Opened 4 years ago

Last modified 15 months ago

#15725 new bug

[hci] PANIC: ACL frame corrupted

Reported by: grexe Owned by: oruizdorantes
Priority: normal Milestone: Unscheduled
Component: Network & Internet/Bluetooth Version: R1/Development
Keywords: Cc:
Blocked By: Blocking: #17178
Platform: All

Description

Using hrev53888, Bluetooth now works on my laptop to the point where it detects my OnePlus 6T, but when trying to pair the device, it crashes.

Lenovo V510-15IKB
PCI, Dual Band Wireless-AC 3165 Plus Bluetooth controller

Due to lack of serial deubugging, I hope it's fine to attach a screenshot...

Attachments (1)

IMG_20200219_130537.jpg (2.8 MB ) - added by grexe 4 years ago.
KDL of Bluetooth driver

Change History (8)

by grexe, 4 years ago

Attachment: IMG_20200219_130537.jpg added

KDL of Bluetooth driver

comment:1 by diver, 4 years ago

Summary: KDL on Bluetooth pairing device[hci] PANIC: ACL frame corrupted

comment:2 by waddlesplash, 4 years ago

Component: DriversNetwork & Internet/Bluetooth
Owner: changed from nobody to oruizdorantes

comment:3 by waddlesplash, 4 years ago

Since your WiFi did not get initialized, it may well be the case that this is caused by that (they live on the same chip.)

However, our Bluetooth code is not very mature, and unfortunately there is nobody around who is an expert in it at the moment, so this is unlikely to get fixed either way in the near future (unless you would care to contribute :)

comment:4 by grexe, 4 years ago

ok thanks, although that is what I feared;) I would certainly care, but I am not a driver developer and certainly not a Bluetooth expert, sadly. I am willing to test out nightly code though if anybody wants to give it a shot.

comment:5 by oruizdorantes, 4 years ago

Hi there, What is the bluetooth version of the device? I think it is shown in preferences or with $ hciconfig.

Searching where I placed that panic, it is related to the size of the packet. Then this device is either something that is way beyond bluetooth 2.0. or not filling properly some size fields.

Second option is that i see missing some code for re-assembling ACL packets(which I see present int BT_EVENT packet type). Looks like for ACL I assumed the whole packet would come in a single bucket, and it is not being the case here...

comment:6 by diver, 3 years ago

Blocking: 17178 added

comment:7 by waddlesplash, 15 months ago

Does this still happen?

Note: See TracTickets for help on using tickets.