Re: [PATCH v3 4/4] driver: bluetooth: hci_qca: SSR(SubSystem Restart)process failed due to tx_idle_timer timeout

[Shuai Zhang <quic_shuaz@quicinc.com>]

Dear,Konrad

On 8/13/2025 7:13 PM, Konrad Dybcio wrote:
> On 8/13/25 5:35 AM, Shuai Zhang wrote:
>> When the SSR (SubSystem Restart) duration exceeds 2 seconds, it triggers
>> host tx_idle_timeout, which sets host TX state to sleep. due to the
> 
> Host? Looks like we're sending a command to the BT chip and that
> triggers the issue

Since the firmware is not downloaded after ssR, the controller does not enter
the ibs_sleep state. If the host enters ibs_sleep, it will send a wake-up 
command before sending any other commands. 

However, the controller only processes IBS wake-up commands when it is already
in ibs_sleep. As a result, it will not respond to the host's wake-up command,
which prevents host from send actual command and leads to a send command timeout.

Therefore, the host must not enter the ibs_sleep state.

> 
>> hardware pulling up bt_en, the firmware is not downloaded after the SSR.
>> As a result, the controller does not enter sleep mode. Consequently,
>> when the host sends a command afterward, it sends 0xFD to the controller,
>> but the controller does not respond, leading to a command timeout.
>>
>> So reset tx_idle_timer after SSR to prevent host enter TX IBS_Sloeep mode.
>>
>> Signed-off-by: Shuai Zhang <quic_shuaz@quicinc.com>
>> ---
>>  drivers/bluetooth/hci_qca.c | 9 +++++++++
>>  1 file changed, 9 insertions(+)
>>
>> diff --git a/drivers/bluetooth/hci_qca.c b/drivers/bluetooth/hci_qca.c
>> index 33c2dd7b8..1fb8eea09 100644
>> --- a/drivers/bluetooth/hci_qca.c
>> +++ b/drivers/bluetooth/hci_qca.c
>> @@ -1668,6 +1668,15 @@ static void qca_hw_error(struct hci_dev *hdev, u8 code)
>>  	if (!test_bit(HCI_QUIRK_NON_PERSISTENT_SETUP, &hdev->quirks)) {
>>  		clear_bit(QCA_SSR_TRIGGERED, &qca->flags);
>>  		clear_bit(QCA_IBS_DISABLED, &qca->flags);
>> +		/*
>> +		 * When the SSR (Sub-System Restart) duration exceeds 2 seconds,
> 
> This comment will become out of date quickly, refer to 'tx_idle_delay' instead
> 

i will update.

> Konrad
> 
>> +		 * it triggers host tx_idle_timeout, which sets host TX state
>> +		 * to sleep. Reset tx_idle_timer after SSR to prevent
>> +		 * host enter TX IBS_Sloeep mode.
>> +		 */
>> +		mod_timer(&qca->tx_idle_timer, jiffies +
>> +			  msecs_to_jiffies(qca->tx_idle_delay));
>> +
>>  		qca->tx_ibs_state = HCI_IBS_TX_AWAKE;
>>  		qca->memdump_state = QCA_MEMDUMP_IDLE;
>>  		msleep(50);

BR,
Shuai