linux-kernel - Re: [PATCH] compiler*.h: Add '__' prefix and suffix to all __attribute_

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:   Tue, 29 Oct 2019 00:03:50 +0100
From:   Luc Van Oostenryck <luc.vanoostenryck@...il.com>
To:     Joe Perches <joe@...ches.com>
Cc:     Miguel Ojeda <miguel.ojeda.sandonis@...il.com>,
        linux-sparse@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>
Subject: Re: [PATCH] compiler*.h: Add '__' prefix and suffix to all
 __attribute__ #defines

On Mon, Oct 28, 2019 at 03:28:17PM -0700, Joe Perches wrote:
> On Mon, 2019-10-28 at 23:15 +0100, Luc Van Oostenryck wrote:
> > On Mon, Oct 28, 2019 at 10:59:47AM -0700, Joe Perches wrote:
> > > On Mon, 2019-10-28 at 18:37 +0100, Miguel Ojeda wrote:
> > > > Just in case: for these ones (i.e. __CHECKER__), did you check if
> > > > sparse handles this syntax? (I don't recall myself if it does).
> > > > 
> > > > Other than that, thanks for the cleanup too! I can pick it up in the
> > > > the compiler-attributes tree and put it in -next.
> > > 
> > > Thanks for asking and no, I did  not until just now.
> > > Turns out sparse does _not_ handle these changes and
> > > the checking fails for these __<changes>__.
> > > 
> > > sparse would have to update parse.c or the __CHECKER__
> > > changes would need to be reverted.
> > > 
> > > Perhaps update parse.c like:
> > 
> > ...
> > 
> > Yes, this was missing. Thanks.
> > Can I have your SoB for this?
> 
> I'm not sure this actually works as there's
> some possible sparse parsing changes in the
> use of __context__.

Yes, indeed. The following shoud be squashed on top of
your patch (not tested yet on linux side):

-- Luc

diff --git a/parse.c b/parse.c
index 4464e2667..4b0a1566c 100644
--- a/parse.c
+++ b/parse.c
@@ -345,6 +345,7 @@ static struct symbol_op goto_op = {
 
 static struct symbol_op __context___op = {
 	.statement = parse_context_statement,
+	.attribute = attribute_context,
 };
 
 static struct symbol_op range_op = {
@@ -537,6 +538,7 @@ static struct init_keyword {
 	{ "while",	NS_KEYWORD, .op = &while_op },
 	{ "do",		NS_KEYWORD, .op = &do_op },
 	{ "goto",	NS_KEYWORD, .op = &goto_op },
+	{ "context",	NS_KEYWORD, .op = &context_op },
 	{ "__context__",NS_KEYWORD, .op = &__context___op },
 	{ "__range__",	NS_KEYWORD, .op = &range_op },
 	{ "asm",	NS_KEYWORD, .op = &asm_op },
@@ -560,8 +562,6 @@ static struct init_keyword {
 	{ "__bitwise__",NS_KEYWORD,	MOD_BITWISE,	.op = &attr_bitwise_op },
 	{ "address_space",NS_KEYWORD,	.op = &address_space_op },
 	{ "__address_space__",NS_KEYWORD,	.op = &address_space_op },
-	{ "context",	NS_KEYWORD,	.op = &context_op },
-	{ "__context__",NS_KEYWORD,	.op = &context_op },
 	{ "designated_init",	NS_KEYWORD,	.op = &designated_init_op },
 	{ "__designated_init__",	NS_KEYWORD,	.op = &designated_init_op },
 	{ "transparent_union",	NS_KEYWORD,	.op = &transparent_union_op },