Patchwork "unsigned expression < 0" always false warning

login
register
mail settings
Submitter Steven Rostedt
Date April 19, 2013, 9:17 p.m.
Message ID <1366406262.9609.123.camel@gandalf.local.home>
Download mbox | patch
Permalink /patch/238112/
State Not Applicable
Headers show

Comments

Steven Rostedt - April 19, 2013, 9:17 p.m.
Does this fix your issue?

-- Steve

From d2802d0739dcc61af5e5ea00773ce7ddead4e9c2 Mon Sep 17 00:00:00 2001
From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
Date: Fri, 19 Apr 2013 17:10:27 -0400
Subject: [PATCH] tracing: Compare to 1 instead of zero for is_signed_type()

The formats of the trace events show if the type of a event field
is signed or not via a macro called is_signed_type(). This does
a trick with the type and compares a -1 to zero after typecasting
to the tested type. If it returns true, it's signed, otherwise
its not. But this unfortunately triggers a warning by gcc:

  warning: comparison of unsigned expression < 0 is always false

As we know it is always false (that's why we do it), this is a
false warning. Luckily for us, the comparison works with a 1 as
well, without giving the warning.

Convert the check to compare (type)-1 < (type)0 to (type)-1 < (type)1
to determine if the type is signed or not.

Link: http://lkml.kernel.org/r/CAErSpo4YXcY9fuOKWYGDkddJwk68kmZTohsmVB6QvrhjboOh1Q@mail.gmail.com

Reported-by: Bjorn Helgaas <bhelgaas@google.com>
Reported-by: Gary Hade <garyhade@us.ibm.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 include/linux/ftrace_event.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Bjorn Helgaas - April 19, 2013, 10:30 p.m.
On Fri, Apr 19, 2013 at 3:17 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
> Does this fix your issue?

Yep, that fixes the warning for me.  Thanks!

Bjorn

> From d2802d0739dcc61af5e5ea00773ce7ddead4e9c2 Mon Sep 17 00:00:00 2001
> From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
> Date: Fri, 19 Apr 2013 17:10:27 -0400
> Subject: [PATCH] tracing: Compare to 1 instead of zero for is_signed_type()
>
> The formats of the trace events show if the type of a event field
> is signed or not via a macro called is_signed_type(). This does
> a trick with the type and compares a -1 to zero after typecasting
> to the tested type. If it returns true, it's signed, otherwise
> its not. But this unfortunately triggers a warning by gcc:
>
>   warning: comparison of unsigned expression < 0 is always false
>
> As we know it is always false (that's why we do it), this is a
> false warning. Luckily for us, the comparison works with a 1 as
> well, without giving the warning.
>
> Convert the check to compare (type)-1 < (type)0 to (type)-1 < (type)1
> to determine if the type is signed or not.
>
> Link: http://lkml.kernel.org/r/CAErSpo4YXcY9fuOKWYGDkddJwk68kmZTohsmVB6QvrhjboOh1Q@mail.gmail.com
>
> Reported-by: Bjorn Helgaas <bhelgaas@google.com>
> Reported-by: Gary Hade <garyhade@us.ibm.com>
> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> ---
>  include/linux/ftrace_event.h |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
> index 4e28b01..34e00fb 100644
> --- a/include/linux/ftrace_event.h
> +++ b/include/linux/ftrace_event.h
> @@ -333,7 +333,7 @@ extern int trace_define_field(struct ftrace_event_call *call, const char *type,
>  extern int trace_add_event_call(struct ftrace_event_call *call);
>  extern void trace_remove_event_call(struct ftrace_event_call *call);
>
> -#define is_signed_type(type)   (((type)(-1)) < (type)0)
> +#define is_signed_type(type)   (((type)(-1)) < (type)1)
>
>  int trace_set_clr_event(const char *system, const char *event, int set);
>
> --
> 1.7.10.4
>
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Gary Hade - April 19, 2013, 10:45 p.m.
On Fri, Apr 19, 2013 at 04:30:14PM -0600, Bjorn Helgaas wrote:
> On Fri, Apr 19, 2013 at 3:17 PM, Steven Rostedt <rostedt@goodmis.org> wrote:
> > Does this fix your issue?
> 
> Yep, that fixes the warning for me.  Thanks!
> 
> Bjorn

It also worked for me.  It got rid of every one the 1401 is_signed_type
associated "warning: comparison of unsigned expression < 0 is always false"
occurrences (during 3.9-rc7 `make W=1` kernel build) mentioned in my
previous email.

Gary

> 
> > From d2802d0739dcc61af5e5ea00773ce7ddead4e9c2 Mon Sep 17 00:00:00 2001
> > From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
> > Date: Fri, 19 Apr 2013 17:10:27 -0400
> > Subject: [PATCH] tracing: Compare to 1 instead of zero for is_signed_type()
> >
> > The formats of the trace events show if the type of a event field
> > is signed or not via a macro called is_signed_type(). This does
> > a trick with the type and compares a -1 to zero after typecasting
> > to the tested type. If it returns true, it's signed, otherwise
> > its not. But this unfortunately triggers a warning by gcc:
> >
> >   warning: comparison of unsigned expression < 0 is always false
> >
> > As we know it is always false (that's why we do it), this is a
> > false warning. Luckily for us, the comparison works with a 1 as
> > well, without giving the warning.
> >
> > Convert the check to compare (type)-1 < (type)0 to (type)-1 < (type)1
> > to determine if the type is signed or not.
> >
> > Link: http://lkml.kernel.org/r/CAErSpo4YXcY9fuOKWYGDkddJwk68kmZTohsmVB6QvrhjboOh1Q@mail.gmail.com
> >
> > Reported-by: Bjorn Helgaas <bhelgaas@google.com>
> > Reported-by: Gary Hade <garyhade@us.ibm.com>
> > Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> > ---
> >  include/linux/ftrace_event.h |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
> > index 4e28b01..34e00fb 100644
> > --- a/include/linux/ftrace_event.h
> > +++ b/include/linux/ftrace_event.h
> > @@ -333,7 +333,7 @@ extern int trace_define_field(struct ftrace_event_call *call, const char *type,
> >  extern int trace_add_event_call(struct ftrace_event_call *call);
> >  extern void trace_remove_event_call(struct ftrace_event_call *call);
> >
> > -#define is_signed_type(type)   (((type)(-1)) < (type)0)
> > +#define is_signed_type(type)   (((type)(-1)) < (type)1)
> >
> >  int trace_set_clr_event(const char *system, const char *event, int set);
> >
> > --
> > 1.7.10.4
> >
> >
> >
> 

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

Patch

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 4e28b01..34e00fb 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -333,7 +333,7 @@  extern int trace_define_field(struct ftrace_event_call *call, const char *type,
 extern int trace_add_event_call(struct ftrace_event_call *call);
 extern void trace_remove_event_call(struct ftrace_event_call *call);
 
-#define is_signed_type(type)	(((type)(-1)) < (type)0)
+#define is_signed_type(type)	(((type)(-1)) < (type)1)
 
 int trace_set_clr_event(const char *system, const char *event, int set);