On Thu, Jan 29, 2015 at 12:27 PM, Vin?cius Tinti [off-list ref]
wrote:

On Thu, Jan 29, 2015 at 3:07 AM, Greg KH [off-list ref] wrote:

quoted

On Thu, Jan 29, 2015 at 02:16:51AM -0200, Vin?cius Tinti wrote:

quoted

On Thu, Jan 29, 2015 at 2:08 AM, Greg KH [off-list ref] wrote:

quoted

On Thu, Jan 29, 2015 at 01:48:43AM -0200, Vin?cius Tinti wrote:

quoted

This is a simple patch that initializes a function with NULL to

avoid some

quoted

quoted

quoted

quoted

compiler warnings. In such cases should I proceed as a normal patch

or it is

quoted

quoted

quoted

quoted

better to send to another ML like to one for trivial patches?

Thanks,

Tinti

quoted

quoted

From a391789bf44afbdbe2a7b3c76301b5ece9f72475 Mon Sep 17 00:00:00

2001

quoted

quoted

quoted

quoted

From: =?UTF-8?q?Vin=C3=ADcius=20Tinti?= <redacted>
Date: Thu, 29 Jan 2015 01:35:34 -0200
Subject: [PATCH] x86: LLVMLinux: Fix uninitialized function do_reloc
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Explicit initializes do_reloc function with NULL. Later the function

is

quoted

quoted

quoted

quoted

either proper initialized of an error issued.

Signed-off-by: Vin?cius Tinti <redacted>
---
 arch/x86/tools/relocs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/tools/relocs.c b/arch/x86/tools/relocs.c
index 0c2fae8..1d533f1 100644
--- a/arch/x86/tools/relocs.c
+++ b/arch/x86/tools/relocs.c

@@ -971,7 +971,7 @@ static void emit_relocs(int as_text, int

use_real_mode)

quoted

quoted

quoted

quoted

      int i;
      int (*write_reloc)(uint32_t, FILE *) = write32;
      int (*do_reloc)(struct section *sec, Elf_Rel *rel, Elf_Sym

*sym,

quoted

quoted

quoted

quoted

-                     const char *symname);
+                     const char *symname) = NULL;

I think you need to get an updated version of the compiler as this

patch

quoted

quoted

quoted

should not be needed at all.  It doesn't cause a warning here for me
without it.

In fact it causes a warning on Clang which complains that:

   arch/x86/tools/relocs.c:977:6: warning: variable 'do_reloc' is used
uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]

I suggest you file a bug with clang, gcc doesn't have this problem at
all as obviously, if you look at the code, that variable can never be
used uninitialized.

I can simply turn down this kind of warning.

quoted

quoted

I think there is not a problem on the current code but to avoid
further problems I believe it is worth to initialize this function
with NULL.
What do you think?

Don't paper over bugs in the compiler with kernel code changes for no
good reason :)

Agreed. But whenever I find a warning in GCC during the build what
should I do with it?
Can I simply send it to the main ml?

If you believe your patch is correct, and if checkpatch.pl doesn't complain
about your patch, then get_maintainers.pl will tell you the people and
mailing lists you should send your patch to.

quoted

thanks,

greg k-h

--
Simplicity is the ultimate sophistication

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies at kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

-- 
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20150129/b15f1fbc/attachment-0001.html