lkml.org 
[lkml]   [2016]   [Apr]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRE: [PATCH 1/7] debugobjects: make fixup functions return bool instead of int
Date
Hi,
> On Fri, 22 Apr 2016, changbin.du@intel.com wrote:
> > From: "Du, Changbin" <changbin.du@intel.com>
> >
> > The object debugging infrastructure core provides some fixup callbacks
> > for the subsystem who use it. These callbacks are called from the debug
> > code whenever a problem in debug_object_init is detected. And
> > debugobjects core suppose them returns 1 when the fixup was successful,
> > otherwise 0. So the return type is boolean.
> >
> > A bad thing is that debug_object_fixup use the return value for
> > arithmetic operation. It confused me that what is the reall return
>
> What's bad about that? The fact that it's used for arithmethic operation or
> that it confused you?
>
It confused me because this is not a common usage. I was confused that what
does he fixup function return? A countable value? But doc says return fixed
or not!
if (fixup)
fixed = fixup(addr, state);
debug_objects_fixups += fixed;
In common,for int return 0 indicates success, negative for fail, positive for something
countable. So I think it is better follow this rule. Here is not of countable, it is Boolean.
So why not this?
if (fixup && fixup(addr, state))
debug_objects_fixups++;

> > Reading over the whole code, I found some place do use the return value
> > incorrectly(see next patch). So why use bool type instead?
>
> Patches which fix a problem need to come first not in the middle of a revamp
> series.
>
Thanks, I am first know of this.

> > + bool (*fixup_init)(void *addr, enum debug_obj_state state);
> > + bool (*fixup_activate)(void *addr, enum debug_obj_state state);
> > + bool (*fixup_destroy)(void *addr, enum debug_obj_state state);
> > + bool (*fixup_free)(void *addr, enum debug_obj_state state);
> > + bool (*fixup_assert_init)(void *addr, enum debug_obj_state state);
>
> So this change will introduce a gazillion of compile warnings because the
> callbacks in the various usage sites are still having 'int' return type.
>
No, I modified all the code who use debugojects API.

> Thanks,
>
> tglx

Thanks,
Du, Changbin

\
 
 \ /
  Last update: 2016-04-22 11:01    [W:0.117 / U:0.220 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site