Re: [PATCH v4 1/2] soc: qcom: smem: Register gunyah watchdog device

[Pavan Kondeti <pavan.kondeti@oss.qualcomm.com>]

On Mon, Nov 03, 2025 at 07:01:51PM -0600, Bjorn Andersson wrote:
> On Mon, Nov 03, 2025 at 04:03:44PM +0530, Hrishabh Rajput wrote:
> > 
> > On 11/2/2025 12:15 AM, Bjorn Andersson wrote:
> > > On Fri, Oct 31, 2025 at 10:18:13AM +0000, Hrishabh Rajput via B4 Relay wrote:
> > > > From: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> > > > 
> > > > To restrict gunyah watchdog initialization to Qualcomm platforms,
> > > > register the watchdog device in the SMEM driver.
> > > > 
> > > > When Gunyah is not present or Gunyah emulates MMIO-based
> > > > watchdog, we expect Qualcomm watchdog or ARM SBSA watchdog device to be
> > > > present in the devicetree. If none of these device nodes are detected,
> > > > we register the SMC-based Gunyah watchdog device.
> > > > 
> > > > Signed-off-by: Hrishabh Rajput <hrishabh.rajput@oss.qualcomm.com>
> > > > ---
> > > >   drivers/soc/qcom/smem.c | 37 +++++++++++++++++++++++++++++++++++++
> > > >   1 file changed, 37 insertions(+)
> > > > 
> > > > diff --git a/drivers/soc/qcom/smem.c b/drivers/soc/qcom/smem.c
> > > > index cf425930539e..40e4749fab02 100644
> > > > --- a/drivers/soc/qcom/smem.c
> > > > +++ b/drivers/soc/qcom/smem.c
> > > > @@ -1118,6 +1118,34 @@ static int qcom_smem_resolve_mem(struct qcom_smem *smem, const char *name,
> > > >   	return 0;
> > > >   }
> > > > +static int register_gunyah_wdt_device(void)
> > > > +{
> > > > +	struct platform_device *gunyah_wdt_dev;
> > > > +	struct device_node *np;
> > > > +
> > > > +	/*
> > > > +	 * When Gunyah is not present or Gunyah is emulating a memory-mapped
> > > > +	 * watchdog, either of Qualcomm watchdog or ARM SBSA watchdog will be
> > > > +	 * present. Skip initialization of SMC-based Gunyah watchdog if that is
> > > > +	 * the case.
> > > E.g. qcom-apq8064.dtsi doesn't define either qcom,kpss-wdt, nor
> > > arm,sbsa-gwdt, does that imply that it implements the Gunyah watchdog?
> > 
> > 
> > It doesn't implement Gunyah watchdog. For platforms like these we've kept a
> > STATUS SMC call in the gunyah_wdt_probe().
> > 
> 
> I think it would be good to make that call before registering the
> platform driver.

Did you mean platform device here? Becase we don't want the gunayh-wdt
module to do anything other than registering the platform driver on
platforms other than qcom.

> 
> > The SMC Call is expected to fail on platforms which do not have support for
> > SMC based Gunyah watchdog, which in turn will fail the probe.
> > 
> 
> Perhaps I'm missing something, just looked quickly and it's been a while
> since I looked at this code, but you're making a HVC (or SMC) call with
> the function:
> 
>   ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, ARM_SMCCC_SMC_32, ARM_SMCCC_OWNER_VENDOR_HYP, 6)
> 
> which doesn't look unique to Gunyah in my eyes.
> 
> If I read correctly, the ARM_SMCCC_SMC_32 is the only bit (literally)
> that differentiates this from being a __vgic_v3_get_gic_config() call in
> KVM, just as an example.
> 

Yes, you are right. This SMCC falls under ARM_SMCCC_OWNER_VENDOR_HYP
space, so it would behave differently on different hypervisors.

Please let me know what you think about this **defensive** approach.

- Move the Gunyah platform device registration to SCM driver.
- Checks to be done before registering the device
  - Make ARM_SMCCC_VENDOR_HYP_CALL_UID_FUNC_ID SMC [1] to verify that we are
    running under gunyah
  - check for the other wdt devices like kpss, sbsa in dT

ARM_SMCCC_VENDOR_HYP_CALL_UID_FUNC_ID SMC will not be supported by all
versions of Gunyah, but atleast it would confirm the definitive presence
of Gunyah.

[1]
https://lore.kernel.org/all/20240222-gunyah-v17-4-1e9da6763d38@quicinc.com/