* [PATCH 0/2] usb: typec: qcom-pmic-typec: properly tear down port on remove and shutdown
@ 2026-11-20 9:59 Neil Armstrong
2026-11-20 9:59 ` [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop Neil Armstrong
2026-11-20 9:59 ` [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown Neil Armstrong
0 siblings, 2 replies; 8+ messages in thread
From: Neil Armstrong @ 2026-11-20 9:59 UTC (permalink / raw)
To: Bryan O'Donoghue, Heikki Krogerus, Greg Kroah-Hartman
Cc: linux-arm-msm, linux-usb, Neil Armstrong
The current driver doesn't leave the port in the same
state as found when probing, and leaves the port enabled
on system shutdown.
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
---
Neil Armstrong (2):
usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop
usb: typec: qcom-pmic-typec: detach the port on shutdown
drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c | 11 +++++++++++
drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h | 1 +
drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c | 2 +-
drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c | 11 +++++++++++
4 files changed, 24 insertions(+), 1 deletion(-)
---
base-commit: adcad7f4e6332cff3d6d7165f41584899f598534
change-id: 20261120-topic-tcpm-qcom-fixes-c9ce1e65c570
Best regards,
--
Neil Armstrong <neil.armstrong@linaro.org>
^ permalink raw reply [flat|nested] 8+ messages in thread* [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop 2026-11-20 9:59 [PATCH 0/2] usb: typec: qcom-pmic-typec: properly tear down port on remove and shutdown Neil Armstrong @ 2026-11-20 9:59 ` Neil Armstrong 2026-11-20 10:03 ` sashiko-bot 2026-11-20 11:52 ` Bryan O'Donoghue 2026-11-20 9:59 ` [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown Neil Armstrong 1 sibling, 2 replies; 8+ messages in thread From: Neil Armstrong @ 2026-11-20 9:59 UTC (permalink / raw) To: Bryan O'Donoghue, Heikki Krogerus, Greg Kroah-Hartman Cc: linux-arm-msm, linux-usb, Neil Armstrong qcom_pmic_typec_pdphy_stop() only terminates the ongoing transmission and clears the RX frame filter, the PD PHY itself stays enabled. Use qcom_pmic_typec_pdphy_disable() so the PD PHY is also turned off before its supply is released, leaving it in the same state as before qcom_pmic_typec_pdphy_start(). Assisted-by: LLM Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> --- drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c index 435e07659d65..08cf65951796 100644 --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c @@ -545,7 +545,7 @@ static void qcom_pmic_typec_pdphy_stop(struct pmic_typec *tcpm) cancel_work_sync(&pmic_typec_pdphy->reset_work); - qcom_pmic_typec_pdphy_reset_on(pmic_typec_pdphy); + qcom_pmic_typec_pdphy_disable(pmic_typec_pdphy); regulator_disable(pmic_typec_pdphy->vdd_pdphy); } -- 2.34.1 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop 2026-11-20 9:59 ` [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop Neil Armstrong @ 2026-11-20 10:03 ` sashiko-bot 2026-11-20 11:52 ` Bryan O'Donoghue 1 sibling, 0 replies; 8+ messages in thread From: sashiko-bot @ 2026-11-20 10:03 UTC (permalink / raw) To: Neil Armstrong; +Cc: linux-usb Thank you for your contribution! Sashiko AI review found 1 potential issue to consider: - [Low] usb: typec: qcom-pmic-typec: missing Fixes trailer for pdphy stop fix -- commit b45aad2ca233d0a40118f71870955c8a4770e204 Author: Neil Armstrong <neil.armstrong@linaro.org> > usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop > > qcom_pmic_typec_pdphy_stop() only terminates the ongoing transmission and > clears the RX frame filter, the PD PHY itself stays enabled. > > Use qcom_pmic_typec_pdphy_disable() so the PD PHY is also turned off > before its supply is released, leaving it in the same state as before > qcom_pmic_typec_pdphy_start(). > > Assisted-by: LLM > Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> [Severity: Low] This fixes qcom_pmic_typec_pdphy_stop() leaving the PD PHY enabled when stopping. Should this include: Fixes: a4422ff22142 ("usb: typec: qcom: Add Qualcomm PMIC Type-C driver") -- Sashiko AI review · https://sashiko.dev/#/patchset/20261120-topic-tcpm-qcom-fixes-v1-0-9a152f906fba@linaro.org?part=1 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop 2026-11-20 9:59 ` [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop Neil Armstrong 2026-11-20 10:03 ` sashiko-bot @ 2026-11-20 11:52 ` Bryan O'Donoghue 1 sibling, 0 replies; 8+ messages in thread From: Bryan O'Donoghue @ 2026-11-20 11:52 UTC (permalink / raw) To: Neil Armstrong, Heikki Krogerus, Greg Kroah-Hartman Cc: linux-arm-msm, linux-usb On 20/11/2026 09:59, Neil Armstrong wrote: > qcom_pmic_typec_pdphy_stop() only terminates the ongoing transmission and > clears the RX frame filter, the PD PHY itself stays enabled. > > Use qcom_pmic_typec_pdphy_disable() so the PD PHY is also turned off > before its supply is released, leaving it in the same state as before > qcom_pmic_typec_pdphy_start(). > > Assisted-by: LLM > Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> > --- > drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c > index 435e07659d65..08cf65951796 100644 > --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c > +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_pdphy.c > @@ -545,7 +545,7 @@ static void qcom_pmic_typec_pdphy_stop(struct pmic_typec *tcpm) > > cancel_work_sync(&pmic_typec_pdphy->reset_work); > > - qcom_pmic_typec_pdphy_reset_on(pmic_typec_pdphy); > + qcom_pmic_typec_pdphy_disable(pmic_typec_pdphy); > > regulator_disable(pmic_typec_pdphy->vdd_pdphy); > } > Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown 2026-11-20 9:59 [PATCH 0/2] usb: typec: qcom-pmic-typec: properly tear down port on remove and shutdown Neil Armstrong 2026-11-20 9:59 ` [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop Neil Armstrong @ 2026-11-20 9:59 ` Neil Armstrong 2026-11-20 10:15 ` sashiko-bot 2026-11-20 11:55 ` Bryan O'Donoghue 1 sibling, 2 replies; 8+ messages in thread From: Neil Armstrong @ 2026-11-20 9:59 UTC (permalink / raw) To: Bryan O'Donoghue, Heikki Krogerus, Greg Kroah-Hartman Cc: linux-arm-msm, linux-usb, Neil Armstrong On reboot or power off the port is left as is: a source partner keeps the negotiated contract, possibly above 5V, and a sink partner stays powered by the OTG boost while nothing services the port anymore. Add a shutdown callback stopping the port and the PD PHY, unregistering the TCPM port, which turns VBUS and VCONN off, and finally disabling the Type-C block so the partner sees a detach. Assisted-by: LLM Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> --- drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c | 11 +++++++++++ drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h | 1 + drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c | 11 +++++++++++ 3 files changed, 23 insertions(+) diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c index 58d1e6aae637..8169db478b8a 100644 --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c @@ -134,6 +134,16 @@ static void qcom_pmic_typec_remove(struct platform_device *pdev) fwnode_handle_put(tcpm->tcpc.fwnode); } +static void qcom_pmic_typec_shutdown(struct platform_device *pdev) +{ + struct pmic_typec *tcpm = platform_get_drvdata(pdev); + + tcpm->pdphy_stop(tcpm); + tcpm->port_stop(tcpm); + tcpm_unregister_port(tcpm->tcpm_port); + tcpm->port_disable(tcpm); +} + static const struct pmic_typec_resources pm8150b_typec_res = { .pdphy_res = &pm8150b_pdphy_res, .port_res = &pm8150b_port_res, @@ -158,6 +168,7 @@ static struct platform_driver qcom_pmic_typec_driver = { }, .probe = qcom_pmic_typec_probe, .remove = qcom_pmic_typec_remove, + .shutdown = qcom_pmic_typec_shutdown, }; module_platform_driver(qcom_pmic_typec_driver); diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h index 3c75820c9187..0818582fe850 100644 --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.h @@ -20,6 +20,7 @@ struct pmic_typec { int (*port_start)(struct pmic_typec *tcpm, struct tcpm_port *tcpm_port); void (*port_stop)(struct pmic_typec *tcpm); + void (*port_disable)(struct pmic_typec *tcpm); }; #define tcpc_to_tcpm(_tcpc_) container_of(_tcpc_, struct pmic_typec, tcpc) diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c index b2ad31e85b5e..4dc302d732e8 100644 --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c @@ -700,6 +700,16 @@ static void qcom_pmic_typec_port_stop(struct pmic_typec *tcpm) disable_delayed_work_sync(&pmic_typec_port->cc_debounce_dwork); } +static void qcom_pmic_typec_port_disable(struct pmic_typec *tcpm) +{ + struct pmic_typec_port *pmic_typec_port = tcpm->pmic_typec_port; + + /* Present neither Rp nor Rd, so a connected partner detects a detach */ + regmap_write(pmic_typec_port->regmap, + pmic_typec_port->base + TYPEC_MODE_CFG_REG, + TYPEC_DISABLE_CMD); +} + int qcom_pmic_typec_port_probe(struct platform_device *pdev, struct pmic_typec *tcpm, const struct pmic_typec_port_resources *res, @@ -779,6 +789,7 @@ int qcom_pmic_typec_port_probe(struct platform_device *pdev, tcpm->port_start = qcom_pmic_typec_port_start; tcpm->port_stop = qcom_pmic_typec_port_stop; + tcpm->port_disable = qcom_pmic_typec_port_disable; return 0; } -- 2.34.1 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown 2026-11-20 9:59 ` [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown Neil Armstrong @ 2026-11-20 10:15 ` sashiko-bot 2026-11-20 10:53 ` Neil Armstrong 2026-11-20 11:55 ` Bryan O'Donoghue 1 sibling, 1 reply; 8+ messages in thread From: sashiko-bot @ 2026-11-20 10:15 UTC (permalink / raw) To: Neil Armstrong; +Cc: linux-usb Thank you for your contribution! Sashiko AI review found 2 potential issues to consider: - [High] usb: typec: qcom: NULL deref in qcom_pmic_typec_shutdown() when VBUS is on - [Medium] usb: typec: qcom-pmic-typec: missing port_disable() in remove and probe unwind -- commit 27e120f037d3faf6c51dac7774569601de458e75 Author: Neil Armstrong <neil.armstrong@linaro.org> usb: typec: qcom-pmic-typec: detach the port on shutdown This commit adds a platform driver shutdown callback that stops the port and PD PHY, unregisters the TCPM port, and disables the Type-C block so that a connected partner sees a detach on reboot or power off. > diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c > index 58d1e6aae637d..8169db478b8aa 100644 > --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c > +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c > @@ -134,6 +134,16 @@ static void qcom_pmic_typec_remove(struct platform_device *pdev) > fwnode_handle_put(tcpm->tcpc.fwnode); > } > > +static void qcom_pmic_typec_shutdown(struct platform_device *pdev) > +{ > + struct pmic_typec *tcpm = platform_get_drvdata(pdev); > + > + tcpm->pdphy_stop(tcpm); > + tcpm->port_stop(tcpm); > + tcpm_unregister_port(tcpm->tcpm_port); [Severity: High] Can calling tcpm_unregister_port() here cause a NULL pointer dereference on port->wq when shutting down with an attached sink partner powered by VBUS? tcpm_unregister_port() destroys port->wq and sets it to NULL before calling tcpm_reset_port() -> tcpm_init_vbus() -> qcom_pmic_typec_port_set_vbus(). When VBUS transitions from enabled to disabled, qcom_pmic_typec_port_set_vbus() calls tcpm_vbus_change(), which queues work onto port->wq: drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c:qcom_pmic_typec_port_set_vbus() { ... ret = qcom_pmic_typec_port_vbus_toggle(pmic_typec_port, on); if (ret) goto done; pmic_typec_port->vbus_enabled = on; tcpm_vbus_change(tcpm->tcpm_port); ... } kthread_queue_work() then dereferences NULL in &worker->lock, crashing before tcpm->port_disable(tcpm) is reached. > + tcpm->port_disable(tcpm); > +} [Severity: Medium] Should tcpm->port_disable(tcpm) also be called in qcom_pmic_typec_remove() and the port_unregister error path in qcom_pmic_typec_probe()? Without it, unbinding the driver or failing probe after tcpm_register_port() leaves TYPEC_MODE_CFG_REG presenting Rp or Rd with no TCPM or PD PHY servicing the port, and a subsequent reboot or power off will not invoke qcom_pmic_typec_shutdown() on the unbound device. -- Sashiko AI review · https://sashiko.dev/#/patchset/20261120-topic-tcpm-qcom-fixes-v1-0-9a152f906fba@linaro.org?part=2 ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown 2026-11-20 10:15 ` sashiko-bot @ 2026-11-20 10:53 ` Neil Armstrong 0 siblings, 0 replies; 8+ messages in thread From: Neil Armstrong @ 2026-11-20 10:53 UTC (permalink / raw) To: sashiko-reviews; +Cc: linux-usb On 11/20/26 11:15, sashiko-bot@kernel.org wrote: > Thank you for your contribution! Sashiko AI review found 2 potential > issues to consider: > - [High] usb: typec: qcom: NULL deref in qcom_pmic_typec_shutdown() when > VBUS is on > - [Medium] usb: typec: qcom-pmic-typec: missing port_disable() in remove > and probe unwind > -- > > commit 27e120f037d3faf6c51dac7774569601de458e75 > Author: Neil Armstrong <neil.armstrong@linaro.org> > > usb: typec: qcom-pmic-typec: detach the port on shutdown > > This commit adds a platform driver shutdown callback that stops the port > and PD PHY, unregisters the TCPM port, and disables the Type-C block so > that a connected partner sees a detach on reboot or power off. > >> diff --git a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c >> index 58d1e6aae637d..8169db478b8aa 100644 >> --- a/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c >> +++ b/drivers/usb/typec/tcpm/qcom/qcom_pmic_typec.c >> @@ -134,6 +134,16 @@ static void qcom_pmic_typec_remove(struct platform_device *pdev) >> fwnode_handle_put(tcpm->tcpc.fwnode); >> } >> >> +static void qcom_pmic_typec_shutdown(struct platform_device *pdev) >> +{ >> + struct pmic_typec *tcpm = platform_get_drvdata(pdev); >> + >> + tcpm->pdphy_stop(tcpm); >> + tcpm->port_stop(tcpm); >> + tcpm_unregister_port(tcpm->tcpm_port); > > [Severity: High] > Can calling tcpm_unregister_port() here cause a NULL pointer dereference > on port->wq when shutting down with an attached sink partner powered by > VBUS? > > tcpm_unregister_port() destroys port->wq and sets it to NULL before > calling tcpm_reset_port() -> tcpm_init_vbus() -> > qcom_pmic_typec_port_set_vbus(). When VBUS transitions from enabled to > disabled, qcom_pmic_typec_port_set_vbus() calls tcpm_vbus_change(), which > queues work onto port->wq: > > drivers/usb/typec/tcpm/qcom/qcom_pmic_typec_port.c:qcom_pmic_typec_port_set_vbus() { > ... > ret = qcom_pmic_typec_port_vbus_toggle(pmic_typec_port, on); > if (ret) > goto done; > > pmic_typec_port->vbus_enabled = on; > tcpm_vbus_change(tcpm->tcpm_port); > ... > } > > kthread_queue_work() then dereferences NULL in &worker->lock, crashing > before tcpm->port_disable(tcpm) is reached. What's the difference with qcom_pmic_typec_remove() ? So the same regression occurs with qcom_pmic_typec_remove() and is not related with this change. I'll try to fix it > >> + tcpm->port_disable(tcpm); >> +} > > [Severity: Medium] > Should tcpm->port_disable(tcpm) also be called in qcom_pmic_typec_remove() > and the port_unregister error path in qcom_pmic_typec_probe()? > > Without it, unbinding the driver or failing probe after > tcpm_register_port() leaves TYPEC_MODE_CFG_REG presenting Rp or Rd with no > TCPM or PD PHY servicing the port, and a subsequent reboot or power off > will not invoke qcom_pmic_typec_shutdown() on the unbound device. > Ok so it should be extended by: - check if not in TYPEC_DISABLE_CMD on probe, and force a detach is enabled before - disable on remove and probe fail to have a coherent state ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown 2026-11-20 9:59 ` [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown Neil Armstrong 2026-11-20 10:15 ` sashiko-bot @ 2026-11-20 11:55 ` Bryan O'Donoghue 1 sibling, 0 replies; 8+ messages in thread From: Bryan O'Donoghue @ 2026-11-20 11:55 UTC (permalink / raw) To: Neil Armstrong, Heikki Krogerus, Greg Kroah-Hartman Cc: linux-arm-msm, linux-usb On 20/11/2026 09:59, Neil Armstrong wrote: > On reboot or power off the port is left as is: a source partner keeps > the negotiated contract, possibly above 5V, and a sink partner stays > powered by the OTG boost while nothing services the port anymore. > > Add a shutdown callback stopping the port and the PD PHY, unregistering > the TCPM port, which turns VBUS and VCONN off, and finally disabling the > Type-C block so the partner sees a detach. > > Assisted-by: LLM > Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org> LGTM Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org> ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2026-11-20 11:55 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-11-20 9:59 [PATCH 0/2] usb: typec: qcom-pmic-typec: properly tear down port on remove and shutdown Neil Armstrong 2026-11-20 9:59 ` [PATCH 1/2] usb: typec: qcom-pmic-typec: pdphy: disable the PD PHY on stop Neil Armstrong 2026-11-20 10:03 ` sashiko-bot 2026-11-20 11:52 ` Bryan O'Donoghue 2026-11-20 9:59 ` [PATCH 2/2] usb: typec: qcom-pmic-typec: detach the port on shutdown Neil Armstrong 2026-11-20 10:15 ` sashiko-bot 2026-11-20 10:53 ` Neil Armstrong 2026-11-20 11:55 ` Bryan O'Donoghue
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.