Timeout waiting for GATT connection to be created |
|||||||||||||||||||||
Issue descriptionWe frequently see a timeout creating a GATT connection during connection attempts. When GATT connections are created successfully, it usually takes about 2-4 seconds. Currently, our timeout is 15 seconds (see [1]), so this should be far more than enough. When this fails, you'll see an error log: "Timed out waiting during SubStatus [waiting for GATT connection to be created]. Destroying connection." [1] https://cs.chromium.org/chromium/src/components/cryptauth/ble/bluetooth_low_energy_weave_client_connection.cc?q=kGattConnectionTimeoutSeconds
,
Oct 3 2017
Yes - it seems to be unrelated to advertising.
,
Oct 10 2017
We should try to get this into M-62. I'm seeing it fairly often.
,
Oct 11 2017
,
Oct 11 2017
I investigated this and found out that there is some kind of race condition in bluetoothd, where if Connect() is being called but the kernel (for some mysterious reason) issues a DEVICE_DISCONNECTED event, bluetoothd would forget to reply the client and the client (Chrome) may reach a bad state so even retrying the connection would always fail again. I have a proposed fix at http://crosreview.com/711374. This will at least make bluetoothd reply that the connection failed rather than doing nothing and timeout. The occasional connection error will still be there, but at least it won't timeout and reach a bad state so the retry will likely be successful.
,
Oct 11 2017
,
Oct 11 2017
Sonny wrote up a CL which he says fixes this bug: https://chromium-review.googlesource.com/c/chromiumos/third_party/bluez/+/711374
,
Oct 11 2017
comment 6, 7, and 8 are actually meant for issue 773062.
,
Oct 13 2017
Attaching logs for an isolated repro (without tether) of this case.
,
Oct 16 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/bluez/+/9fca8b3288156fb64cbcdd7a8ad56cd5c50a3926 commit 9fca8b3288156fb64cbcdd7a8ad56cd5c50a3926 Author: Sonny Sasaka <sonnysasaka@chromium.org> Date: Mon Oct 16 21:14:02 2017 CHROMIUM: Fix race condition between device connection and disconnection When Connect() is called and waiting for return, dev_disconnected may be called due to MGMT_EV_DEVICE_DISCONNECTED event from kernel. In that case reply to client that the connection failed otherwise the dbus method will timeout because bluetoothd never replies. BUG= chromium:771016 TEST=Simulate a lot of Connect() to bluetooth devices and check that error is returned from bluetoothd rather than dbus timeout when this race condition happens. Change-Id: Ica16f54b793986cd17cb450a9a1d15935d6f5386 Reviewed-on: https://chromium-review.googlesource.com/711374 Commit-Ready: Sonny Sasaka <sonnysasaka@chromium.org> Tested-by: Sonny Sasaka <sonnysasaka@chromium.org> Reviewed-by: Rahul Chaturvedi <rkc@chromium.org> Reviewed-by: Miao-chen Chou <mcchou@chromium.org> [modify] https://crrev.com/9fca8b3288156fb64cbcdd7a8ad56cd5c50a3926/src/device.c
,
Oct 16 2017
,
Oct 16 2017
This bug requires manual review: We are only 0 days from stable. Please contact the milestone owner if you have questions. Owners: amineer@(Android), cmasso@(iOS), bhthompson@(ChromeOS), abdulsyed@(Desktop) For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Oct 16 2017
Consider this merge approved for 62 once it has been verified on a ToT canary build and merged into 63.
,
Oct 17 2017
Your change meets the bar and is auto-approved for M63. Please go ahead and merge the CL to branch 3239 manually. Please contact milestone owner if you have questions. Owners: cmasso@(Android), cmasso@(iOS), gkihumba@(ChromeOS), govind@(Desktop) For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Oct 19 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/bluez/+/9fbe131ad414fb23672f29379047d59d387a9267 commit 9fbe131ad414fb23672f29379047d59d387a9267 Author: Sonny Sasaka <sonnysasaka@chromium.org> Date: Thu Oct 19 17:38:42 2017 CHROMIUM: Fix race condition between device connection and disconnection When Connect() is called and waiting for return, dev_disconnected may be called due to MGMT_EV_DEVICE_DISCONNECTED event from kernel. In that case reply to client that the connection failed otherwise the dbus method will timeout because bluetoothd never replies. BUG= chromium:771016 TEST=Simulate a lot of Connect() to bluetooth devices and check that error is returned from bluetoothd rather than dbus timeout when this race condition happens. Change-Id: Ica16f54b793986cd17cb450a9a1d15935d6f5386 Reviewed-on: https://chromium-review.googlesource.com/711374 Commit-Ready: Sonny Sasaka <sonnysasaka@chromium.org> Tested-by: Sonny Sasaka <sonnysasaka@chromium.org> Reviewed-by: Rahul Chaturvedi <rkc@chromium.org> Reviewed-by: Miao-chen Chou <mcchou@chromium.org> (cherry picked from commit 9fca8b3288156fb64cbcdd7a8ad56cd5c50a3926) Reviewed-on: https://chromium-review.googlesource.com/724639 Trybot-Ready: Miao-chen Chou <mcchou@chromium.org> Commit-Queue: Miao-chen Chou <mcchou@chromium.org> [modify] https://crrev.com/9fbe131ad414fb23672f29379047d59d387a9267/src/device.c
,
Oct 23 2017
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Oct 24 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/bluez/+/b6a3b370d7057f67677ff5d1b2876c5352d3a01a commit b6a3b370d7057f67677ff5d1b2876c5352d3a01a Author: Sonny Sasaka <sonnysasaka@chromium.org> Date: Tue Oct 24 17:55:08 2017 CHROMIUM: Fix race condition between device connection and disconnection When Connect() is called and waiting for return, dev_disconnected may be called due to MGMT_EV_DEVICE_DISCONNECTED event from kernel. In that case reply to client that the connection failed otherwise the dbus method will timeout because bluetoothd never replies. BUG= chromium:771016 TEST=Simulate a lot of Connect() to bluetooth devices and check that error is returned from bluetoothd rather than dbus timeout when this race condition happens. Change-Id: Ica16f54b793986cd17cb450a9a1d15935d6f5386 Reviewed-on: https://chromium-review.googlesource.com/711374 Commit-Ready: Sonny Sasaka <sonnysasaka@chromium.org> Tested-by: Sonny Sasaka <sonnysasaka@chromium.org> Reviewed-by: Rahul Chaturvedi <rkc@chromium.org> Reviewed-by: Miao-chen Chou <mcchou@chromium.org> (cherry picked from commit 9fca8b3288156fb64cbcdd7a8ad56cd5c50a3926) Reviewed-on: https://chromium-review.googlesource.com/733940 Reviewed-by: Sonny Sasaka <sonnysasaka@chromium.org> Commit-Queue: Sonny Sasaka <sonnysasaka@chromium.org> [modify] https://crrev.com/b6a3b370d7057f67677ff5d1b2876c5352d3a01a/src/device.c
,
Oct 24 2017
,
Oct 27 2017
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Sep 17
|
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by sonnysasaka@chromium.org
, Oct 3 2017