linux-kernel - Re: [PATCH] Bluetooth: Simplify / fix return values from tk

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <6456552C-5910-4D77-9607-14D9D1FA38FD@holtmann.org>
Date:   Mon, 6 Apr 2020 14:06:10 +0200
From:   Marcel Holtmann <marcel@...tmann.org>
To:     Guenter Roeck <linux@...ck-us.net>
Cc:     Alain Michaud <alainmichaud@...gle.com>,
        Johan Hedberg <johan.hedberg@...il.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        BlueZ <linux-bluetooth@...r.kernel.org>,
        netdev <netdev@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Sonny Sasaka <sonnysasaka@...omium.org>
Subject: Re: [PATCH] Bluetooth: Simplify / fix return values from tk_request

Hi Guenter,

>>> Some static checker run by 0day reports a variableScope warning.
>>> 
>>> net/bluetooth/smp.c:870:6: warning:
>>>        The scope of the variable 'err' can be reduced. [variableScope]
>>> 
>>> There is no need for two separate variables holding return values.
>>> Stick with the existing variable. While at it, don't pre-initialize
>>> 'ret' because it is set in each code path.
>>> 
>>> tk_request() is supposed to return a negative error code on errors,
>>> not a bluetooth return code. The calling code converts the return
>>> value to SMP_UNSPECIFIED if needed.
>>> 
>>> Fixes: 92516cd97fd4 ("Bluetooth: Always request for user confirmation for Just Works")
>>> Cc: Sonny Sasaka <sonnysasaka@...omium.org>
>>> Signed-off-by: Guenter Roeck <linux@...ck-us.net>
>>> ---
>>> net/bluetooth/smp.c | 9 ++++-----
>>> 1 file changed, 4 insertions(+), 5 deletions(-)
>>> 
>>> diff --git a/net/bluetooth/smp.c b/net/bluetooth/smp.c
>>> index d0b695ee49f6..30e8626dd553 100644
>>> --- a/net/bluetooth/smp.c
>>> +++ b/net/bluetooth/smp.c
>>> @@ -854,8 +854,7 @@ static int tk_request(struct l2cap_conn *conn, u8 remote_oob, u8 auth,
>>>        struct l2cap_chan *chan = conn->smp;
>>>        struct smp_chan *smp = chan->data;
>>>        u32 passkey = 0;
>>> -       int ret = 0;
>>> -       int err;
>>> +       int ret;
>>> 
>>>        /* Initialize key for JUST WORKS */
>>>        memset(smp->tk, 0, sizeof(smp->tk));
>>> @@ -887,12 +886,12 @@ static int tk_request(struct l2cap_conn *conn, u8 remote_oob, u8 auth,
>>>        /* If Just Works, Continue with Zero TK and ask user-space for
>>>         * confirmation */
>>>        if (smp->method == JUST_WORKS) {
>>> -               err = mgmt_user_confirm_request(hcon->hdev, &hcon->dst,
>>> +               ret = mgmt_user_confirm_request(hcon->hdev, &hcon->dst,
>>>                                                hcon->type,
>>>                                                hcon->dst_type,
>>>                                                passkey, 1);
>>> -               if (err)
>>> -                       return SMP_UNSPECIFIED;
>>> +               if (ret)
>>> +                       return ret;
>> I think there may be some miss match between expected types of error
>> codes here.  The SMP error code type seems to be expected throughout
>> this code base, so this change would propagate a potential negative
>> value while the rest of the SMP protocol expects strictly positive
>> error codes.
>> 
> 
> Up to the patch introducing the SMP_UNSPECIFIED return value, tk_request()
> returned negative error codes, and all callers convert it to SMP_UNSPECIFIED.
> 
> If tk_request() is supposed to return SMP_UNSPECIFIED on error, it should
> be returned consistently, and its callers don't have to convert it again.

maybe we need to fix that initial patch then.

Regards

Marcel