linux-kernel - Re: [PATCH v2 1/2] iio: gyro: Add driver support for ADXRS290

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: <5cb55101-af5c-b6a2-d770-9717f8a463cc@gmail.com>
Date:   Wed, 22 Jul 2020 15:10:06 +0530
From:   Nishant Malpani <nish.malpani25@...il.com>
To:     Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     Jonathan Cameron <jic23@...nel.org>,
        "Bogdan, Dragos" <dragos.bogdan@...log.com>,
        darius.berghe@...log.com,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-iio <linux-iio@...r.kernel.org>
Subject: Re: [PATCH v2 1/2] iio: gyro: Add driver support for ADXRS290


On 22/07/20 3:08 am, Andy Shevchenko wrote:
> On Tue, Jul 21, 2020 at 11:35 PM Nishant Malpani
> <nish.malpani25@...il.com> wrote:
>> On 22/07/20 1:16 am, Andy Shevchenko wrote:
>>> On Tue, Jul 21, 2020 at 9:20 PM Nishant Malpani
>>> <nish.malpani25@...il.com> wrote:
> 
> ...
> 
>>>> +               *vals = (const int *)adxrs290_lpf_3db_freq_tbl;
>>>
>>> Why casting?
>>>
>> adxrs290_lpf_3db_freq_tbl is of type (int *)[2], right? Without the
>> casting, an incompatible-pointer-type error is thrown.
>>
>>> ...
>>>
>>>> +               *vals = (const int *)adxrs290_hpf_3db_freq_tbl;
>>>
>>> Ditto.
>>>
>> See above comment.
> 
> Can't you declare table as const int?
> 
I'm not sure I understand you completely here; do you mean const int *? 
So, an array of alternate integer and fractional parts? I suppose that's 
possible but we'd be introducing unwanted complexity I feel - for 
example, currently the index of the 3db frequency in the table is used 
to directly map & set bits in the filter register corresponding to that 
frequency but with the approach you share, we'd have to apply a 
transformation (div by 2) to set the same bits in the filter register. 
Do you think the added complexity justifies the removal of the casting?

> ...
> 
>>>> +       /* max transition time to measurement mode */
>>>> +       msleep_interruptible(ADXRS290_MAX_TRANSITION_TIME_MS);
>>>
>>> I'm not sure what the point of interruptible variant here?
>>>
>> I referred Documentation/timers/timers-howto.rst for this.
>> My reasoning was shaped to use the interruptible variant because the
>> transition settles in a time *less than* 100ms and since 100ms is quite
>> a huge time to sleep, it should be interrupted in case a signal arrives.
> 
> This is probe of the device,
> What are the expectations here?
> 
I fail to understand why this can't be used in the probe() but perhaps 
in a routine to standby/resume. Could you please elaborate?

With regards,
Nishant Malpani