linux-kernel - Re: [PATCH][RESEND] do not redefine userspace's NULL #define

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sat, 14 Apr 2012 08:43:08 +0200
From:	Lubos Lunak <l.lunak@...e.cz>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Peter Seebach <peter.seebach@...driver.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org, Arnd Bergmann <arnd@...db.de>
Subject: Re: [PATCH][RESEND] do not redefine userspace's NULL #define

On Saturday 14 of April 2012, Linus Torvalds wrote:
> On Fri, Apr 13, 2012 at 4:18 PM, Lubos Lunak <l.lunak@...e.cz> wrote:
> >> imagine replacing the kernel ((void *)0) with __null.
> >
> >  __null apparently exists only with g++, C does not have the stronger
> > type safety that prevents ((void*)0) from being usable in C++
>
> Please don't continue to spread this total bogosity.
>
> The reason C++ cannot use "(void *)0" has nothing to do with "stronger
> type safety". That's a total idiotic lie by C++ apologists, and I hate
> hearing it repeated over and over again.
>
> And it really *is* a lie. The C++ type system isn't even "stronger",

$ cat b.c
void foo()
    {
    int* a;
    void* b;
    char c;
    a = b = &c;
    }
$ gcc -Wall b.c -c
$ g++ -Wall b.c -c
b.c: In function ‘void foo()’:
b.c:6:14: error: invalid conversion from ‘void*’ to ‘int*’

> it's just different, and it's actively *broken* wrt NULL. Always has
> been.
>
> The sane thing to do for C++ would always have been to recognize that
> "(void *)0" is not a "void pointer" - it's just NULL.

 As much as I agree that it was stupid to copy the special-casing of 0 from C 
and they should have instead special-cased (void*)0, the reality is that it 
took them until C++11 to add that to the standard, and even there they did it 
differently, so C++ now has to live with the legacy of NULL being possibly 
implemented in a stupid way.

 The reality is also that g++ people recognized this quite some time back, 
implemented NULL in a sane way, and kernel headers still replace that 
implementation with one that, as you yourself say, is technically wrong. So 
how about my patch that fixes that? The rest is a discussion about things 
that should have been but are not and is completely pointless by now.

8<-----
headers: do not redefine userspace's NULL #define when compiling C++ code

GCC's NULL is actually __null when compiling C++, which allows detecting some
questionable NULL usage and warn about it.

Signed-off-by: Luboš Luňák <l.lunak@...e.cz>
---
 include/linux/stddef.h |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/include/linux/stddef.h b/include/linux/stddef.h
index 6a40c76..a75f4b9 100644
--- a/include/linux/stddef.h
+++ b/include/linux/stddef.h
@@ -3,12 +3,18 @@
 
 #include <linux/compiler.h>
 
-#undef NULL
 #if defined(__cplusplus)
-#define NULL 0
+#ifndef NULL
+#ifdef __GNUG__
+#define NULL __null
 #else
-#define NULL ((void *)0)
+#define NULL 0
 #endif
+#endif
+#else /* __cplusplus */
+#undef NULL
+#define NULL ((void *)0)
+#endif /* __cplusplus */
 
 #ifdef __KERNEL__
 
-- 
1.7.7


-- 
 Lubos Lunak
 l.lunak@...e.cz
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/