From: Kees Cook <keescook@chromium.org>
To: linux-kernel@vger.kernel.org
Cc: Kees Cook <keescook@chromium.org>,
Emese Revfy <re.emese@gmail.com>, Arnd Bergmann <arnd@arndb.de>,
Josh Triplett <josh@joshtriplett.org>,
pageexec@freemail.hu, yamada.masahiro@socionext.com,
minipli@ld-linux.so, linux@armlinux.org.uk,
catalin.marinas@arm.com, linux@rasmusvillemoes.dk,
david.brown@linaro.org, benh@kernel.crashing.org,
tglx@linutronix.de, akpm@linux-foundation.org,
jlayton@poochiereds.net, sam@ravnborg.org,
kernel-hardening@lists.openwall.com
Subject: [PATCH v5 4/4] initify: Mark functions with the __unverified_nocapture attribute
Date: Tue, 31 Jan 2017 12:24:22 -0800 [thread overview]
Message-ID: <1485894263-91051-5-git-send-email-keescook@chromium.org> (raw)
In-Reply-To: <1485894263-91051-1-git-send-email-keescook@chromium.org>
From: Emese Revfy <re.emese@gmail.com>
The initify plugin attempts to analyze function arguments that have been
marked correctly with the __nocapture attribute. However, due to code
complexity, this is not always possible to verify. As a result, some
__nocapture attributes need to be marked as excluded from the automatic
verification. To do this, the __unverified_nocapture attribute is added.
It is intedned to only be used on function parameters that are difficult
for the plugin to analyze.
Signed-off-by: Emese Revfy <re.emese@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
---
include/linux/compiler-gcc.h | 8 ++++++++
include/linux/compiler.h | 4 ++++
lib/vsprintf.c | 4 ++--
3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
index cd4e9ffb00a7..fc0495e849ff 100644
--- a/include/linux/compiler-gcc.h
+++ b/include/linux/compiler-gcc.h
@@ -204,9 +204,17 @@
* been unmapped from memory. In order to identify functions that are confirmed
* to not capture their arguments, the __nocapture() attribute is used so that
* initify can better identify candidate variables.
+ *
+ * Arguments marked in this way are verified by the plugin, but sometimes
+ * code complexity and other limitiations will cause initify to not be able
+ * to check it correctly. For these cases, the __unverified_nocapture
+ * attribute can be added to disable this checking, overriding the plugin
+ * logic for cases that have been manually verified. This should not need
+ * to be used very often.
*/
#ifdef INITIFY_PLUGIN
#define __nocapture(...) __attribute__((nocapture(__VA_ARGS__)))
+#define __unverified_nocapture(...) __attribute__((unverified_nocapture(__VA_ARGS__)))
#endif
/*
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index 8b3dcc790bb6..1bde420f07bb 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -437,6 +437,10 @@ static __always_inline void __write_once_size(volatile void *p, void *res, int s
# define __nocapture(...)
#endif
+#ifndef __unverified_nocapture
+# define __unverified_nocapture(...)
+#endif
+
/*
* Tell gcc if a function is cold. The compiler will assume any path
* directly leading to the call is unlikely.
diff --git a/lib/vsprintf.c b/lib/vsprintf.c
index a192761d338a..cb964b51f9f8 100644
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c
@@ -118,7 +118,7 @@ long long simple_strtoll(const char *cp, char **endp, unsigned int base)
}
EXPORT_SYMBOL(simple_strtoll);
-static noinline_for_stack __nocapture(1)
+static noinline_for_stack __nocapture(1) __unverified_nocapture(1)
int skip_atoi(const char **s)
{
int i = 0;
@@ -1570,7 +1570,7 @@ int kptr_restrict __read_mostly;
* function pointers are really function descriptors, which contain a
* pointer to the real address.
*/
-static noinline_for_stack __nocapture(1)
+static noinline_for_stack __nocapture(1) __unverified_nocapture(1)
char *pointer(const char *fmt, char *buf, char *end, void *ptr,
struct printf_spec spec)
{
--
2.7.4
prev parent reply other threads:[~2017-01-31 20:32 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-31 20:24 [PATCH v5 0/4] Introduce the initify gcc plugin Kees Cook
2017-01-31 20:24 ` [PATCH v5 1/4] gcc-plugins: Add " Kees Cook
2017-03-27 7:38 ` [kernel-hardening] " Andrew Donnellan
2017-03-27 16:14 ` Kees Cook
2017-03-27 19:31 ` Rasmus Villemoes
2017-03-27 19:33 ` Kees Cook
2017-03-28 7:49 ` [kernel-hardening] " Arnd Bergmann
2017-03-28 19:03 ` Kees Cook
2017-03-28 20:31 ` Arnd Bergmann
2017-03-28 3:03 ` Andrew Donnellan
2017-03-28 19:00 ` Kees Cook
2017-03-29 8:32 ` Arnd Bergmann
2017-03-29 8:56 ` Arnd Bergmann
2017-03-30 11:04 ` Arnd Bergmann
2017-01-31 20:24 ` [PATCH v5 2/4] util: Move type casts into is_kernel_rodata Kees Cook
2017-01-31 20:24 ` [PATCH v5 3/4] initify: Mark functions with the __nocapture attribute Kees Cook
2017-01-31 20:24 ` Kees Cook [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1485894263-91051-5-git-send-email-keescook@chromium.org \
--to=keescook@chromium.org \
--cc=akpm@linux-foundation.org \
--cc=arnd@arndb.de \
--cc=benh@kernel.crashing.org \
--cc=catalin.marinas@arm.com \
--cc=david.brown@linaro.org \
--cc=jlayton@poochiereds.net \
--cc=josh@joshtriplett.org \
--cc=kernel-hardening@lists.openwall.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@armlinux.org.uk \
--cc=linux@rasmusvillemoes.dk \
--cc=minipli@ld-linux.so \
--cc=pageexec@freemail.hu \
--cc=re.emese@gmail.com \
--cc=sam@ravnborg.org \
--cc=tglx@linutronix.de \
--cc=yamada.masahiro@socionext.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).