irqchip/gic-v4: Correctly deal with set_affinity on lazily-mapped VPEs

Zenghui points out that a recent change to the way set_affinity is
handled for VPEs has the potential to return an error if the VPE
hasn't been mapped yet (because the guest hasn't emited a MAPTI
command yet), affecting GICv4.0 implementations that rely on the
ITSList feature.

Fix this by making the set_affinity succeed in this case, and
return early, without trying to touch the HW.

Fixes: 1442ee0011983 ("irqchip/gic-v4: Don't allow a VMOVP on a dying VPE")
Reported-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/all/20241027102220.1858558-1-maz@kernel.org
Link: https://lore.kernel.org/r/aab45cd3-e5ca-58cf-e081-e32a17f5b4e7@huawei.com

diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
index ab597e7..52f625e 100644 (file)
--- a/drivers/irqchip/irq-gic-v3-its.c
+++ b/drivers/irqchip/irq-gic-v3-its.c
@@ -3810,8 +3810,18 @@ static int its_vpe_set_affinity(struct irq_data *d,
         * Check if we're racing against a VPE being destroyed, for
         * which we don't want to allow a VMOVP.
         */
-       if (!atomic_read(&vpe->vmapp_count))
-               return -EINVAL;
+       if (!atomic_read(&vpe->vmapp_count)) {
+               if (gic_requires_eager_mapping())
+                       return -EINVAL;
+
+               /*
+                * If we lazily map the VPEs, this isn't an error and
+                * we can exit cleanly.
+                */
+               cpu = cpumask_first(mask_val);
+               irq_data_update_effective_affinity(d, cpumask_of(cpu));
+               return IRQ_SET_MASK_OK_DONE;
+       }
 
        /*
         * Changing affinity is mega expensive, so let's be as lazy as