2021-08-18 21:41:32

by Kees Cook

[permalink] [raw]
Subject: [PATCH v2 1/7] Compiler Attributes: Add __alloc_size() for better bounds checking

GCC and Clang can use the "alloc_size" attribute to better inform the
results of __builtin_object_size() (for compile-time constant values).
Clang can additionally use alloc_size to inform the results of
__builtin_dynamic_object_size() (for run-time values).

Because GCC sees the frequent use of struct_size() as an allocator size
argument, and notices it can return SIZE_MAX (the overflow indication),
it complains about these call sites may overflow (since SIZE_MAX is
greater than the default -Walloc-size-larger-than=PTRDIFF_MAX). This
isn't helpful since we already know a SIZE_MAX will be caught at run-time
(this was an intentional design). Instead, just disable this check as
it is both a false positive and redundant. (Clang does not have this
warning option.)

Cc: Miguel Ojeda <[email protected]>
Cc: Nathan Chancellor <[email protected]>
Cc: Nick Desaulniers <[email protected]>
Cc: [email protected]
Signed-off-by: Kees Cook <[email protected]>
---
Makefile | 6 +++++-
include/linux/compiler_attributes.h | 6 ++++++
2 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/Makefile b/Makefile
index 72f9e2b0202c..34cffcdfd5dc 100644
--- a/Makefile
+++ b/Makefile
@@ -1078,9 +1078,13 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
# Another good warning that we'll want to enable eventually
KBUILD_CFLAGS += $(call cc-disable-warning, restrict)

-# Enabled with W=2, disabled by default as noisy
ifdef CONFIG_CC_IS_GCC
+# Enabled with W=2, disabled by default as noisy
KBUILD_CFLAGS += -Wno-maybe-uninitialized
+
+# The allocators already balk at large sizes, so silence the compiler
+# warnings for bounds checks involving those possible values.
+KBUILD_CFLAGS += -Wno-alloc-size-larger-than
endif

# disable invalid "can't wrap" optimizations for signed / pointers
diff --git a/include/linux/compiler_attributes.h b/include/linux/compiler_attributes.h
index 67c5667f8042..203b0ac62d15 100644
--- a/include/linux/compiler_attributes.h
+++ b/include/linux/compiler_attributes.h
@@ -54,6 +54,12 @@
#define __aligned(x) __attribute__((__aligned__(x)))
#define __aligned_largest __attribute__((__aligned__))

+/*
+ * gcc: https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#index-alloc_005fsize-function-attribute
+ * clang: https://clang.llvm.org/docs/AttributeReference.html#alloc-size
+ */
+#define __alloc_size(x, ...) __attribute__((__alloc_size__(x, ## __VA_ARGS__)))
+
/*
* Note: users of __always_inline currently do not write "inline" themselves,
* which seems to be required by gcc to apply the attribute according
--
2.30.2


2021-08-18 21:52:48

by Nathan Chancellor

[permalink] [raw]
Subject: Re: [PATCH v2 1/7] Compiler Attributes: Add __alloc_size() for better bounds checking

On 8/18/2021 2:40 PM, Kees Cook wrote:
> GCC and Clang can use the "alloc_size" attribute to better inform the
> results of __builtin_object_size() (for compile-time constant values).
> Clang can additionally use alloc_size to inform the results of
> __builtin_dynamic_object_size() (for run-time values).
>
> Because GCC sees the frequent use of struct_size() as an allocator size
> argument, and notices it can return SIZE_MAX (the overflow indication),
> it complains about these call sites may overflow (since SIZE_MAX is
> greater than the default -Walloc-size-larger-than=PTRDIFF_MAX). This
> isn't helpful since we already know a SIZE_MAX will be caught at run-time
> (this was an intentional design). Instead, just disable this check as
> it is both a false positive and redundant. (Clang does not have this
> warning option.)
>
> Cc: Miguel Ojeda <[email protected]>
> Cc: Nathan Chancellor <[email protected]>
> Cc: Nick Desaulniers <[email protected]>
> Cc: [email protected]
> Signed-off-by: Kees Cook <[email protected]>

Reviewed-by: Nathan Chancellor <[email protected]>

> ---
> Makefile | 6 +++++-
> include/linux/compiler_attributes.h | 6 ++++++
> 2 files changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/Makefile b/Makefile
> index 72f9e2b0202c..34cffcdfd5dc 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -1078,9 +1078,13 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
> # Another good warning that we'll want to enable eventually
> KBUILD_CFLAGS += $(call cc-disable-warning, restrict)
>
> -# Enabled with W=2, disabled by default as noisy
> ifdef CONFIG_CC_IS_GCC
> +# Enabled with W=2, disabled by default as noisy
> KBUILD_CFLAGS += -Wno-maybe-uninitialized
> +
> +# The allocators already balk at large sizes, so silence the compiler
> +# warnings for bounds checks involving those possible values.
> +KBUILD_CFLAGS += -Wno-alloc-size-larger-than
> endif
>
> # disable invalid "can't wrap" optimizations for signed / pointers
> diff --git a/include/linux/compiler_attributes.h b/include/linux/compiler_attributes.h
> index 67c5667f8042..203b0ac62d15 100644
> --- a/include/linux/compiler_attributes.h
> +++ b/include/linux/compiler_attributes.h
> @@ -54,6 +54,12 @@
> #define __aligned(x) __attribute__((__aligned__(x)))
> #define __aligned_largest __attribute__((__aligned__))
>
> +/*
> + * gcc: https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#index-alloc_005fsize-function-attribute
> + * clang: https://clang.llvm.org/docs/AttributeReference.html#alloc-size
> + */
> +#define __alloc_size(x, ...) __attribute__((__alloc_size__(x, ## __VA_ARGS__)))
> +
> /*
> * Note: users of __always_inline currently do not write "inline" themselves,
> * which seems to be required by gcc to apply the attribute according
>

2021-08-18 23:21:41

by Andrew Morton

[permalink] [raw]
Subject: Re: [PATCH v2 1/7] Compiler Attributes: Add __alloc_size() for better bounds checking

On Wed, 18 Aug 2021 14:40:15 -0700 Kees Cook <[email protected]> wrote:

> GCC and Clang can use the "alloc_size" attribute to better inform the
> results of __builtin_object_size() (for compile-time constant values).
> Clang can additionally use alloc_size to inform the results of
> __builtin_dynamic_object_size() (for run-time values).
>
> Because GCC sees the frequent use of struct_size() as an allocator size
> argument, and notices it can return SIZE_MAX (the overflow indication),
> it complains about these call sites may overflow (since SIZE_MAX is
> greater than the default -Walloc-size-larger-than=PTRDIFF_MAX). This
> isn't helpful since we already know a SIZE_MAX will be caught at run-time
> (this was an intentional design). Instead, just disable this check as
> it is both a false positive and redundant. (Clang does not have this
> warning option.)
>
> ...
>
> --- a/Makefile
> +++ b/Makefile
> @@ -1078,9 +1078,13 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
> # Another good warning that we'll want to enable eventually
> KBUILD_CFLAGS += $(call cc-disable-warning, restrict)
>
> -# Enabled with W=2, disabled by default as noisy
> ifdef CONFIG_CC_IS_GCC
> +# Enabled with W=2, disabled by default as noisy
> KBUILD_CFLAGS += -Wno-maybe-uninitialized
> +
> +# The allocators already balk at large sizes, so silence the compiler
> +# warnings for bounds checks involving those possible values.
> +KBUILD_CFLAGS += -Wno-alloc-size-larger-than
> endif
>
> # disable invalid "can't wrap" optimizations for signed / pointers

Makefile has changed. I did this:

--- a/Makefile~compiler-attributes-add-__alloc_size-for-better-bounds-checking
+++ a/Makefile
@@ -1003,6 +1003,12 @@ KBUILD_CFLAGS += $(call cc-disable-warni
# Enabled with W=2, disabled by default as noisy
KBUILD_CFLAGS += $(call cc-disable-warning, maybe-uninitialized)

+ifdef CONFIG_CC_IS_GCC
+# The allocators already balk at large sizes, so silence the compiler
+# warnings for bounds checks involving those possible values.
+KBUILD_CFLAGS += -Wno-alloc-size-larger-than
+endif
+
# disable invalid "can't wrap" optimizations for signed / pointers
KBUILD_CFLAGS += -fno-strict-overflow

_

2021-08-19 00:06:54

by Miguel Ojeda

[permalink] [raw]
Subject: Re: [PATCH v2 1/7] Compiler Attributes: Add __alloc_size() for better bounds checking

On Wed, Aug 18, 2021 at 11:40 PM Kees Cook <[email protected]> wrote:
>
> GCC and Clang can use the "alloc_size" attribute to better inform the
> results of __builtin_object_size() (for compile-time constant values).
> Clang can additionally use alloc_size to inform the results of
> __builtin_dynamic_object_size() (for run-time values).
>
> Because GCC sees the frequent use of struct_size() as an allocator size
> argument, and notices it can return SIZE_MAX (the overflow indication),
> it complains about these call sites may overflow (since SIZE_MAX is
> greater than the default -Walloc-size-larger-than=PTRDIFF_MAX). This
> isn't helpful since we already know a SIZE_MAX will be caught at run-time
> (this was an intentional design). Instead, just disable this check as
> it is both a false positive and redundant. (Clang does not have this
> warning option.)

Thanks!

Reviewed-by: Miguel Ojeda <[email protected]>

Cheers,
Miguel

2021-08-19 07:07:23

by Kees Cook

[permalink] [raw]
Subject: Re: [PATCH v2 1/7] Compiler Attributes: Add __alloc_size() for better bounds checking

On Wed, Aug 18, 2021 at 04:19:12PM -0700, Andrew Morton wrote:
> On Wed, 18 Aug 2021 14:40:15 -0700 Kees Cook <[email protected]> wrote:
>
> > GCC and Clang can use the "alloc_size" attribute to better inform the
> > results of __builtin_object_size() (for compile-time constant values).
> > Clang can additionally use alloc_size to inform the results of
> > __builtin_dynamic_object_size() (for run-time values).
> >
> > Because GCC sees the frequent use of struct_size() as an allocator size
> > argument, and notices it can return SIZE_MAX (the overflow indication),
> > it complains about these call sites may overflow (since SIZE_MAX is
> > greater than the default -Walloc-size-larger-than=PTRDIFF_MAX). This
> > isn't helpful since we already know a SIZE_MAX will be caught at run-time
> > (this was an intentional design). Instead, just disable this check as
> > it is both a false positive and redundant. (Clang does not have this
> > warning option.)
> >
> > ...
> >
> > --- a/Makefile
> > +++ b/Makefile
> > @@ -1078,9 +1078,13 @@ KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
> > # Another good warning that we'll want to enable eventually
> > KBUILD_CFLAGS += $(call cc-disable-warning, restrict)
> >
> > -# Enabled with W=2, disabled by default as noisy
> > ifdef CONFIG_CC_IS_GCC
> > +# Enabled with W=2, disabled by default as noisy
> > KBUILD_CFLAGS += -Wno-maybe-uninitialized
> > +
> > +# The allocators already balk at large sizes, so silence the compiler
> > +# warnings for bounds checks involving those possible values.
> > +KBUILD_CFLAGS += -Wno-alloc-size-larger-than
> > endif
> >
> > # disable invalid "can't wrap" optimizations for signed / pointers
>
> Makefile has changed. I did this:
>
> --- a/Makefile~compiler-attributes-add-__alloc_size-for-better-bounds-checking
> +++ a/Makefile
> @@ -1003,6 +1003,12 @@ KBUILD_CFLAGS += $(call cc-disable-warni
> # Enabled with W=2, disabled by default as noisy
> KBUILD_CFLAGS += $(call cc-disable-warning, maybe-uninitialized)
>
> +ifdef CONFIG_CC_IS_GCC
> +# The allocators already balk at large sizes, so silence the compiler
> +# warnings for bounds checks involving those possible values.
> +KBUILD_CFLAGS += -Wno-alloc-size-larger-than
> +endif
> +
> # disable invalid "can't wrap" optimizations for signed / pointers
> KBUILD_CFLAGS += -fno-strict-overflow

Oh, er, where did "Makefile: remove stale cc-option checks" go? Ah, I
see now:
https://lore.kernel.org/mm-commits/20210814215814.W_qqW%[email protected]/T/#u

Looks like I just happened to pick the wrong linux-next. ;)

Thanks for the fix-up!

--
Kees Cook