Bug 213923 - Bluetooth: AC9560 doesn't respond any more after suspend multiple times
Summary: Bluetooth: AC9560 doesn't respond any more after suspend multiple times
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Bluetooth (show other bugs)
Hardware: All Linux
: P1 high
Assignee: linux-bluetooth@vger.kernel.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2021-07-30 05:37 UTC by KobaKo
Modified: 2021-08-25 02:22 UTC (History)
2 users (show)

See Also:
Kernel Version: v5.14-rc3
Subsystem:
Regression: No
Bisected commit-id:


Attachments
lsusb (1.14 KB, text/plain)
2021-07-30 05:37 UTC, KobaKo
Details
lspci (28.65 KB, text/plain)
2021-07-30 05:38 UTC, KobaKo
Details
dmesg_20210802 (100.49 KB, text/plain)
2021-08-02 08:28 UTC, KobaKo
Details
btmon_20210802 (675 bytes, text/plain)
2021-08-02 09:06 UTC, KobaKo
Details
syslog_20210820 (589.72 KB, text/plain)
2021-08-20 01:13 UTC, KobaKo
Details
syslog_20210825 (1.56 MB, text/plain)
2021-08-25 02:18 UTC, KobaKo
Details

Description KobaKo 2021-07-30 05:37:50 UTC
Created attachment 298119 [details]
lsusb

1. Turn BT off.
2. suspend(deep) and resume 10times.
3. can't open BT again
#[failed]rfkill unblock bluetooth
#[failed]sudo hciconfig hci0 up.
4. status of hci is always down.
#sudo hciconfig -a

Once issue appeare, get these errors
[ 877.194032] Bluetooth: hci0: urb 0000000061b9a002 failed to resubmit (113)
[ 886.941327] Bluetooth: hci0: Failed to read MSFT supported features (-110)

it's related to mstf feature.
1. Remove MSTF from configuration, issue wouldn't be triggered.
2. revert this patch, issue wouldn't be observed.
4a37682c6b59c8888acf93117362d761c5923a69 Bluetooth: advmon offload MSFT handle
Comment 1 KobaKo 2021-07-30 05:38:14 UTC
Created attachment 298121 [details]
lspci
Comment 2 KobaKo 2021-08-02 08:28:06 UTC
Created attachment 298149 [details]
dmesg_20210802
Comment 3 KobaKo 2021-08-02 09:06:38 UTC
Created attachment 298151 [details]
btmon_20210802
Comment 4 KobaKo 2021-08-06 07:18:06 UTC
Here's a full btmon log during reproducing issue.
https://drive.google.com/file/d/1wSn1wugdp73ntt0tsYoxfdu-SRgumECv/view?usp=sharing

btmon_20210802 is only contained btmon log after turn on bt.
Comment 5 Chethan T N 2021-08-18 15:28:15 UTC
Would you please help to provide more details on the test setup and test case.

Setup details:
Linux distribution : Ubuntu/Fedora  ??
kernel version: 5.xx??
Model name: ??

From the logs shared looks like one every suspend/resume Bluetooth device is getting enumerated and FW download sequence is initiated.

May i know why the above behavior ?

And also noticed that in most of the suspend/resume flow during the FW download sequence itself the system is entering the suspend state.

//System entering SUSPEND State 
37213	2021-08-06 12:39:52.879966	host	controller	252				Sent Intel Secure Send

//System RESUME state
37222	2021-08-06 12:40:16.529947	host	controller	3				Sent Intel Read Version 

May i know why the above behavior?
Comment 6 KobaKo 2021-08-18 17:08:37 UTC
@Chethan,
Im using Ubuntu to reproduce on JSL platform.
the issue could be observed since 5.14rc3 only with AC9560.
With ax210, there's no issue.

The suspend&resume is a normal behavior when you don't use notebook any more and  suspend notebook for the power saving.
User wouldn't know if bt fw is downloading when user want to suspend the his notebook.

I just use a scripts to run suspend&resume automatically.
#for i in $(seq 10);do sleep 10;sudo rtcwake -m mem -s 10;done

you also can suspend by pushing suspend button manually.

Could you share what's correct hw_variant of ac9560?
0x11 or 0x13?
Comment 7 KobaKo 2021-08-20 01:13:28 UTC
Created attachment 298375 [details]
syslog_20210820
Comment 8 KobaKo 2021-08-25 02:18:04 UTC
Created attachment 298461 [details]
syslog_20210825
Comment 9 KobaKo 2021-08-25 02:22:19 UTC
Comment on attachment 298461 [details]
syslog_20210825

this is log that reproduced the issue

Note You need to log in before you can comment on or make changes to this bug.