RFCOMM not releasing properly

[Peyrega]

I'm using stock raspian updated to lastest version on a raspberry pi 2 (1GB) device, together with a Inateck BTA-CSR4B5 bluetooth adapter.
The goal is to do serial communication over bluetooth.
Device are paired and I can achieve the communication properly with commands :

sudo rfcomm listen rfcomm0 1

upon deconnection (e.g. from the remote side), issueing the same command again fails to allow the remote side to reconnect.
If I first type :

sudo hciconfig hci0 reset

and then reissue the rfcomm listen command, it starts working again...

However, from time to time, this stops working completely, I get an error message :

Can't bind RFCOMM socket: Address already in use

and the result of dmesg shows the following :

[ 82.996877] ------------[ cut here ]------------
[ 82.996931] WARNING: CPU: 0 PID: 32 at drivers/tty/tty_port.c:144 tty_port_destructor+0x80/0x88()
[ 82.996944] Modules linked in: rfcomm snd_bcm2835 snd_pcm snd_seq snd_seq_device snd_timer snd ecb btusb evdev joydev bluetooth rfkill uio_pdrv_genirq uio
[ 82.997044] CPU: 0 PID: 32 Comm: kworker/0:1 Not tainted 3.18.8-v7+ #764
[ 82.997066] Workqueue: events release_one_tty
[ 82.997111] [<80016070>] (unwind_backtrace) from [<80012938>] (show_stack+0x20/0x24)
[ 82.997142] [<80012938>] (show_stack) from [<8051b1f0>] (dump_stack+0x98/0xd8)
[ 82.997216] [<8051b1f0>] (dump_stack) from [<800243e8>] (warn_slowpath_common+0x7c/0xa0)
[ 82.997246] [<800243e8>] (warn_slowpath_common) from [<800244c8>] (warn_slowpath_null+0x2c/0x34)
[ 82.997277] [<800244c8>] (warn_slowpath_null) from [<80322b48>] (tty_port_destructor+0x80/0x88)
[ 82.997335] [<80322b48>] (tty_port_destructor) from [<80322b9c>] (tty_port_put+0x4c/0x50)
[ 82.997395] [<80322b9c>] (tty_port_put) from [<7f159f10>] (rfcomm_tty_cleanup+0x5c/0x60 [rfcomm])
[ 82.997452] [<7f159f10>] (rfcomm_tty_cleanup [rfcomm]) from [<80318fd0>] (release_one_tty+0x3c/0xa0)
[ 82.997482] [<80318fd0>] (release_one_tty) from [<80039a50>] (process_one_work+0x138/0x408)
[ 82.997510] [<80039a50>] (process_one_work) from [<8003a494>] (worker_thread+0x4c/0x4d0)
[ 82.997535] [<8003a494>] (worker_thread) from [<8003ebc4>] (kthread+0xe0/0xfc)
[ 82.997586] [<8003ebc4>] (kthread) from [<8000eb88>] (ret_from_fork+0x14/0x20)
[ 82.997603] ---[ end trace 0d5fd4688197372d ]---

which looks like there is a crash in some destuctor (old C++ souvenirs...)

rfcomm -a

shows

rfcomm0: 00:1A:7D:DA:71:13 -> 40:2C:F4:DB:C9:73 channel 1 closed [reuse-dlc release-on-hup ]

This is quite reproducible, and while googling around, I found a few links that show similar things :

https://bugzilla.redhat.com/show_bug.cgi?id=1075265

I'm quite new to bug repports and be happy to provide missing information that could help spotting and solving this.
Regards

[Peyrega (mathieu-peyrega-d)]

it seems that a proposed patch (not sure this is related, but I would think so)

https://groups.google.com/forum/#!topic ... GQKKO37epg

is not applied in current raspian version :

https://github.com/raspberrypi/linux/bl ... tty_port.c
where the reported (supposed ?) buggy source code is the unpatched version

[Jim Sievert (james-sievert)]

I also found that an rfcomm device was not properly released. I see the following in dmesg:

Nov 4 17:24:22 live kernel: [75732.976759] ------------[ cut here ]------------
Nov 4 17:24:22 live kernel: [75732.976793] WARNING: CPU: 2 PID: 47 at /build/linux-rwKq_7/linux-3.13.0/drivers/tty/tty_port.c:145 tty_port_destructor+0x83/0x90()
Nov 4 17:24:22 live kernel: [75732.976803] Modules linked in: ipt_MASQUERADE iptable_nat nf_nat_ipv4 nf_nat ipt_REJECT xt_CHECKSUM iptable_mangle bridge stp llc ip6table_filter ip6_tables ebtable_nat ebtables xt_owner nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack xt_tcpudp iptable_filter ip_tables x_tables autofs4 dm_crypt arc4 iwlmvm(OX) mac80211(OX) snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm dm_multipath intel_rapl kvm_intel scsi_dh snd_page_alloc kvm snd_seq_midi snd_seq_midi_event snd_rawmidi crct10dif_pclmul crc32_pclmul snd_seq ghash_clmulni_intel cryptd snd_seq_device snd_timer iwlwifi(OX) snd btusb(OX) rfcomm(OX) btbcm(OX) btintel(OX) bnep(OX) cfg80211(OX) cmac soundcore shpchp bluetooth(OX) compat(OX) i2c_scmi parport_pc ppdev lp mac_hid parport i2c_dev coretemp nfsd auth_rpcgss nfs_acl nfs lockd sunrpc fscache squashfs overlayfs nls_iso8859_1 hid_generic usbhid hid usb_storage igb dca ahci ptp i915(OX) pps_core sdhci_pci drm_mipi_dsi(OX) i2c_algo_bit drm_kms_helper(OX) sdhci libahci drm(OX) video(OX)
Nov 4 17:24:22 live kernel: [75732.977143] CPU: 2 PID: 47 Comm: kworker/2:1 Tainted: G OX 3.13.0-54-generic #0
Nov 4 17:24:22 live kernel: [75732.977154] Hardware name: N/A N/A/COMe-cBTi6R, BIOS BSCIA_PLX041.105.E38 X64 07/28/2015
Nov 4 17:24:22 live kernel: [75732.977173] Workqueue: events release_one_tty
Nov 4 17:24:22 live kernel: [75732.977184] 0000000000000009 ffff880273bb7d60 ffffffff817232a4 0000000000000000
Nov 4 17:24:22 live kernel: [75732.977207] ffff880273bb7d98 ffffffff8106784d ffff8801f438c56c ffff8801f438c400
Nov 4 17:24:22 live kernel: [75732.977227] ffff88027281bc00 0000000000000000 0000000000000080 ffff880273bb7da8
Nov 4 17:24:22 live kernel: [75732.977248] Call Trace:
Nov 4 17:24:22 live kernel: [75732.977273] [<ffffffff817232a4>] dump_stack+0x45/0x56
Nov 4 17:24:22 live kernel: [75732.977295] [<ffffffff8106784d>] warn_slowpath_common+0x7d/0xa0
Nov 4 17:24:22 live kernel: [75732.977314] [<ffffffff8106792a>] warn_slowpath_null+0x1a/0x20
Nov 4 17:24:22 live kernel: [75732.977333] [<ffffffff81455d33>] tty_port_destructor+0x83/0x90
Nov 4 17:24:22 live kernel: [75732.977351] [<ffffffff8145664c>] tty_port_put+0x2c/0x30
Nov 4 17:24:22 live kernel: [75732.977396] [<ffffffffa04460c8>] rfcomm_tty_cleanup+0x68/0x70 [rfcomm]
Nov 4 17:24:22 live kernel: [75732.977417] [<ffffffff8144b4d4>] release_one_tty+0x34/0xc0
Nov 4 17:24:22 live kernel: [75732.977439] [<ffffffff81083b22>] process_one_work+0x182/0x450
Nov 4 17:24:22 live kernel: [75732.977459] [<ffffffff81084911>] worker_thread+0x121/0x410
Nov 4 17:24:22 live kernel: [75732.977479] [<ffffffff810847f0>] ? rescuer_thread+0x430/0x430
Nov 4 17:24:22 live kernel: [75732.977498] [<ffffffff8108b702>] kthread+0xd2/0xf0
Nov 4 17:24:22 live kernel: [75732.977518] [<ffffffff...