Skip to contentSkip to conversation

usb: typec: tipd: Restore generic TPS6598x contract interrupts

PATCH In review Vincent Cloutier 3 participants 4 messages Apr 2 – Apr 3

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 usb: typec: tipd: Restore generic TPS6598x contract interrupts

From: Vincent Cloutier <[email protected]>

The generic TPS6598x interrupt handler still relies on PP_SWITCH_CHANGED, NEW_CONTRACT_AS_CONSUMER, HARD_RESET, and STATUS_UPDATE, but the irq_mask1 refactor only kept POWER_STATUS_UPDATE, DATA_STATUS_UPDATE, and PLUG_EVENT in tps6598x_data.

On the librem5 that leaves PD partners stuck at the 500 mA fallback because the active contract is never refreshed after attach.

Restore the missing interrupt bits so the existing handler paths are reachable again. This fixes USB-C charging negotiation on the librem5:
after a replug the TPS6598x source power supply reports 3 A instead of 500 mA and the BQ25890 input limit follows suit.

Fixes b3dddff502c5 ("usb: typec: tipd: Move initial irq mask to tipd_data")
---
 drivers/usb/typec/tipd/core.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/usb/typec/tipd/core.c +5 -1
index 84ee5687bb27..83f2fec6e34e 100644
@@ -2395,7 +2395,11 @@ static const struct tipd_data tps6598x_data = {
 	.irq_handler = tps6598x_interrupt,
 	.irq_mask1 = TPS_REG_INT_POWER_STATUS_UPDATE |
 		     TPS_REG_INT_DATA_STATUS_UPDATE |
-		     TPS_REG_INT_PLUG_EVENT,
+		     TPS_REG_INT_PLUG_EVENT |
+		     TPS_REG_INT_PP_SWITCH_CHANGED |
+		     TPS_REG_INT_NEW_CONTRACT_AS_CONSUMER |
+		     TPS_REG_INT_HARD_RESET |
+		     TPS_REG_INT_STATUS_UPDATE,
 	.tps_struct_size = sizeof(struct tps6598x),
 	.register_port = tps6598x_register_port,
 	.unregister_port = tps6598x_unregister_port,
Signature · 1 line
2.53.0
maintainer Acked-by on v1 1/1
#1 lore raw

Wed, Apr 01, 2026 at 08:09:50PM -0400, Vincent Cloutier kirjoitti:

Quoted 11 lines

From: Vincent Cloutier <[email protected]>

The generic TPS6598x interrupt handler still relies on PP_SWITCH_CHANGED, NEW_CONTRACT_AS_CONSUMER, HARD_RESET, and STATUS_UPDATE, but the irq_mask1 refactor only kept POWER_STATUS_UPDATE, DATA_STATUS_UPDATE, and PLUG_EVENT in tps6598x_data.

On the librem5 that leaves PD partners stuck at the 500 mA fallback because the active contract is never refreshed after attach.

Quoted 8 lines

Restore the missing interrupt bits so the existing handler paths are reachable again. This fixes USB-C charging negotiation on the librem5:
after a replug the TPS6598x source power supply reports 3 A instead of 500 mA and the BQ25890 input limit follows suit.

Fixes b3dddff502c5 ("usb: typec: tipd: Move initial irq mask to tipd_data")
Quoted 21 lines
---
 drivers/usb/typec/tipd/core.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/usb/typec/tipd/core.c +5 -1
index 84ee5687bb27..83f2fec6e34e 100644
@@ -2395,7 +2395,11 @@ static const struct tipd_data tps6598x_data = {
 	.irq_handler = tps6598x_interrupt,
 	.irq_mask1 = TPS_REG_INT_POWER_STATUS_UPDATE |
 		     TPS_REG_INT_DATA_STATUS_UPDATE |
-		     TPS_REG_INT_PLUG_EVENT,
+		     TPS_REG_INT_PLUG_EVENT |
+		     TPS_REG_INT_PP_SWITCH_CHANGED |
+		     TPS_REG_INT_NEW_CONTRACT_AS_CONSUMER |
+		     TPS_REG_INT_HARD_RESET |
+		     TPS_REG_INT_STATUS_UPDATE,
 	.tps_struct_size = sizeof(struct tps6598x),
 	.register_port = tps6598x_register_port,
 	.unregister_port = tps6598x_unregister_port,
Signature · 1 line
heikki
#2 lore raw
In reply to Vincent Cloutier
Quoted 30 lines · On czwartek, 2 kwietnia 2026 02:09:50 czas środkowoeuropejski letni Vincent…

From: Vincent Cloutier <[email protected]>

The generic TPS6598x interrupt handler still relies on PP_SWITCH_CHANGED, NEW_CONTRACT_AS_CONSUMER, HARD_RESET, and STATUS_UPDATE, but the irq_mask1 refactor only kept POWER_STATUS_UPDATE, DATA_STATUS_UPDATE, and PLUG_EVENT in tps6598x_data.

On the librem5 that leaves PD partners stuck at the 500 mA fallback because the active contract is never refreshed after attach.

Restore the missing interrupt bits so the existing handler paths are reachable again. This fixes USB-C charging negotiation on the librem5:
after a replug the TPS6598x source power supply reports 3 A instead of 500 mA and the BQ25890 input limit follows suit.

Fixes b3dddff502c5 ("usb: typec: tipd: Move initial irq mask to tipd_data")
---
 drivers/usb/typec/tipd/core.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/usb/typec/tipd/core.c +0 -0
index 84ee5687bb27..83f2fec6e34e 100644
@@ -2395,7 +2395,11 @@ static const struct tipd_data tps6598x_data = {
 	.irq_handler = tps6598x_interrupt,
 	.irq_mask1 = TPS_REG_INT_POWER_STATUS_UPDATE |

drivers/usb/typec/tipd/core.c static const struct tipd_data tps6598x_data = { · line 2405

Quoted 10 lines
drivers/usb/typec/tipd/core.c +5 -1
 		     TPS_REG_INT_DATA_STATUS_UPDATE |
-		     TPS_REG_INT_PLUG_EVENT,
+		     TPS_REG_INT_PLUG_EVENT |
+		     TPS_REG_INT_PP_SWITCH_CHANGED |
+		     TPS_REG_INT_NEW_CONTRACT_AS_CONSUMER |
+		     TPS_REG_INT_HARD_RESET |
+		     TPS_REG_INT_STATUS_UPDATE,
 	.tps_struct_size = sizeof(struct tps6598x),
 	.register_port = tps6598x_register_port,
 	.unregister_port = tps6598x_unregister_port,

This driver has never handled these interrupts (as seen in the commit you referenced), so these should be added in patches that make it handle them.

You likely got confused by the patches that still stay in the downstream Librem 5 tree. We should get them cleaned up and mainlined, but until that happens this patch doesn't make much sense here I'm afraid.

S.

#3 lore raw
Quoted 50 lines · Sebastian Krzyszkowiak

On czwartek, 2 kwietnia 2026 02:09:50 czas środkowoeuropejski letni Vincent

Quoted 30 lines · Cloutier

From: Vincent Cloutier <[email protected]>

The generic TPS6598x interrupt handler still relies on PP_SWITCH_CHANGED, NEW_CONTRACT_AS_CONSUMER, HARD_RESET, and STATUS_UPDATE, but the irq_mask1 refactor only kept POWER_STATUS_UPDATE, DATA_STATUS_UPDATE, and PLUG_EVENT in tps6598x_data.

On the librem5 that leaves PD partners stuck at the 500 mA fallback because the active contract is never refreshed after attach.

Restore the missing interrupt bits so the existing handler paths are reachable again. This fixes USB-C charging negotiation on the librem5:
after a replug the TPS6598x source power supply reports 3 A instead of 500 mA and the BQ25890 input limit follows suit.

Fixes b3dddff502c5 ("usb: typec: tipd: Move initial irq mask to tipd_data")

---
drivers/usb/typec/tipd/core.c | 6 +++++-

1 file changed, 5 insertions(+), 1 deletion(-)
drivers/usb/typec/tipd/core.c +0 -0
index 84ee5687bb27..83f2fec6e34e 100644
@@ -2395,7 +2395,11 @@ static const struct tipd_data tps6598x_data = {
   .irq_handler = tps6598x_interrupt,
   .irq_mask1 = TPS_REG_INT_POWER_STATUS_UPDATE |

drivers/usb/typec/tipd/core.c static const struct tipd_data tps6598x_data = { · line 2405

Quoted 10 lines
drivers/usb/typec/tipd/core.c +5 -1
            TPS_REG_INT_DATA_STATUS_UPDATE |
-             TPS_REG_INT_PLUG_EVENT,
+             TPS_REG_INT_PLUG_EVENT |
+             TPS_REG_INT_PP_SWITCH_CHANGED |
+             TPS_REG_INT_NEW_CONTRACT_AS_CONSUMER |
+             TPS_REG_INT_HARD_RESET |
+             TPS_REG_INT_STATUS_UPDATE,
   .tps_struct_size = sizeof(struct tps6598x),
   .register_port = tps6598x_register_port,
   .unregister_port = tps6598x_unregister_port,

This driver has never handled these interrupts (as seen in the commit you referenced), so these should be added in patches that make it handle them.

Quoted 3 lines

You likely got confused by the patches that still stay in the downstream Librem 5 tree. We should get them cleaned up and mainlined, but until that happens this patch doesn't make much sense here I'm afraid.

I’m afraid that’s exactly what happened… I have learned a lesson here.

Quoted 4 lines

S.

Sorry about this noise and thank you all for your time!

Vincent