6.6-stable review patch. If anyone has any objections, please let me know.
------------------
From: RD Babiera <rdbabiera@xxxxxxxxxx>
commit bf20c69cf3cf9c6445c4925dd9a8a6ca1b78bfdf upstream.
When a Fast Role Swap control message attempt results in a transition
to ERROR_RECOVERY, the TCPC can still queue a TCPM_SOURCING_VBUS event.
If the event is queued but processed after the tcpm_reset_port() call
in the PORT_RESET state, then the following occurs:
1. tcpm_reset_port() calls tcpm_init_vbus() to reset the vbus sourcing and
sinking state
2. tcpm_pd_event_handler() turns VBUS on before the port is in the default
state.
3. The port resolves as a sink. In the SNK_DISCOVERY state,
tcpm_set_charge() cannot set vbus to charge.
Clear pd events within PORT_RESET to get rid of non-applicable events.
Fixes: b17dd57118fe ("staging: typec: tcpm: Improve role swap with non PD capable partners")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: RD Babiera <rdbabiera@xxxxxxxxxx>
Reviewed-by: Heikki Krogerus <heikki.krogerus@xxxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20240423202715.3375827-2-rdbabiera@xxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/usb/typec/tcpm/tcpm.c | 1 +
1 file changed, 1 insertion(+)
--- a/drivers/usb/typec/tcpm/tcpm.c
+++ b/drivers/usb/typec/tcpm/tcpm.c
@@ -4859,6 +4859,7 @@ static void run_state_machine(struct tcp
break;
case PORT_RESET:
tcpm_reset_port(port);
+ port->pd_events = 0;
if (port->self_powered)
tcpm_set_cc(port, TYPEC_CC_OPEN);
else
