kernelnewbies.kernelnewbies.org archive mirror
 help / color / mirror / Atom feed
* BUILD_BUG_ON_ZERO
@ 2011-02-25 10:50 Zhang Meng
  2011-02-25 16:33 ` BUILD_BUG_ON_ZERO Dave Hylands
  0 siblings, 1 reply; 3+ messages in thread
From: Zhang Meng @ 2011-02-25 10:50 UTC (permalink / raw)
  To: kernelnewbies

Hi ~List,

Could anybody explain the macro below? what does it mean?

Thanks in advance.

#define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))


-- 
Yours sincerely
ZhangMeng

^ permalink raw reply	[flat|nested] 3+ messages in thread

* BUILD_BUG_ON_ZERO
  2011-02-25 10:50 BUILD_BUG_ON_ZERO Zhang Meng
@ 2011-02-25 16:33 ` Dave Hylands
  2011-02-26  1:14   ` BUILD_BUG_ON_ZERO Meng Zhang
  0 siblings, 1 reply; 3+ messages in thread
From: Dave Hylands @ 2011-02-25 16:33 UTC (permalink / raw)
  To: kernelnewbies

Hi Zhang,

On Fri, Feb 25, 2011 at 3:50 AM, Zhang Meng <jammy.linux@gmail.com> wrote:
> Hi ~List,
>
> Could anybody explain the macro below? what does it mean?
>
> Thanks in advance.
>
> #define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))

This is also known as a compile time assert. I think that this
particular variant has to be used inside a function.

! is just negation and produces a zero or 1 result. !! just does it
twice, so that a non-zero value coming in becomes 1, and a zero value
remains as zero.

If e evaluates to false (zero) then -!!(e) evaluates to zero;
if e evaluates to true (non-zero) then -!!(e) evaluates to -1.

Declaring a bit field with a size of -1 will cause a compiler error.
I'm actually surprised that declaring a bitfield of size 0 works.

The typical declarations of this I've seen usually use arrays and
arrange for the size to be -1 or 1 (which is generally more portable).
When you use the array style declaration, you can use it outside a
function as well.

Dave Hylands

^ permalink raw reply	[flat|nested] 3+ messages in thread

* BUILD_BUG_ON_ZERO
  2011-02-25 16:33 ` BUILD_BUG_ON_ZERO Dave Hylands
@ 2011-02-26  1:14   ` Meng Zhang
  0 siblings, 0 replies; 3+ messages in thread
From: Meng Zhang @ 2011-02-26  1:14 UTC (permalink / raw)
  To: kernelnewbies

On Sat, Feb 26, 2011 at 12:33 AM, Dave Hylands <dhylands@gmail.com> wrote:
> Hi Zhang,
>
> On Fri, Feb 25, 2011 at 3:50 AM, Zhang Meng <jammy.linux@gmail.com> wrote:
>> Hi ~List,
>>
>> Could anybody explain the macro below? what does it mean?
>>
>> Thanks in advance.
>>
>> #define BUILD_BUG_ON_ZERO(e) (sizeof(struct { int:-!!(e); }))
>
> This is also known as a compile time assert. I think that this
> particular variant has to be used inside a function.
>
> ! is just negation and produces a zero or 1 result. !! just does it
> twice, so that a non-zero value coming in becomes 1, and a zero value
> remains as zero.
>
> If e evaluates to false (zero) then -!!(e) evaluates to zero;
> if e evaluates to true (non-zero) then -!!(e) evaluates to -1.
>
> Declaring a bit field with a size of -1 will cause a compiler error.
> I'm actually surprised that declaring a bitfield of size 0 works.
>
> The typical declarations of this I've seen usually use arrays and
> arrange for the size to be -1 or 1 (which is generally more portable).
> When you use the array style declaration, you can use it outside a
> function as well.
>
> Dave Hylands
>


Thanks Dave.

My dilemma also comes from declaring a anonymous bitfield of size 0.

The following one is easy to understand anyway.

#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1)


-- 
Yours sincerely
ZhangMeng

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2011-02-26  1:14 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-02-25 10:50 BUILD_BUG_ON_ZERO Zhang Meng
2011-02-25 16:33 ` BUILD_BUG_ON_ZERO Dave Hylands
2011-02-26  1:14   ` BUILD_BUG_ON_ZERO Meng Zhang

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).