From: Marcelo Tosatti <mtosatti@redhat.com>
To: Avi Kivity <avi@redhat.com>, kvm@vger.kernel.org
Cc: Gregory Haskins <gregory.haskins@gmail.com>,
Marcelo Tosatti <mtosatti@redhat.com>
Subject: [patch 3/4] KVM: introduce irq_lock, use it to protect ioapic
Date: Wed, 20 May 2009 15:48:44 -0300 [thread overview]
Message-ID: <20090520185134.748295218@localhost.localdomain> (raw)
In-Reply-To: 20090520184841.954066003@localhost.localdomain
[-- Attachment #1: ioapic-lock --]
[-- Type: text/plain, Size: 3320 bytes --]
Subject says it all.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Index: kvm-irqlock/include/linux/kvm_host.h
===================================================================
--- kvm-irqlock.orig/include/linux/kvm_host.h
+++ kvm-irqlock/include/linux/kvm_host.h
@@ -123,7 +123,6 @@ struct kvm_kernel_irq_routing_entry {
};
struct kvm {
- struct mutex lock; /* protects the vcpus array and APIC accesses */
spinlock_t mmu_lock;
struct rw_semaphore slots_lock;
struct mm_struct *mm; /* userspace tied to this vm */
@@ -132,6 +131,12 @@ struct kvm {
KVM_PRIVATE_MEM_SLOTS];
struct kvm_vcpu *vcpus[KVM_MAX_VCPUS];
struct list_head vm_list;
+ struct mutex lock; /*
+ * - protects mmio_bus, pio_bus.
+ * - protects a few concurrent ioctl's (FIXME).
+ * - protects concurrent create_vcpu, but
+ * kvm->vcpus walkers do it locklessly (FIXME).
+ */
struct kvm_io_bus mmio_bus;
struct kvm_io_bus pio_bus;
struct list_head irqfds;
@@ -143,6 +148,7 @@ struct kvm {
struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;
#endif
+ struct mutex irq_lock; /* protects high level irq logic, ioapic */
#ifdef CONFIG_HAVE_KVM_IRQCHIP
struct list_head irq_routing; /* of kvm_kernel_irq_routing_entry */
struct hlist_head mask_notifier_list;
Index: kvm-irqlock/virt/kvm/ioapic.c
===================================================================
--- kvm-irqlock.orig/virt/kvm/ioapic.c
+++ kvm-irqlock/virt/kvm/ioapic.c
@@ -225,6 +225,10 @@ static int ioapic_in_range(struct kvm_io
{
struct kvm_ioapic *ioapic = (struct kvm_ioapic *)this->private;
+ /*
+ * Lockless check for ioapic->base_address on the hazy assumption
+ * it does not change during the lifetime of a VM.
+ */
return ((addr >= ioapic->base_address &&
(addr < ioapic->base_address + IOAPIC_MEM_LENGTH)));
}
@@ -238,6 +242,7 @@ static void ioapic_mmio_read(struct kvm_
ioapic_debug("addr %lx\n", (unsigned long)addr);
ASSERT(!(addr & 0xf)); /* check alignment */
+ mutex_lock(&ioapic->kvm->irq_lock);
addr &= 0xff;
switch (addr) {
case IOAPIC_REG_SELECT:
@@ -264,6 +269,7 @@ static void ioapic_mmio_read(struct kvm_
default:
printk(KERN_WARNING "ioapic: wrong length %d\n", len);
}
+ mutex_unlock(&ioapic->kvm->irq_lock);
}
static void ioapic_mmio_write(struct kvm_io_device *this, gpa_t addr, int len,
@@ -275,6 +281,8 @@ static void ioapic_mmio_write(struct kvm
ioapic_debug("ioapic_mmio_write addr=%p len=%d val=%p\n",
(void*)addr, len, val);
ASSERT(!(addr & 0xf)); /* check alignment */
+
+ mutex_lock(&ioapic->kvm->irq_lock);
if (len == 4 || len == 8)
data = *(u32 *) val;
else {
@@ -300,6 +308,7 @@ static void ioapic_mmio_write(struct kvm
default:
break;
}
+ mutex_unlock(&ioapic->kvm->irq_lock);
}
void kvm_ioapic_reset(struct kvm_ioapic *ioapic)
Index: kvm-irqlock/virt/kvm/kvm_main.c
===================================================================
--- kvm-irqlock.orig/virt/kvm/kvm_main.c
+++ kvm-irqlock/virt/kvm/kvm_main.c
@@ -985,6 +985,7 @@ static struct kvm *kvm_create_vm(void)
kvm_io_bus_init(&kvm->pio_bus);
INIT_LIST_HEAD(&kvm->irqfds);
mutex_init(&kvm->lock);
+ mutex_init(&kvm->irq_lock);
kvm_io_bus_init(&kvm->mmio_bus);
init_rwsem(&kvm->slots_lock);
atomic_set(&kvm->users_count, 1);
--
next prev parent reply other threads:[~2009-05-20 18:54 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-05-18 16:56 [patch 0/4] move irq protection role to separate kvm->irq_lock Marcelo Tosatti
2009-05-18 16:56 ` [patch 1/4] KVM: x86: grab pic lock in kvm_pic_clear_isr_ack Marcelo Tosatti
2009-05-18 16:56 ` [patch 2/4] KVM: move coalesced_mmio locking to its own device Marcelo Tosatti
2009-05-20 12:06 ` Avi Kivity
2009-05-20 14:09 ` Marcelo Tosatti
2009-05-20 14:29 ` Avi Kivity
2009-05-20 15:13 ` Marcelo Tosatti
2009-05-20 15:22 ` Marcelo Tosatti
2009-05-20 15:22 ` Avi Kivity
2009-05-20 18:48 ` [patch 0/4] move irq protection role to separate lock v2 Marcelo Tosatti
2009-05-20 18:48 ` [patch 1/4] KVM: x86: grab pic lock in kvm_pic_clear_isr_ack Marcelo Tosatti
2009-05-20 18:48 ` [patch 2/4] KVM: move coalesced_mmio locking to its own device Marcelo Tosatti
2009-05-24 14:04 ` Avi Kivity
2009-05-25 11:04 ` Marcelo Tosatti
2009-05-26 11:24 ` Avi Kivity
2009-05-26 13:15 ` Marcelo Tosatti
2009-05-26 13:27 ` Avi Kivity
2009-05-20 18:48 ` Marcelo Tosatti [this message]
2009-05-24 14:10 ` [patch 3/4] KVM: introduce irq_lock, use it to protect ioapic Avi Kivity
2009-05-25 11:11 ` Marcelo Tosatti
2009-05-26 11:33 ` Avi Kivity
2009-05-28 4:45 ` [patch 0/4] move irq protection role to separate lock v3 Marcelo Tosatti
2009-05-28 4:45 ` [patch 1/4] KVM: x86: grab pic lock in kvm_pic_clear_isr_ack Marcelo Tosatti
2009-05-28 4:45 ` [patch 2/4] KVM: move coalesced_mmio locking to its own device Marcelo Tosatti
2009-05-31 12:14 ` Avi Kivity
2009-06-01 21:23 ` Marcelo Tosatti
2009-06-01 21:43 ` Avi Kivity
2009-06-04 18:08 ` [patch 0/4] move irq protection role to separate lock v4 Marcelo Tosatti
2009-06-04 18:08 ` [patch 1/4] KVM: x86: grab pic lock in kvm_pic_clear_isr_ack Marcelo Tosatti
2009-06-04 18:08 ` [patch 2/4] KVM: move coalesced_mmio locking to its own device Marcelo Tosatti
2009-06-04 18:08 ` [patch 3/4] KVM: introduce irq_lock, use it to protect ioapic Marcelo Tosatti
2009-06-04 18:08 ` [patch 4/4] KVM: switch irq injection/acking data structures to irq_lock Marcelo Tosatti
2009-06-08 9:18 ` [patch 0/4] move irq protection role to separate lock v4 Avi Kivity
2009-05-28 4:45 ` [patch 3/4] KVM: introduce irq_lock, use it to protect ioapic Marcelo Tosatti
2009-05-28 4:45 ` [patch 4/4] KVM: switch irq injection/acking data structures to irq_lock Marcelo Tosatti
2009-05-20 18:48 ` [patch 4/4] KVM: switch irq injection/acking " Marcelo Tosatti
2009-05-21 4:50 ` [patch 0/4] move irq protection role to separate lock v2 Marcelo Tosatti
2009-05-21 6:55 ` Christian Bornträger
2009-05-21 7:26 ` Avi Kivity
2009-05-21 14:32 ` Marcelo Tosatti
2009-05-21 15:02 ` Avi Kivity
2009-05-18 16:56 ` [patch 3/4] KVM: introduce irq_lock, use it protect ioapic Marcelo Tosatti
2009-05-20 12:11 ` Avi Kivity
2009-05-20 14:11 ` Marcelo Tosatti
2009-05-20 14:29 ` Avi Kivity
2009-05-18 16:56 ` [patch 4/4] KVM: switch irq injection/acking to irq_lock Marcelo Tosatti
2009-05-20 12:15 ` Gregory Haskins
2009-05-20 14:16 ` Marcelo Tosatti
2009-05-24 14:53 ` Avi Kivity
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20090520185134.748295218@localhost.localdomain \
--to=mtosatti@redhat.com \
--cc=avi@redhat.com \
--cc=gregory.haskins@gmail.com \
--cc=kvm@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox