KVM: PPC: Book3S HV: Lockless tlbie for HPT hcalls

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 c17b98cf60 ("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>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>

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;

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;

arch/powerpc/kvm/book3s_hv_rm_mmu.c

@@ -435,24 +435,6 @@ 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)
 {
@@ -464,8 +446,6 @@ 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) {
@@ -484,7 +464,6 @@ 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");