[Xen-devel] [PATCH] x86/vtd: Fix S3 resume following c/s 650c31d3af

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

c/s 650c31d3af "x86/IRQ: fix locking around vector management" adjusted the
locking in adjust_irq_affinity().

The S3 path ends up here via iommu_resume() before interrupts are enabled, at
which point spin_lock_irq() which fail ASSERT(local_irq_is_enabled()); but
with no working console.

Use spin_lock_irqsave() instead to cope with interrupts already being
disabled.

Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
---
CC: Jan Beulich <JBeulich@xxxxxxxx>
CC: Wei Liu <wl@xxxxxxx>
CC: Roger Pau Monné <roger.pau@xxxxxxxxxx>
CC: Jun Nakajima <jun.nakajima@xxxxxxxxx>
CC: Kevin Tian <kevin.tian@xxxxxxxxx>

I'm fairly confident that the AMD side of things is fine, because
enable_iommu() is encompased by a spin_lock_irqsave() block.
---
 xen/drivers/passthrough/vtd/iommu.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/xen/drivers/passthrough/vtd/iommu.c 
b/xen/drivers/passthrough/vtd/iommu.c
index 5d72270c5b..defa74fae3 100644
--- a/xen/drivers/passthrough/vtd/iommu.c
+++ b/xen/drivers/passthrough/vtd/iommu.c
@@ -2135,15 +2135,16 @@ static void adjust_irq_affinity(struct acpi_drhd_unit 
*drhd)
                              : NUMA_NO_NODE;
     const cpumask_t *cpumask = NULL;
     struct irq_desc *desc;
+    unsigned long flags;
 
     if ( node < MAX_NUMNODES && node_online(node) &&
          cpumask_intersects(&node_to_cpumask(node), &cpu_online_map) )
         cpumask = &node_to_cpumask(node);
 
     desc = irq_to_desc(drhd->iommu->msi.irq);
-    spin_lock_irq(&desc->lock);
+    spin_lock_irqsave(&desc->lock, flags);
     dma_msi_set_affinity(desc, cpumask);
-    spin_unlock_irq(&desc->lock);
+    spin_unlock_irqrestore(&desc->lock, flags);
 }
 
 static int adjust_vtd_irq_affinities(void)
-- 
2.11.0


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel