Hi Denis,
I wasn't able to reproduce yesterday.
From traces I kept, with problematic SIM card, I had this trace when
problem appeared:
../git/drivers/qmimodem/sim.c:get_card_status() info1->app_state:0x6:
OFONO_SIM_PASSWORD_INVALID
I had a look at QMI lib and app_state to 0x6 means
QMI_UIM_CARD_APPLICATION_STATE_ILLEGAL. Other unhandled states are :
- QMI_UIM_CARD_APPLICATION_STATE_UNKNOWN
- QMI_UIM_CARD_APPLICATION_STATE_DETECTED
- QMI_UIM_CARD_APPLICATION_STATE_PIN1_BLOCKED
With current driver code, all these states will lead to
OFONO_SIM_PASSWORD_INVALID. I don't know if a retry will help if we have
any of these states. It would be easy to make a retry for these states
as this is already implemented for
QMI_UIM_CARD_APPLICATION_STATE_CHECK_PERSONALIZATION_STATE state.
Do you want me to change the patch to implement that?
On my side, my concern is more: what "PinRequired" property (in
SimManager properties) should be if we fail to get PIN state (in QMI
driver or any other driver (with AT+CPIN? replying ERROR for example))?
I think that with current code it will be "none" and I would prefer
something like "unknown", "error" or "invalid" to distinguish a SIM with
entered or disabled PIN from a SIM where driver is not able to get PIN
state. What do you think?
Best Regards,
Christophe Ronco
On 08/30/2017 08:29 PM, Denis Kenzior wrote:
Hi Christophe,
On 08/29/2017 07:29 AM, Christophe Ronco wrote:
This will avoid a crash seen in sim_pin_query_cb (src/sim.c) if pin type
is OFONO_SIM_PASSWORD_INVALID.
---
drivers/qmimodem/sim-legacy.c | 5 ++++-
drivers/qmimodem/sim.c | 6 +++++-
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/qmimodem/sim-legacy.c
b/drivers/qmimodem/sim-legacy.c
index 318b1ae..d16b2eb 100644
--- a/drivers/qmimodem/sim-legacy.c
+++ b/drivers/qmimodem/sim-legacy.c
@@ -211,7 +211,10 @@ static void get_pin_status_cb(struct qmi_result
*result, void *user_data)
data->retries[OFONO_SIM_PASSWORD_SIM_PUK2] = pin->unblock_retries;
done:
- CALLBACK_WITH_SUCCESS(cb, pin_type, cbd->data);
+ if (pin_type == OFONO_SIM_PASSWORD_INVALID)
+ CALLBACK_WITH_FAILURE(cb, -1, cbd->data);
+ else
+ CALLBACK_WITH_SUCCESS(cb, pin_type, cbd->data);
}
static void qmi_query_passwd_state(struct ofono_sim *sim,
diff --git a/drivers/qmimodem/sim.c b/drivers/qmimodem/sim.c
index 8d40028..7425d5e 100644
--- a/drivers/qmimodem/sim.c
+++ b/drivers/qmimodem/sim.c
@@ -483,7 +483,11 @@ static void query_passwd_state_cb(struct
qmi_result *result,
case GET_CARD_STATUS_RESULT_OK:
DBG("passwd state %d", sim_stat.passwd_state);
data->retry_count = 0;
- CALLBACK_WITH_SUCCESS(cb, sim_stat.passwd_state, cbd->data);
+ if (sim_stat.passwd_state == OFONO_SIM_PASSWORD_INVALID)
+ CALLBACK_WITH_FAILURE(cb, -1, cbd->data);
+ else
+ CALLBACK_WITH_SUCCESS(cb, sim_stat.passwd_state,
+ cbd->data);
So we get back a CARD_STATUS_RESULT_OK, yet the password state is not
valid. It would seem there's a more serious issue here? Is is that
we need to keep retrying to get the password state? Is this related to
Personalization state?
break;
case GET_CARD_STATUS_RESULT_TEMP_ERROR:
data->retry_count++;
Regards,
-Denis
_______________________________________________
ofono mailing list
ofono@ofono.org
https://lists.ofono.org/mailman/listinfo/ofono
