On Fri, Mar 24, 2023 at 03:05:53PM +0100, neil.armstrong@xxxxxxxxxx wrote: > On 24/03/2023 15:01, Dmitry Rokosov wrote: > > In some meson boards, secure monitor device has children, for example, > > power secure controller. By default, secure monitor isn't the bus in terms > > of device tree subsystem, so the of_platform initialization code doesn't > > populate its device tree data. As a result, secure monitor's children > > aren't probed at all. > > > > Run the 'of_platform_populate()' routine manually to resolve such issues. > > > > Signed-off-by: Dmitry Rokosov <ddrokosov@xxxxxxxxxxxxxx> > > --- > > drivers/firmware/meson/meson_sm.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/firmware/meson/meson_sm.c b/drivers/firmware/meson/meson_sm.c > > index 77aa5c6398aa..b79d0e316cb1 100644 > > --- a/drivers/firmware/meson/meson_sm.c > > +++ b/drivers/firmware/meson/meson_sm.c > > @@ -316,7 +316,7 @@ static int __init meson_sm_probe(struct platform_device *pdev) > > if (sysfs_create_group(&pdev->dev.kobj, &meson_sm_sysfs_attr_group)) > > goto out_in_base; > > - return 0; > > + return devm_of_platform_populate(dev); > > You should check return and jump to out_in_base on error instead. > > Neil > Ah, exactly. There is one direct non-devm ioremap above. I hurried up, sorry. > > out_in_base: > > iounmap(fw->sm_shmem_in_base); > -- Thank you, Dmitry
