Prior to introducing GICv3 eSPI support, writes to these registers were
ignored. Now the behavior is changed and we inject fault to the guests.
According to documentation, eSPI registers are treated as RES0, when
GICD_TYPER.ESPI is 0 (this would be the case when CONFIG_GICV3_ESPI=n or
nr_espis is 0). Restore the previous behavior by checking whether eSPI
is in use or not.

Fixes: 93eecb2c4b88 ("xen/arm: vgic-v3: add emulation of GICv3.1 eSPI registers")
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
---
Changes in v2:
 - only ignore when eSPI is not in use
---
 xen/arch/arm/vgic-v3.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/xen/arch/arm/vgic-v3.c b/xen/arch/arm/vgic-v3.c
index 8b1c8eef8024..77aab5c3c293 100644
--- a/xen/arch/arm/vgic-v3.c
+++ b/xen/arch/arm/vgic-v3.c
@@ -884,9 +884,18 @@ static int __vgic_v3_distr_common_mmio_write(const char *name, struct vcpu *v,
                    "%pv: %s: unhandled word write %#"PRIregister" to ISACTIVER%d\n",
                    v, name, r, reg - GICD_ISACTIVER);
         else
+        {
+#ifdef CONFIG_GICV3_ESPI
+            if ( !v->domain->arch.vgic.nr_espis )
+                goto write_ignore;
+
             printk(XENLOG_G_ERR
                    "%pv: %s: unhandled word write %#"PRIregister" to ISACTIVER%dE\n",
                    v, name, r, reg - GICD_ISACTIVERnE);
+#else
+            goto write_ignore;
+#endif
+        }
         return 0;
 
     case VRANGE32(GICD_ICACTIVER, GICD_ICACTIVERN):
-- 
2.43.0

Hi,

On 28/11/2025 11:49, Michal Orzel wrote:
> Prior to introducing GICv3 eSPI support, writes to these registers were
> ignored. Now the behavior is changed and we inject fault to the guests.
> According to documentation, eSPI registers are treated as RES0, when
> GICD_TYPER.ESPI is 0 (this would be the case when CONFIG_GICV3_ESPI=n or
> nr_espis is 0). Restore the previous behavior by checking whether eSPI
> is in use or not.
> 
> Fixes: 93eecb2c4b88 ("xen/arm: vgic-v3: add emulation of GICv3.1 eSPI registers")
> Signed-off-by: Michal Orzel <michal.orzel@amd.com>

Reviewed-by: Julien Grall <jgrall@amazon.com>

Cheers,

-- 
Julien Grall