All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefano Stabellini <sstabellini@kernel.org>
To: julien.grall@arm.com
Cc: xen-devel@lists.xenproject.org, sstabellini@kernel.org
Subject: [PATCH v5 2/3] xen/arm: move setting of new target vcpu to vgic_migrate_irq
Date: Wed,  1 Mar 2017 14:15:44 -0800	[thread overview]
Message-ID: <1488406545-26164-2-git-send-email-sstabellini@kernel.org> (raw)
In-Reply-To: <1488406545-26164-1-git-send-email-sstabellini@kernel.org>

Move the atomic write of rank->vcpu, which sets the new vcpu target, to
vgic_migrate_irq, at the beginning of the lock protected area (protected
by the vgic lock).

This code movement reduces race conditions between vgic_migrate_irq and
setting rank->vcpu on one pcpu and gic_update_one_lr on another pcpu.

When gic_update_one_lr and vgic_migrate_irq take the same vgic lock,
there are no more race conditions with this patch. When vgic_migrate_irq
is called multiple times while GIC_IRQ_GUEST_MIGRATING is already set, a
race condition still exists because in that case gic_update_one_lr and
vgic_migrate_irq take different vgic locks.

Signed-off-by: Stefano Stabellini <sstabellini@kernel.org>
---
 xen/arch/arm/vgic-v2.c     |  5 ++---
 xen/arch/arm/vgic-v3.c     |  4 +---
 xen/arch/arm/vgic.c        | 15 ++++++++++-----
 xen/include/asm-arm/vgic.h |  3 ++-
 4 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/xen/arch/arm/vgic-v2.c b/xen/arch/arm/vgic-v2.c
index 0674f7b..43b4ac3 100644
--- a/xen/arch/arm/vgic-v2.c
+++ b/xen/arch/arm/vgic-v2.c
@@ -158,10 +158,9 @@ static void vgic_store_itargetsr(struct domain *d, struct vgic_irq_rank *rank,
         {
             vgic_migrate_irq(d->vcpu[old_target],
                              d->vcpu[new_target],
-                             virq);
+                             virq,
+                             &rank->vcpu[offset]);
         }
-
-        write_atomic(&rank->vcpu[offset], new_target);
     }
 }
 
diff --git a/xen/arch/arm/vgic-v3.c b/xen/arch/arm/vgic-v3.c
index 0679e76..a6fc6cb 100644
--- a/xen/arch/arm/vgic-v3.c
+++ b/xen/arch/arm/vgic-v3.c
@@ -151,9 +151,7 @@ static void vgic_store_irouter(struct domain *d, struct vgic_irq_rank *rank,
 
     /* Only migrate the IRQ if the target vCPU has changed */
     if ( new_vcpu != old_vcpu )
-        vgic_migrate_irq(old_vcpu, new_vcpu, virq);
-
-    write_atomic(&rank->vcpu[offset], new_vcpu->vcpu_id);
+        vgic_migrate_irq(old_vcpu, new_vcpu, virq, &rank->vcpu[offset]);
 }
 
 static inline bool vgic_reg64_check_access(struct hsr_dabt dabt)
diff --git a/xen/arch/arm/vgic.c b/xen/arch/arm/vgic.c
index 67d75a6..a323e7e 100644
--- a/xen/arch/arm/vgic.c
+++ b/xen/arch/arm/vgic.c
@@ -237,7 +237,8 @@ static int vgic_get_virq_priority(struct vcpu *v, unsigned int virq)
     return priority;
 }
 
-void vgic_migrate_irq(struct vcpu *old, struct vcpu *new, unsigned int irq)
+void vgic_migrate_irq(struct vcpu *old, struct vcpu *new,
+                      unsigned int irq, uint8_t *t_vcpu)
 {
     unsigned long flags;
     struct pending_irq *p = irq_to_pending(old, irq);
@@ -246,13 +247,17 @@ void vgic_migrate_irq(struct vcpu *old, struct vcpu *new, unsigned int irq)
     if ( p->desc == NULL )
         return;
 
-    /* migration already in progress, no need to do anything */
-    if ( test_bit(GIC_IRQ_GUEST_MIGRATING, &p->status) )
-        return;
-
     perfc_incr(vgic_irq_migrates);
 
     spin_lock_irqsave(&old->arch.vgic.lock, flags);
+    write_atomic(t_vcpu, new->vcpu_id);
+
+    /* migration already in progress, no need to do anything */
+    if ( test_bit(GIC_IRQ_GUEST_MIGRATING, &p->status) )
+    {
+        spin_unlock_irqrestore(&old->arch.vgic.lock, flags);
+        return;
+    }
 
     if ( list_empty(&p->inflight) )
     {
diff --git a/xen/include/asm-arm/vgic.h b/xen/include/asm-arm/vgic.h
index 894c3f1..5d8b8c8 100644
--- a/xen/include/asm-arm/vgic.h
+++ b/xen/include/asm-arm/vgic.h
@@ -314,7 +314,8 @@ extern int vcpu_vgic_free(struct vcpu *v);
 extern bool vgic_to_sgi(struct vcpu *v, register_t sgir,
                         enum gic_sgi_mode irqmode, int virq,
                         const struct sgi_target *target);
-extern void vgic_migrate_irq(struct vcpu *old, struct vcpu *new, unsigned int irq);
+extern void vgic_migrate_irq(struct vcpu *old, struct vcpu *new, unsigned int irq,
+                             uint8_t *t_vcpu);
 
 /* Reserve a specific guest vIRQ */
 extern bool vgic_reserve_virq(struct domain *d, unsigned int virq);
-- 
1.9.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

  reply	other threads:[~2017-03-01 22:15 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-01 22:15 [PATCH v5 0/3] xen/arm: remove race conditions in irq migration Stefano Stabellini
2017-03-01 22:15 ` [PATCH v5 1/3] arm: remove irq from inflight, then change physical affinity Stefano Stabellini
2017-03-01 22:15   ` Stefano Stabellini [this message]
2017-03-03 17:38     ` [PATCH v5 2/3] xen/arm: move setting of new target vcpu to vgic_migrate_irq Julien Grall
2017-03-29 23:48       ` Stefano Stabellini
2017-03-31 15:29         ` Julien Grall
2017-03-01 22:15   ` [PATCH v5 3/3] xen/arm: vgic_migrate_irq: do not race against GIC_IRQ_GUEST_MIGRATING Stefano Stabellini
2017-03-03 17:47     ` Julien Grall
2017-03-29 23:47       ` Stefano Stabellini
2017-03-31 16:02         ` Julien Grall
2017-03-31 20:24           ` Stefano Stabellini
2017-04-03 11:03             ` Julien Grall
2017-04-03 21:24               ` Stefano Stabellini
2017-03-01 23:24   ` [PATCH v5 1/3] arm: remove irq from inflight, then change physical affinity Julien Grall
2017-03-03 17:04     ` Julien Grall
2017-03-03 19:34       ` Stefano Stabellini
2017-03-03 19:38         ` Julien Grall
2017-03-22 23:59           ` Stefano Stabellini

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=1488406545-26164-2-git-send-email-sstabellini@kernel.org \
    --to=sstabellini@kernel.org \
    --cc=julien.grall@arm.com \
    --cc=xen-devel@lists.xenproject.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.