On Mon, 22 Nov 2021 13:10:32 +0000,
Geert Uytterhoeven [off-list ref] wrote:

Hi Marc,

On Mon, Nov 22, 2021 at 11:30 AM Marc Zyngier [off-list ref] wrote:

quoted

Since 041284181226 ("of/irq: Allow matching of an interrupt-map local
to an interrupt controller"), a handful of interrupt controllers have
stopped working correctly. This is due to the DT exposing a non-sensical
interrupt-map property, and their drivers relying on the kernel ignoring
this property.

Since we cannot realistically fix this terrible behaviour, add a quirk
for the limited set of devices that have implemented this monster,
and document that this is a pretty bad practice.

Cc: Rob Herring <robh@kernel.org>
Cc: John Crispin <john@phrozen.org>
Cc: Biwen Li <redacted>
Cc: Chris Brandt <chris.brandt@renesas.com>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Marc Zyngier <maz@kernel.org>

Thanks for your patch!

quoted

--- a/drivers/of/irq.c
+++ b/drivers/of/irq.c

@@ -76,6 +76,36 @@ struct device_node *of_irq_find_parent(struct device_node *child)
 }
 EXPORT_SYMBOL_GPL(of_irq_find_parent);

+/*
+ * These interrupt controllers abuse interrupt-map for unspeakable
+ * reasons and rely on the core code to *ignore* it (the drivers do
+ * their own parsing of the property).
+ *
+ * If you think of adding to the list for something *new*, think
+ * again. There is a high chance that you will be sent back to the
+ * drawing board.
+ */
+static const char * const of_irq_imap_abusers[] = {
+       "CBEA,platform-spider-pic",
+       "sti,platform-spider-pic",
+       "realtek,rtl-intc",
+       "fsl,ls1021a-extirq",
+       "fsl,ls1043a-extirq",
+       "fsl,ls1088a-extirq",
+       "renesas,rza1-irqc",
+};

Are you sure "renesas,rza1-irqc" handles this wrong? How should it
be handled instead? I read the other thread[1], but didn't became
any wiser: interrupts are mapped one-to-one with the RZ/A1 IRQC.

In both v5.15 and v5.16-rc1, interrupts seem to work fine on RSK+RZA1
and RZA2MEVB, both with gpio-keys and when used as a wake-up interrupt.

This is odd. 5.16-rc1 should actively breaks the behaviour, as each
interrupt is directly routed to the GIC. Here's an extract of the DT
for r7s9210:

    interrupt-map = <0 0 &gic GIC_SPI 4 IRQ_TYPE_LEVEL_HIGH>,
                    <1 0 &gic GIC_SPI 5 IRQ_TYPE_LEVEL_HIGH>,
                    <2 0 &gic GIC_SPI 6 IRQ_TYPE_LEVEL_HIGH>,
                    <3 0 &gic GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>,
                    <4 0 &gic GIC_SPI 8 IRQ_TYPE_LEVEL_HIGH>,
                    <5 0 &gic GIC_SPI 9 IRQ_TYPE_LEVEL_HIGH>,
                    <6 0 &gic GIC_SPI 10 IRQ_TYPE_LEVEL_HIGH>,
                    <7 0 &gic GIC_SPI 11 IRQ_TYPE_LEVEL_HIGH>;

I expect v5.16-rc1 to honour the routing described here and not
involve rza1-irqc, because that's what the DT says.

With this patch applied, I see double keypresses with evtest: when
pressing a key, I get a key-down event, immediately followed by a
key-up event. When releasing the key, I again get two events.

Good (v5.15 or v5.16-rc1):

    Event: time 1637585631.288990, type 1 (EV_KEY), code 2 (KEY_1), value 1
    Event: time 1637585631.288990, -------------- SYN_REPORT ------------
    Event: time 1637585631.499924, type 1 (EV_KEY), code 2 (KEY_1), value 0
    Event: time 1637585631.499924, -------------- SYN_REPORT ------------

Bad (v5.16-rc1 + this patch):

    Event: time 1637585341.946647, type 1 (EV_KEY), code 2 (KEY_1), value 1
    Event: time 1637585341.946647, -------------- SYN_REPORT ------------
    Event: time 1637585341.960256, type 1 (EV_KEY), code 2 (KEY_1), value 0
    Event: time 1637585341.960256, -------------- SYN_REPORT ------------
    Event: time 1637585342.146775, type 1 (EV_KEY), code 2 (KEY_1), value 1
    Event: time 1637585342.146775, -------------- SYN_REPORT ------------
    Event: time 1637585342.160092, type 1 (EV_KEY), code 2 (KEY_1), value 0
    Event: time 1637585342.160092, -------------- SYN_REPORT ------------

Is there any chance you could trace whether rza1-irqc gets called at
all when setting up and handling the interrupt?

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.