V) {
+				vcpu->arch.slb[nr].orige = slbee | i;
+				vcpu->arch.slb[nr].origv = slbev;
+				nr++;
+			}
 		}
+		vcpu->arch.slb_max = nr;
+		mtslb(0, 0, 0);
+		slb_invalidate(6);
 	}
 }
 
@@ -234,10 +291,18 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 
 	mtspr(SPRN_AMOR, ~0UL);
 
-	if (cpu_has_feature(CPU_FTR_P9_RADIX_PREFETCH_BUG))
-		__mtmsrd(msr & ~(MSR_IR|MSR_DR|MSR_RI), 0);
+	local_paca->kvm_hstate.in_guest = KVM_GUEST_MODE_GUEST_HV_FAST;
+	if (kvm_is_radix(kvm)) {
+		if (cpu_has_feature(CPU_FTR_P9_RADIX_PREFETCH_BUG))
+			__mtmsrd(msr & ~(MSR_IR|MSR_DR|MSR_RI), 0);
+		switch_mmu_to_guest_radix(kvm, vcpu, lpcr);
+		if (!cpu_has_feature(CPU_FTR_P9_RADIX_PREFETCH_BUG))
+			__mtmsrd(0, 1); /* clear RI */
 
-	switch_mmu_to_guest_radix(kvm, vcpu, lpcr);
+	} else {
+		__mtmsrd(msr & ~(MSR_IR|MSR_DR|MSR_RI), 0);
+		switch_mmu_to_guest_hpt(kvm, vcpu, lpcr);
+	}
 
 	/*
 	 * P9 suppresses the HDEC exception when LPCR[HDICE] = 0,
@@ -245,9 +310,6 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 	 */
 	mtspr(SPRN_HDEC, hdec);
 
-	if (!cpu_has_feature(CPU_FTR_P9_RADIX_PREFETCH_BUG))
-		__mtmsrd(0, 1); /* clear RI */
-
 	mtspr(SPRN_DAR, vcpu->arch.shregs.dar);
 	mtspr(SPRN_DSISR, vcpu->arch.shregs.dsisr);
 	mtspr(SPRN_SRR0, vcpu->arch.shregs.srr0);
@@ -255,10 +317,7 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 
 	accumulate_time(vcpu, &vcpu->arch.guest_time);
 
-	local_paca->kvm_hstate.in_guest = KVM_GUEST_MODE_GUEST_HV_FAST;
 	kvmppc_p9_enter_guest(vcpu);
-	// Radix host and guest means host never runs with guest MMU state
-	local_paca->kvm_hstate.in_guest = KVM_GUEST_MODE_NONE;
 
 	accumulate_time(vcpu, &vcpu->arch.rm_intr);
 
@@ -359,8 +418,6 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 #endif
 	}
 
-	radix_clear_slb();
-
 	accumulate_time(vcpu, &vcpu->arch.rm_exit);
 
 	/* Advance host PURR/SPURR by the amount used by guest */
@@ -394,11 +451,14 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 		mtspr(SPRN_DAWRX1, host_dawrx1);
 	}
 
-	/*
-	 * Since this is radix, do a eieio; tlbsync; ptesync sequence in
-	 * case we interrupted the guest between a tlbie and a ptesync.
-	 */
-	asm volatile("eieio; tlbsync; ptesync");
+	if (kvm_is_radix(kvm)) {
+		/*
+		 * Since this is radix, do a eieio; tlbsync; ptesync sequence
+		 * in case we interrupted the guest between a tlbie and a
+		 * ptesync.
+		 */
+		asm volatile("eieio; tlbsync; ptesync");
+	}
 
 	/*
 	 * cp_abort is required if the processor supports local copy-paste
@@ -425,7 +485,9 @@  int kvmhv_vcpu_entry_p9(struct kvm_vcpu *vcpu, u64 time_limit, unsigned long lpc
 	/* HDEC must be at least as large as DEC, so decrementer_max fits */
 	mtspr(SPRN_HDEC, decrementer_max);
 
+	save_clear_guest_mmu(kvm, vcpu);
 	switch_mmu_to_host_radix(kvm, host_pidr);
+	local_paca->kvm_hstate.in_guest = KVM_GUEST_MODE_NONE;
 
 	/*
 	 * If we are in real mode, only switch MMU on after the MMU is
diff --git a/arch/powerpc/kvm/book3s_hv_rm_mmu.c b/arch/powerpc/kvm/book3s_hv_rm_mmu.c
index 8cc73abbf42b..f487ebb3a70a 100644
--- a/arch/powerpc/kvm/book3s_hv_rm_mmu.c
+++ b/arch/powerpc/kvm/book3s_hv_rm_mmu.c
@@ -57,6 +57,10 @@  static int global_invalidates(struct kvm *kvm)
 	else
 		global = 1;
 
+	/* LPID has been switched to host if in virt mode so can't do local */
+	if (!global && (mfmsr() & (MSR_IR|MSR_DR)))
+		global = 1;
+
 	if (!global) {
 		/* any other core might now have stale TLB entries... */
 		smp_wmb();
diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 564ca9feef35..5cbc26ebcf8e 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
@@ -885,14 +885,11 @@  ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
 	cmpdi	r3, 512		/* 1 microsecond */
 	blt	hdec_soon
 
-	/* For hash guest, clear out and reload the SLB */
-BEGIN_MMU_FTR_SECTION
-	/* Radix host won't have populated the SLB, so no need to clear */
+	/* Clear out and reload the SLB */
 	li	r6, 0
 	slbmte	r6, r6
 	PPC_SLBIA(6)
 	ptesync
-END_MMU_FTR_SECTION_IFCLR(MMU_FTR_TYPE_RADIX)
 
 	/* Load up guest SLB entries (N.B. slb_max will be 0 for radix) */
 	lwz	r5,VCPU_SLB_MAX(r4)
@@ -1394,9 +1391,6 @@  guest_exit_cont:		/* r9 = vcpu, r12 = trap, r13 = paca */
 	stw	r5,VCPU_SLB_MAX(r9)
 
 	/* load host SLB entries */
-BEGIN_MMU_FTR_SECTION
-	b	guest_bypass
-END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
 	ld	r8,PACA_SLBSHADOWPTR(r13)
 
 	.rept	SLB_NUM_BOLTED
@@ -3148,10 +3142,6 @@  END_FTR_SECTION_IFSET(CPU_FTR_DAWR1)
 	PPC_SLBIA(6)
 	ptesync
 
-BEGIN_MMU_FTR_SECTION
-	b	4f
-END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
-
 	/* load host SLB entries */
 	ld	r8, PACA_SLBSHADOWPTR(r13)
 	.rept	SLB_NUM_BOLTED
@@ -3165,7 +3155,7 @@  END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
 3:	addi	r8, r8, 16
 	.endr
 
-4:	lwz	r7, KVM_HOST_LPID(r10)
+	lwz	r7, KVM_HOST_LPID(r10)
 	mtspr	SPRN_LPID, r7
 	mtspr	SPRN_PID, r0
 	ld	r8, KVM_HOST_LPCR(r10)