Re: [PATCH] usb: typec: tcpci: add alert.fault handling
Heikki Krogerus <[email protected]> Wed, 5 Aug 2026 14:17:01 +0200
| Newsgroups | org.kernel.vger.linux-usb,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <[email protected]> |
On Fri, Jul 31, 2026 at 10:34:59AM +0200, Marco Felsch wrote: > Add minimal handling to clear the ALERT.Fault bit which may be set in > case of an over-current or over-voltage detection. At the moment only > the IRQ clear support is added since the TCPC Spec mentions that the > handling of such events is vendor specific. > > Signed-off-by: Marco Felsch <[email protected]> > --- > drivers/usb/typec/tcpm/tcpci.c | 16 +++++++++++++++- > 1 file changed, 15 insertions(+), 1 deletion(-) > > diff --git a/drivers/usb/typec/tcpm/tcpci.c b/drivers/usb/typec/tcpm/tcpci.c > index a56e31b20c21..d2d5bf7ca618 100644 > --- a/drivers/usb/typec/tcpm/tcpci.c > +++ b/drivers/usb/typec/tcpm/tcpci.c > @@ -692,7 +692,8 @@ static int tcpci_init(struct tcpc_dev *tcpc) > > reg = TCPC_ALERT_TX_SUCCESS | TCPC_ALERT_TX_FAILED | > TCPC_ALERT_TX_DISCARDED | TCPC_ALERT_RX_STATUS | > - TCPC_ALERT_RX_HARD_RST | TCPC_ALERT_CC_STATUS; > + TCPC_ALERT_RX_HARD_RST | TCPC_ALERT_CC_STATUS | > + TCPC_ALERT_FAULT; > if (tcpci->controls_vbus) > reg |= TCPC_ALERT_POWER_STATUS; > /* Enable VSAFE0V status interrupt when detecting VSAFE0V is supported */ > @@ -720,6 +721,19 @@ irqreturn_t tcpci_irq(struct tcpci *tcpci) > irq_ret = status & tcpci->alert_mask; > > process_status: > + /* > + * Handle nested fault before clearing the alert status to not > + * re-trigger a fault alert immediately again. > + */ Shouldn't that explanation in the comment be also in the commit message? That sounds like the "why" part to me. > + if (status & TCPC_ALERT_FAULT) { > + unsigned int fault; > + > + /* TODO: Add a proper over-current and over-voltage handling */ > + ret = regmap_read(tcpci->regmap, TCPC_FAULT_STATUS, &fault); > + if (!ret) > + regmap_write(tcpci->regmap, TCPC_FAULT_STATUS, fault); > + } > + > /* > * Clear alert status for everything except RX_STATUS, which shouldn't > * be cleared until we have successfully retrieved message. > -- > 2.47.3 -- heikki