linux-kernel - Re: [PATCH] ARM: Add support for STACKLEAK gcc plugin

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] [day] [month] [year] [list]
Date: Fri, 21 Jun 2024 13:00:48 +0200
From: Ard Biesheuvel <ardb@...nel.org>
To: Jinjie Ruan <ruanjinjie@...wei.com>
Cc: linux@...linux.org.uk, arnd@...db.de, afd@...com, 
	akpm@...ux-foundation.org, rmk+kernel@...linux.org.uk, 
	linus.walleij@...aro.org, eric.devolder@...cle.com, robh@...nel.org, 
	kees@...nel.org, masahiroy@...nel.org, palmer@...osinc.com, 
	samitolvanen@...gle.com, xiao.w.wang@...el.com, alexghiti@...osinc.com, 
	nathan@...nel.org, jan.kiszka@...mens.com, 
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org, 
	linux-efi@...r.kernel.org
Subject: Re: [PATCH] ARM: Add support for STACKLEAK gcc plugin

On Thu, 20 Jun 2024 at 15:14, Jinjie Ruan <ruanjinjie@...wei.com> wrote:
>
> Add the STACKLEAK gcc plugin to arm32 by adding the helper used by
> stackleak common code: on_thread_stack(). It initialize the stack with the
> poison value before returning from system calls which improves the kernel
> security. Additionally, this disables the plugin in EFI stub code and
> decompress code, which are out of scope for the protection.
>
> Before the test on Qemu versatilepb board:
>         # echo STACKLEAK_ERASING  > /sys/kernel/debug/provoke-crash/DIRECT
>         lkdtm: Performing direct entry STACKLEAK_ERASING
>         lkdtm: XFAIL: stackleak is not supported on this arch (HAVE_ARCH_STACKLEAK=n)
>
> After:
>         # echo STACKLEAK_ERASING  > /sys/kernel/debug/provoke-crash/DIRECT
>         lkdtm: Performing direct entry STACKLEAK_ERASING
>         lkdtm: stackleak stack usage:
>           high offset: 80 bytes
>           current:     280 bytes
>           lowest:      696 bytes
>           tracked:     696 bytes
>           untracked:   192 bytes
>           poisoned:    7220 bytes
>           low offset:  4 bytes
>         lkdtm: OK: the rest of the thread stack is properly erased
>
> Signed-off-by: Jinjie Ruan <ruanjinjie@...wei.com>
> ---
>  arch/arm/Kconfig                      | 1 +
>  arch/arm/boot/compressed/Makefile     | 1 +
>  arch/arm/include/asm/stacktrace.h     | 5 +++++
>  arch/arm/kernel/entry-common.S        | 3 +++
>  drivers/firmware/efi/libstub/Makefile | 3 ++-
>  5 files changed, 12 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> index 036381c5d42f..b211b7f5a138 100644
> --- a/arch/arm/Kconfig
> +++ b/arch/arm/Kconfig
> @@ -86,6 +86,7 @@ config ARM
>         select HAVE_ARCH_PFN_VALID
>         select HAVE_ARCH_SECCOMP
>         select HAVE_ARCH_SECCOMP_FILTER if AEABI && !OABI_COMPAT
> +       select HAVE_ARCH_STACKLEAK
>         select HAVE_ARCH_THREAD_STRUCT_WHITELIST
>         select HAVE_ARCH_TRACEHOOK
>         select HAVE_ARCH_TRANSPARENT_HUGEPAGE if ARM_LPAE
> diff --git a/arch/arm/boot/compressed/Makefile b/arch/arm/boot/compressed/Makefile
> index 6bca03c0c7f0..945b5975fce2 100644
> --- a/arch/arm/boot/compressed/Makefile
> +++ b/arch/arm/boot/compressed/Makefile
> @@ -9,6 +9,7 @@ OBJS            =
>
>  HEAD   = head.o
>  OBJS   += misc.o decompress.o
> +CFLAGS_decompress.o += $(DISABLE_STACKLEAK_PLUGIN)
>  ifeq ($(CONFIG_DEBUG_UNCOMPRESS),y)
>  OBJS   += debug.o
>  AFLAGS_head.o += -DDEBUG
> diff --git a/arch/arm/include/asm/stacktrace.h b/arch/arm/include/asm/stacktrace.h
> index 360f0d2406bf..a9b4b72ed241 100644
> --- a/arch/arm/include/asm/stacktrace.h
> +++ b/arch/arm/include/asm/stacktrace.h
> @@ -26,6 +26,11 @@ struct stackframe {
>  #endif
>  };
>
> +static inline bool on_thread_stack(void)
> +{
> +       return !(((unsigned long)(current->stack) ^ current_stack_pointer) & ~(THREAD_SIZE - 1));

Can we make this a bit more legible, without the redundant parens?
Perhaps something like the below?


unsigned long delta = current_stack_pointer ^ (unsigned long)current->stack;

return delta < THREAD_SIZE;


Other than that, this looks fine to me

Acked-by: Ard Biesheuvel <ardb@...nel.org>


> +}
> +
>  static __always_inline
>  void arm_get_current_stackframe(struct pt_regs *regs, struct stackframe *frame)
>  {
> diff --git a/arch/arm/kernel/entry-common.S b/arch/arm/kernel/entry-common.S
> index 5c31e9de7a60..f379c852dcb7 100644
> --- a/arch/arm/kernel/entry-common.S
> +++ b/arch/arm/kernel/entry-common.S
> @@ -119,6 +119,9 @@ no_work_pending:
>
>         ct_user_enter save = 0
>
> +#ifdef CONFIG_GCC_PLUGIN_STACKLEAK
> +       bl      stackleak_erase_on_task_stack
> +#endif
>         restore_user_regs fast = 0, offset = 0
>  ENDPROC(ret_to_user_from_irq)
>  ENDPROC(ret_to_user)
> diff --git a/drivers/firmware/efi/libstub/Makefile b/drivers/firmware/efi/libstub/Makefile
> index 06f0428a723c..20d8a491f25f 100644
> --- a/drivers/firmware/efi/libstub/Makefile
> +++ b/drivers/firmware/efi/libstub/Makefile
> @@ -27,7 +27,8 @@ cflags-$(CONFIG_ARM64)                += -fpie $(DISABLE_STACKLEAK_PLUGIN) \
>  cflags-$(CONFIG_ARM)           += -DEFI_HAVE_STRLEN -DEFI_HAVE_STRNLEN \
>                                    -DEFI_HAVE_MEMCHR -DEFI_HAVE_STRRCHR \
>                                    -DEFI_HAVE_STRCMP -fno-builtin -fpic \
> -                                  $(call cc-option,-mno-single-pic-base)
> +                                  $(call cc-option,-mno-single-pic-base) \
> +                                  $(DISABLE_STACKLEAK_PLUGIN)
>  cflags-$(CONFIG_RISCV)         += -fpic -DNO_ALTERNATIVE -mno-relax
>  cflags-$(CONFIG_LOONGARCH)     += -fpie
>
> --
> 2.34.1
>
>