Re: [PATCH v1] irqchip/irq-mvebu-icu: Fix irq_set_type for sei and nsr

On Mon, Jan 20, 2025 at 09:45:43AM +0100, Thomas Gleixner wrote:

quoted

On Fri, Jan 17 2025 at 19:03, Stefan Eichenberger wrote:

quoted

It seems that irq_chip_set_type_parent is not failing. By adding some
debug messages to kernel/irq/manage.c I found that irqd_get_trigger_type
returns IRQF_TRIGGER_RISING even though it should return
IRQF_TRIGGER_HIGH according to the device tree. Maybe this was fixed
between 6.10 and 6.12 but I need to analyze that again in more detail.

The trigger type is fixed up in mvebu_icu_irq_domain_translate(), but
that should be the same with the new code in mvebu_icu_translate().

Can you instrument mvebu_icu_translate() and validate that it is

    1) Invoked at all

    2) The type fixup is done:

       if (msi_data->subset_data->icu_group == ICU_GRP_SEI)
              *type = IRQ_TYPE_EDGE_RISING;

Thanks a lot for pointing me in the right direction. I think I found
what is going on. My instrumentation showed that we still end up in that
function but the translation does not work. I think the reason is that
the msi_data conversion is wrong. The following patch would solve the
issue:

diff --git a/drivers/irqchip/irq-mvebu-icu.c b/drivers/irqchip/irq-mvebu-icu.c
index b337f6c05f18..4eebed39880a 100644
--- a/drivers/irqchip/irq-mvebu-icu.c
+++ b/drivers/irqchip/irq-mvebu-icu.c

@@ -68,7 +68,8 @@ static int mvebu_icu_translate(struct irq_domain *d, struct irq_fwspec *fwspec,
                               unsigned long *hwirq, unsigned int *type)
 {
        unsigned int param_count = static_branch_unlikely(&legacy_bindings) ? 3 : 2;
-       struct mvebu_icu_msi_data *msi_data = d->host_data;
+       struct msi_domain_info *info = d->host_data;
+       struct mvebu_icu_msi_data *msi_data = info->chip_data;
        struct mvebu_icu *icu = msi_data->icu;

        /* Check the count of the parameters in dt */

Do you agree with that fix? If so I will prepare a proper patch.