Merge tag 'kvm-s390-next-5.13-1' of...

Merge tag 'kvm-s390-next-5.13-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD

KVM: s390: Updates for 5.13

- properly handle MVPG in nesting KVM (vsie)
- allow to forward the yield_to hypercall (diagnose 9c)
- fixes

Documentation/virt/kvm/s390-diag.rst

@@ -84,3 +84,36 @@ If the function code specifies 0x501, breakpoint functions may be performed.
 This function code is handled by userspace.
 
 This diagnose function code has no subfunctions and uses no parameters.
+
+
+DIAGNOSE function code 'X'9C - Voluntary Time Slice Yield
+---------------------------------------------------------
+
+General register 1 contains the target CPU address.
+
+In a guest of a hypervisor like LPAR, KVM or z/VM using shared host CPUs,
+DIAGNOSE with function code 0x9c may improve system performance by
+yielding the host CPU on which the guest CPU is running to be assigned
+to another guest CPU, preferably the logical CPU containing the specified
+target CPU.
+
+
+DIAG 'X'9C forwarding
++++++++++++++++++++++
+
+The guest may send a DIAGNOSE 0x9c in order to yield to a certain
+other vcpu. An example is a Linux guest that tries to yield to the vcpu
+that is currently holding a spinlock, but not running.
+
+However, on the host the real cpu backing the vcpu may itself not be
+running.
+Forwarding the DIAGNOSE 0x9c initially sent by the guest to yield to
+the backing cpu will hopefully cause that cpu, and thus subsequently
+the guest's vcpu, to be scheduled.
+
+
+diag9c_forwarding_hz
+    KVM kernel parameter allowing to specify the maximum number of DIAGNOSE
+    0x9c forwarding per second in the purpose of avoiding a DIAGNOSE 0x9c
+    forwarding storm.
+    A value of 0 turns the forwarding off.

arch/s390/include/asm/kvm_host.h

@@ -454,6 +454,7 @@ struct kvm_vcpu_stat {
 	u64 diagnose_44;
 	u64 diagnose_9c;
 	u64 diagnose_9c_ignored;
+	u64 diagnose_9c_forward;
 	u64 diagnose_258;
 	u64 diagnose_308;
 	u64 diagnose_500;

arch/s390/include/asm/smp.h

@@ -63,5 +63,6 @@ extern void __noreturn cpu_die(void);
 extern void __cpu_die(unsigned int cpu);
 extern int __cpu_disable(void);
 extern void schedule_mcck_handler(void);
+void notrace smp_yield_cpu(int cpu);
 
 #endif /* __ASM_SMP_H */

arch/s390/kernel/smp.c

@@ -429,6 +429,7 @@ void notrace smp_yield_cpu(int cpu)
 	asm volatile("diag %0,0,0x9c"
 		     : : "d" (pcpu_devices[cpu].address));
 }
+EXPORT_SYMBOL_GPL(smp_yield_cpu);
 
 /*
  * Send cpus emergency shutdown signal. This gives the cpus the

arch/s390/kvm/diag.c

@@ -150,6 +150,19 @@ static int __diag_time_slice_end(struct kvm_vcpu *vcpu)
 	return 0;
 }
 
+static int forward_cnt;
+static unsigned long cur_slice;
+
+static int diag9c_forwarding_overrun(void)
+{
+	/* Reset the count on a new slice */
+	if (time_after(jiffies, cur_slice)) {
+		cur_slice = jiffies;
+		forward_cnt = diag9c_forwarding_hz / HZ;
+	}
+	return forward_cnt-- <= 0 ? 1 : 0;
+}
+
 static int __diag_time_slice_end_directed(struct kvm_vcpu *vcpu)
 {
 	struct kvm_vcpu *tcpu;
@@ -167,10 +180,22 @@ static int __diag_time_slice_end_directed(struct kvm_vcpu *vcpu)
 	if (!tcpu)
 		goto no_yield;
 
-	/* target already running */
-	if (READ_ONCE(tcpu->cpu) >= 0)
+	/* target guest VCPU already running */
+	if (READ_ONCE(tcpu->cpu) >= 0) {
+		if (!diag9c_forwarding_hz || diag9c_forwarding_overrun())
 			goto no_yield;
 
+		/* target host CPU already running */
+		if (!vcpu_is_preempted(tcpu->cpu))
+			goto no_yield;
+		smp_yield_cpu(tcpu->cpu);
+		VCPU_EVENT(vcpu, 5,
+			   "diag time slice end directed to %d: yield forwarded",
+			   tid);
+		vcpu->stat.diagnose_9c_forward++;
+		return 0;
+	}
+
 	if (kvm_vcpu_yield_to(tcpu) <= 0)
 		goto no_yield;
 

arch/s390/kvm/gaccess.c

@@ -976,7 +976,9 @@ int kvm_s390_check_low_addr_prot_real(struct kvm_vcpu *vcpu, unsigned long gra)
  * kvm_s390_shadow_tables - walk the guest page table and create shadow tables
  * @sg: pointer to the shadow guest address space structure
  * @saddr: faulting address in the shadow gmap
- * @pgt: pointer to the page table address result
+ * @pgt: pointer to the beginning of the page table for the given address if
+ *	 successful (return value 0), or to the first invalid DAT entry in
+ *	 case of exceptions (return value > 0)
  * @fake: pgt references contiguous guest memory block, not a pgtable
  */
 static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
@@ -1034,6 +1036,7 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
 			rfte.val = ptr;
 			goto shadow_r2t;
 		}
+		*pgt = ptr + vaddr.rfx * 8;
 		rc = gmap_read_table(parent, ptr + vaddr.rfx * 8, &rfte.val);
 		if (rc)
 			return rc;
@@ -1060,6 +1063,7 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
 			rste.val = ptr;
 			goto shadow_r3t;
 		}
+		*pgt = ptr + vaddr.rsx * 8;
 		rc = gmap_read_table(parent, ptr + vaddr.rsx * 8, &rste.val);
 		if (rc)
 			return rc;
@@ -1087,6 +1091,7 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
 			rtte.val = ptr;
 			goto shadow_sgt;
 		}
+		*pgt = ptr + vaddr.rtx * 8;
 		rc = gmap_read_table(parent, ptr + vaddr.rtx * 8, &rtte.val);
 		if (rc)
 			return rc;
@@ -1123,6 +1128,7 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
 			ste.val = ptr;
 			goto shadow_pgt;
 		}
+		*pgt = ptr + vaddr.sx * 8;
 		rc = gmap_read_table(parent, ptr + vaddr.sx * 8, &ste.val);
 		if (rc)
 			return rc;
@@ -1157,6 +1163,8 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
  * @vcpu: virtual cpu
  * @sg: pointer to the shadow guest address space structure
  * @saddr: faulting address in the shadow gmap
+ * @datptr: will contain the address of the faulting DAT table entry, or of
+ *	    the valid leaf, plus some flags
  *
  * Returns: - 0 if the shadow fault was successfully resolved
  *	    - > 0 (pgm exception code) on exceptions while faulting
@@ -1165,11 +1173,11 @@ static int kvm_s390_shadow_tables(struct gmap *sg, unsigned long saddr,
  *	    - -ENOMEM if out of memory
  */
 int kvm_s390_shadow_fault(struct kvm_vcpu *vcpu, struct gmap *sg,
-			  unsigned long saddr)
+			  unsigned long saddr, unsigned long *datptr)
 {
 	union vaddress vaddr;
 	union page_table_entry pte;
-	unsigned long pgt;
+	unsigned long pgt = 0;
 	int dat_protection, fake;
 	int rc;
 
@@ -1191,8 +1199,20 @@ int kvm_s390_shadow_fault(struct kvm_vcpu *vcpu, struct gmap *sg,
 		pte.val = pgt + vaddr.px * PAGE_SIZE;
 		goto shadow_page;
 	}
-	if (!rc)
-		rc = gmap_read_table(sg->parent, pgt + vaddr.px * 8, &pte.val);
+
+	switch (rc) {
+	case PGM_SEGMENT_TRANSLATION:
+	case PGM_REGION_THIRD_TRANS:
+	case PGM_REGION_SECOND_TRANS:
+	case PGM_REGION_FIRST_TRANS:
+		pgt |= PEI_NOT_PTE;
+		break;
+	case 0:
+		pgt += vaddr.px * 8;
+		rc = gmap_read_table(sg->parent, pgt, &pte.val);
+	}
+	if (datptr)
+		*datptr = pgt | dat_protection * PEI_DAT_PROT;
 	if (!rc && pte.i)
 		rc = PGM_PAGE_TRANSLATION;
 	if (!rc && pte.z)

arch/s390/kvm/gaccess.h

@@ -18,17 +18,14 @@
 
 /**
  * kvm_s390_real_to_abs - convert guest real address to guest absolute address
- * @vcpu - guest virtual cpu
+ * @prefix - guest prefix
  * @gra - guest real address
  *
  * Returns the guest absolute address that corresponds to the passed guest real
- * address @gra of a virtual guest cpu by applying its prefix.
+ * address @gra of by applying the given prefix.
  */
-static inline unsigned long kvm_s390_real_to_abs(struct kvm_vcpu *vcpu,
-						 unsigned long gra)
+static inline unsigned long _kvm_s390_real_to_abs(u32 prefix, unsigned long gra)
 {
-	unsigned long prefix  = kvm_s390_get_prefix(vcpu);
-
 	if (gra < 2 * PAGE_SIZE)
 		gra += prefix;
 	else if (gra >= prefix && gra < prefix + 2 * PAGE_SIZE)
@@ -36,6 +33,43 @@ static inline unsigned long kvm_s390_real_to_abs(struct kvm_vcpu *vcpu,
 	return gra;
 }
 
+/**
+ * kvm_s390_real_to_abs - convert guest real address to guest absolute address
+ * @vcpu - guest virtual cpu
+ * @gra - guest real address
+ *
+ * Returns the guest absolute address that corresponds to the passed guest real
+ * address @gra of a virtual guest cpu by applying its prefix.
+ */
+static inline unsigned long kvm_s390_real_to_abs(struct kvm_vcpu *vcpu,
+						 unsigned long gra)
+{
+	return _kvm_s390_real_to_abs(kvm_s390_get_prefix(vcpu), gra);
+}
+
+/**
+ * _kvm_s390_logical_to_effective - convert guest logical to effective address
+ * @psw: psw of the guest
+ * @ga: guest logical address
+ *
+ * Convert a guest logical address to an effective address by applying the
+ * rules of the addressing mode defined by bits 31 and 32 of the given PSW
+ * (extendended/basic addressing mode).
+ *
+ * Depending on the addressing mode, the upper 40 bits (24 bit addressing
+ * mode), 33 bits (31 bit addressing mode) or no bits (64 bit addressing
+ * mode) of @ga will be zeroed and the remaining bits will be returned.
+ */
+static inline unsigned long _kvm_s390_logical_to_effective(psw_t *psw,
+							   unsigned long ga)
+{
+	if (psw_bits(*psw).eaba == PSW_BITS_AMODE_64BIT)
+		return ga;
+	if (psw_bits(*psw).eaba == PSW_BITS_AMODE_31BIT)
+		return ga & ((1UL << 31) - 1);
+	return ga & ((1UL << 24) - 1);
+}
+
 /**
  * kvm_s390_logical_to_effective - convert guest logical to effective address
  * @vcpu: guest virtual cpu
@@ -52,13 +86,7 @@ static inline unsigned long kvm_s390_real_to_abs(struct kvm_vcpu *vcpu,
 static inline unsigned long kvm_s390_logical_to_effective(struct kvm_vcpu *vcpu,
 							  unsigned long ga)
 {
-	psw_t *psw = &vcpu->arch.sie_block->gpsw;
-
-	if (psw_bits(*psw).eaba == PSW_BITS_AMODE_64BIT)
-		return ga;
-	if (psw_bits(*psw).eaba == PSW_BITS_AMODE_31BIT)
-		return ga & ((1UL << 31) - 1);
-	return ga & ((1UL << 24) - 1);
+	return _kvm_s390_logical_to_effective(&vcpu->arch.sie_block->gpsw, ga);
 }
 
 /*
@@ -359,7 +387,11 @@ void ipte_unlock(struct kvm_vcpu *vcpu);
 int ipte_lock_held(struct kvm_vcpu *vcpu);
 int kvm_s390_check_low_addr_prot_real(struct kvm_vcpu *vcpu, unsigned long gra);
 
+/* MVPG PEI indication bits */
+#define PEI_DAT_PROT 2
+#define PEI_NOT_PTE 4
+
 int kvm_s390_shadow_fault(struct kvm_vcpu *vcpu, struct gmap *shadow,
-			  unsigned long saddr);
+			  unsigned long saddr, unsigned long *datptr);
 
 #endif /* __KVM_S390_GACCESS_H */

arch/s390/kvm/kvm-s390.c

@@ -158,6 +158,7 @@ struct kvm_stats_debugfs_item debugfs_entries[] = {
 	VCPU_STAT("instruction_diag_44", diagnose_44),
 	VCPU_STAT("instruction_diag_9c", diagnose_9c),
 	VCPU_STAT("diag_9c_ignored", diagnose_9c_ignored),
+	VCPU_STAT("diag_9c_forward", diagnose_9c_forward),
 	VCPU_STAT("instruction_diag_258", diagnose_258),
 	VCPU_STAT("instruction_diag_308", diagnose_308),
 	VCPU_STAT("instruction_diag_500", diagnose_500),
@@ -185,6 +186,11 @@ static bool use_gisa  = true;
 module_param(use_gisa, bool, 0644);
 MODULE_PARM_DESC(use_gisa, "Use the GISA if the host supports it.");
 
+/* maximum diag9c forwarding per second */
+unsigned int diag9c_forwarding_hz;
+module_param(diag9c_forwarding_hz, uint, 0644);
+MODULE_PARM_DESC(diag9c_forwarding_hz, "Maximum diag9c forwarding per second, 0 to turn off");
+
 /*
  * For now we handle at most 16 double words as this is what the s390 base
  * kernel handles and stores in the prefix page. If we ever need to go beyond
@@ -4542,7 +4548,7 @@ int kvm_s390_vcpu_start(struct kvm_vcpu *vcpu)
 		/*
 		 * As we are starting a second VCPU, we have to disable
 		 * the IBS facility on all VCPUs to remove potentially
-		 * oustanding ENABLE requests.
+		 * outstanding ENABLE requests.
 		 */
 		__disable_ibs_on_all_vcpus(vcpu->kvm);
 	}

arch/s390/kvm/kvm-s390.h

@@ -471,4 +471,12 @@ void kvm_s390_reinject_machine_check(struct kvm_vcpu *vcpu,
  * @kvm: the KVM guest
  */
 void kvm_s390_vcpu_crypto_reset_all(struct kvm *kvm);
+
+/**
+ * diag9c_forwarding_hz
+ *
+ * Set the maximum number of diag9c forwarding per second
+ */
+extern unsigned int diag9c_forwarding_hz;
+
 #endif

arch/s390/kvm/vsie.c

@@ -417,11 +417,6 @@ static void unshadow_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
 		memcpy((void *)((u64)scb_o + 0xc0),
 		       (void *)((u64)scb_s + 0xc0), 0xf0 - 0xc0);
 		break;
-	case ICPT_PARTEXEC:
-		/* MVPG only */
-		memcpy((void *)((u64)scb_o + 0xc0),
-		       (void *)((u64)scb_s + 0xc0), 0xd0 - 0xc0);
-		break;
 	}
 
 	if (scb_s->ihcpu != 0xffffU)
@@ -620,10 +615,10 @@ static int map_prefix(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
 	/* with mso/msl, the prefix lies at offset *mso* */
 	prefix += scb_s->mso;
 
-	rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap, prefix);
+	rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap, prefix, NULL);
 	if (!rc && (scb_s->ecb & ECB_TE))
 		rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap,
-					   prefix + PAGE_SIZE);
+					   prefix + PAGE_SIZE, NULL);
 	/*
 	 * We don't have to mprotect, we will be called for all unshadows.
 	 * SIE will detect if protection applies and trigger a validity.
@@ -914,7 +909,7 @@ static int handle_fault(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
 				    current->thread.gmap_addr, 1);
 
 	rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap,
-				   current->thread.gmap_addr);
+				   current->thread.gmap_addr, NULL);
 	if (rc > 0) {
 		rc = inject_fault(vcpu, rc,
 				  current->thread.gmap_addr,
@@ -936,7 +931,7 @@ static void handle_last_fault(struct kvm_vcpu *vcpu,
 {
 	if (vsie_page->fault_addr)
 		kvm_s390_shadow_fault(vcpu, vsie_page->gmap,
-				      vsie_page->fault_addr);
+				      vsie_page->fault_addr, NULL);
 	vsie_page->fault_addr = 0;
 }
 
@@ -983,6 +978,98 @@ static int handle_stfle(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
 	return 0;
 }
 
+/*
+ * Get a register for a nested guest.
+ * @vcpu the vcpu of the guest
+ * @vsie_page the vsie_page for the nested guest
+ * @reg the register number, the upper 4 bits are ignored.
+ * returns: the value of the register.
+ */
+static u64 vsie_get_register(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page, u8 reg)
+{
+	/* no need to validate the parameter and/or perform error handling */
+	reg &= 0xf;
+	switch (reg) {
+	case 15:
+		return vsie_page->scb_s.gg15;
+	case 14:
+		return vsie_page->scb_s.gg14;
+	default:
+		return vcpu->run->s.regs.gprs[reg];
+	}
+}
+
+static int vsie_handle_mvpg(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
+{
+	struct kvm_s390_sie_block *scb_s = &vsie_page->scb_s;
+	unsigned long pei_dest, pei_src, src, dest, mask, prefix;
+	u64 *pei_block = &vsie_page->scb_o->mcic;
+	int edat, rc_dest, rc_src;
+	union ctlreg0 cr0;
+
+	cr0.val = vcpu->arch.sie_block->gcr[0];
+	edat = cr0.edat && test_kvm_facility(vcpu->kvm, 8);
+	mask = _kvm_s390_logical_to_effective(&scb_s->gpsw, PAGE_MASK);
+	prefix = scb_s->prefix << GUEST_PREFIX_SHIFT;
+
+	dest = vsie_get_register(vcpu, vsie_page, scb_s->ipb >> 20) & mask;
+	dest = _kvm_s390_real_to_abs(prefix, dest) + scb_s->mso;
+	src = vsie_get_register(vcpu, vsie_page, scb_s->ipb >> 16) & mask;
+	src = _kvm_s390_real_to_abs(prefix, src) + scb_s->mso;
+
+	rc_dest = kvm_s390_shadow_fault(vcpu, vsie_page->gmap, dest, &pei_dest);
+	rc_src = kvm_s390_shadow_fault(vcpu, vsie_page->gmap, src, &pei_src);
+	/*
+	 * Either everything went well, or something non-critical went wrong
+	 * e.g. because of a race. In either case, simply retry.
+	 */
+	if (rc_dest == -EAGAIN || rc_src == -EAGAIN || (!rc_dest && !rc_src)) {
+		retry_vsie_icpt(vsie_page);
+		return -EAGAIN;
+	}
+	/* Something more serious went wrong, propagate the error */
+	if (rc_dest < 0)
+		return rc_dest;
+	if (rc_src < 0)
+		return rc_src;
+
+	/* The only possible suppressing exception: just deliver it */
+	if (rc_dest == PGM_TRANSLATION_SPEC || rc_src == PGM_TRANSLATION_SPEC) {
+		clear_vsie_icpt(vsie_page);
+		rc_dest = kvm_s390_inject_program_int(vcpu, PGM_TRANSLATION_SPEC);
+		WARN_ON_ONCE(rc_dest);
+		return 1;
+	}
+
+	/*
+	 * Forward the PEI intercept to the guest if it was a page fault, or
+	 * also for segment and region table faults if EDAT applies.
+	 */
+	if (edat) {
+		rc_dest = rc_dest == PGM_ASCE_TYPE ? rc_dest : 0;
+		rc_src = rc_src == PGM_ASCE_TYPE ? rc_src : 0;
+	} else {
+		rc_dest = rc_dest != PGM_PAGE_TRANSLATION ? rc_dest : 0;
+		rc_src = rc_src != PGM_PAGE_TRANSLATION ? rc_src : 0;
+	}
+	if (!rc_dest && !rc_src) {
+		pei_block[0] = pei_dest;
+		pei_block[1] = pei_src;
+		return 1;
+	}
+
+	retry_vsie_icpt(vsie_page);
+
+	/*
+	 * The host has edat, and the guest does not, or it was an ASCE type
+	 * exception. The host needs to inject the appropriate DAT interrupts
+	 * into the guest.
+	 */
+	if (rc_dest)
+		return inject_fault(vcpu, rc_dest, dest, 1);
+	return inject_fault(vcpu, rc_src, src, 0);
+}
+
 /*
  * Run the vsie on a shadow scb and a shadow gmap, without any further
  * sanity checks, handling SIE faults.
@@ -1071,6 +1158,10 @@ static int do_vsie_run(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
 		if ((scb_s->ipa & 0xf000) != 0xf000)
 			scb_s->ipa += 0x1000;
 		break;
+	case ICPT_PARTEXEC:
+		if (scb_s->ipa == 0xb254)
+			rc = vsie_handle_mvpg(vcpu, vsie_page);
+		break;
 	}
 	return rc;
 }