This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Why was the reallocarray function not added to glibc?


On 10/04/17 16:24, Zack Weinberg wrote:
> On Mon, Apr 10, 2017 at 6:15 AM, Szabolcs Nagy <szabolcs.nagy@arm.com> wrote:
>>
>> the idiomatic unsigned mul overflow check is
>>
>>   if (b && -1/b < a)
>>     return 0;
>>   c = a*b;
> 
> I've never seen this construct before; it is not obvious to me why it
> checks for unsigned mul overflow.  It would need, at the very least,
> an explanation in comments.  Also, this...
> 
> #include <stdlib.h>
> 
> void *reallocarray(void *ptr, size_t size, size_t nmemb)
> {
>     if (!size || !nmemb) {
>         free(ptr);
>         return 0;
>     }
>     if (-1/nmemb < size) {
>         return 0;
>     }
>     return realloc(ptr, size * nmemb);
> }
> 
> ... generates code with a divide instruction in it, using either gcc
> 6.3.0 or clang 3.8.1, so I do not think we can use it yet.

gcc-7 gets it right though (single mul)
and this code is portable with no hacks.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]