ACKcmnt: [xenial] SRU: Bluetooth: increase timeout for le auto connections

[Kleber Souza]

On 11/17/17 11:23, Konrad Zapałowicz wrote:
> On Fri, Nov 17, 2017 at 10:57 AM, Kleber Souza
> <kleber.souza at canonical.com> wrote:
>> On 11/17/17 08:47, Konrad Zapalowicz wrote:
>>> On 11/14, Kleber Souza wrote:
>>>> On 11/10/17 13:11, Konrad Zapalowicz wrote:
>>>>> SRU Justification:
>>>>>
>>>>> [Impact]
>>>>>
>>>>> Not possible to pair BLE remote devices such as sensors and such with
>>>>> Ubuntu when the host system uses BT+WiFi combo SoC. The connection
>>>>> attempt is disturbed with a timeout.
>>>>>
>>>>> [Fix]
>>>>>
>>>>> It has been fixed by increasing the timeout value from 2 seconds to
>>>>> 4 seconds. It is enough for the events to reach the stack (measured
>>>>> that it takes between 3 and 3.5 seconds).
>>>>>
>>>>> [Testcase]
>>>>>
>>>>> Tested with the device that failed to connect to Ubuntu Core gateway.
>>>>> It fails w/o the patch, it connects just fine with the patch applied.
>>>>>
>>>>> [Regression Potential]
>>>>>
>>>>> Very small. The increased timeout is taken into consideration only for
>>>>> new and scan report triggered connections. It will not make any already
>>>>> working device to fail to pair.
>>>>>
>>>>> [Other Info]
>>>>>
>>>>> The bug has been discussed and fixed here:
>>>>>
>>>>> https://marc.info/?l=linux-bluetooth&m=150824844606937&w=2
>>>>>
>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git/commit/?id=1f01d8be0e6a04bd682a55f6d50c14c1679e7571
>>>>>
>>>>> The patch has been accepted by the upstream and will be a part of the
>>>>> next kernel release. Currently in the bluetooth-next tree.
>>>>>
>>>>> [Patch]
>>>>>
>>>>> From 2d5a130b6cf6cbf992f9a7a66dcab8ab4e8e15f9 Mon Sep 17 00:00:00 2001
>>>>> From: =?UTF-8?q?Konrad=20Zapa=C5=82owicz?= <konrad.zapalowicz at canonical.com>
>>>>> Date: Tue, 17 Oct 2017 15:53:49 +0200
>>>>> Subject: [PATCH] [xenial] (upstream) Bluetooth: increase timeout for le auto
>>>>>  connections
>>>>> MIME-Version: 1.0
>>>>> Content-Type: text/plain; charset=UTF-8
>>>>> Content-Transfer-Encoding: 8bit
>>>>>
>>>>> BugLink: https://bugs.launchpad.net/bugs/1731467
>>>>>
>>>>> This patch increases the connection timeout for LE connections that are
>>>>> triggered by the advertising report to 4 seconds.
>>>>>
>>>>> It has been observed that devices equipped with wifi+bt combo SoC fail
>>>>> to create a connection with BLE devices due to their coexistence issues.
>>>>> Increasing this timeout gives them enough time to complete the
>>>>> connection with success.
>>>>>
>>>>> Signed-off-by: Konrad Zapałowicz <konrad.zapalowicz at canonical.com>
>>>>> Signed-off-by: Marcel Holtmann <marcel at holtmann.org>
>>>>> (cherry picked from commit 1f01d8be0e6a04bd682a55f6d50c14c1679e7571)
>>>>> ---
>>>>>  include/net/bluetooth/hci.h | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/include/net/bluetooth/hci.h b/include/net/bluetooth/hci.h
>>>>> index 0205b80..630f6d7 100644
>>>>> --- a/include/net/bluetooth/hci.h
>>>>> +++ b/include/net/bluetooth/hci.h
>>>>> @@ -266,7 +266,7 @@ enum {
>>>>>  #define HCI_AUTO_OFF_TIMEOUT       msecs_to_jiffies(2000)  /* 2 seconds */
>>>>>  #define HCI_POWER_OFF_TIMEOUT      msecs_to_jiffies(5000)  /* 5 seconds */
>>>>>  #define HCI_LE_CONN_TIMEOUT        msecs_to_jiffies(20000) /* 20 seconds */
>>>>> -#define HCI_LE_AUTOCONN_TIMEOUT    msecs_to_jiffies(2000)  /* 2 seconds */
>>>>> +#define HCI_LE_AUTOCONN_TIMEOUT    msecs_to_jiffies(4000)  /* 4 seconds */
>>>>>
>>>>>  /* HCI data types */
>>>>>  #define HCI_COMMAND_PKT            0x01
>>>>>
>>>>
>>>> As pointed out by Seth, there are some issues with this patch
>>>> submission, but we can fix them while applying it. Please follow these
>>>> guidelines when sending future patches.
>>>>
>>>> When submitting a patch to fix a bug in the kernel, the Launchpad bug
>>>> needs to be marked as affecting the "linux (Ubuntu)" package and
>>>> nominated for the relevant series (Xenial in this case). I have fixed it
>>>> on LP:#1731467.
>>>>
>>>> Given that the issues pointed out by Seth is fixed while applying the patch:
>>>> Acked-by: Kleber Sacilotto de Souza <kleber.souza at canonical.com>
>>>
>>> Thanks. So it has been merged, right?
>>>
>>> /Konrad
>>>
>>
>> Hi,
>>
>> The patch hasn't been applied to the Xenial tree yet, when that's done
>> we will send a reply with a "APPLIED" tag on the subject. However, it's
>> got the necessary ACK's on the mailing list and the next SRU cycle
>> begins next week, so it will be applied soon and the fix will become
>> available on the next kernel update.
> 
> December 11th, right?

Right, that's the scheduled date for the next SRU cycle going to -updates.

> 
> Sorry for pushing however I need to notify the customer about when the change
> will be available. This one is highly expected.
> 
> Thanks,
> Konrad
> 
>>
>> Regards,
>> Kleber