On Thu, Jan 29, 2015 at 3:07 AM, Greg KH <greg@kroah.com> wrote:
On Thu, Jan 29, 2015 at 02:16:51AM -0200, Vinícius Tinti wrote:
On Thu, Jan 29, 2015 at 2:08 AM, Greg KH <greg@kroah.com> wrote:
On Thu, Jan 29, 2015 at 01:48:43AM -0200, Vinícius Tinti wrote:
This is a simple patch that initializes a function with NULL to avoid some compiler warnings. In such cases should I proceed as a normal patch or it is better to send to another ML like to one for trivial patches?
Thanks,
Tinti
From a391789bf44afbdbe2a7b3c76301b5ece9f72475 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Vin=C3=ADcius=20Tinti?= <viniciustinti@gmail.com> 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 either proper initialized of an error issued.
Signed-off-by: Vinícius Tinti <viniciustinti@gmail.com> --- 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) int i; int (*write_reloc)(uint32_t, FILE *) = write32; int (*do_reloc)(struct section *sec, Elf_Rel *rel, Elf_Sym *sym, - const char *symname); + const char *symname) = NULL;
I think you need to get an updated version of the compiler as this patch 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.
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?
thanks,
greg k-h
-- Simplicity is the ultimate sophistication