On Fri, Jul 17, 2026 at 03:39:55PM +0530, Neha Malcom Francis wrote: > On Thu, 16 Jul 2026 14:15:51 +0530, Varadarajan Narayanan > <[email protected]> wrote: > > diff --git a/drivers/misc/qcom_geni.c b/drivers/misc/qcom_geni.c > > index a62ae6a2478..f3133f85885 100644 > > --- a/drivers/misc/qcom_geni.c > > +++ b/drivers/misc/qcom_geni.c > > @@ -414,6 +468,9 @@ static int geni_se_of_to_plat(struct udevice *dev) > > > > if (proto == GENI_SE_INVALID_PROTO) > > plat->need_firmware_load = true; > > + > > + if (readl(res.start + SE_HW_PARAM_2) & GENI_USE_MINICORES) > > + plat->is_mini_core = true; > > Add IS_ENABLED(CONFIG_QCOM_GENI_MINICORE) to prevent accidental NULL > dereferencing in case CONFIG not set.
Not sure if I'm missing something. If plat was NULL, it would have crashed while doing 'plat->need_firmware_load = false' towards the beginning of the function itself. > > diff --git a/include/soc/qcom/minicore.h b/include/soc/qcom/minicore.h > > new file mode 100644 > > index 00000000000..9f2071ed86b > > --- /dev/null > > +++ b/include/soc/qcom/minicore.h > > @@ -0,0 +1,25 @@ > > [ ... skip 16 lines ... ] > > +}; > > + > > +#if IS_ENABLED(CONFIG_QCOM_GENI_MINICORE) > > +extern struct qup_mini_core_info qup_mini_cores[]; > > +#else > > +struct qup_mini_core_info *qup_mini_cores; > > else #define qup_mini_cores NULL Ok. > > diff --git a/include/soc/qcom/qup-fw-load.h b/include/soc/qcom/qup-fw-load.h > > index a67a93c72a4..df8a615b46b 100644 > > --- a/include/soc/qcom/qup-fw-load.h > > +++ b/include/soc/qcom/qup-fw-load.h > > @@ -14,6 +14,7 @@ > > #define GENI_INIT_CFG_REVISION 0x0 > > #define GENI_S_INIT_CFG_REVISION 0x4 > > #define GENI_FORCE_DEFAULT_REG 0x20 > > +#define GENI_OUTPUT_CTRL 0x24 > > Redefinition from geni-se.h ? Will remove. Thanks Varada
