Return-Path: <cyrus@holtmann.org>
MIME-Version: 1.0
In-Reply-To: <1611692.3irn0nrenv@leonov>
References: <BD5E516FF3711A41B93DC005170007BB010E5DB57E78@EMBSVR01.ember.local>
	<1831684.talzOfVBnx@leonov>
	<BD5E516FF3711A41B93DC005170007BB010E5DB57E91@EMBSVR01.ember.local>
	<1611692.3irn0nrenv@leonov>
Date: Thu, 28 May 2015 12:41:14 +0300
Message-ID: <CABBYNZLwhyhhi=r-YrrKD92fdteQoh7qtUkaU-p7dmzR-XXruQ@mail.gmail.com>
Subject: Re: Issue: Sending ATT Signed Write Command via btgatt-client
From: Luiz Augusto von Dentz <luiz.dentz@gmail.com>
To: Szymon Janc <szymon.janc@tieto.com>
Cc: Ankur Patel <ankur@embertec.com>,
	"linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>
Content-Type: text/plain; charset=UTF-8
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi,

On Thu, May 28, 2015 at 12:15 PM, Szymon Janc <szymon.janc@tieto.com> wrote:
> Hi Ankur,
>
> On Thursday 28 of May 2015 18:07:48 Ankur Patel wrote:
>> > Hi Ankur,
>> >
>> > On Thursday 28 of May 2015 16:47:09 Ankur Patel wrote:
>> > > Hello Folks,
>> > >
>> > > I would like to Send Signed Write command to modify a characteristic
>> > > on the Host. It failed into the Error: Failed to initiate write
>> > > without response procedure
>> > >
>> > > Information on Controller (May be helpful):
>> > > root@imx28evk:~# cat /proc/crypto
>> > > name         : ecb(aes)
>> > > driver       : ecb-aes-dcp
>> > > module       : kernel
>> > > priority     : 400
>> > > refcnt       : 1
>> > > selftest     : passed
>> > > type         : ablkcipher
>> > > async        : yes
>> > > blocksize    : 16
>> > > min keysize  : 16
>> > > max keysize  : 32
>> > > ivsize       : 0
>> > > geniv        : <default>
>> > >
>> > > root@imx28evk:~# btmgmt info
>> > > Index list with 1 item
>> > > hci0:   Primary controller
>> > >
>> > >         addr 1C:BA:8C:86:81:21 version 6 manufacturer 13 class 0x000000
>> > >         supported settings: powered connectable fast-connectable
>> > >
>> > > discoverable bondable link-security ssp br/edr hs le advertising
>> > > debug-keys privacy current settings: powered connectable discoverable
>> > > bondable link-security br/edr le advertising name testbt
>> > >
>> > >         short name
>> > >
>> > > root@imx28evk:~/tools# btgatt-client -d 00:1B:DC:07:2E:85 Connecting
>> > > to device... Done Service Added - UUID:
>> > > 00001800-0000-1000-8000-00805f9b34fb start: 0x0001
>> > > end: 0x000d Service Added - UUID: 00001801-0000-1000-8000-00805f9b34fb
>> > > start: 0x0014 end: 0x0016 [GATT client]# GATT discovery procedures
>> > > complete [GATT client]# ...
>> > > [GATT client]# set-sign-key -c efcdab8967452301efcdab8967452301 [GATT
>> > > client]# write-value -s -w 0xd 01 Failed to initiate write without
>> > > response procedure
>> > >
>> > > I tried to trace the error as:
>> > > bt_gatt_client_write_without_response(..) -> bt_att_send(..) ->
>> > > create_att_send_op(..) -> encode_pdu(..) -> bt_crypto_sign_att(..) ->
>> > > As
>> > > att->crypto is NULL it returns false.
>> > >
>> > > I tried to check the reason for att->crypto to be NULL. Found that
>> > > while initialization of the btgatt-client, under bt_crypto_new() ->
>> > > crypto->ecb_aes = ecb_aes_setup();. It fails under ecb_aes_setup
>> >
>> > function.
>> >
>> > > From the output of /proc/crypto, the kernel has the support. I am
>> > > wondering how to enable the crypto?
>> >
>> > Most likely following options are missing in your kernel configuration:
>> > CONFIG_CRYPTO_CMAC
>> > CONFIG_CRYPTO_USER_API
>> > CONFIG_CRYPTO_USER_API_HASH
>> > CONFIG_CRYPTO_USER_API_SKCIPHER
>>
>> It Worked!!. Thank you Szymon for quick response.
>>
>> Now I am able to send the Signed Write Command but without Authentication
>> Signature. If I understood it correctly, 1) Set the Signature using  [GATT
>> client]# set-sign-key -c efcdab8967452301efcdab8967452301 2) Execute
>> command: write-value -s -w 0xd 01  to send write command.
>>
>>  I couldn't able to see the signature on the Sniffer.
>> In addition on the Controller I see below output of root@imx28evk:~#Btmon:
>>
>> < ACL Data TX: Handle 1025 flags 0x00 dlen 20
>>
>>
>>               [hci0] 2394.331806 ATT: Signed Write Command (0xd2) len 15
>>         0d 00 01 03 00 00 00 47 8a ec df 50 fc a5 ba     .......G...P...
>
> This looks OK to me.
> 0d 00    01     03 00 00 00 47 8a ec df 50 fc a5 ba
> handle   data   signature
>
>
> Check CoreSpec 4.2 Vol3, Part F, p. 3.4.5.4 (pdf page 2194) for details.

We could probably improve btmon to print out the handle, data and
signature separately.


-- 
Luiz Augusto von Dentz