On 2025-08-12 15:00, Jason Andryuk wrote:
VIRQs come in 3 flavors, per-VPU, per-domain, and global. The existing
tracking of VIRQs is handled by per-cpu variables virq_to_irq.
The issue is that bind_virq_to_irq() sets the per_cpu virq_to_irq at
registration time - typically CPU 0. Later, the interrupt can migrate,
and info->cpu is updated. When calling unbind_from_irq(), the per-cpu
virq_to_irq is cleared for a different cpu. If bind_virq_to_irq() is
called again with CPU 0, the stale irq is returned.
Change the virq_to_irq tracking to use CPU 0 for per-domain and global
VIRQs. As there can be at most one of each, there is no need for
per-vcpu tracking. Also, per-domain and global VIRQs need to be
registered on CPU 0 and can later move, so this matches the expectation.
Fixes: e46cdb66c8fc ("xen: event channels")
Cc: sta...@vger.kernel.org
Signed-off-by: Jason Andryuk <jason.andr...@amd.com>
---
Fixes is the introduction of the virq_to_irq per-cpu array.
This was found with the out-of-tree argo driver during suspend/resume.
On suspend, the per-domain VIRQ_ARGO is unbound. On resume, the driver
attempts to bind VIRQ_ARGO. The stale irq is returned, but the
WARN_ON(info == NULL || info->type != IRQT_VIRQ) in bind_virq_to_irq()
triggers for NULL info. The bind fails and execution continues with the
driver trying to clean up by unbinding. This eventually faults over the
NULL info.
---
drivers/xen/events/events_base.c | 17 ++++++++++++++++-
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/xen/events/events_base.c b/drivers/xen/events/events_base.c
index 41309d38f78c..a27e4d7f061e 100644
--- a/drivers/xen/events/events_base.c
+++ b/drivers/xen/events/events_base.c
@@ -159,7 +159,19 @@ static DEFINE_MUTEX(irq_mapping_update_lock);
static LIST_HEAD(xen_irq_list_head);
-/* IRQ <-> VIRQ mapping. */
+static bool is_per_vcpu_virq(int virq) {
+ switch (virq) {
+ case VIRQ_TIMER:
+ case VIRQ_DEBUG:
+ case VIRQ_XENOPROF:
+ case VIRQ_XENPMU:
+ return true;
+ default:
+ return false;
+ }
+}
+
+/* IRQ <-> VIRQ mapping. Global/Domain virqs are tracked in cpu 0. */
static DEFINE_PER_CPU(int [NR_VIRQS], virq_to_irq) = {[0 ... NR_VIRQS-1] =
-1};
/* IRQ <-> IPI mapping */
@@ -974,6 +986,9 @@ static void __unbind_from_irq(struct irq_info *info,
unsigned int irq)
switch (info->type) {
case IRQT_VIRQ:
+ if (!is_per_vcpu_virq(virq_from_irq(info)))
+ cpu = 0;
+
per_cpu(virq_to_irq, cpu)[virq_from_irq(info)] = -1;
break;
case IRQT_IPI:
Thinking about it a little more, bind_virq_to_irq() should ensure cpu ==
0 for per-domain and global VIRQs to ensure the property holds. Also
virq_to_irq accesses should go through wrappers to ensure all accesses
are handled consistently.
I'll send a v2.
Regards,
Jason
