On Sat, 21 Apr 2012, Dan Carpenter wrote:

> On Fri, Apr 20, 2012 at 09:14:44PM +0300, Kalle Valo wrote:
>> Dan Carpenter <dan.carpenter@oracle.com> writes:
>>
>>> On Fri, Apr 20, 2012 at 06:57:00PM +1000, Julian Calaby wrote:
>>>>> -       struct mib_local *m = kmalloc(sizeof(struct mib_phy), GFP_KERNEL);
>>>>> +       struct mib_local *m = kmalloc(sizeof(struct mib_local), GFP_KERNEL);
>>>>
>>>> Would it be better practice to use sizeof(*m)?
>>>>
>>>
>>> That was my temptation as well...  But I decided to make it match
>>> with the surrounding code.  I'm happy to resend if people want.
>>
>> IMHO sizeof(*m) is better and I tend to use it.
>>
>> Related to this: I have a bad habit of sometimes dropping '*' from
>> sizeof()? Is there a tool which could spot that?
>>
>
> That's what I was working on for Smatch when I sent this patch.
>
> The odd thing is that I can't find any bugs like this in the kernel.
> If sizeof(foo) is less than sizeof(*foo), which is probably the
> normal case, then these get caught early on in testing.
>
> Still I think people must have done manual audits as well...  It
> feels too clean to be natural.

Looking for x = ... sizeof(x) ... I get 9 reports.  In most cases it looks 
like sizeof(x) is coincidentally the same as the size that is wanted.  Two 
cases that look like they could have some noticible effect are:

arch/xtensa/platforms/iss/network.c, line 789
drivers/block/cciss.c, line 4211

I will send patches for those two.

julia