Re: [PATCH v2 1/1] usb: typec: ucsi: Check capabilities before cable and identity discovery

From: neil . armstrong
Date: Fri Mar 15 2024 - 13:23:10 EST


Hi,

On 15/03/2024 18:18, Jameson Thies wrote:
Check the UCSI_CAP_GET_PD_MESSAGE bit before sending GET_PD_MESSAGE to
discover partner and cable identity, check UCSI_CAP_CABLE_DETAILS before
sending GET_CABLE_PROPERTY to discover the cable and check
UCSI_CAP_ALT_MODE_DETAILS before registering the a cable plug. Additionally,
move 8 bits from reserved_1 to features in the ucsi_capability struct. This
makes the field 16 bits, still 8 short of the 24 bits allocated for it in
UCSI v3.0, but it will not overflow because UCSI only defines 14 bits in
bmOptionalFeatures.

Fixes: 38ca416597b0 ("usb: typec: ucsi: Register cables based on GET_CABLE_PROPERTY")
Link: https://lore.kernel.org/linux-usb/44e8142f-d9b3-487b-83fe-39deadddb492@xxxxxxxxxx
Suggested-by: Neil Armstrong <neil.armstrong@xxxxxxxxxx>
Signed-off-by: Jameson Thies <jthies@xxxxxxxxxx>
---
Confirmed a device which supports GET_PD_MESSAGE, GET_CABLE_PROPERTY and
GET_ALTERNATE_MODES still requested identity and cable information.

drivers/usb/typec/ucsi/ucsi.c | 34 +++++++++++++++++++++-------------
drivers/usb/typec/ucsi/ucsi.h | 5 +++--
2 files changed, 24 insertions(+), 15 deletions(-)

diff --git a/drivers/usb/typec/ucsi/ucsi.c b/drivers/usb/typec/ucsi/ucsi.c
index cf52cb34d2859..958dc82989b60 100644
--- a/drivers/usb/typec/ucsi/ucsi.c
+++ b/drivers/usb/typec/ucsi/ucsi.c
@@ -1133,17 +1133,21 @@ static int ucsi_check_cable(struct ucsi_connector *con)
if (ret < 0)
return ret;
- ret = ucsi_get_cable_identity(con);
- if (ret < 0)
- return ret;
+ if (con->ucsi->cap.features & UCSI_CAP_GET_PD_MESSAGE) {
+ ret = ucsi_get_cable_identity(con);
+ if (ret < 0)
+ return ret;
+ }
- ret = ucsi_register_plug(con);
- if (ret < 0)
- return ret;
+ if (con->ucsi->cap.features & UCSI_CAP_ALT_MODE_DETAILS) {
+ ret = ucsi_register_plug(con);
+ if (ret < 0)
+ return ret;
- ret = ucsi_register_altmodes(con, UCSI_RECIPIENT_SOP_P);
- if (ret < 0)
- return ret;
+ ret = ucsi_register_altmodes(con, UCSI_RECIPIENT_SOP_P);
+ if (ret < 0)
+ return ret;
+ }
return 0;
}
@@ -1189,8 +1193,10 @@ static void ucsi_handle_connector_change(struct work_struct *work)
ucsi_register_partner(con);
ucsi_partner_task(con, ucsi_check_connection, 1, HZ);
ucsi_partner_task(con, ucsi_check_connector_capability, 1, HZ);
- ucsi_partner_task(con, ucsi_get_partner_identity, 1, HZ);
- ucsi_partner_task(con, ucsi_check_cable, 1, HZ);
+ if (con->ucsi->cap.features & UCSI_CAP_GET_PD_MESSAGE)
+ ucsi_partner_task(con, ucsi_get_partner_identity, 1, HZ);
+ if (con->ucsi->cap.features & UCSI_CAP_CABLE_DETAILS)
+ ucsi_partner_task(con, ucsi_check_cable, 1, HZ);
if (UCSI_CONSTAT_PWR_OPMODE(con->status.flags) ==
UCSI_CONSTAT_PWR_OPMODE_PD)
@@ -1589,8 +1595,10 @@ static int ucsi_register_port(struct ucsi *ucsi, struct ucsi_connector *con)
ucsi_register_partner(con);
ucsi_pwr_opmode_change(con);
ucsi_port_psy_changed(con);
- ucsi_get_partner_identity(con);
- ucsi_check_cable(con);
+ if (con->ucsi->cap.features & UCSI_CAP_GET_PD_MESSAGE)
+ ucsi_get_partner_identity(con);
+ if (con->ucsi->cap.features & UCSI_CAP_CABLE_DETAILS)
+ ucsi_check_cable(con);
}
/* Only notify USB controller if partner supports USB data */
diff --git a/drivers/usb/typec/ucsi/ucsi.h b/drivers/usb/typec/ucsi/ucsi.h
index 32daf5f586505..0e7c92eb1b227 100644
--- a/drivers/usb/typec/ucsi/ucsi.h
+++ b/drivers/usb/typec/ucsi/ucsi.h
@@ -206,7 +206,7 @@ struct ucsi_capability {
#define UCSI_CAP_ATTR_POWER_OTHER BIT(10)
#define UCSI_CAP_ATTR_POWER_VBUS BIT(14)
u8 num_connectors;
- u8 features;
+ u16 features;
#define UCSI_CAP_SET_UOM BIT(0)
#define UCSI_CAP_SET_PDM BIT(1)
#define UCSI_CAP_ALT_MODE_DETAILS BIT(2)
@@ -215,7 +215,8 @@ struct ucsi_capability {
#define UCSI_CAP_CABLE_DETAILS BIT(5)
#define UCSI_CAP_EXT_SUPPLY_NOTIFICATIONS BIT(6)
#define UCSI_CAP_PD_RESET BIT(7)
- u16 reserved_1;
+#define UCSI_CAP_GET_PD_MESSAGE BIT(8)
+ u8 reserved_1;
u8 num_alt_modes;
u8 reserved_2;
u16 bc_version;

It looks fine:
Reviewed-by: Neil Armstrong <neil.armstrong@xxxxxxxxxx>

I'll run a test on the affected platforms in the next days and
report if it fixes the regression.

Thanks,
Neil