Re: [PATCH] lib: Fix bug in int_sqrt() for 64 bit longs

[Peter Williams <pwil3058@bigpond.net.au>]

Eric Dumazet wrote:
> Peter Williams a écrit :
> 
>> The implementation of int_sqrt() assumes that longs have 32 bits.  On 
>> systems that have 64 bit longs this will result in gross errors when 
>> the argument to the function is greater than 2^32 - 1 on such systems. 
>> I doubt whether any such use is currently made of int_sqrt() but the 
>> attached patch fixes the problem anyway.
>>
>> Signed-off-by: Peter Williams <pwil3058@bigpond.com.au>
>>
>>
>> ------------------------------------------------------------------------
>>
>> Index: GIT-warnings/lib/int_sqrt.c
>> ===================================================================
>> --- GIT-warnings.orig/lib/int_sqrt.c    2005-10-25 13:55:22.000000000 
>> +1000
>> +++ GIT-warnings/lib/int_sqrt.c    2006-01-06 14:29:19.000000000 +1100
>> @@ -15,7 +15,7 @@ unsigned long int_sqrt(unsigned long x)
>>      op = x;
>>      res = 0;
>>  
>> -    one = 1 << 30;
>> +    one = 1 << (BITS_PER_LONG - 2);
>>      while (one > op)
>>          one >>= 2;
>>  
> 
> 
> Are you sure it works ?
> 
> I would have writen :
> 
> one = 1L << (BITS_PER_LONG - 2);

I think you're right as just using 1 would make the expression an 
integer variable and the compiler would complain about shifting too far 
unless integers were the same size as longs.  (I don't have a 64 bit 
system to confirm that on which is why it snuck by me. :-()

Thanks
Peter
-- 
Peter Williams                                   pwil3058@bigpond.net.au

"Learning, n. The kind of ignorance distinguishing the studious."
  -- Ambrose Bierce