linux-kernel - Re: [PATCH] iio: common: scmi_iio: use kcalloc() instead of kzalloc()

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: <48b58e81-9d1e-497c-abcd-84deaed98ab1@vivo.com>
Date: Tue, 19 Aug 2025 20:00:03 +0800
From: Qianfeng Rong <rongqianfeng@...o.com>
To: Andy Shevchenko <andy.shevchenko@...il.com>
Cc: Jyoti Bhayana <jbhayana@...gle.com>, Jonathan Cameron <jic23@...nel.org>,
 David Lechner <dlechner@...libre.com>, Nuno Sá
 <nuno.sa@...log.com>, Andy Shevchenko <andy@...nel.org>,
 linux-iio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iio: common: scmi_iio: use kcalloc() instead of kzalloc()


在 2025/8/19 18:13, Andy Shevchenko 写道:
> [You don't often get email from andy.shevchenko@...il.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
>
> On Tue, Aug 19, 2025 at 1:08 PM Qianfeng Rong <rongqianfeng@...o.com> wrote:
>> 在 2025/8/19 17:46, Andy Shevchenko 写道:
>>> On Tue, Aug 19, 2025 at 11:56 AM Qianfeng Rong <rongqianfeng@...o.com> wrote:
> ...
>
>>> While this change is correct...
>>>>           sensor->freq_avail =
>>>> -               devm_kzalloc(&iio_dev->dev,
>>>> -                            sizeof(*sensor->freq_avail) *
>>>> -                                    (sensor->sensor_info->intervals.count * 2),
>>>> +               devm_kcalloc(&iio_dev->dev,
>>>> +                            sensor->sensor_info->intervals.count * 2,
>>> ...I would also switch this to use array_size() instead of explicit
>>> multiplication as it will check for boundaries that are not static in
>>> this case.
>> I don't understand what "will check for boundaries that are not static in
>> this case" means. Could you explain it to me?
> intervals.count may be anything and of any type. Compiler may or may
> not proof that it holds the value less than size_t / 2 (which may be
> == int / 2 on 32-bit platforms). That's why it's better to use
> array_size(intervals.count, 2),
Thanks for taking the time to explain! The devm_kmalloc_array() function
implements a check function similar to array_size() by default.


devm_kmalloc_array():

static inline void *devm_kmalloc_array(struct device *dev, size_t n, 
size_t size, gfp_t flags)
{
     size_t bytes;

     if (unlikely(check_mul_overflow(n, size, &bytes)))
         return NULL;

     return devm_kmalloc(dev, bytes, flags);
}
static inline void *devm_kcalloc(struct device *dev, size_t n, size_t 
size, gfp_t flags)
{
     return devm_kmalloc_array(dev, n, size, flags | __GFP_ZERO);
}


array_size():

static inline size_t __must_check size_mul(size_t factor1, size_t factor2)
{
     size_t bytes;

     if (check_mul_overflow(factor1, factor2, &bytes))
         return SIZE_MAX;

     return bytes;
}

#define array_size(a, b)    size_mul(a, b)

>
>> I've experimented with the following command and found that kmalloc_array()
>> generates fewer instructions than kmalloc(array_size()):
>> objdump -dSl --prefix-addresses <changed module>.o
> It's about the second parameter in devm_kcalloc(), and of course it
> may generate less instructions but it's irrelevant to my comment.
>
>
Best regards,
Qianfeng