Re: [PATCH] usb: typec: hd3ss3220: Propagate regmap_read() errors from attached-state check
From: Ian Rogers
Date: Fri Nov 20 2026 - 12:28:04 EST
On Fri, Nov 20, 2026 at 8:05 AM Greg Kroah-Hartman
<gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Thu, Nov 19, 2026 at 09:06:49AM -0800, Ian Rogers wrote:
> > hd3ss3220_get_attached_state() has return type 'enum usb_role' (an
> > unsigned enum) yet returns negative regmap_read() error codes directly.
> > Neither hd3ss3220_set_role() nor output_poll_execute() checks for errors;
> > on a transient I2C failure (such as -EIO, cast to 0xfffffffb),
> > hd3ss3220_set_role() passes the bogus value to usb_role_switch_set_role()
> > and disables VBUS power. On role-switch providers without a .get callback
> > (where sw->role caches the value passed to usb_role_switch_set_role()),
> > reading /sys/class/usb_role/*/role subsequently indexes usb_roles[role]
> > out-of-bounds.
> >
> > Return an int status from hd3ss3220_get_attached_state(), pass the
> > resolved enum usb_role via an output pointer, and leave the current role
> > and VBUS state unchanged on I2C read errors.
> >
> > Found via an x86_64 LLVM=1 (Clang 21.1.8) allyesconfig audit with
> > -Wimplicit-int-enum-cast enabled; fix and changelog drafted with an AI
> > coding assistant. Build-tested under LLVM=1 W=1 (-Werror) and
> > -Wimplicit-int-enum-cast; not tested on TI HD3SS3220 hardware.
> >
> > Fixes: 1c48c759ef4b ("usb: typec: driver for TI HD3SS3220 USB Type-C DRP port controller")
> > Signed-off-by: Ian Rogers <irogers@xxxxxxxxxx>
> > ---
> > drivers/usb/typec/hd3ss3220.c | 22 +++++++++++++---------
> > 1 file changed, 13 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/usb/typec/hd3ss3220.c b/drivers/usb/typec/hd3ss3220.c
> > index 60d6de9ba339..212c3624961c 100644
> > --- a/drivers/usb/typec/hd3ss3220.c
> > +++ b/drivers/usb/typec/hd3ss3220.c
> > @@ -163,10 +163,10 @@ static int hd3ss3220_set_source_pref(struct hd3ss3220 *hd3ss3220, int prefer_rol
> > src_pref);
> > }
> >
> > -static enum usb_role hd3ss3220_get_attached_state(struct hd3ss3220 *hd3ss3220)
> > +static int hd3ss3220_get_attached_state(struct hd3ss3220 *hd3ss3220,
> > + enum usb_role *attached_state)
> > {
> > unsigned int reg_val;
> > - enum usb_role attached_state;
> > int ret;
> >
> > ret = regmap_read(hd3ss3220->regmap, HD3SS3220_REG_CN_STAT_CTRL,
> > @@ -176,17 +176,17 @@ static enum usb_role hd3ss3220_get_attached_state(struct hd3ss3220 *hd3ss3220)
> >
> > switch (reg_val & HD3SS3220_REG_CN_STAT_CTRL_ATTACHED_STATE_MASK) {
> > case HD3SS3220_REG_CN_STAT_CTRL_AS_DFP:
> > - attached_state = USB_ROLE_HOST;
> > + *attached_state = USB_ROLE_HOST;
> > break;
> > case HD3SS3220_REG_CN_STAT_CTRL_AS_UFP:
> > - attached_state = USB_ROLE_DEVICE;
> > + *attached_state = USB_ROLE_DEVICE;
> > break;
> > default:
> > - attached_state = USB_ROLE_NONE;
> > + *attached_state = USB_ROLE_NONE;
> > break;
> > }
> >
> > - return attached_state;
> > + return 0;
> > }
> >
> > static int hd3ss3220_try_role(struct typec_port *port, int role)
> > @@ -231,7 +231,10 @@ static void hd3ss3220_regulator_control(struct hd3ss3220 *hd3ss3220, bool on)
> >
> > static void hd3ss3220_set_role(struct hd3ss3220 *hd3ss3220)
> > {
> > - enum usb_role role_state = hd3ss3220_get_attached_state(hd3ss3220);
> > + enum usb_role role_state;
> > +
> > + if (hd3ss3220_get_attached_state(hd3ss3220, &role_state) < 0)
> > + return;
>
> No error propagation backward?
Good point - hd3ss3220_probe() can fail probe and hd3ss3220_irq() can
return IRQ_NONE without clearing INT_STATUS if hd3ss3220_set_role()
returns an int error (the third caller, output_poll_execute(), is a
void delayed_work callback). I'll propagate the error out of
hd3ss3220_set_role() in v2.
> >
> > usb_role_switch_set_role(hd3ss3220->role_sw, role_state);
> >
> > @@ -258,9 +261,10 @@ static void output_poll_execute(struct work_struct *work)
> > struct hd3ss3220 *hd3ss3220 = container_of(delayed_work,
> > struct hd3ss3220,
> > output_poll_work);
> > - enum usb_role role_state = hd3ss3220_get_attached_state(hd3ss3220);
> > + enum usb_role role_state;
> >
> > - if (hd3ss3220->role_state != role_state)
> > + if (hd3ss3220_get_attached_state(hd3ss3220, &role_state) == 0 &&
>
> Why test for 0?
I'll change this to '!hd3ss3220_get_attached_state(...)' in v2 (and
add Cc: stable@xxxxxxxxxxxxxxx per Heikki's note).
Thanks,
Ian
> thanks,
>
> greg k-h