mbox series

[RFC,0/3] Add device STM32L4x5 GPIO

Message ID 20240112170635.303226-1-ines.varhol@telecom-paris.fr (mailing list archive)
Headers show
Series Add device STM32L4x5 GPIO | expand

Message

Inès Varhol Jan. 12, 2024, 5:05 p.m. UTC
This patch adds a new device STM32L4x5 GPIO device and is part
of a series implementing the STM32L4x5 with a few peripherals.

This is RFC as the tests need to be corrected and completed.

The way the short-circuits are handled in the code currently :
(0) The model is simplified (it detects pins driven internally and
externally, not actual short-circuits)
(1) It reacts by ignoring external driving and writes
a `qemu_log_mask` about it
(2) The model is tested by using a fake register
`GPIO_DISCONNECTED_PINS` which is quite practical.
However the tests could disconnect pins and check if a pin is
disconnected in other ways (like setting all pins in push-pull
output to disconnect them), should I unmap this attribute?

Some context and more details for (0)

This code uses a simplified model. Instead of checking for
short-circuits each time the driving (internal and external) changes,
some configurations aren't allowed :
- push-pull with external driving
- open-drain with pin set high

Concretely, the pins configured as output can't be set externally
when in push-pull mode, or set high when in open-drain mode.
Conversely, when input/output mode or push-pull/open-drain mode
is changed, the problematic pins driven externally are "disconnected"
and the external value isn't considered anymore.

I saw sifive_gpio.c uses a similar model :
```
/* Pin both driven externally and internally */
if (output_en && in_mask) {
    qemu_log_mask(LOG_GUEST_ERROR, "GPIO pin %zu short circuited\n", i);
}
```

But nrf51_gpio.c actually checks for short-circuits :
```
if (connected_out && connected_in && out != in) {
    /* Pin both driven externally and internally */
    qemu_log_mask(LOG_GUEST_ERROR,
                    "GPIO pin %zu short circuited\n", i);
}
```

Based-on: 20240109194438.70934-1-ines.varhol@telecom-paris.fr
([PATCH v4 0/3] Add device STM32L4x5 SYSCFG)

Signed-off-by: Arnaud Minier <arnaud.minier@telecom-paris.fr>
Signed-off-by: Inès Varhol <ines.varhol@telecom-paris.fr>

Inès Varhol (3):
  hw/gpio: Implement STM32L4x5 GPIO
  hw/arm: Connect STM32L4x5 GPIO to STM32L4x5 SoC
  tests/qtest: Add STM32L4x5 GPIO QTest testcase

 docs/system/arm/b-l475e-iot01a.rst |   2 +-
 hw/arm/Kconfig                     |   3 +-
 hw/arm/stm32l4x5_soc.c             |  62 +++-
 hw/gpio/Kconfig                    |   3 +
 hw/gpio/meson.build                |   1 +
 hw/gpio/stm32l4x5_gpio.c           | 520 +++++++++++++++++++++++++++++
 hw/gpio/trace-events               |   6 +
 include/hw/arm/stm32l4x5_soc.h     |   9 +
 include/hw/gpio/stm32l4x5_gpio.h   |  79 +++++
 tests/qtest/meson.build            |   3 +-
 tests/qtest/stm32l4x5_gpio-test.c  | 319 ++++++++++++++++++
 11 files changed, 991 insertions(+), 16 deletions(-)
 create mode 100644 hw/gpio/stm32l4x5_gpio.c
 create mode 100644 include/hw/gpio/stm32l4x5_gpio.h
 create mode 100644 tests/qtest/stm32l4x5_gpio-test.c