[PATCH v5 0/4] KVM: s390: Introducing kvm_arch_set_irq_inatomic Fast Inject

[PATCH v5 0/4] KVM: s390: Introducing kvm_arch_set_irq_inatomic Fast Inject

[Douglas Freimuth]

S390 needs this series of three patches in order to enable a non-blocking
path for irqfd injection on s390 via kvm_arch_set_irq_inatomic(). Before
these changes, kvm_arch_set_irq_inatomic() would just return -EWOULDBLOCK
and place all interrupts on the global work queue, which must subsequently
be processed by a different thread. This series of patches implements an
s390 version of inatomic and is relevant to virtio-blk and virtio-net and
was tested against virtio-pci and virtio-ccw.

The inatomic fast path cannot lose control since it is running with
interrupts disabled. This meant making the following changes that exist on
the slow path today. First, the adapter_indicators page needs to be mapped
since it is accessed with interrupts disabled, so we added map/unmap
functions. Second, access to shared resources between the fast and slow
paths needed to be changed from mutex and semaphores to raw_spin_lock's.
Finally, the memory allocation on the slow path utilizes GFP_KERNEL_ACCOUNT
but we had to implement the fast path with GFP_ATOMIC allocation. Each of
these enhancements were required to prevent blocking on the fast inject
path. In order to handle the RT kernel case, raw spin locks were added.

Statistical counters have been added to enable analysis of irq injection on
the fast path and slow path including io_390_inatomic, io_flic_inject_airq,
io_set_adapter_int and io_390_inatomic_adapter_masked. And counters have
been added to analyze map/unmap of the adapter_indicator
pages in non-Secure Execution environments and to track fencing of Fast
Inject in Secure Execution environments. In order to take advantage of this
kernel series with virtio-pci, a QEMU that includes the
's390x/pci: set kvm_msi_via_irqfd_allowed' fix is needed.  Additionally,
the guest xml needs a thread pool and threads explicitly assigned per disk
device using the common way of defining threads for disks.

Patch 1 enables map/unmap of adapter indicator pages but for Secure
Execution environments it avoids the long term mapping.

v4->v5: Replace get_map_page using pin functions FOLL_WRITE | FOLL_LONGTERM
v4->v5: Add shared function for destroy_adapters, unmap_all_adapters_pv
v4->v5: Upon failed injection on fast path clear the summary bit only
v4->v5: If allocation fails in inatomic, clear summary bit and do slow path
v4->v5: Change fi->lock to a raw spin lock to handle the RT case

Douglas Freimuth (4):
  KVM: s390: Add map/unmap ioctl and clean mappings post-guest
  KVM: s390: Enable adapter_indicators_set to use mapped pages
  KVM: s390: Change the fi->lock to a raw_spinlock for RT case
  KVM: s390: Introducing kvm_arch_set_irq_inatomic fast inject

 arch/s390/include/asm/kvm_host.h |  13 +-
 arch/s390/kvm/intercept.c        |   4 +-
 arch/s390/kvm/interrupt.c        | 497 +++++++++++++++++++++++++------
 arch/s390/kvm/kvm-s390.c         |  31 +-
 arch/s390/kvm/kvm-s390.h         |   5 +-
 5 files changed, 444 insertions(+), 106 deletions(-)

-- 
2.52.0

[PATCH v5 1/4] KVM: s390: Add map/unmap ioctl and clean mappings post-guest

[Douglas Freimuth]

S390 needs map/unmap ioctls, which map the adapter set
indicator pages, so the pages can be accessed when interrupts are
disabled. The mappings are cleaned up when the guest is removed.

Map/Unmap ioctls are fenced in order to avoid the longterm pinning
in Secure Execution environments. In Secure Execution
environments the path of execution available before this patch is followed.

Statistical counters to count map/unmap functions for adapter indicator
pages are added. The counters can be used to analyze
map/unmap functions in non-Secure Execution environments and similarly
can be used to analyze Secure Execution environments where the counters
will not be incremented as the adapter indicator pages are not mapped.

Signed-off-by: Douglas Freimuth <freimuth@linux.ibm.com>
---
 arch/s390/include/asm/kvm_host.h |   5 +
 arch/s390/kvm/interrupt.c        | 190 ++++++++++++++++++++++++++-----
 arch/s390/kvm/kvm-s390.c         |   8 ++
 arch/s390/kvm/kvm-s390.h         |   2 +
 4 files changed, 176 insertions(+), 29 deletions(-)

diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 8a4f4a39f7a2..fbb2406b31d2 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -448,6 +448,8 @@ struct kvm_vcpu_arch {
 struct kvm_vm_stat {
 	struct kvm_vm_stat_generic generic;
 	u64 inject_io;
+	u64 io_390_adapter_map;
+	u64 io_390_adapter_unmap;
 	u64 inject_float_mchk;
 	u64 inject_pfault_done;
 	u64 inject_service_signal;
@@ -479,6 +481,9 @@ struct s390_io_adapter {
 	bool masked;
 	bool swap;
 	bool suppressible;
+	raw_spinlock_t maps_lock;
+	struct list_head maps;
+	unsigned int nr_maps;
 };
 
 #define MAX_S390_IO_ADAPTERS ((MAX_ISC + 1) * 8)
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 07f59c3b9a7b..a9b418996225 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -2429,6 +2429,9 @@ static int register_io_adapter(struct kvm_device *dev,
 	if (!adapter)
 		return -ENOMEM;
 
+	INIT_LIST_HEAD(&adapter->maps);
+	raw_spin_lock_init(&adapter->maps_lock);
+	adapter->nr_maps = 0;
 	adapter->id = adapter_info.id;
 	adapter->isc = adapter_info.isc;
 	adapter->maskable = adapter_info.maskable;
@@ -2453,12 +2456,151 @@ int kvm_s390_mask_adapter(struct kvm *kvm, unsigned int id, bool masked)
 	return ret;
 }
 
+static struct page *pin_map_page(struct kvm *kvm, u64 uaddr,
+				 unsigned int gup_flags)
+{
+	struct mm_struct *mm = kvm->mm;
+	struct page *page = NULL;
+	int locked = 1;
+
+	if (mmget_not_zero(mm)) {
+		mmap_read_lock(mm);
+		pin_user_pages_remote(mm, uaddr, 1, FOLL_WRITE | gup_flags,
+				      &page, &locked);
+		if (locked)
+			mmap_read_unlock(mm);
+		mmput(mm);
+	}
+
+	return page;
+}
+
+static int kvm_s390_adapter_map(struct kvm *kvm, unsigned int id, __u64 addr)
+{
+	struct s390_io_adapter *adapter = get_io_adapter(kvm, id);
+	struct s390_map_info *map;
+	unsigned long flags;
+	__u64 host_addr;
+	int ret, idx;
+
+	if (!adapter || !addr)
+		return -EINVAL;
+
+	map = kzalloc_obj(*map, GFP_KERNEL_ACCOUNT);
+	if (!map)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&map->list);
+	idx = srcu_read_lock(&kvm->srcu);
+	host_addr = gpa_to_hva(kvm, addr);
+	if (kvm_is_error_hva(host_addr)) {
+		srcu_read_unlock(&kvm->srcu, idx);
+		ret = -EFAULT;
+		goto out;
+	}
+	srcu_read_unlock(&kvm->srcu, idx);
+	map->guest_addr = addr;
+	map->addr = host_addr;
+	map->page = pin_map_page(kvm, host_addr, FOLL_LONGTERM);
+	if (!map->page) {
+		ret = -EINVAL;
+		goto out;
+	}
+	raw_spin_lock_irqsave(&adapter->maps_lock, flags);
+	if (adapter->nr_maps < MAX_S390_ADAPTER_MAPS) {
+		list_add_tail(&map->list, &adapter->maps);
+		adapter->nr_maps++;
+		ret = 0;
+	} else {
+		ret = -EINVAL;
+	}
+	raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+	if (ret)
+		unpin_user_page(map->page);
+out:
+	if (ret)
+		kfree(map);
+	return ret;
+}
+
+static int kvm_s390_adapter_unmap(struct kvm *kvm, unsigned int id, __u64 addr)
+{
+	struct s390_io_adapter *adapter = get_io_adapter(kvm, id);
+	struct s390_map_info *map, *tmp, *map_to_free;
+	struct page *map_page_to_put = NULL;
+	u64 map_addr_to_mark = 0;
+	unsigned long flags;
+	int found = 0, idx;
+
+	if (!adapter || !addr)
+		return -EINVAL;
+
+	raw_spin_lock_irqsave(&adapter->maps_lock, flags);
+	list_for_each_entry_safe(map, tmp, &adapter->maps, list) {
+		if (map->guest_addr == addr) {
+			found = 1;
+			adapter->nr_maps--;
+			list_del(&map->list);
+			map_page_to_put = map->page;
+			map_addr_to_mark = map->guest_addr;
+			map_to_free = map;
+			break;
+		}
+	}
+	raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+
+	if (found) {
+		kfree(map_to_free);
+		idx = srcu_read_lock(&kvm->srcu);
+		mark_page_dirty(kvm, map_addr_to_mark >> PAGE_SHIFT);
+		set_page_dirty_lock(map_page_to_put);
+		srcu_read_unlock(&kvm->srcu, idx);
+		unpin_user_page(map_page_to_put);
+	}
+
+	return found ? 0 : -ENOENT;
+}
+
+void kvm_s390_unmap_all_adapters(struct kvm *kvm)
+{
+	struct s390_map_info *map, *tmp;
+	unsigned long flags;
+	int i, idx;
+
+	for (i = 0; i < MAX_S390_IO_ADAPTERS; i++) {
+		struct s390_io_adapter *adapter = kvm->arch.adapters[i];
+		LIST_HEAD(local_list);
+
+		if (!adapter)
+			continue;
+
+		raw_spin_lock_irqsave(&adapter->maps_lock, flags);
+		list_splice_init(&adapter->maps, &local_list);
+		adapter->nr_maps = 0;
+		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+
+		list_for_each_entry_safe(map, tmp, &local_list, list) {
+			list_del(&map->list);
+			idx = srcu_read_lock(&kvm->srcu);
+			mark_page_dirty(kvm, map->guest_addr >> PAGE_SHIFT);
+			set_page_dirty_lock(map->page);
+			srcu_read_unlock(&kvm->srcu, idx);
+			unpin_user_page(map->page);
+			kfree(map);
+		}
+	}
+}
+
 void kvm_s390_destroy_adapters(struct kvm *kvm)
 {
 	int i;
 
-	for (i = 0; i < MAX_S390_IO_ADAPTERS; i++)
+	kvm_s390_unmap_all_adapters(kvm);
+
+	for (i = 0; i < MAX_S390_IO_ADAPTERS; i++) {
 		kfree(kvm->arch.adapters[i]);
+		kvm->arch.adapters[i] = NULL;
+	}
 }
 
 static int modify_io_adapter(struct kvm_device *dev,
@@ -2480,14 +2622,22 @@ static int modify_io_adapter(struct kvm_device *dev,
 		if (ret > 0)
 			ret = 0;
 		break;
-	/*
-	 * The following operations are no longer needed and therefore no-ops.
-	 * The gpa to hva translation is done when an IRQ route is set up. The
-	 * set_irq code uses get_user_pages_remote() to do the actual write.
-	 */
 	case KVM_S390_IO_ADAPTER_MAP:
 	case KVM_S390_IO_ADAPTER_UNMAP:
-		ret = 0;
+		/* If in Secure Execution mode do not long term pin. */
+		mutex_lock(&dev->kvm->lock);
+		if (kvm_s390_pv_is_protected(dev->kvm)) {
+			mutex_unlock(&dev->kvm->lock);
+			return 0;
+		}
+		if (req.type == KVM_S390_IO_ADAPTER_MAP) {
+			dev->kvm->stat.io_390_adapter_map++;
+			ret = kvm_s390_adapter_map(dev->kvm, req.id, req.addr);
+		} else {
+			dev->kvm->stat.io_390_adapter_unmap++;
+			ret = kvm_s390_adapter_unmap(dev->kvm, req.id, req.addr);
+		}
+		mutex_unlock(&dev->kvm->lock);
 		break;
 	default:
 		ret = -EINVAL;
@@ -2733,24 +2883,6 @@ static unsigned long get_ind_bit(__u64 addr, unsigned long bit_nr, bool swap)
 	return swap ? (bit ^ (BITS_PER_LONG - 1)) : bit;
 }
 
-static struct page *get_map_page(struct kvm *kvm, u64 uaddr)
-{
-	struct mm_struct *mm = kvm->mm;
-	struct page *page = NULL;
-	int locked = 1;
-
-	if (mmget_not_zero(mm)) {
-		mmap_read_lock(mm);
-		get_user_pages_remote(mm, uaddr, 1, FOLL_WRITE,
-				      &page, &locked);
-		if (locked)
-			mmap_read_unlock(mm);
-		mmput(mm);
-	}
-
-	return page;
-}
-
 static int adapter_indicators_set(struct kvm *kvm,
 				  struct s390_io_adapter *adapter,
 				  struct kvm_s390_adapter_int *adapter_int)
@@ -2760,10 +2892,10 @@ static int adapter_indicators_set(struct kvm *kvm,
 	struct page *ind_page, *summary_page;
 	void *map;
 
-	ind_page = get_map_page(kvm, adapter_int->ind_addr);
+	ind_page = pin_map_page(kvm, adapter_int->ind_addr, 0);
 	if (!ind_page)
 		return -1;
-	summary_page = get_map_page(kvm, adapter_int->summary_addr);
+	summary_page = pin_map_page(kvm, adapter_int->summary_addr, 0);
 	if (!summary_page) {
 		put_page(ind_page);
 		return -1;
@@ -2784,8 +2916,8 @@ static int adapter_indicators_set(struct kvm *kvm,
 	set_page_dirty_lock(summary_page);
 	srcu_read_unlock(&kvm->srcu, idx);
 
-	put_page(ind_page);
-	put_page(summary_page);
+	unpin_user_page(ind_page);
+	unpin_user_page(summary_page);
 	return summary_set ? 0 : 1;
 }
 
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index e09960c2e6ed..74f453f039a3 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -68,6 +68,8 @@
 const struct kvm_stats_desc kvm_vm_stats_desc[] = {
 	KVM_GENERIC_VM_STATS(),
 	STATS_DESC_COUNTER(VM, inject_io),
+	STATS_DESC_COUNTER(VM, io_390_adapter_map),
+	STATS_DESC_COUNTER(VM, io_390_adapter_unmap),
 	STATS_DESC_COUNTER(VM, inject_float_mchk),
 	STATS_DESC_COUNTER(VM, inject_pfault_done),
 	STATS_DESC_COUNTER(VM, inject_service_signal),
@@ -2497,6 +2499,11 @@ static int kvm_s390_pv_dmp(struct kvm *kvm, struct kvm_pv_cmd *cmd,
 	return r;
 }
 
+static void kvm_s390_unmap_all_adapters_pv(struct kvm *kvm)
+{
+	kvm_s390_unmap_all_adapters(kvm);
+}
+
 static int kvm_s390_handle_pv(struct kvm *kvm, struct kvm_pv_cmd *cmd)
 {
 	const bool need_lock = (cmd->cmd != KVM_PV_ASYNC_CLEANUP_PERFORM);
@@ -2513,6 +2520,7 @@ static int kvm_s390_handle_pv(struct kvm *kvm, struct kvm_pv_cmd *cmd)
 		if (kvm_s390_pv_is_protected(kvm))
 			break;
 
+		kvm_s390_unmap_all_adapters_pv(kvm);
 		mmap_write_lock(kvm->mm);
 		/*
 		 * Disable creation of new THPs. Existing THPs can stay, they
diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h
index dc0573b7aa4b..7ba885cb6bd1 100644
--- a/arch/s390/kvm/kvm-s390.h
+++ b/arch/s390/kvm/kvm-s390.h
@@ -560,6 +560,8 @@ void kvm_s390_gisa_disable(struct kvm *kvm);
 void kvm_s390_gisa_enable(struct kvm *kvm);
 int __init kvm_s390_gib_init(u8 nisc);
 void kvm_s390_gib_destroy(void);
+void kvm_s390_unmap_all_adapters(struct kvm *kvm);
+
 
 /* implemented in guestdbg.c */
 void kvm_s390_backup_guest_per_regs(struct kvm_vcpu *vcpu);
-- 
2.52.0

[PATCH v5 2/4] KVM: s390: Enable adapter_indicators_set to use mapped pages

[Douglas Freimuth]

The S390 adapter_indicators_set function needs to be able to use mapped
pages so that worked can be processed,on a fast path when interrupts are
disabled. If adapter indicator pages are not mapped then local mapping is
done on a slow path as it is prior to this patch. For example, Secure
Execution environments will take the local mapping path as it does prior to
this patch.

Signed-off-by: Douglas Freimuth <freimuth@linux.ibm.com>
---
 arch/s390/kvm/interrupt.c | 94 ++++++++++++++++++++++++++++-----------
 1 file changed, 69 insertions(+), 25 deletions(-)

diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index a9b418996225..12d8d38c260d 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -2883,41 +2883,85 @@ static unsigned long get_ind_bit(__u64 addr, unsigned long bit_nr, bool swap)
 	return swap ? (bit ^ (BITS_PER_LONG - 1)) : bit;
 }
 
+static struct s390_map_info *get_map_info(struct s390_io_adapter *adapter,
+					  u64 addr)
+{
+	struct s390_map_info *map;
+
+	if (!adapter)
+		return NULL;
+
+	list_for_each_entry(map, &adapter->maps, list) {
+		if (map->addr == addr)
+			return map;
+	}
+	return NULL;
+}
+
 static int adapter_indicators_set(struct kvm *kvm,
 				  struct s390_io_adapter *adapter,
 				  struct kvm_s390_adapter_int *adapter_int)
 {
 	unsigned long bit;
 	int summary_set, idx;
-	struct page *ind_page, *summary_page;
+	struct s390_map_info *ind_info, *summary_info;
 	void *map;
+	struct page *ind_page, *summary_page;
+	unsigned long flags;
 
-	ind_page = pin_map_page(kvm, adapter_int->ind_addr, 0);
-	if (!ind_page)
-		return -1;
-	summary_page = pin_map_page(kvm, adapter_int->summary_addr, 0);
-	if (!summary_page) {
-		put_page(ind_page);
-		return -1;
+	raw_spin_lock_irqsave(&adapter->maps_lock, flags);
+	ind_info = get_map_info(adapter, adapter_int->ind_addr);
+	if (!ind_info) {
+		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+		ind_page = pin_map_page(kvm, adapter_int->ind_addr, 0);
+		if (!ind_page)
+			return -1;
+		idx = srcu_read_lock(&kvm->srcu);
+		map = page_address(ind_page);
+		bit = get_ind_bit(adapter_int->ind_addr,
+				  adapter_int->ind_offset, adapter->swap);
+		set_bit(bit, map);
+		mark_page_dirty(kvm, adapter_int->ind_gaddr >> PAGE_SHIFT);
+		set_page_dirty_lock(ind_page);
+		srcu_read_unlock(&kvm->srcu, idx);
+	} else {
+		map = page_address(ind_info->page);
+		bit = get_ind_bit(ind_info->addr, adapter_int->ind_offset, adapter->swap);
+		set_bit(bit, map);
+		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+	}
+	raw_spin_lock_irqsave(&adapter->maps_lock, flags);
+	summary_info = get_map_info(adapter, adapter_int->summary_addr);
+	if (!summary_info) {
+		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
+		summary_page = pin_map_page(kvm, adapter_int->summary_addr, 0);
+		if (!summary_page) {
+			if (!ind_info) {
+				WARN_ON_ONCE(!ind_page);
+				unpin_user_page(ind_page);
+			}
+			return -1;
+		}
+		idx = srcu_read_lock(&kvm->srcu);
+		map = page_address(summary_page);
+		bit = get_ind_bit(adapter_int->summary_addr,
+				  adapter_int->summary_offset, adapter->swap);
+		summary_set = test_and_set_bit(bit, map);
+		mark_page_dirty(kvm, adapter_int->summary_gaddr >> PAGE_SHIFT);
+		set_page_dirty_lock(summary_page);
+		srcu_read_unlock(&kvm->srcu, idx);
+	} else {
+		map = page_address(summary_info->page);
+		bit = get_ind_bit(summary_info->addr, adapter_int->summary_offset,
+				  adapter->swap);
+		summary_set = test_and_set_bit(bit, map);
+		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
 	}
 
-	idx = srcu_read_lock(&kvm->srcu);
-	map = page_address(ind_page);
-	bit = get_ind_bit(adapter_int->ind_addr,
-			  adapter_int->ind_offset, adapter->swap);
-	set_bit(bit, map);
-	mark_page_dirty(kvm, adapter_int->ind_gaddr >> PAGE_SHIFT);
-	set_page_dirty_lock(ind_page);
-	map = page_address(summary_page);
-	bit = get_ind_bit(adapter_int->summary_addr,
-			  adapter_int->summary_offset, adapter->swap);
-	summary_set = test_and_set_bit(bit, map);
-	mark_page_dirty(kvm, adapter_int->summary_gaddr >> PAGE_SHIFT);
-	set_page_dirty_lock(summary_page);
-	srcu_read_unlock(&kvm->srcu, idx);
-
-	unpin_user_page(ind_page);
-	unpin_user_page(summary_page);
+	if (!ind_info)
+		unpin_user_page(ind_page);
+	if (!summary_info)
+		unpin_user_page(summary_page);
 	return summary_set ? 0 : 1;
 }
 
-- 
2.52.0

[PATCH v5 3/4] KVM: s390: Change the fi->lock to a raw_spinlock for RT case

[Douglas Freimuth]

s390 needs to maintain support for an RT kernel. This requires the
floating interrupt lock, fi->lock to be changed to a raw spin lock 
since the fi->lock maybe called with interrupts disabled in __inject_io.

Signed-off-by: Douglas Freimuth <freimuth@linux.ibm.com>
---
 arch/s390/include/asm/kvm_host.h |  2 +-
 arch/s390/kvm/intercept.c        |  4 +-
 arch/s390/kvm/interrupt.c        | 68 ++++++++++++++++----------------
 arch/s390/kvm/kvm-s390.c         |  2 +-
 4 files changed, 38 insertions(+), 38 deletions(-)

diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index fbb2406b31d2..9dd8a4986592 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -353,7 +353,7 @@ struct kvm_s390_local_interrupt {
 struct kvm_s390_float_interrupt {
 	unsigned long pending_irqs;
 	unsigned long masked_irqs;
-	spinlock_t lock;
+	raw_spinlock_t lock;
 	struct list_head lists[FIRQ_LIST_COUNT];
 	int counters[FIRQ_MAX_COUNT];
 	struct kvm_s390_mchk_info mchk;
diff --git a/arch/s390/kvm/intercept.c b/arch/s390/kvm/intercept.c
index 39aff324203e..6e9ad58c0e90 100644
--- a/arch/s390/kvm/intercept.c
+++ b/arch/s390/kvm/intercept.c
@@ -518,7 +518,7 @@ static int handle_pv_sclp(struct kvm_vcpu *vcpu)
 {
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	/*
 	 * 2 cases:
 	 * a: an sccb answering interrupt was already pending or in flight.
@@ -534,7 +534,7 @@ static int handle_pv_sclp(struct kvm_vcpu *vcpu)
 	fi->srv_signal.ext_params |= 0x43000;
 	set_bit(IRQ_PEND_EXT_SERVICE, &fi->pending_irqs);
 	clear_bit(IRQ_PEND_EXT_SERVICE, &fi->masked_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	return 0;
 }
 
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 12d8d38c260d..49ccdeccc70c 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -625,7 +625,7 @@ static int __must_check __deliver_machine_check(struct kvm_vcpu *vcpu)
 	int deliver = 0;
 	int rc = 0;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	spin_lock(&li->lock);
 	if (test_bit(IRQ_PEND_MCHK_EX, &li->pending_irqs) ||
 	    test_bit(IRQ_PEND_MCHK_REP, &li->pending_irqs)) {
@@ -654,7 +654,7 @@ static int __must_check __deliver_machine_check(struct kvm_vcpu *vcpu)
 		deliver = 1;
 	}
 	spin_unlock(&li->lock);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	if (deliver) {
 		VCPU_EVENT(vcpu, 3, "deliver: machine check mcic 0x%llx",
@@ -942,10 +942,10 @@ static int __must_check __deliver_service(struct kvm_vcpu *vcpu)
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
 	struct kvm_s390_ext_info ext;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	if (test_bit(IRQ_PEND_EXT_SERVICE, &fi->masked_irqs) ||
 	    !(test_bit(IRQ_PEND_EXT_SERVICE, &fi->pending_irqs))) {
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return 0;
 	}
 	ext = fi->srv_signal;
@@ -954,7 +954,7 @@ static int __must_check __deliver_service(struct kvm_vcpu *vcpu)
 	clear_bit(IRQ_PEND_EXT_SERVICE_EV, &fi->pending_irqs);
 	if (kvm_s390_pv_cpu_is_protected(vcpu))
 		set_bit(IRQ_PEND_EXT_SERVICE, &fi->masked_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	if (!ext.ext_params)
 		return 0;
@@ -973,16 +973,16 @@ static int __must_check __deliver_service_ev(struct kvm_vcpu *vcpu)
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
 	struct kvm_s390_ext_info ext;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	if (!(test_bit(IRQ_PEND_EXT_SERVICE_EV, &fi->pending_irqs))) {
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return 0;
 	}
 	ext = fi->srv_signal;
 	/* only clear the event bits */
 	fi->srv_signal.ext_params &= ~SCCB_EVENT_PENDING;
 	clear_bit(IRQ_PEND_EXT_SERVICE_EV, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	VCPU_EVENT(vcpu, 4, "%s", "deliver: sclp parameter event");
 	vcpu->stat.deliver_service_signal++;
@@ -998,7 +998,7 @@ static int __must_check __deliver_pfault_done(struct kvm_vcpu *vcpu)
 	struct kvm_s390_interrupt_info *inti;
 	int rc = 0;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	inti = list_first_entry_or_null(&fi->lists[FIRQ_LIST_PFAULT],
 					struct kvm_s390_interrupt_info,
 					list);
@@ -1008,7 +1008,7 @@ static int __must_check __deliver_pfault_done(struct kvm_vcpu *vcpu)
 	}
 	if (list_empty(&fi->lists[FIRQ_LIST_PFAULT]))
 		clear_bit(IRQ_PEND_PFAULT_DONE, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	if (inti) {
 		trace_kvm_s390_deliver_interrupt(vcpu->vcpu_id,
@@ -1040,7 +1040,7 @@ static int __must_check __deliver_virtio(struct kvm_vcpu *vcpu)
 	struct kvm_s390_interrupt_info *inti;
 	int rc = 0;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	inti = list_first_entry_or_null(&fi->lists[FIRQ_LIST_VIRTIO],
 					struct kvm_s390_interrupt_info,
 					list);
@@ -1058,7 +1058,7 @@ static int __must_check __deliver_virtio(struct kvm_vcpu *vcpu)
 	}
 	if (list_empty(&fi->lists[FIRQ_LIST_VIRTIO]))
 		clear_bit(IRQ_PEND_VIRTIO, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	if (inti) {
 		rc  = put_guest_lc(vcpu, EXT_IRQ_CP_SERVICE,
@@ -1119,7 +1119,7 @@ static int __must_check __deliver_io(struct kvm_vcpu *vcpu,
 
 	fi = &vcpu->kvm->arch.float_int;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	isc = irq_type_to_isc(irq_type);
 	isc_list = &fi->lists[isc];
 	inti = list_first_entry_or_null(isc_list,
@@ -1146,7 +1146,7 @@ static int __must_check __deliver_io(struct kvm_vcpu *vcpu,
 	}
 	if (list_empty(isc_list))
 		clear_bit(irq_type, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 
 	if (inti) {
 		rc = __do_deliver_io(vcpu, &(inti->io));
@@ -1663,7 +1663,7 @@ static struct kvm_s390_interrupt_info *get_io_int(struct kvm *kvm,
 	u16 id = (schid & 0xffff0000U) >> 16;
 	u16 nr = schid & 0x0000ffffU;
 
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	list_for_each_entry(iter, isc_list, list) {
 		if (schid && (id != iter->io.subchannel_id ||
 			      nr != iter->io.subchannel_nr))
@@ -1673,10 +1673,10 @@ static struct kvm_s390_interrupt_info *get_io_int(struct kvm *kvm,
 		fi->counters[FIRQ_CNTR_IO] -= 1;
 		if (list_empty(isc_list))
 			clear_bit(isc_to_irq_type(isc), &fi->pending_irqs);
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return iter;
 	}
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	return NULL;
 }
 
@@ -1771,7 +1771,7 @@ static int __inject_service(struct kvm *kvm,
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 
 	kvm->stat.inject_service_signal++;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	fi->srv_signal.ext_params |= inti->ext.ext_params & SCCB_EVENT_PENDING;
 
 	/* We always allow events, track them separately from the sccb ints */
@@ -1791,7 +1791,7 @@ static int __inject_service(struct kvm *kvm,
 	fi->srv_signal.ext_params |= inti->ext.ext_params & SCCB_MASK;
 	set_bit(IRQ_PEND_EXT_SERVICE, &fi->pending_irqs);
 out:
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	kfree(inti);
 	return 0;
 }
@@ -1802,15 +1802,15 @@ static int __inject_virtio(struct kvm *kvm,
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 
 	kvm->stat.inject_virtio++;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	if (fi->counters[FIRQ_CNTR_VIRTIO] >= KVM_S390_MAX_VIRTIO_IRQS) {
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return -EBUSY;
 	}
 	fi->counters[FIRQ_CNTR_VIRTIO] += 1;
 	list_add_tail(&inti->list, &fi->lists[FIRQ_LIST_VIRTIO]);
 	set_bit(IRQ_PEND_VIRTIO, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	return 0;
 }
 
@@ -1820,16 +1820,16 @@ static int __inject_pfault_done(struct kvm *kvm,
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 
 	kvm->stat.inject_pfault_done++;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	if (fi->counters[FIRQ_CNTR_PFAULT] >=
 		(ASYNC_PF_PER_VCPU * KVM_MAX_VCPUS)) {
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return -EBUSY;
 	}
 	fi->counters[FIRQ_CNTR_PFAULT] += 1;
 	list_add_tail(&inti->list, &fi->lists[FIRQ_LIST_PFAULT]);
 	set_bit(IRQ_PEND_PFAULT_DONE, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	return 0;
 }
 
@@ -1840,11 +1840,11 @@ static int __inject_float_mchk(struct kvm *kvm,
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 
 	kvm->stat.inject_float_mchk++;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	fi->mchk.cr14 |= inti->mchk.cr14 & (1UL << CR_PENDING_SUBCLASS);
 	fi->mchk.mcic |= inti->mchk.mcic;
 	set_bit(IRQ_PEND_MCHK_REP, &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	kfree(inti);
 	return 0;
 }
@@ -1873,9 +1873,9 @@ static int __inject_io(struct kvm *kvm, struct kvm_s390_interrupt_info *inti)
 	}
 
 	fi = &kvm->arch.float_int;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	if (fi->counters[FIRQ_CNTR_IO] >= KVM_S390_MAX_FLOAT_IRQS) {
-		spin_unlock(&fi->lock);
+		raw_spin_unlock(&fi->lock);
 		return -EBUSY;
 	}
 	fi->counters[FIRQ_CNTR_IO] += 1;
@@ -1890,7 +1890,7 @@ static int __inject_io(struct kvm *kvm, struct kvm_s390_interrupt_info *inti)
 	list = &fi->lists[FIRQ_LIST_IO_ISC_0 + isc];
 	list_add_tail(&inti->list, list);
 	set_bit(isc_to_irq_type(isc), &fi->pending_irqs);
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	return 0;
 }
 
@@ -2181,7 +2181,7 @@ void kvm_s390_clear_float_irqs(struct kvm *kvm)
 	if (!kvm_s390_pv_is_protected(kvm))
 		fi->masked_irqs = 0;
 	mutex_unlock(&kvm->lock);
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	fi->pending_irqs = 0;
 	memset(&fi->srv_signal, 0, sizeof(fi->srv_signal));
 	memset(&fi->mchk, 0, sizeof(fi->mchk));
@@ -2189,7 +2189,7 @@ void kvm_s390_clear_float_irqs(struct kvm *kvm)
 		clear_irq_list(&fi->lists[i]);
 	for (i = 0; i < FIRQ_MAX_COUNT; i++)
 		fi->counters[i] = 0;
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 	kvm_s390_gisa_clear(kvm);
 };
 
@@ -2235,7 +2235,7 @@ static int get_all_floating_irqs(struct kvm *kvm, u8 __user *usrbuf, u64 len)
 		}
 	}
 	fi = &kvm->arch.float_int;
-	spin_lock(&fi->lock);
+	raw_spin_lock(&fi->lock);
 	for (i = 0; i < FIRQ_LIST_COUNT; i++) {
 		list_for_each_entry(inti, &fi->lists[i], list) {
 			if (n == max_irqs) {
@@ -2272,7 +2272,7 @@ static int get_all_floating_irqs(struct kvm *kvm, u8 __user *usrbuf, u64 len)
 }
 
 out:
-	spin_unlock(&fi->lock);
+	raw_spin_unlock(&fi->lock);
 out_nolock:
 	if (!ret && n > 0) {
 		if (copy_to_user(usrbuf, buf, sizeof(struct kvm_s390_irq) * n))
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 74f453f039a3..d8011b6d6801 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -3263,7 +3263,7 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 	}
 
 	mutex_init(&kvm->arch.float_int.ais_lock);
-	spin_lock_init(&kvm->arch.float_int.lock);
+	raw_spin_lock_init(&kvm->arch.float_int.lock);
 	for (i = 0; i < FIRQ_LIST_COUNT; i++)
 		INIT_LIST_HEAD(&kvm->arch.float_int.lists[i]);
 	init_waitqueue_head(&kvm->arch.ipte_wq);
-- 
2.52.0

[PATCH v5 4/4] KVM: s390: Introducing kvm_arch_set_irq_inatomic fast inject

[Douglas Freimuth]

S390 needs a fast path for irq injection, and along those lines we
introduce kvm_arch_set_irq_inatomic. Instead of placing all interrupts on
the global work queue as it does today, this patch provides a fast path for
irq injection.

The inatomic fast path cannot lose control since it is running with
interrupts disabled. This meant making the following changes that exist on
the slow path today. First, the adapter_indicators page needs to be mapped
since it is accessed with interrupts disabled, so we added map/unmap
functions. Second, access to shared resources between the fast and slow
paths needed to be changed from mutex and semaphores to raw_spin_lock's.
Finally, the memory allocation on the slow path utilizes GFP_KERNEL_ACCOUNT
but we had to implement the fast path with GFP_ATOMIC allocation. Each of
these enhancements were required to prevent blocking on the fast inject
path.

Fencing of Fast Inject in Secure Execution environments is enabled in the
patch series by not mapping adapter indicator pages. In Secure Execution
environments the path of execution available before this patch is followed.

Statistical counters have been added to enable analysis of irq injection on
the fast path and slow path including io_390_inatomic, io_flic_inject_airq,
io_set_adapter_int and io_390_inatomic_adapter_masked.

Signed-off-by: Douglas Freimuth <freimuth@linux.ibm.com>
---
 arch/s390/include/asm/kvm_host.h |   6 +-
 arch/s390/kvm/interrupt.c        | 163 +++++++++++++++++++++++++++----
 arch/s390/kvm/kvm-s390.c         |  21 +++-
 arch/s390/kvm/kvm-s390.h         |   3 +-
 4 files changed, 170 insertions(+), 23 deletions(-)

diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 9dd8a4986592..b485dee4c766 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -359,7 +359,7 @@ struct kvm_s390_float_interrupt {
 	struct kvm_s390_mchk_info mchk;
 	struct kvm_s390_ext_info srv_signal;
 	int last_sleep_cpu;
-	struct mutex ais_lock;
+	raw_spinlock_t ais_lock;
 	u8 simm;
 	u8 nimm;
 };
@@ -450,6 +450,10 @@ struct kvm_vm_stat {
 	u64 inject_io;
 	u64 io_390_adapter_map;
 	u64 io_390_adapter_unmap;
+	u64 io_390_inatomic;
+	u64 io_flic_inject_airq;
+	u64 io_set_adapter_int;
+	u64 io_390_inatomic_adapter_masked;
 	u64 inject_float_mchk;
 	u64 inject_pfault_done;
 	u64 inject_service_signal;
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 49ccdeccc70c..1c79ad072fce 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -1966,15 +1966,10 @@ static int __inject_vm(struct kvm *kvm, struct kvm_s390_interrupt_info *inti)
 }
 
 int kvm_s390_inject_vm(struct kvm *kvm,
-		       struct kvm_s390_interrupt *s390int)
+		       struct kvm_s390_interrupt *s390int, struct kvm_s390_interrupt_info *inti)
 {
-	struct kvm_s390_interrupt_info *inti;
 	int rc;
 
-	inti = kzalloc_obj(*inti, GFP_KERNEL_ACCOUNT);
-	if (!inti)
-		return -ENOMEM;
-
 	inti->type = s390int->type;
 	switch (inti->type) {
 	case KVM_S390_INT_VIRTIO:
@@ -2010,6 +2005,7 @@ int kvm_s390_inject_vm(struct kvm *kvm,
 				 2);
 
 	rc = __inject_vm(kvm, inti);
+	/* memory allocation is done by the caller and inti is passed in, we free it here */
 	if (rc)
 		kfree(inti);
 	return rc;
@@ -2287,6 +2283,7 @@ static int flic_ais_mode_get_all(struct kvm *kvm, struct kvm_device_attr *attr)
 {
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 	struct kvm_s390_ais_all ais;
+	unsigned long flags;
 
 	if (attr->attr < sizeof(ais))
 		return -EINVAL;
@@ -2294,10 +2291,10 @@ static int flic_ais_mode_get_all(struct kvm *kvm, struct kvm_device_attr *attr)
 	if (!test_kvm_facility(kvm, 72))
 		return -EOPNOTSUPP;
 
-	mutex_lock(&fi->ais_lock);
+	raw_spin_lock_irqsave(&fi->ais_lock, flags);
 	ais.simm = fi->simm;
 	ais.nimm = fi->nimm;
-	mutex_unlock(&fi->ais_lock);
+	raw_spin_unlock_irqrestore(&fi->ais_lock, flags);
 
 	if (copy_to_user((void __user *)attr->addr, &ais, sizeof(ais)))
 		return -EFAULT;
@@ -2674,6 +2671,7 @@ static int modify_ais_mode(struct kvm *kvm, struct kvm_device_attr *attr)
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 	struct kvm_s390_ais_req req;
 	int ret = 0;
+	unsigned long flags;
 
 	if (!test_kvm_facility(kvm, 72))
 		return -EOPNOTSUPP;
@@ -2690,7 +2688,7 @@ static int modify_ais_mode(struct kvm *kvm, struct kvm_device_attr *attr)
 				       2 : KVM_S390_AIS_MODE_SINGLE :
 				       KVM_S390_AIS_MODE_ALL, req.mode);
 
-	mutex_lock(&fi->ais_lock);
+	raw_spin_lock_irqsave(&fi->ais_lock, flags);
 	switch (req.mode) {
 	case KVM_S390_AIS_MODE_ALL:
 		fi->simm &= ~AIS_MODE_MASK(req.isc);
@@ -2703,7 +2701,7 @@ static int modify_ais_mode(struct kvm *kvm, struct kvm_device_attr *attr)
 	default:
 		ret = -EINVAL;
 	}
-	mutex_unlock(&fi->ais_lock);
+	raw_spin_unlock_irqrestore(&fi->ais_lock, flags);
 
 	return ret;
 }
@@ -2717,25 +2715,33 @@ static int kvm_s390_inject_airq(struct kvm *kvm,
 		.parm = 0,
 		.parm64 = isc_to_int_word(adapter->isc),
 	};
+	struct kvm_s390_interrupt_info *inti;
+	unsigned long flags;
+
 	int ret = 0;
 
+	inti = kzalloc_obj(*inti, GFP_KERNEL_ACCOUNT);
+	if (!inti)
+		return -ENOMEM;
+
 	if (!test_kvm_facility(kvm, 72) || !adapter->suppressible)
-		return kvm_s390_inject_vm(kvm, &s390int);
+		return kvm_s390_inject_vm(kvm, &s390int, inti);
 
-	mutex_lock(&fi->ais_lock);
+	raw_spin_lock_irqsave(&fi->ais_lock, flags);
 	if (fi->nimm & AIS_MODE_MASK(adapter->isc)) {
 		trace_kvm_s390_airq_suppressed(adapter->id, adapter->isc);
+		kfree(inti);
 		goto out;
 	}
 
-	ret = kvm_s390_inject_vm(kvm, &s390int);
+	ret = kvm_s390_inject_vm(kvm, &s390int, inti);
 	if (!ret && (fi->simm & AIS_MODE_MASK(adapter->isc))) {
 		fi->nimm |= AIS_MODE_MASK(adapter->isc);
 		trace_kvm_s390_modify_ais_mode(adapter->isc,
 					       KVM_S390_AIS_MODE_SINGLE, 2);
 	}
 out:
-	mutex_unlock(&fi->ais_lock);
+	raw_spin_unlock_irqrestore(&fi->ais_lock, flags);
 	return ret;
 }
 
@@ -2744,6 +2750,8 @@ static int flic_inject_airq(struct kvm *kvm, struct kvm_device_attr *attr)
 	unsigned int id = attr->attr;
 	struct s390_io_adapter *adapter = get_io_adapter(kvm, id);
 
+	kvm->stat.io_flic_inject_airq++;
+
 	if (!adapter)
 		return -EINVAL;
 
@@ -2754,6 +2762,7 @@ static int flic_ais_mode_set_all(struct kvm *kvm, struct kvm_device_attr *attr)
 {
 	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
 	struct kvm_s390_ais_all ais;
+	unsigned long flags;
 
 	if (!test_kvm_facility(kvm, 72))
 		return -EOPNOTSUPP;
@@ -2761,10 +2770,10 @@ static int flic_ais_mode_set_all(struct kvm *kvm, struct kvm_device_attr *attr)
 	if (copy_from_user(&ais, (void __user *)attr->addr, sizeof(ais)))
 		return -EFAULT;
 
-	mutex_lock(&fi->ais_lock);
+	raw_spin_lock_irqsave(&fi->ais_lock, flags);
 	fi->simm = ais.simm;
 	fi->nimm = ais.nimm;
-	mutex_unlock(&fi->ais_lock);
+	raw_spin_unlock_irqrestore(&fi->ais_lock, flags);
 
 	return 0;
 }
@@ -2930,6 +2939,7 @@ static int adapter_indicators_set(struct kvm *kvm,
 		set_bit(bit, map);
 		raw_spin_unlock_irqrestore(&adapter->maps_lock, flags);
 	}
+
 	raw_spin_lock_irqsave(&adapter->maps_lock, flags);
 	summary_info = get_map_info(adapter, adapter_int->summary_addr);
 	if (!summary_info) {
@@ -2965,6 +2975,44 @@ static int adapter_indicators_set(struct kvm *kvm,
 	return summary_set ? 0 : 1;
 }
 
+static int adapter_indicators_set_fast(struct kvm *kvm,
+				       struct s390_io_adapter *adapter,
+				       struct kvm_s390_adapter_int *adapter_int,
+				       int setbit)
+{
+	unsigned long bit;
+	int summary_set;
+	struct s390_map_info *ind_info, *summary_info;
+	void *map;
+
+	raw_spin_lock(&adapter->maps_lock);
+	ind_info = get_map_info(adapter, adapter_int->ind_addr);
+	if (!ind_info) {
+		raw_spin_unlock(&adapter->maps_lock);
+		return -EWOULDBLOCK;
+	}
+	map = page_address(ind_info->page);
+	bit = get_ind_bit(ind_info->addr, adapter_int->ind_offset, adapter->swap);
+	if (setbit)
+		set_bit(bit, map);
+	summary_info = get_map_info(adapter, adapter_int->summary_addr);
+	if (!summary_info) {
+		raw_spin_unlock(&adapter->maps_lock);
+		return -EWOULDBLOCK;
+	}
+	map = page_address(summary_info->page);
+	bit = get_ind_bit(summary_info->addr, adapter_int->summary_offset,
+			  adapter->swap);
+	/* If setbit then set summary bit. Else if falling back to the slow path */
+	/* with setbit==0 then clear the summary bit so the slow path re-injects */
+	if (setbit)
+		summary_set = test_and_set_bit(bit, map);
+	else
+		summary_set = test_and_clear_bit(bit, map);
+	raw_spin_unlock(&adapter->maps_lock);
+	return summary_set ? 0 : 1;
+}
+
 /*
  * < 0 - not injected due to error
  * = 0 - coalesced, summary indicator already active
@@ -2977,6 +3025,8 @@ static int set_adapter_int(struct kvm_kernel_irq_routing_entry *e,
 	int ret;
 	struct s390_io_adapter *adapter;
 
+	kvm->stat.io_set_adapter_int++;
+
 	/* We're only interested in the 0->1 transition. */
 	if (!level)
 		return 0;
@@ -3045,7 +3095,6 @@ int kvm_set_routing_entry(struct kvm *kvm,
 	int idx;
 
 	switch (ue->type) {
-	/* we store the userspace addresses instead of the guest addresses */
 	case KVM_IRQ_ROUTING_S390_ADAPTER:
 		if (kvm_is_ucontrol(kvm))
 			return -EINVAL;
@@ -3636,3 +3685,83 @@ int __init kvm_s390_gib_init(u8 nisc)
 out:
 	return rc;
 }
+
+/*
+ * kvm_arch_set_irq_inatomic: fast-path for irqfd injection
+ */
+int kvm_arch_set_irq_inatomic(struct kvm_kernel_irq_routing_entry *e,
+			      struct kvm *kvm, int irq_source_id, int level,
+			      bool line_status)
+{
+	int ret, setbit;
+	struct s390_io_adapter *adapter;
+	struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
+	struct kvm_s390_interrupt_info *inti;
+	struct kvm_s390_interrupt s390int = {
+			.type = KVM_S390_INT_IO(1, 0, 0, 0),
+			.parm = 0,
+	};
+
+	kvm->stat.io_390_inatomic++;
+
+	/* We're only interested in the 0->1 transition. */
+	if (!level)
+		return -EWOULDBLOCK;
+	if (e->type != KVM_IRQ_ROUTING_S390_ADAPTER)
+		return -EWOULDBLOCK;
+
+	adapter = get_io_adapter(kvm, e->adapter.adapter_id);
+	if (!adapter)
+		return -EWOULDBLOCK;
+
+	s390int.parm64 = isc_to_int_word(adapter->isc);
+	setbit = 1;
+	ret = adapter_indicators_set_fast(kvm, adapter, &e->adapter, setbit);
+	if (ret < 0)
+		return -EWOULDBLOCK;
+	if (!ret || adapter->masked) {
+		kvm->stat.io_390_inatomic_adapter_masked++;
+		return 0;
+	}
+
+	inti = kzalloc_obj(*inti, GFP_ATOMIC);
+	if (!inti) {
+		setbit = 0;
+		adapter_indicators_set_fast(kvm, adapter, &e->adapter, setbit);
+		return -EWOULDBLOCK;
+	}
+
+	if (!test_kvm_facility(kvm, 72) || !adapter->suppressible) {
+		ret = kvm_s390_inject_vm(kvm, &s390int, inti);
+		if (ret == 0) {
+			return ret;
+		} else {
+			setbit = 0;
+			adapter_indicators_set_fast(kvm, adapter, &e->adapter, setbit);
+			return -EWOULDBLOCK;
+		}
+	}
+
+	raw_spin_lock(&fi->ais_lock);
+	if (fi->nimm & AIS_MODE_MASK(adapter->isc)) {
+		trace_kvm_s390_airq_suppressed(adapter->id, adapter->isc);
+		kfree(inti);
+		goto out;
+	}
+
+	ret = kvm_s390_inject_vm(kvm, &s390int, inti);
+	if (!ret && (fi->simm & AIS_MODE_MASK(adapter->isc))) {
+		fi->nimm |= AIS_MODE_MASK(adapter->isc);
+		trace_kvm_s390_modify_ais_mode(adapter->isc,
+					       KVM_S390_AIS_MODE_SINGLE, 2);
+	} else if (ret) {
+		raw_spin_unlock(&fi->ais_lock);
+		setbit = 0;
+		adapter_indicators_set_fast(kvm, adapter, &e->adapter, setbit);
+		return -EWOULDBLOCK;
+	}
+
+out:
+	raw_spin_unlock(&fi->ais_lock);
+	return 0;
+}
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index d8011b6d6801..11b62fa8634f 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -70,6 +70,10 @@ const struct kvm_stats_desc kvm_vm_stats_desc[] = {
 	STATS_DESC_COUNTER(VM, inject_io),
 	STATS_DESC_COUNTER(VM, io_390_adapter_map),
 	STATS_DESC_COUNTER(VM, io_390_adapter_unmap),
+	STATS_DESC_COUNTER(VM, io_390_inatomic),
+	STATS_DESC_COUNTER(VM, io_flic_inject_airq),
+	STATS_DESC_COUNTER(VM, io_set_adapter_int),
+	STATS_DESC_COUNTER(VM, io_390_inatomic_adapter_masked),
 	STATS_DESC_COUNTER(VM, inject_float_mchk),
 	STATS_DESC_COUNTER(VM, inject_pfault_done),
 	STATS_DESC_COUNTER(VM, inject_service_signal),
@@ -2856,6 +2860,7 @@ int kvm_arch_vm_ioctl(struct file *filp, unsigned int ioctl, unsigned long arg)
 	void __user *argp = (void __user *)arg;
 	struct kvm_device_attr attr;
 	int r;
+	struct kvm_s390_interrupt_info *inti;
 
 	switch (ioctl) {
 	case KVM_S390_INTERRUPT: {
@@ -2864,7 +2869,10 @@ int kvm_arch_vm_ioctl(struct file *filp, unsigned int ioctl, unsigned long arg)
 		r = -EFAULT;
 		if (copy_from_user(&s390int, argp, sizeof(s390int)))
 			break;
-		r = kvm_s390_inject_vm(kvm, &s390int);
+		inti = kzalloc_obj(*inti, GFP_KERNEL_ACCOUNT);
+		if (!inti)
+			return -ENOMEM;
+		r = kvm_s390_inject_vm(kvm, &s390int, inti);
 		break;
 	}
 	case KVM_CREATE_IRQCHIP: {
@@ -3262,7 +3270,7 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 		mutex_unlock(&kvm->lock);
 	}
 
-	mutex_init(&kvm->arch.float_int.ais_lock);
+	raw_spin_lock_init(&kvm->arch.float_int.ais_lock);
 	raw_spin_lock_init(&kvm->arch.float_int.lock);
 	for (i = 0; i < FIRQ_LIST_COUNT; i++)
 		INIT_LIST_HEAD(&kvm->arch.float_int.lists[i]);
@@ -4384,19 +4392,24 @@ int kvm_s390_try_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clo
 }
 
 static void __kvm_inject_pfault_token(struct kvm_vcpu *vcpu, bool start_token,
-				      unsigned long token)
+				     unsigned long token)
 {
 	struct kvm_s390_interrupt inti;
 	struct kvm_s390_irq irq;
+	struct kvm_s390_interrupt_info *inti_mem = NULL;
 
 	if (start_token) {
 		irq.u.ext.ext_params2 = token;
 		irq.type = KVM_S390_INT_PFAULT_INIT;
 		WARN_ON_ONCE(kvm_s390_inject_vcpu(vcpu, &irq));
 	} else {
+		inti_mem = kzalloc_obj(*inti_mem, GFP_KERNEL_ACCOUNT);
+		if (WARN_ON_ONCE(!inti_mem))
+			return;
+
 		inti.type = KVM_S390_INT_PFAULT_DONE;
 		inti.parm64 = token;
-		WARN_ON_ONCE(kvm_s390_inject_vm(vcpu->kvm, &inti));
+		WARN_ON_ONCE(kvm_s390_inject_vm(vcpu->kvm, &inti, inti_mem));
 	}
 }
 
diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h
index 7ba885cb6bd1..6d2842fb71a3 100644
--- a/arch/s390/kvm/kvm-s390.h
+++ b/arch/s390/kvm/kvm-s390.h
@@ -376,7 +376,8 @@ int __must_check kvm_s390_deliver_pending_interrupts(struct kvm_vcpu *vcpu);
 void kvm_s390_clear_local_irqs(struct kvm_vcpu *vcpu);
 void kvm_s390_clear_float_irqs(struct kvm *kvm);
 int __must_check kvm_s390_inject_vm(struct kvm *kvm,
-				    struct kvm_s390_interrupt *s390int);
+				    struct kvm_s390_interrupt *s390int,
+				    struct kvm_s390_interrupt_info *inti);
 int __must_check kvm_s390_inject_vcpu(struct kvm_vcpu *vcpu,
 				      struct kvm_s390_irq *irq);
 static inline int kvm_s390_inject_prog_irq(struct kvm_vcpu *vcpu,
-- 
2.52.0