Re: [PATCH] x86, vsyscall: Fix build warning in vsyscall_64.c

[Andy Lutomirski <luto@MIT.EDU>]

On 06/14/2011 01:43 PM, Rakib Mullick wrote:
> On Tue, Jun 14, 2011 at 12:06 AM, Andrew Lutomirski<luto@mit.edu>  wrote:
>> On Mon, Jun 13, 2011 at 4:45 AM, Rakib Mullick<rakib.mullick@gmail.com>  wrote:
>>> On Mon, Jun 13, 2011 at 10:54 AM, Rakib Mullick<rakib.mullick@gmail.com>  wrote:
>>>> On Mon, Jun 13, 2011 at 8:52 AM, Andrew Lutomirski<luto@mit.edu>  wrote:
>>>>> On Sun, Jun 12, 2011 at 1:12 AM, Rakib Mullick<rakib.mullick@gmail.com>  wrote:
>>>>>> On Sat, Jun 11, 2011 at 5:01 PM, Andrew Lutomirski<luto@mit.edu>  wrote:
>>>>>>> On Sat, Jun 11, 2011 at 3:31 AM, Rakib Mullick<rakib.mullick@gmail.com>  wrote:
>>>>>
>>>>> I think there are three separate issues here:
>>>>>
>>>>> 1. Can ret be used uninitialized?  I say no, even as seen by the
>>>>> compiler.  If vsyscall_nr is 0, 1, or 2, then ret is initialized.  If
>>>>> vsyscall_nr is 3, then the BUG gets hit.  BUG is defined as some
>>>>> assembly magic followed by unreachable(), and the compiler is supposed
>>>>> to know that code after unreachable() is qunreachable.  So how can ret
>>>>> be used uninitialized?
>>>>>
>>>> I don't have much knowledge of advance assembly, so I really don't
>>>> understand that part - how BUG handles this. If it really makes sure
>>>> that, it will handle it properly then I think you can drop this patch.
>>>>
>>>>> What version of gcc do you have?  gcc (GCC) 4.6.0 20110530 (Red Hat
>>>>> 4.6.0-9) does not produce this warning.
>>>>>
>>>> Currently, I'm replying from outside my home. I'll let you know when
>>>> I'm back home.
>>>>
>>> Here is my GCC version - gcc version 4.5.1 20100924 (Red Hat 4.5.1-4)
>>> (GCC). I'm using Fedora 14.
>>
>> I also have gcc (GCC) 4.5.1 20100924 (Red Hat 4.5.1-4) on another box,
>> and I still can't reproduce this.
>>
>> Can you tell me which git revision you're building and send me your
>> .config and the output of:
>>
> I'm using 3.0.0-rc2 (lastly I pulled tip tree 3 days ago). I've
> attached the .config (config.log).
>
>> $ touch arch/x86/kernel/vsyscall_64.o
>> $ make V=1 arch/x86/kernel/vsyscall_64.o
>>
> Output of the above steps are attached (vsyscall_64.log). Hope that will help.

Aha!  You have CONFIG_BUG=n.  I'm not sure that fixing warnings for that 
case is worth it (or is even a good idea).

Can you try this patch, though:

Signed-off-by: Andy Lutomirski <luto@mit.edu>

diff --git a/include/asm-generic/bug.h b/include/asm-generic/bug.h
index dfb0ec6..f4083f4 100644
--- a/include/asm-generic/bug.h
+++ b/include/asm-generic/bug.h
@@ -107,11 +107,11 @@ extern void warn_slowpath_null(const char *file, 
const int line);

  #else /* !CONFIG_BUG */
  #ifndef HAVE_ARCH_BUG
-#define BUG() do {} while(0)
+#define BUG() do { unreachable(); } while(0)
  #endif

  #ifndef HAVE_ARCH_BUG_ON
-#define BUG_ON(condition) do { if (condition) ; } while(0)
+#define BUG_ON(condition) do { if (condition) unreachable(); } while(0)
  #endif

  #ifndef HAVE_ARCH_WARN_ON

It may silence a lot of warnings, although it'll come at a cost of 
increased code size with CONFIG_BUG=n on older gcc.  On newer GCC, 
you'll get possibly faster and smaller code.

--Andy