Re: [PATCH v6 14/16] wifi: ath12k: Switch to generic PAS TZ APIs

[Sumit Garg <sumit.garg@kernel.org>]

On Tue, May 19, 2026 at 07:34:08AM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] The transition from `QCOM_SCM` to the generic `QCOM_PAS` framework introduces an asynchronous probing bug where `ath12k` fails to check if the PAS backend is available, causing permanent probe failures.
> --
> 
> commit 9d619dc2a0265a4c4b222e1e790293c324c3c591
> Author: Sumit Garg <sumit.garg@oss.qualcomm.com>
> 
> wifi: ath12k: Switch to generic PAS TZ APIs
> 
> This commit switches the ath12k client driver over to the generic PAS TZ
> APIs. This abstracts the TZ implementation, allowing the driver to support
> multiple backends such as the QTEE based SCM PAS service or OP-TEE based
> PAS service.
> 
> > diff --git a/drivers/net/wireless/ath/ath12k/ahb.c b/drivers/net/wireless/ath/ath12k/ahb.c
> > --- a/drivers/net/wireless/ath/ath12k/ahb.c
> > +++ b/drivers/net/wireless/ath/ath12k/ahb.c
> 
> [ ... ]
> 
> > @@ -420,7 +420,7 @@ static int ath12k_ahb_power_up(struct ath12k_base *ab)
> >  
> >  	if (ab_ahb->scm_auth_enabled) {
> >  		/* Authenticate FW image using peripheral ID */
> > -		ret = qcom_scm_pas_auth_and_reset(pasid);
> > +		ret = qcom_pas_auth_and_reset(pasid);
> >  		if (ret) {
> >  			ath12k_err(ab, "failed to boot the remote processor %d\n", ret);
> >  			goto err_fw2;
> 
> Does this change introduce an asynchronous probing failure?
> 
> Since the generic QCOM_PAS layer allows its backends (like QCOM_SCM or
> QCOM_PAS_TEE) to be built as loadable modules, the ath12k_ahb driver might
> probe before the backend has registered its operations.
> 
> If that happens, qcom_pas_auth_and_reset() returns -ENODEV:
> 
> drivers/firmware/qcom/qcom_pas.c:qcom_pas_auth_and_reset() {
> 	if (!ops_ptr)
> 		return -ENODEV;
> 
> 	return ops_ptr->auth_and_reset(ops_ptr->dev, pas_id);
> }
> 
> Because this returns -ENODEV instead of -EPROBE_DEFER, the ath12k_ahb driver
> will treat this as a fatal hardware initialization error. The kernel driver
> core will not retry the probe, resulting in a permanent failure.
> 
> Should the driver be updated to check if the PAS backend is available and
> explicitly return -EPROBE_DEFER if it is not yet registered?

QCOM_SCM was already allowed to be built as a module, so not in scope of
this patch-set.

-Sumit

> 
> -- 
> Sashiko AI review · https://sashiko.dev/#/patchset/20260518072856.22790-1-sumit.garg@kernel.org?part=14