diff mbox series

[4/4,v5] fortify: Add Clang support

Message ID 20220202003033.704951-5-keescook@chromium.org (mailing list archive)
State Superseded
Commit 0b27a15c23f17d6ca9a01505f38b477178fabb52
Headers show
Series fortify: Add Clang support | expand

Commit Message

Kees Cook Feb. 2, 2022, 12:30 a.m. UTC
Enable FORTIFY_SOURCE support for Clang:

Use the new __pass_object_size and __overloadable attributes so
that Clang will have appropriate visibility into argument types such
that __builtin_object_size(p, 1) will behave correctly. Additional
details here:
    https://github.com/llvm/llvm-project/issues/53516
    https://github.com/ClangBuiltLinux/linux/issues/1401

Use the new __diagnose_as attribute to make sure no compile-time
diagnostic warnings are lost due to the effectively renamed string
functions.

Redefine strlen() as a macro that tests for being a constant expression
so that strlen() can still be used in static initializers, which was
lost when adding __pass_object_size and __overloadable.

Finally, a bug with __builtin_constant_p() of globally defined variables
was fixed in Clang 13, so FORTIFY support must depend on that version or
later. Additional details here:
    https://bugs.llvm.org/show_bug.cgi?id=41459
    commit a52f8a59aef4 ("fortify: Explicitly disable Clang support")

Cc: Miguel Ojeda <ojeda@kernel.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: George Burgess IV <gbiv@google.com>
Cc: llvm@lists.linux.dev
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 include/linux/fortify-string.h | 52 ++++++++++++++++++++++++----------
 security/Kconfig               |  2 +-
 2 files changed, 38 insertions(+), 16 deletions(-)

Comments

Nick Desaulniers Feb. 2, 2022, 9:22 p.m. UTC | #1
On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
>
> --- a/security/Kconfig
> +++ b/security/Kconfig
> @@ -179,7 +179,7 @@ config FORTIFY_SOURCE
>         depends on ARCH_HAS_FORTIFY_SOURCE
>         # https://bugs.llvm.org/show_bug.cgi?id=50322
>         # https://bugs.llvm.org/show_bug.cgi?id=41459
> -       depends on !CC_IS_CLANG
> +       depends on !CC_IS_CLANG || CLANG_VERSION >= 130000

Are these comments still relevant, and is the clang version still correct?

In https://lore.kernel.org/llvm/CANiq72n1d7ouKNi+pbsy7chsg0DfCXxez27qqtS9XE1n3m5=8Q@mail.gmail.com/
Miguel notes that diagnose_as only exists in clang-14+.  If this
series relies on diagnose_as, then should this version check be for
clang-14+ rather than clang-13+?

https://bugs.llvm.org/show_bug.cgi?id=50322 is still open, but doesn't
signify why there's a version check. It makes sense if there's no
version check, but I'm not sure it's still relevant to this Kconfig
option after your series.

https://bugs.llvm.org/show_bug.cgi?id=41459 was fixed in clang-13, but
it was also backported to the clang 12.0.1 release.  Is it still
relevant if we're gated on diagnose_as from clang-14?

Perhaps a single comment, about the diagnose_as attribute or a link to
https://reviews.llvm.org/rGbc5f2d12cadce765620efc56a1ca815221db47af or
whatever, and updating the version check to be against clang-14 would
be more precise?
Nick Desaulniers Feb. 2, 2022, 9:27 p.m. UTC | #2
On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
>
> +#define BOS    const __pass_object_size(1)
> +#define BOS0   const __pass_object_size(0)

A dumb bikeshed, but would you mind naming these BOS1 and BOS0, and
perhaps consider adding a comment or pointer or link to something that
describes why we use the two different modes?  I recognize that the
code already uses the two different modes already without comments,
but this might be a nice place to point folks like myself to so that
in a month or so when I forget what the difference is between modes
(again), we have a shorter trail of breadcrumbs.
Kees Cook Feb. 3, 2022, 3:15 a.m. UTC | #3
On Wed, Feb 02, 2022 at 01:22:09PM -0800, Nick Desaulniers wrote:
> On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > --- a/security/Kconfig
> > +++ b/security/Kconfig
> > @@ -179,7 +179,7 @@ config FORTIFY_SOURCE
> >         depends on ARCH_HAS_FORTIFY_SOURCE
> >         # https://bugs.llvm.org/show_bug.cgi?id=50322
> >         # https://bugs.llvm.org/show_bug.cgi?id=41459
> > -       depends on !CC_IS_CLANG
> > +       depends on !CC_IS_CLANG || CLANG_VERSION >= 130000
> 
> Are these comments still relevant, and is the clang version still correct?

Oh, good call. I thought the version was still correct (more below),
but yes, the comments need adjusting.

> In https://lore.kernel.org/llvm/CANiq72n1d7ouKNi+pbsy7chsg0DfCXxez27qqtS9XE1n3m5=8Q@mail.gmail.com/
> Miguel notes that diagnose_as only exists in clang-14+.  If this
> series relies on diagnose_as, then should this version check be for
> clang-14+ rather than clang-13+?

It doesn't rely on it; this is just taking advantage of an improvement.

> https://bugs.llvm.org/show_bug.cgi?id=50322 is still open, but doesn't
> signify why there's a version check. It makes sense if there's no
> version check, but I'm not sure it's still relevant to this Kconfig
> option after your series.

With __overloadable, this probably ended up going away.

> https://bugs.llvm.org/show_bug.cgi?id=41459 was fixed in clang-13, but
> it was also backported to the clang 12.0.1 release.  Is it still
> relevant if we're gated on diagnose_as from clang-14?

Ah-ha! I missed that this got backported. Looks like 12.0.1 and later
have this fixed. That's excellent!

> Perhaps a single comment, about the diagnose_as attribute or a link to
> https://reviews.llvm.org/rGbc5f2d12cadce765620efc56a1ca815221db47af or
> whatever, and updating the version check to be against clang-14 would
> be more precise?

Yup, I will rework this after double-checking 12.0.1 builds.

Thanks!
Kees Cook Feb. 3, 2022, 3:18 a.m. UTC | #4
On Wed, Feb 02, 2022 at 01:27:11PM -0800, Nick Desaulniers wrote:
> On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > +#define BOS    const __pass_object_size(1)
> > +#define BOS0   const __pass_object_size(0)
> 
> A dumb bikeshed, but would you mind naming these BOS1 and BOS0, and
> perhaps consider adding a comment or pointer or link to something that
> describes why we use the two different modes?  I recognize that the
> code already uses the two different modes already without comments,
> but this might be a nice place to point folks like myself to so that
> in a month or so when I forget what the difference is between modes
> (again), we have a shorter trail of breadcrumbs.

Sure, I can do that. My expectation was to entirely eliminate mode 0
usage in the future.

Though now that things are so close, I'll just do some builds with the
last few users switched over. But maybe memcmp() was a pain? I'll go
check...
Nick Desaulniers Feb. 3, 2022, 10:13 p.m. UTC | #5
On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
>
> diff --git a/include/linux/fortify-string.h b/include/linux/fortify-string.h
> index c45159dbdaa1..5482536d3197 100644
> --- a/include/linux/fortify-string.h
> +++ b/include/linux/fortify-string.h
> @@ -2,7 +2,9 @@
>  #ifndef _LINUX_FORTIFY_STRING_H_
>  #define _LINUX_FORTIFY_STRING_H_
>
> -#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline))
> +#include <linux/const.h>
> +
> +#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline)) __overloadable

Sorry, I just noticed this line (already) uses a mix of open coding
__attribute__ and not. Would you mind also please changing the open
coded gnu_inline to simply __gnu_inline to make the entire line
consistent?
Miguel Ojeda Feb. 3, 2022, 10:28 p.m. UTC | #6
On Thu, Feb 3, 2022 at 11:13 PM Nick Desaulniers
<ndesaulniers@google.com> wrote:
>
> Sorry, I just noticed this line (already) uses a mix of open coding
> __attribute__ and not. Would you mind also please changing the open
> coded gnu_inline to simply __gnu_inline to make the entire line
> consistent?

+1

Cheers,
Miguel
Kees Cook Feb. 4, 2022, 12:28 a.m. UTC | #7
On Thu, Feb 03, 2022 at 02:13:33PM -0800, Nick Desaulniers wrote:
> On Tue, Feb 1, 2022 at 4:30 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > diff --git a/include/linux/fortify-string.h b/include/linux/fortify-string.h
> > index c45159dbdaa1..5482536d3197 100644
> > --- a/include/linux/fortify-string.h
> > +++ b/include/linux/fortify-string.h
> > @@ -2,7 +2,9 @@
> >  #ifndef _LINUX_FORTIFY_STRING_H_
> >  #define _LINUX_FORTIFY_STRING_H_
> >
> > -#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline))
> > +#include <linux/const.h>
> > +
> > +#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline)) __overloadable
> 
> Sorry, I just noticed this line (already) uses a mix of open coding
> __attribute__ and not. Would you mind also please changing the open
> coded gnu_inline to simply __gnu_inline to make the entire line
> consistent?

Ah yeah, thanks. Fixed.
diff mbox series

Patch

diff --git a/include/linux/fortify-string.h b/include/linux/fortify-string.h
index c45159dbdaa1..5482536d3197 100644
--- a/include/linux/fortify-string.h
+++ b/include/linux/fortify-string.h
@@ -2,7 +2,9 @@ 
 #ifndef _LINUX_FORTIFY_STRING_H_
 #define _LINUX_FORTIFY_STRING_H_
 
-#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline))
+#include <linux/const.h>
+
+#define __FORTIFY_INLINE extern __always_inline __attribute__((gnu_inline)) __overloadable
 #define __RENAME(x) __asm__(#x)
 
 void fortify_panic(const char *name) __noreturn __cold;
@@ -50,7 +52,11 @@  extern char *__underlying_strncpy(char *p, const char *q, __kernel_size_t size)
 #define __underlying_strncpy	__builtin_strncpy
 #endif
 
-__FORTIFY_INLINE char *strncpy(char *p, const char *q, __kernel_size_t size)
+#define BOS	const __pass_object_size(1)
+#define BOS0	const __pass_object_size(0)
+
+__FORTIFY_INLINE __diagnose_as(__builtin_strncpy, 1, 2, 3)
+char *strncpy(char * BOS p, const char *q, __kernel_size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 1);
 
@@ -61,7 +67,8 @@  __FORTIFY_INLINE char *strncpy(char *p, const char *q, __kernel_size_t size)
 	return __underlying_strncpy(p, q, size);
 }
 
-__FORTIFY_INLINE char *strcat(char *p, const char *q)
+__FORTIFY_INLINE __diagnose_as(__builtin_strcat, 1, 2)
+char *strcat(char * BOS p, const char *q)
 {
 	size_t p_size = __builtin_object_size(p, 1);
 
@@ -73,7 +80,7 @@  __FORTIFY_INLINE char *strcat(char *p, const char *q)
 }
 
 extern __kernel_size_t __real_strnlen(const char *, __kernel_size_t) __RENAME(strnlen);
-__FORTIFY_INLINE __kernel_size_t strnlen(const char *p, __kernel_size_t maxlen)
+__FORTIFY_INLINE __kernel_size_t strnlen(const char * BOS p, __kernel_size_t maxlen)
 {
 	size_t p_size = __builtin_object_size(p, 1);
 	size_t p_len = __compiletime_strlen(p);
@@ -93,8 +100,16 @@  __FORTIFY_INLINE __kernel_size_t strnlen(const char *p, __kernel_size_t maxlen)
 	return ret;
 }
 
-/* defined after fortified strnlen to reuse it. */
-__FORTIFY_INLINE __kernel_size_t strlen(const char *p)
+/*
+ * Defined after fortified strnlen to reuse it. However, it must still be
+ * possible for strlen() to be used on compile-time strings for use in
+ * static initializers (i.e. as a constant expression).
+ */
+#define strlen(p)							\
+	__builtin_choose_expr(__is_constexpr(__builtin_strlen(p)),	\
+		__builtin_strlen(p), __fortify_strlen(p))
+__FORTIFY_INLINE __diagnose_as(__builtin_strlen, 1)
+__kernel_size_t __fortify_strlen(const char * BOS p)
 {
 	__kernel_size_t ret;
 	size_t p_size = __builtin_object_size(p, 1);
@@ -110,7 +125,7 @@  __FORTIFY_INLINE __kernel_size_t strlen(const char *p)
 
 /* defined after fortified strlen to reuse it */
 extern size_t __real_strlcpy(char *, const char *, size_t) __RENAME(strlcpy);
-__FORTIFY_INLINE size_t strlcpy(char *p, const char *q, size_t size)
+__FORTIFY_INLINE size_t strlcpy(char * BOS p, const char * BOS q, size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 1);
 	size_t q_size = __builtin_object_size(q, 1);
@@ -137,7 +152,7 @@  __FORTIFY_INLINE size_t strlcpy(char *p, const char *q, size_t size)
 
 /* defined after fortified strnlen to reuse it */
 extern ssize_t __real_strscpy(char *, const char *, size_t) __RENAME(strscpy);
-__FORTIFY_INLINE ssize_t strscpy(char *p, const char *q, size_t size)
+__FORTIFY_INLINE ssize_t strscpy(char * BOS p, const char * BOS q, size_t size)
 {
 	size_t len;
 	/* Use string size rather than possible enclosing struct size. */
@@ -183,7 +198,8 @@  __FORTIFY_INLINE ssize_t strscpy(char *p, const char *q, size_t size)
 }
 
 /* defined after fortified strlen and strnlen to reuse them */
-__FORTIFY_INLINE char *strncat(char *p, const char *q, __kernel_size_t count)
+__FORTIFY_INLINE __diagnose_as(__builtin_strncat, 1, 2, 3)
+char *strncat(char * BOS p, const char * BOS q, __kernel_size_t count)
 {
 	size_t p_len, copy_len;
 	size_t p_size = __builtin_object_size(p, 1);
@@ -354,7 +370,7 @@  __FORTIFY_INLINE void fortify_memcpy_chk(__kernel_size_t size,
 		memmove)
 
 extern void *__real_memscan(void *, int, __kernel_size_t) __RENAME(memscan);
-__FORTIFY_INLINE void *memscan(void *p, int c, __kernel_size_t size)
+__FORTIFY_INLINE void *memscan(void * BOS0 p, int c, __kernel_size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 0);
 
@@ -365,7 +381,8 @@  __FORTIFY_INLINE void *memscan(void *p, int c, __kernel_size_t size)
 	return __real_memscan(p, c, size);
 }
 
-__FORTIFY_INLINE int memcmp(const void *p, const void *q, __kernel_size_t size)
+__FORTIFY_INLINE __diagnose_as(__builtin_memcmp, 1, 2, 3)
+int memcmp(const void * BOS0 p, const void * BOS0 q, __kernel_size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 0);
 	size_t q_size = __builtin_object_size(q, 0);
@@ -381,7 +398,8 @@  __FORTIFY_INLINE int memcmp(const void *p, const void *q, __kernel_size_t size)
 	return __underlying_memcmp(p, q, size);
 }
 
-__FORTIFY_INLINE void *memchr(const void *p, int c, __kernel_size_t size)
+__FORTIFY_INLINE __diagnose_as(__builtin_memchr, 1, 2, 3)
+void *memchr(const void * BOS0 p, int c, __kernel_size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 0);
 
@@ -393,7 +411,7 @@  __FORTIFY_INLINE void *memchr(const void *p, int c, __kernel_size_t size)
 }
 
 void *__real_memchr_inv(const void *s, int c, size_t n) __RENAME(memchr_inv);
-__FORTIFY_INLINE void *memchr_inv(const void *p, int c, size_t size)
+__FORTIFY_INLINE void *memchr_inv(const void * BOS0 p, int c, size_t size)
 {
 	size_t p_size = __builtin_object_size(p, 0);
 
@@ -405,7 +423,7 @@  __FORTIFY_INLINE void *memchr_inv(const void *p, int c, size_t size)
 }
 
 extern void *__real_kmemdup(const void *src, size_t len, gfp_t gfp) __RENAME(kmemdup);
-__FORTIFY_INLINE void *kmemdup(const void *p, size_t size, gfp_t gfp)
+__FORTIFY_INLINE void *kmemdup(const void * BOS0 p, size_t size, gfp_t gfp)
 {
 	size_t p_size = __builtin_object_size(p, 0);
 
@@ -417,7 +435,8 @@  __FORTIFY_INLINE void *kmemdup(const void *p, size_t size, gfp_t gfp)
 }
 
 /* Defined after fortified strlen to reuse it. */
-__FORTIFY_INLINE char *strcpy(char *p, const char *q)
+__FORTIFY_INLINE __diagnose_as(__builtin_strcpy, 1, 2)
+char *strcpy(char * BOS p, const char * BOS q)
 {
 	size_t p_size = __builtin_object_size(p, 1);
 	size_t q_size = __builtin_object_size(q, 1);
@@ -446,4 +465,7 @@  __FORTIFY_INLINE char *strcpy(char *p, const char *q)
 #undef __underlying_strncat
 #undef __underlying_strncpy
 
+#undef BOS0
+#undef BOS
+
 #endif /* _LINUX_FORTIFY_STRING_H_ */
diff --git a/security/Kconfig b/security/Kconfig
index 0b847f435beb..1a25a567965f 100644
--- a/security/Kconfig
+++ b/security/Kconfig
@@ -179,7 +179,7 @@  config FORTIFY_SOURCE
 	depends on ARCH_HAS_FORTIFY_SOURCE
 	# https://bugs.llvm.org/show_bug.cgi?id=50322
 	# https://bugs.llvm.org/show_bug.cgi?id=41459
-	depends on !CC_IS_CLANG
+	depends on !CC_IS_CLANG || CLANG_VERSION >= 130000
 	help
 	  Detect overflows of buffers in common string and memory functions
 	  where the compiler can determine and validate the buffer sizes.