linux-kernel - Re: [PATCH v9] mmc: Export host capabilities to debugfs.

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]

Date:   Mon, 12 Mar 2018 19:04:17 +0530
From:   Harish Jenny K N <harish_kandiga@...tor.com>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Avri Altman <Avri.Altman@....com>,
        "ulf.hansson@...aro.org" <ulf.hansson@...aro.org>,
        "linus.walleij@...aro.org" <linus.walleij@...aro.org>,
        "adrian.hunter@...el.com" <adrian.hunter@...el.com>,
        "shawn.lin@...k-chips.com" <shawn.lin@...k-chips.com>
CC:     "linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "Vladimir_Zapolskiy@...tor.com" <Vladimir_Zapolskiy@...tor.com>
Subject: Re: [PATCH v9] mmc: Export host capabilities to debugfs.



On Monday 12 March 2018 06:33 PM, Andy Shevchenko wrote:
> On Mon, 2018-03-12 at 18:00 +0530, Harish Jenny K N wrote:
>
>
>>>>>>> +	for_each_set_bit(bit, (const unsigned long *)&caps,
>>>>>>> BITS_PER_LONG)
>>>>>>> +	for_each_set_bit(bit, (const unsigned long
>>>>>>> *)&caps2,
>>>>>>> BITS_PER_LONG)
>>>>>> Explicit casting is not needed anymore in both cases.
>>>>> Also maybe use sizeof(mmc_host_capabilities) instead of
>>>>> BITS_PER_LONG?
>>>> You mean sizeof(caps) and not sizeof(mmc_host_capabilities) .
>>>> Right ?
>>> meant ARRAY_SIZE(mmc_host_capabilities)
>> ARRAY_SIZE(mmc_host_capabilities) will be 32 and this will be my old
>> change for which I got a comment from Andy Shevchenko asking me to
>> replace u32 with unsigned long.
>>
>> This is the old comment:
>>
>>>> +    int size = sizeof(u32) * BITS_PER_BYTE;
>>> This is redundant. Use BITS_PER_LONG (why's that, see below) in the
> There is nothing about ARRAY_SIZE().
> ARRAY_SIZE() will work quite good as well.

Yes. Definitely it works. I was only mentioning about changing u32 to unsigned long for caps and passing BITS_PER_LONG  as size in for_each_set_bit.
>
>> for_each_set_bit().
>>
>>>> +    for_each_set_bit(bit, (const unsigned long *)&caps, size)
>>>> +    for_each_set_bit(bit, (const unsigned long *)&caps2, size)
>>> These are UB cases.
>>> Fix is simple, replace u32 by unsigned long in (1) above.
>> Note: Without typecasting &caps to(const unsigned long *) will give
>> compilation error in this case.
> What kind of?
>

Sorry for I was not clear that the compilation error is for the following case
> u32 caps = host->caps;
> for_each_set_bit(bit, &caps, ARRAY_SIZE(mmc_host_capabilities))


example: compilation for arm gives this error for the following lines:

drivers/mmc/core/debugfs.c: In function 'mmc_caps_show':
./include/linux/bitops.h:41:30: error: passing argument 1 of '_find_first_bit_le' from incompatible pointer type [-Werror=incompatible-pointer-types]
  for ((bit) = find_first_bit((addr), (size));  \
                              ^
./arch/arm/include/asm/bitops.h:202:50: note: in definition of macro 'find_first_bit'
 #define find_first_bit(p,sz)  _find_first_bit_le(p,sz)
                                                  ^
drivers/mmc/core/debugfs.c:318:2: note: in expansion of macro 'for_each_set_bit'
  for_each_set_bit(bit, &caps, ARRAY_SIZE(mmc_host_capabilities))
  ^
./arch/arm/include/asm/bitops.h:165:12: note: expected 'const long unsigned int *' but argument is of type 'u32 * {aka unsigned int *}'
 extern int _find_first_bit_le(const unsigned long *p, unsigned size);
            ^





Thanks,
Harish Jenny K N