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

[Konrad Zapałowicz]

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?

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