[PATCH v3] KVM: PPC: Book3S HV: lockless tlbie for HPT hcalls
From: Nicholas Piggin <npiggin@gmail.com>
Date: 2018-05-17 06:59:22
Subsystem:
kernel virtual machine for powerpc (kvm/powerpc), linux for powerpc (32-bit and 64-bit), the rest · Maintainers:
Madhavan Srinivasan, Linus Torvalds
tlbies to an LPAR do not have to be serialised since POWER4/PPC970,
after which the MMU_FTR_LOCKLESS_TLBIE feature was introduced to
avoid tlbie locking.
Since commit c17b98cf6028 ("KVM: PPC: Book3S HV: Remove code for
PPC970 processors"), KVM no longer supports processors that do not
have this feature, so the tlbie locking can be removed completely.
A sanity check for the feature is put in kvmppc_mmu_hv_init.
Testing was done on a POWER9 system in HPT mode, with a -smp 32 guest
in HPT mode. 32 instances of the powerpc fork benchmark from selftests
were run with --fork, and the results measured.
Without this patch, total throughput was about 13.5K/sec, and this is
the top of the host profile:
74.52% [k] do_tlbies
2.95% [k] kvmppc_book3s_hv_page_fault
1.80% [k] calc_checksum
1.80% [k] kvmppc_vcpu_run_hv
1.49% [k] kvmppc_run_core
After this patch, throughput was about 51K/sec, with this profile:
21.28% [k] do_tlbies
5.26% [k] kvmppc_run_core
4.88% [k] kvmppc_book3s_hv_page_fault
3.30% [k] _raw_spin_lock_irqsave
3.25% [k] gup_pgd_range
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
Changes since v2:
- Removed the locking code completely (thanks mpe)
- Updated changelog (thanks mpe)
arch/powerpc/include/asm/kvm_host.h | 1 -
arch/powerpc/kvm/book3s_64_mmu_hv.c | 3 +++
arch/powerpc/kvm/book3s_hv_rm_mmu.c | 21 ---------------------
3 files changed, 3 insertions(+), 22 deletions(-)
diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h
index 17498e9a26e4..7756b0c6da75 100644
--- a/arch/powerpc/include/asm/kvm_host.h
+++ b/arch/powerpc/include/asm/kvm_host.h@@ -269,7 +269,6 @@ struct kvm_arch { unsigned long host_lpcr; unsigned long sdr1; unsigned long host_sdr1; - int tlbie_lock; unsigned long lpcr; unsigned long vrma_slb_v; int mmu_ready;
diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
index a670fa5fbe50..37cd6434d1c8 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c@@ -272,6 +272,9 @@ int kvmppc_mmu_hv_init(void) if (!cpu_has_feature(CPU_FTR_HVMODE)) return -EINVAL; + if (!mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE)) + return -EINVAL; + /* POWER7 has 10-bit LPIDs (12-bit in POWER8) */ host_lpid = mfspr(SPRN_LPID); rsvd_lpid = LPID_RSVD;
diff --git a/arch/powerpc/kvm/book3s_hv_rm_mmu.c b/arch/powerpc/kvm/book3s_hv_rm_mmu.c
index 78e6a392330f..89d909b3b881 100644
--- a/arch/powerpc/kvm/book3s_hv_rm_mmu.c
+++ b/arch/powerpc/kvm/book3s_hv_rm_mmu.c@@ -434,24 +434,6 @@ static inline int is_mmio_hpte(unsigned long v, unsigned long r) (HPTE_R_KEY_HI | HPTE_R_KEY_LO)); } -static inline int try_lock_tlbie(unsigned int *lock) -{ - unsigned int tmp, old; - unsigned int token = LOCK_TOKEN; - - asm volatile("1:lwarx %1,0,%2\n" - " cmpwi cr0,%1,0\n" - " bne 2f\n" - " stwcx. %3,0,%2\n" - " bne- 1b\n" - " isync\n" - "2:" - : "=&r" (tmp), "=&r" (old) - : "r" (lock), "r" (token) - : "cc", "memory"); - return old == 0; -} - static void do_tlbies(struct kvm *kvm, unsigned long *rbvalues, long npages, int global, bool need_sync) {
@@ -463,8 +445,6 @@ static void do_tlbies(struct kvm *kvm, unsigned long *rbvalues, * the RS field, this is backwards-compatible with P7 and P8. */ if (global) { - while (!try_lock_tlbie(&kvm->arch.tlbie_lock)) - cpu_relax(); if (need_sync) asm volatile("ptesync" : : : "memory"); for (i = 0; i < npages; ++i) {
@@ -483,7 +463,6 @@ static void do_tlbies(struct kvm *kvm, unsigned long *rbvalues, } asm volatile("eieio; tlbsync; ptesync" : : : "memory"); - kvm->arch.tlbie_lock = 0; } else { if (need_sync) asm volatile("ptesync" : : : "memory");
--
2.17.0