Re: [PATCH RFC 5/6] watchdog: qcom-wdt: add support to read the restart reason from IMEM

[Krzysztof Kozlowski <krzk@kernel.org>]

On 08/04/2025 10:49, Kathiravan Thirumoorthy wrote:
> When the system boots up after a watchdog reset, the EXPIRED_STATUS bit
> in the WDT_STS register is cleared. To identify if the system was restarted
> due to WDT expiry, bootloaders update the information in the IMEM region.
> Update the driver to read the restart reason from IMEM and populate the
> bootstatus accordingly.
> 
> For backward compatibility, keep the EXPIRED_STATUS bit check. Add a new
> function qcom_wdt_get_restart_reason() to read the restart reason from
> IMEM.
> 
> Signed-off-by: Kathiravan Thirumoorthy <kathiravan.thirumoorthy@oss.qualcomm.com>
> ---
>  drivers/watchdog/qcom-wdt.c | 40 +++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 39 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/watchdog/qcom-wdt.c b/drivers/watchdog/qcom-wdt.c
> index 006f9c61aa64fd2b4ee9db493aeb54c8fafac818..54d6eaa132ab9f63e1312a69ad51b7a14f78fe2d 100644
> --- a/drivers/watchdog/qcom-wdt.c
> +++ b/drivers/watchdog/qcom-wdt.c
> @@ -9,6 +9,7 @@
>  #include <linux/kernel.h>
>  #include <linux/module.h>
>  #include <linux/of.h>
> +#include <linux/of_address.h>
>  #include <linux/platform_device.h>
>  #include <linux/watchdog.h>
>  
> @@ -22,6 +23,8 @@ enum wdt_reg {
>  
>  #define QCOM_WDT_ENABLE		BIT(0)
>  
> +#define NON_SECURE_WDT_RESET	0x5
> +
>  static const u32 reg_offset_data_apcs_tmr[] = {
>  	[WDT_RST] = 0x38,
>  	[WDT_EN] = 0x40,
> @@ -187,6 +190,39 @@ static const struct qcom_wdt_match_data match_data_kpss = {
>  	.max_tick_count = 0xFFFFFU,
>  };
>  
> +static int  qcom_wdt_get_restart_reason(struct qcom_wdt *wdt)
> +{
> +	struct device_node *np;
> +	struct resource imem;
> +	void __iomem *base;
> +	int ret;
> +
> +	np = of_find_compatible_node(NULL, NULL, "qcom,restart-reason-info");
> +	if (!np)

That's not how you express dependencies between devices.

> +		return -ENOENT;
> +
> +	ret = of_address_to_resource(np, 0, &imem);
> +	of_node_put(np);
> +	if (ret < 0) {
> +		dev_err(wdt->wdd.parent, "can't translate OF node address\n");
> +		return ret;
> +	}
> +
> +	base = ioremap(imem.start, resource_size(&imem));
> +	if (!base) {
> +		dev_err(wdt->wdd.parent, "failed to map restart reason info region\n");
> +		return -ENOMEM;
> +	}
> +
> +	memcpy_fromio(&ret, base, sizeof(ret));
> +	iounmap(base);

All this is wrong usage of syscon API, missing devlinks, messing up with
other device's address space.


Best regards,
Krzysztof