[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ea91bcfdf3ae32fb0f7ac4f866c63dbf5bbaa0ab.camel@intel.com>
Date: Thu, 3 Aug 2023 04:22:55 +0000
From: "Huang, Kai" <kai.huang@...el.com>
To: "linux-sgx@...r.kernel.org" <linux-sgx@...r.kernel.org>,
"jarkko@...nel.org" <jarkko@...nel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Van Bulck, Jo" <jo.vanbulck@...kuleuven.be>
CC: "dave.hansen@...ux.intel.com" <dave.hansen@...ux.intel.com>
Subject: Re: [PATCH 4/5] selftests/sgx: Ensure expected enclave data buffer
size and placement.
On Mon, 2023-07-24 at 18:58 +0200, Jo Van Bulck wrote:
> Do not declare the enclave data buffer static to ensure it is not optimized
> away by the compiler, even when not used entirely by the test enclave code.
The "encl_buffer" array is initialized to 1 explicitly, which means it should be
in .data section. As Jarkko commented, can you add more information about in
what cases it can be optimized away?
> Use -fPIE to make the compiler access the non-static buffer with
> RIP-relative addressing.
>
I am not quite following how does "RIP-relative addressing" fix any problem? Is
there any hard relationship between "RIP-relative addressing" and the problem
that you are having?
> Place the enclave data buffer in a separate
> section that is explicitly placed at the start of the .data segment in the
> linker script, as expected by the external tests manipulating page
> permissions.
So this change is not to fix the problem that "the compiler may optimize away
the encl_buffer array", correct?
>
> Signed-off-by: Jo Van Bulck <jo.vanbulck@...kuleuven.be>
> ---
> tools/testing/selftests/sgx/Makefile | 2 +-
> tools/testing/selftests/sgx/test_encl.c | 5 +++--
> tools/testing/selftests/sgx/test_encl.lds | 1 +
> 3 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/tools/testing/selftests/sgx/Makefile b/tools/testing/selftests/sgx/Makefile
> index 50aab6b57da3..c5483445ba28 100644
> --- a/tools/testing/selftests/sgx/Makefile
> +++ b/tools/testing/selftests/sgx/Makefile
> @@ -13,7 +13,7 @@ endif
>
> INCLUDES := -I$(top_srcdir)/tools/include
> HOST_CFLAGS := -Wall -Werror -g $(INCLUDES) -fPIC -z noexecstack
> -ENCL_CFLAGS := -Wall -Werror -static -nostdlib -nostartfiles -fPIC \
> +ENCL_CFLAGS := -Wall -Werror -static -nostdlib -nostartfiles -fPIE \
> -fno-stack-protector -mrdrnd $(INCLUDES)
>
> TEST_CUSTOM_PROGS := $(OUTPUT)/test_sgx
> diff --git a/tools/testing/selftests/sgx/test_encl.c b/tools/testing/selftests/sgx/test_encl.c
> index aba301abefb8..5c274e517d13 100644
> --- a/tools/testing/selftests/sgx/test_encl.c
> +++ b/tools/testing/selftests/sgx/test_encl.c
> @@ -7,9 +7,10 @@
> /*
> * Data buffer spanning two pages that will be placed first in .data
> * segment. Even if not used internally the second page is needed by
> - * external test manipulating page permissions.
> + * external test manipulating page permissions. Do not declare this
> + * buffer as static, so the compiler cannot optimize it out.
> */
> -static uint8_t encl_buffer[8192] = { 1 };
> +uint8_t __attribute__((section(".data.encl_buffer"))) encl_buffer[8192];
>
> enum sgx_enclu_function {
> EACCEPT = 0x5,
> diff --git a/tools/testing/selftests/sgx/test_encl.lds b/tools/testing/selftests/sgx/test_encl.lds
> index ca659db2a534..79b1e41d8d24 100644
> --- a/tools/testing/selftests/sgx/test_encl.lds
> +++ b/tools/testing/selftests/sgx/test_encl.lds
> @@ -24,6 +24,7 @@ SECTIONS
> } : text
>
> .data : {
> + *(.data.encl_buffer)
> *(.data*)
> } : data
>
Powered by blists - more mailing lists