Hi,
> De: Philippe <phi...@linaro.org>
> Envoyé: mercredi 7 février 2024 23:02 CET
>
> Hi Inès,
>
> (this is now commit 52671f69f7).
>
> On 9/1/24 17:06, Inès Varhol wrote:
> > Tested-by: Philippe Mathieu-Daudé <phi...@linaro.org>
> > Reviewed-by: Philippe Mathieu-Daudé <phi...@linaro.org>
> > Reviewed-by: Alistair Francis <alistair.fran...@wdc.com>
> > Signed-off-by: Arnaud Minier <arnaud.min...@telecom-paris.fr>
> > Signed-off-by: Inès Varhol <ines.var...@telecom-paris.fr>
> > ---
> > hw/arm/Kconfig | 1 +
> > hw/arm/stm32l4x5_soc.c | 52 +++++++++++++++++++++++++++++++++-
> > include/hw/arm/stm32l4x5_soc.h | 3 ++
> > 3 files changed, 55 insertions(+), 1 deletion(-)
>
>
> > +#define NUM_EXTI_IRQ 40
> > +/* Match exti line connections with their CPU IRQ number */
> > +/* See Vector Table (Reference Manual p.396) */
> > +static const int exti_irq[NUM_EXTI_IRQ] = {
> > + 6, /* GPIO[0] */
> > + 7, /* GPIO[1] */
> > + 8, /* GPIO[2] */
> > + 9, /* GPIO[3] */
> > + 10, /* GPIO[4] */
> > + 23, 23, 23, 23, 23, /* GPIO[5..9] */
> > + 40, 40, 40, 40, 40, 40, /* GPIO[10..15] */
>
> I'm sorry because I missed that earlier, and I'm surprised
> you aren't chasing weird bugs. Due to how QEMU IRQs are
> implemented, we can not wire multiple input lines to the same
> output without using an intermediate "OR gate". We model it
> as TYPE_OR_IRQ. See the comment in "hw/qdev-core.h" added in
> commit cd07d7f9f5 ("qdev: Document GPIO related functions"):
Better fixing it now than later :)
I must admit I didn't pay attention to the particularity of EXTI5 to 15.
Current exti tests don't even use these lines, a testcase will have
to be added. Otherwise we mostly ran executables using GPIOs as output,
so no weird bugs encountered.
Thank you for noticing !
Ines
>
> * It is not valid to try to connect one outbound GPIO to multiple
> * qemu_irqs at once, or to connect multiple outbound GPIOs to the
> * same qemu_irq. (Warning: there is no assertion or other guard to
> * catch this error: the model will just not do the right thing.)
> * Instead, for fan-out you can use the TYPE_SPLIT_IRQ device: connect
> * a device's outbound GPIO to the splitter's input, and connect each
> * of the splitter's outputs to a different device. For fan-in you
> * can use the TYPE_OR_IRQ device, which is a model of a logical OR
> * gate with multiple inputs and one output.
>
> So for example for the GPIO[10..15] you need to create a 6-line
> OR gate as (totally untested):
>
> /* 6-line OR IRQ gate */
> Object *orgate40 = object_new(TYPE_OR_IRQ);
> object_property_set_int(orgate40, "num-lines", 6, &error_fatal);
> qdev_realize(DEVICE(orgate), NULL, &error_fatal);
>
> /* OR gate -> IRQ #40 */
> qdev_connect_gpio_out(DEVICE(orgate40), 0,
> qdev_get_gpio_in(armv7m, 40));
>
> /* EXTI GPIO[10..15] -> OR gate */
> for (unsigned i = 0; i < 6; i++) {
> sysbus_connect_irq(SYS_BUS_DEVICE(&s->exti), 10 + i,
> qdev_get_gpio_in(DEVICE(orgate40), i));
> }
>
> > + 1, /* PVD */
> > + 67, /* OTG_FS_WKUP, Direct */
> > + 41, /* RTC_ALARM */
> > + 2, /* RTC_TAMP_STAMP2/CSS_LSE */
> > + 3, /* RTC wakeup timer */
> > + 63, /* COMP1 */
> > + 63, /* COMP2 */
> > + 31, /* I2C1 wakeup, Direct */
> > + 33, /* I2C2 wakeup, Direct */
> > + 72, /* I2C3 wakeup, Direct */
> > + 37, /* USART1 wakeup, Direct */
> > + 38, /* USART2 wakeup, Direct */
> > + 39, /* USART3 wakeup, Direct */
> > + 52, /* UART4 wakeup, Direct */
> > + 53, /* UART4 wakeup, Direct */
> > + 70, /* LPUART1 wakeup, Direct */
> > + 65, /* LPTIM1, Direct */
> > + 66, /* LPTIM2, Direct */
> > + 76, /* SWPMI1 wakeup, Direct */
> > + 1, /* PVM1 wakeup */
> > + 1, /* PVM2 wakeup */
> > + 1, /* PVM3 wakeup */
> > + 1, /* PVM4 wakeup */
> > + 78 /* LCD wakeup, Direct */
> > +};
>
> > + busdev = SYS_BUS_DEVICE(&s->exti);
> > + if (!sysbus_realize(busdev, errp)) {
> > + return;
> > + }
> > + sysbus_mmio_map(busdev, 0, EXTI_ADDR);
> > + for (unsigned i = 0; i < NUM_EXTI_IRQ; i++) {
> > + sysbus_connect_irq(busdev, i, qdev_get_gpio_in(armv7m, exti_irq[i]));
>
> ^^^^^^^^^^
> > + }
> Regards,
>
> Phil.
