[RFC][PATCH] fix casting constant to _Bool

[RFC][PATCH] fix casting constant to _Bool

[Xi Wang]

Casting to _Bool requires a zero test rather than truncation.

Signed-off-by: Xi Wang <xi.wang@gmail.com>
---
A simple example is:

  _Bool x = 0x200;

x should have been true rather than false.

BTW, this patch also disables the warning "cast truncates bits from
constant value".  Does that sound good?

A more serious problem is something like:

  _Bool foo(int x) { return x; }

sparse emits:

  scast.1     %r2 <- (32) %arg1

which makes sparse-llvm generate:

  %R2 = trunc i32 %0 to i1

My experimental backend "splay" has to treat _Bool specifically to
generate:

  %0 = icmp ne i32 %x, 0

Should we leave the conversion job to backends, or should we just fix
the sparse IR (e.g., emitting setne rather than scast for casts to _Bool)?
---
 expand.c |    6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/expand.c b/expand.c
index 63a9075..ee818a4 100644
--- a/expand.c
+++ b/expand.c
@@ -92,6 +92,12 @@ void cast_value(struct expression *expr, struct symbol *newtype,
 	value = get_longlong(old);
 
 Int:
+	// _Bool requires a zero test rather than truncation.
+	if (is_bool_type(newtype)) {
+		expr->value = value ? 1 : 0;
+		return;
+	}
+
 	// Truncate it to the new size
 	signmask = 1ULL << (new_size-1);
 	mask = signmask | (signmask-1);
-- 
1.7.9.5

Re: [RFC][PATCH] fix casting constant to _Bool

[Pekka Enberg]

On Sat, Jun 16, 2012 at 9:55 AM, Xi Wang <xi.wang@gmail.com> wrote:
> Casting to _Bool requires a zero test rather than truncation.
>
> Signed-off-by: Xi Wang <xi.wang@gmail.com>
> ---
> A simple example is:
>
>  _Bool x = 0x200;
>
> x should have been true rather than false.
>
> BTW, this patch also disables the warning "cast truncates bits from
> constant value".  Does that sound good?

That's unfortunate.

> A more serious problem is something like:
>
>  _Bool foo(int x) { return x; }
>
> sparse emits:
>
>  scast.1     %r2 <- (32) %arg1
>
> which makes sparse-llvm generate:
>
>  %R2 = trunc i32 %0 to i1
>
> My experimental backend "splay" has to treat _Bool specifically to
> generate:
>
>  %0 = icmp ne i32 %x, 0
>
> Should we leave the conversion job to backends, or should we just fix
> the sparse IR (e.g., emitting setne rather than scast for casts to _Bool)?

It'd be best if the backend didn't have to do anything special here.
Can we have it both ways? Not disable the warning but still deal with
it in the frontend?

> ---
>  expand.c |    6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/expand.c b/expand.c
> index 63a9075..ee818a4 100644
> --- a/expand.c
> +++ b/expand.c
> @@ -92,6 +92,12 @@ void cast_value(struct expression *expr, struct symbol *newtype,
>        value = get_longlong(old);
>
>  Int:
> +       // _Bool requires a zero test rather than truncation.
> +       if (is_bool_type(newtype)) {
> +               expr->value = value ? 1 : 0;
> +               return;
> +       }
> +
>        // Truncate it to the new size
>        signmask = 1ULL << (new_size-1);
>        mask = signmask | (signmask-1);
> --
> 1.7.9.5
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC][PATCH] fix casting constant to _Bool

[Linus Torvalds]

On Fri, Jun 15, 2012 at 11:55 PM, Xi Wang <xi.wang@gmail.com> wrote:
> Casting to _Bool requires a zero test rather than truncation.

Ack. _Bool is very special (and rather non-C-like).

> BTW, this patch also disables the warning "cast truncates bits from
> constant value".  Does that sound good?

For cast to _Bool, that is correct. Since it doesn't truncate the
value, the upper bits are not ignored as they are with other casts.

               Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC][PATCH] fix casting constant to _Bool

[Dan Carpenter]

On Sat, Jun 16, 2012 at 09:32:50AM -0700, Linus Torvalds wrote:
> On Fri, Jun 15, 2012 at 11:55 PM, Xi Wang <xi.wang@gmail.com> wrote:
> > Casting to _Bool requires a zero test rather than truncation.
> 
> Ack. _Bool is very special (and rather non-C-like).
> 
> > BTW, this patch also disables the warning "cast truncates bits from
> > constant value".  Does that sound good?
> 
> For cast to _Bool, that is correct. Since it doesn't truncate the
> value, the upper bits are not ignored as they are with other casts.
> 

It's pretty rare to see a cast to bool that truncates the value.
My feeling is that most times they are bugs.  I just saw one of
these a couple weeks ago:

http://www.spinics.net/lists/linux-iio/msg05573.html

I wish the error message could be fixed so it says 1 instead of 0.

old: warning: cast truncates bits from constant value (2 becomes 0)
new: warning: cast truncates bits from constant value (2 becomes 1)

regards,
dan carpenter
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC][PATCH] fix casting constant to _Bool

[Linus Torvalds]

On Sat, Jun 16, 2012 at 9:59 AM, Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
> It's pretty rare to see a cast to bool that truncates the value.
> My feeling is that most times they are bugs.  I just saw one of
> these a couple weeks ago:
>
> http://www.spinics.net/lists/linux-iio/msg05573.html
>
> I wish the error message could be fixed so it says 1 instead of 0.

Hmm - I could imagine havinbg a warning, but it shouldn't be about truncation.

Maybe something like "Odd constant _Bool cast (%lld -> 1)"? It will
always result in 1, since the only value that results in 0 is 0 (and
isn't odd).

               Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [RFC][PATCH] fix casting constant to _Bool

[Xi Wang]

On Jun 16, 2012, at 1:20 PM, Linus Torvalds wrote:
> Maybe something like "Odd constant _Bool cast (%lld -> 1)"? It will
> always result in 1, since the only value that results in 0 is 0 (and
> isn't odd).

Sounds good.  I will send a v2.

- xi