userspace: compartmentalized app memory organization

Summary: revised attempt at addressing issue 6290.  The
following provides an alternative to using
CONFIG_APPLICATION_MEMORY by compartmentalizing data into
Memory Domains.  Dependent on MPU limitations, supports
compartmentalized Memory Domains for 1...N logical
applications.  This is considered an initial attempt at
designing flexible compartmentalized Memory Domains for
multiple logical applications and, with the provided python
script and edited CMakeLists.txt, provides support for power
of 2 aligned MPU architectures.

Overview: The current patch uses qualifiers to group data into
subsections.  The qualifier usage allows for dynamic subsection
creation and affords the developer a large amount of flexibility
in the grouping, naming, and size of the resulting partitions and
domains that are built on these subsections. By additional macro
calls, functions are created that help calculate the size,
address, and permissions for the subsections and enable the
developer to control application data in specified partitions and
memory domains.

Background: Initial attempts focused on creating a single
section in the linker script that then contained internally
grouped variables/data to allow MPU/MMU alignment and protection.
This did not provide additional functionality beyond
CONFIG_APPLICATION_MEMORY as we were unable to reliably group
data or determine their grouping via exported linker symbols.
Thus, the resulting decision was made to dynamically create
subsections using the current qualifier method. An attempt to
group the data by object file was tested, but found that this
broke applications such as ztest where two object files are
created: ztest and main.  This also creates an issue of grouping
the two object files together in the same memory domain while
also allowing for compartmenting other data among threads.

Because it is not possible to know a) the name of the partition
and thus the symbol in the linker, b) the size of all the data
in the subsection, nor c) the overall number of partitions
created by the developer, it was not feasible to align the
subsections at compile time without using dynamically generated
linker script for MPU architectures requiring power of 2
alignment.

In order to provide support for MPU architectures that require a
power of 2 alignment, a python script is run at build prior to
when linker_priv_stacks.cmd is generated.  This script scans the
built object files for all possible partitions and the names given
to them. It then generates a linker file (app_smem.ld) that is
included in the main linker.ld file.  This app_smem.ld allows the
compiler and linker to then create each subsection and align to
the next power of 2.

Usage:
 - Requires: app_memory/app_memdomain.h .
 - _app_dmem(id) marks a variable to be placed into a data
section for memory partition id.
 - _app_bmem(id) marks a variable to be placed into a bss
section for memory partition id.
 - These are seen in the linker.map as "data_smem_id" and
"data_smem_idb".
 - To create a k_mem_partition, call the macro
app_mem_partition(part0) where "part0" is the name then used to
refer to that partition. This macro only creates a function and
necessary data structures for the later "initialization".
 - To create a memory domain for the partition, the macro
app_mem_domain(dom0) is called where "dom0" is the name then
used for the memory domain.
 - To initialize the partition (effectively adding the partition
to a linked list), init_part_part0() is called. This is followed
by init_app_memory(), which walks all partitions in the linked
list and calculates the sizes for each partition.
 - Once the partition is initialized, the domain can be
initialized with init_domain_dom0(part0) which initializes the
domain with partition part0.
 - After the domain has been initialized, the current thread
can be added using add_thread_dom0(k_current_get()).
 - The code used in ztests ans kernel/init has been added under
a conditional #ifdef to isolate the code from other tests.
The userspace test CMakeLists.txt file has commands to insert
the CONFIG_APP_SHARED_MEM definition into the required build
targets.
  Example:
        /* create partition at top of file outside functions */
        app_mem_partition(part0);
        /* create domain */
        app_mem_domain(dom0);
        _app_dmem(dom0) int var1;
        _app_bmem(dom0) static volatile int var2;

        int main()
        {
                init_part_part0();
                init_app_memory();
                init_domain_dom0(part0);
                add_thread_dom0(k_current_get());
                ...
        }

 - If multiple partitions are being created, a variadic
preprocessor macro can be used as provided in
app_macro_support.h:

        FOR_EACH(app_mem_partition, part0, part1, part2);

or, for multiple domains, similarly:

        FOR_EACH(app_mem_domain, dom0, dom1);

Similarly, the init_part_* can also be used in the macro:

        FOR_EACH(init_part, part0, part1, part2);

Testing:
 - This has been successfully tested on qemu_x86 and the
ARM frdm_k64f board.  It compiles and builds power of 2
aligned subsections for the linker script on the 96b_carbon
boards.  These power of 2 alignments have been checked by
hand and are viewable in the zephyr.map file that is
produced during build. However, due to a shortage of
available MPU regions on the 96b_carbon board, we are unable
to test this.
 - When run on the 96b_carbon board, the test suite will
enter execution, but each individaul test will fail due to
an MPU FAULT.  This is expected as the required number of
MPU regions exceeds the number allowed due to the static
allocation. As the MPU driver does not detect this issue,
the fault occurs because the data being accessed has been
placed outside the active MPU region.
 - This now compiles successfully for the ARC boards
em_starterkit_em7d and em_starterkit_em7d_v22. However,
as we lack ARC hardware to run this build on, we are unable
to test this build.

Current known issues:
1) While the script and edited CMakeLists.txt creates the
ability to align to the next power of 2, this does not
address the shortage of available MPU regions on certain
devices (e.g. 96b_carbon).  In testing the APB and PPB
regions were commented out.
2) checkpatch.pl lists several issues regarding the
following:
a) Complex macros. The FOR_EACH macros as defined in
app_macro_support.h are listed as complex macros needing
parentheses.  Adding parentheses breaks their
functionality, and we have otherwise been unable to
resolve the reported error.
b) __aligned() preferred. The _app_dmem_pad() and
_app_bmem_pad() macros give warnings that __aligned()
is preferred. Prior iterations had this implementation,
which resulted in errors due to "complex macros".
c) Trailing semicolon. The macro init_part(name) has
a trailing semicolon as the semicolon is needed for the
inlined macro call that is generated when this macro
expands.

Update: updated to alternative CONFIG_APPLCATION_MEMORY.
Added config option CONFIG_APP_SHARED_MEM to enable a new section
app_smem to contain the shared memory component.  This commit
seperates the Kconfig definition from the definition used for the
conditional code.  The change is in response to changes in the
way the build system treats definitions.  The python script used
to generate a linker script for app_smem was also midified to
simplify the alignment directives.  A default linker script
app_smem.ld was added to remove the conditional includes dependency
on CONFIG_APP_SHARED_MEM.  By addining the default linker script
the prebuild stages link properly prior to the python script running

Signed-off-by: Joshua Domagalski <jedomag@tycho.nsa.gov>
Signed-off-by: Shawn Mosley <smmosle@tycho.nsa.gov>
This commit is contained in:
Shawn Mosley 2018-04-26 10:14:02 -04:00 committed by Andrew Boie
commit 573f32b6d2
25 changed files with 717 additions and 29 deletions

View file

@ -632,6 +632,9 @@ endif() # CONFIG_APPLICATION_MEMORY
# Declare MPU userspace dependencies before the linker scripts to make
# sure the order of dependencies are met
if(CONFIG_CPU_HAS_MPU AND CONFIG_USERSPACE)
if(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT AND CONFIG_APP_SHARED_MEM )
set(APP_SMEM_DEP app_smem_linker)
endif()
if(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT AND CONFIG_APPLICATION_MEMORY)
set(ALIGN_SIZING_DEP app_sizing_prebuilt linker_app_sizing_script)
endif()
@ -1008,8 +1011,33 @@ configure_file(
$ENV{ZEPHYR_BASE}/include/arch/arm/cortex_m/scripts/app_data_alignment.ld
${PROJECT_BINARY_DIR}/include/generated/app_data_alignment.ld)
configure_file(
$ENV{ZEPHYR_BASE}/include/arch/arm/cortex_m/scripts/app_smem.ld
${PROJECT_BINARY_DIR}/include/generated/app_smem.ld)
if(CONFIG_CPU_HAS_MPU AND CONFIG_USERSPACE)
if(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT AND CONFIG_APP_SHARED_MEM)
set(GEN_APP_SMEM $ENV{ZEPHYR_BASE}/scripts/gen_app_smem.py)
set(APP_SMEM_LD "${PROJECT_BINARY_DIR}/include/generated/app_smem.ld")
set(OBJ_FILE_DIR "${PROJECT_BINARY_DIR}/../")
add_custom_target(
${APP_SMEM_DEP} ALL
DEPENDS zephyr_prebuilt
)
add_custom_command(
TARGET ${APP_SMEM_DEP}
COMMAND ${PYTHON_EXECUTABLE} ${GEN_APP_SMEM}
-d ${OBJ_FILE_DIR}
-o ${APP_SMEM_LD}
WORKING_DIRECTORY ${PROJECT_BINARY_DIR}/
COMMENT "Generating power of 2 aligned app_smem linker section"
)
endif()
if(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT AND CONFIG_APPLICATION_MEMORY)
construct_add_custom_command_for_linker_pass(linker_app_sizing custom_command)
@ -1038,7 +1066,7 @@ if(CONFIG_CPU_HAS_MPU AND CONFIG_USERSPACE)
add_executable( app_sizing_prebuilt misc/empty_file.c)
target_link_libraries(app_sizing_prebuilt ${TOPT} ${PROJECT_BINARY_DIR}/linker_app_sizing.cmd ${zephyr_lnk})
set_property(TARGET app_sizing_prebuilt PROPERTY LINK_DEPENDS ${PROJECT_BINARY_DIR}/linker_app_sizing.cmd)
add_dependencies( app_sizing_prebuilt linker_app_sizing_script offsets)
add_dependencies( app_sizing_prebuilt linker_app_sizing_script offsets )
add_custom_command(
TARGET app_sizing_prebuilt
@ -1119,7 +1147,7 @@ if(GKOF OR GKSF)
add_executable( kernel_elf misc/empty_file.c ${GKSF})
target_link_libraries(kernel_elf ${GKOF} ${TOPT} ${PROJECT_BINARY_DIR}/linker_pass_final.cmd ${zephyr_lnk})
set_property(TARGET kernel_elf PROPERTY LINK_DEPENDS ${PROJECT_BINARY_DIR}/linker_pass_final.cmd)
add_dependencies( kernel_elf ${ALIGN_SIZING_DEP} ${PRIV_STACK_DEP} linker_pass_final_script)
add_dependencies( kernel_elf ${ALIGN_SIZING_DEP} ${PRIV_STACK_DEP} ${APP_SMEM_DEP} linker_pass_final_script)
else()
set(logical_target_for_zephyr_elf zephyr_prebuilt)
# Use the prebuilt elf as the final elf since we don't have a

View file

@ -25,7 +25,10 @@ MMU_BOOT_REGION((u32_t)&_image_rom_start, (u32_t)&_image_rom_size,
MMU_BOOT_REGION((u32_t)&__app_ram_start, (u32_t)&__app_ram_size,
MMU_ENTRY_WRITE | MMU_ENTRY_USER | MMU_ENTRY_EXECUTE_DISABLE);
#endif
#ifdef CONFIG_APP_SHARED_MEM
MMU_BOOT_REGION((u32_t)&_app_smem_start, (u32_t)&_app_smem_size,
MMU_ENTRY_WRITE | MMU_ENTRY_USER | MMU_ENTRY_EXECUTE_DISABLE);
#endif
/* __kernel_ram_size includes all unused memory, which is used for heaps.
* User threads cannot access this unless granted at runtime. This is done
* automatically for stacks.

View file

@ -190,3 +190,4 @@ for execution after the kernel starts:
memory_domain.rst
mpu_stack_objects.rst
mpu_userspace.rst
usermode_sharedmem.rst

View file

@ -0,0 +1,104 @@
.. _usermode_sharedmem:
Application Shared Memory
#########################
.. note::
In this document, we will cover the basic usage of enabling shared
memory using a template around app_memory subsystem.
Overview
********
The use of subsystem app_memory in userspace allows control of
shared memory between threads. The foundation of the implementation
consists of memory domains and partitions. Memory partitions are created
and used in the definition of variable to group them into a
common space. The memory partitions are linked to domains
that are then assigned to a thead. The process allows selective
access to memory from a thread and sharing of memory between two
threads by assigning a partion to two different domains. By using
the shared memory template, code to protect memory can be used
on different platform without the application needing to implement
specific handlers for each platform. Note the developer should understand
the hardware limitations in context to the maximum number of memory
partitions available to a thread. Specifically processors with MPU's
cannot support the same number of partitions as a MMU.
This specific implementation adds a wrapper to simplify the programmers
task of using the app_memmory subsystem through the use of macros and
a python script to generate the linker script. The linker script provides
the proper alignment for processors requiring power of two boundaries.
Without the wrapper, a developer is required to implement custom
linker scripts for each processor the project.
The general usage is as follows. Define CONFIG_APP_SHARED_MEM=y in the
proj.conf file in the project folder. Include app_memory/app_memdomain.h
in the userspace source file. Mark the variable to be placed in
a memory partition. The two markers are for data and bss respectivly:
_app_dmem(id) and _app_bmem(id). The id is used as the partition name.
The resulting section name can be seen in the linker.map as
"data_smem_id" and "data_smem_idb".
To create a k_mem_partition, call the macro app_mem_partition(part0)
where "part0" is the name then used to refer to that partition.
This macro only creates a function and necessary data structures for
the later "initialization".
To create a memory domain for the partition, the macro app_mem_domain(dom0)
is called where "dom0" is the name then used for the memory domain.
To initialize the partition (effectively adding the partition
to a linked list), init_part_part0() is called. This is followed
by init_app_memory(), which walks all partitions in the linked
list and calculates the sizes for each partition.
Once the partition is initialized, the domain can be
initialized with init_domain_dom0(part0) which initializes the
domain with partition part0.
After the domain has been initialized, the current thread
can be added using add_thread_dom0(k_current_get()).
Example:
.. code-block:: c
/* create partition at top of file outside functions */
app_mem_partition(part0);
/* create domain */
app_mem_domain(dom0);
/* assign variables to the domain */
_app_dmem(dom0) int var1;
_app_bmem(dom0) static volatile int var2;
int main()
{
init_part_part0();
init_app_memory();
init_domain_dom0(part0);
add_thread_dom0(k_current_get());
...
}
If multiple partitions are being created, a variadic
preprocessor macro can be used as provided in
app_macro_support.h:
.. code-block:: c
FOR_EACH(app_mem_partition, part0, part1, part2);
or, for multiple domains, similarly:
.. code-block:: c
FOR_EACH(app_mem_domain, dom0, dom1);
Similarly, the init_part_* can also be used in the macro:
.. code-block:: c
FOR_EACH(init_part, part0, part1, part2);

View file

@ -0,0 +1,134 @@
#ifndef _APP_MEMDOMAIN__H_
#define _APP_MEMDOMAIN__H_
#include <linker/linker-defs.h>
#include <misc/dlist.h>
#include <kernel.h>
#if defined(CONFIG_X86)
#define MEM_DOMAIN_ALIGN_SIZE _STACK_BASE_ALIGN
#elif defined(STACK_ALIGN)
#define MEM_DOMAIN_ALIGN_SIZE STACK_ALIGN
#else
#error "Not implemented for this architecture"
#endif
/*
* There has got to be a better way of doing this. This
* tries to ensure that a) each subsection has a
* data_smem_#id_b part and b) that each k_mem_partition
* matches the page size or MPU region. If there is no
* data_smem_#id_b subsection, then the size calculations
* will fail. Additionally, if each k_mem_partition does
* not match the page size or MPU region, then the
* partition will fail to be created.
* checkpatch.pl complains that __aligned(size) is
* preferred, but, if implemented, then complains about
* complex macro without parentheses.
*/
#define _app_dmem_pad(id) \
__attribute__((aligned(MEM_DOMAIN_ALIGN_SIZE), \
section("data_smem_" #id)))
#define _app_bmem_pad(id) \
__attribute__((aligned(MEM_DOMAIN_ALIGN_SIZE), \
section("data_smem_" #id "b")))
/*
* Qualifier to collect any object preceded with _app
* and place into section "data_smem_".
* _app_dmem(#) is for variables meant to be stored in .data .
* _app_bmem(#) is intended for static variables that are
* initialized to zero.
*/
#define _app_dmem(id) \
__attribute__((section("data_smem_" #id)))
#define _app_bmem(id) \
__attribute__((section("data_smem_" #id "b")))
/*
* Creation of a struct to save start addresses, sizes, and
* a pointer to a k_mem_partition. It also adds a linked
* list node.
*/
struct app_region {
char *dmem_start;
char *bmem_start;
u32_t smem_size;
u32_t dmem_size;
u32_t bmem_size;
struct k_mem_partition *partition;
sys_dnode_t lnode;
};
/*
* Declares a partition and provides a function to add the
* partition to the linke dlist and initialize the partition.
*/
#define appmem_partition(name) \
extern char *data_smem_##name; \
extern char *data_smem_##name##b; \
_app_dmem_pad(name) char name##_dmem_pad; \
_app_bmem_pad(name) char name##_bmem_pad; \
__kernel struct k_mem_partition mem_domain_##name; \
__kernel struct app_region name; \
static inline void appmem_init_part_##name(void) \
{ \
name.dmem_start = (char *)&data_smem_##name; \
name.bmem_start = (char *)&data_smem_##name##b; \
sys_dlist_append(&app_mem_list, &name.lnode); \
mem_domain_##name.start = (u32_t) name.dmem_start; \
mem_domain_##name.attr = K_MEM_PARTITION_P_RW_U_RW; \
name.partition = &mem_domain_##name; \
}
/*
* A wrapper around the k_mem_domain_* functions. Goal here was
* to a) differentiate these operations from the k_mem_domain*
* functions, and b) to simply the usage and handling of data
* types (i.e. app_region, k_mem_domain, etc).
*/
#define appmem_domain(name) \
__kernel struct k_mem_domain domain_##name; \
static inline void appmem_add_thread_##name(k_tid_t thread) \
{ \
k_mem_domain_add_thread(&domain_##name, thread); \
} \
static inline void appmem_rm_thread_##name(k_tid_t thread) \
{ \
k_mem_domain_remove_thread(thread); \
} \
static inline void appmem_add_part_##name(struct app_region region) \
{ \
k_mem_domain_add_partition(&domain_##name, \
&region.partition[0]); \
} \
static inline void appmem_rm_part_##name(struct app_region region) \
{ \
k_mem_domain_remove_partition(&domain_##name, \
&region.partition[0]); \
} \
static inline void appmem_init_domain_##name(struct app_region region) \
{ \
k_mem_domain_init(&domain_##name, 1, &region.partition); \
}
/*
* The following allows the FOR_EACH macro to call each partition's
* appmem_init_part_##name . Note: semicolon needed or else compiler
* complains as semicolon needed for function call once expanded by
* macro.
*/
#define appmem_init_part(name) \
appmem_init_part_##name();
extern sys_dlist_t app_mem_list;
extern void app_bss_zero(void);
extern void app_calc_size(void);
extern void appmem_init_app_memory(void);
#endif /* _APP_MEMDOMAIN__H_ */

View file

@ -127,6 +127,23 @@ SECTIONS {
GROUP_START(RAMABLE_REGION)
#include <app_data_alignment.ld>
SECTION_PROLOGUE(_APP_SMEM_SECTION_NAME, (OPTIONAL),)
{
MPU_MIN_SIZE_ALIGN
_image_ram_start = .;
_app_smem_start = .;
#if defined(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT)
#include <app_smem.ld>
#else
APP_SMEM_SECTION()
#endif
MPU_MIN_SIZE_ALIGN
_app_smem_end = .;
} GROUP_DATA_LINK_IN(RAMABLE_REGION, RAMABLE_REGION)
_app_smem_size = _app_smem_end - _app_smem_start;
_app_smem_rom_start = LOADADDR(_APP_SMEM_SECTION_NAME);
#ifdef CONFIG_APPLICATION_MEMORY
SECTION_DATA_PROLOGUE(_APP_DATA_SECTION_NAME, (OPTIONAL),)
{

View file

@ -0,0 +1,2 @@
/* space holder */
APP_SMEM_SECTION()

View file

@ -219,15 +219,29 @@ SECTIONS
}
#endif
#include <app_data_alignment.ld>
SECTION_PROLOGUE(_APP_SMEM_SECTION_NAME, (OPTIONAL),)
{
. = ALIGN(4);
_image_ram_start = .;
_app_smem_start = .;
#if defined(CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT)
#include <app_smem.ld>
#else
APP_SMEM_SECTION()
#endif
_app_smem_end = .;
_app_smem_size = _app_smem_end - _app_smem_start;
. = ALIGN(4);
} GROUP_DATA_LINK_IN(RAMABLE_REGION, ROMABLE_REGION)
_app_smem_rom_start = LOADADDR(_APP_SMEM_SECTION_NAME);
#ifdef CONFIG_APPLICATION_MEMORY
SECTION_DATA_PROLOGUE(_APP_DATA_SECTION_NAME, (OPTIONAL),)
{
#include <app_data_alignment.ld>
__app_ram_start = .;
__app_data_ram_start = .;
_image_ram_start = .;
APP_INPUT_SECTION(.data)
APP_INPUT_SECTION(".data.*")
__app_data_ram_end = .;
@ -271,9 +285,6 @@ SECTIONS
*/
. = ALIGN(4);
__bss_start = .;
#ifndef CONFIG_APPLICATION_MEMORY
_image_ram_start = .;
#endif
__kernel_ram_start = .;
KERNEL_INPUT_SECTION(.bss)

View file

@ -158,13 +158,24 @@ SECTIONS
/* RAMABLE_REGION */
GROUP_START(RAMABLE_REGION)
/* APP SHARED MEMORY REGION */
SECTION_PROLOGUE(_APP_SMEM_SECTION_NAME, (OPTIONAL),)
{
_image_ram_start = .;
_app_smem_start = .;
APP_SMEM_SECTION()
MMU_PAGE_ALIGN
_app_smem_end = .;
} GROUP_DATA_LINK_IN(RAMABLE_REGION, RAMABLE_REGION)
_app_smem_size = _app_smem_end - _app_smem_start;
_app_smem_rom_start = LOADADDR(_APP_SMEM_SECTION_NAME);
#ifdef CONFIG_APPLICATION_MEMORY
SECTION_DATA_PROLOGUE(_APP_DATA_SECTION_NAME, (OPTIONAL),)
{
#ifndef CONFIG_XIP
MMU_PAGE_ALIGN
#endif
_image_ram_start = .;
__app_ram_start = .;
__app_data_ram_start = .;
APP_INPUT_SECTION(.data)
@ -213,9 +224,6 @@ SECTIONS
* a multiple of 4 bytes.
*/
. = ALIGN(4);
#ifndef CONFIG_APPLICATION_MEMORY
_image_ram_start = .;
#endif
__kernel_ram_start = .;
__bss_start = .;

View file

@ -130,10 +130,12 @@
UTIL_LISTIFY(NUM_KERNEL_OBJECT_FILES, X, sect)
#define APP_INPUT_SECTION(sect) \
*(EXCLUDE_FILE (UTIL_LISTIFY(NUM_KERNEL_OBJECT_FILES, Y, ~)) sect)
#define APP_SMEM_SECTION() KEEP(*(SORT(data_smem_[_a-zA-Z0-9]*)))
#else
#define KERNEL_INPUT_SECTION(sect) *(sect)
#define APP_INPUT_SECTION(sect) *(sect)
#define APP_SMEM_SECTION() KEEP(*(SORT(data_smem_[_a-zA-Z0-9]*)))
#endif
@ -171,6 +173,15 @@ GDATA(__data_num_words)
#else /* ! _ASMLANGUAGE */
#include <zephyr/types.h>
/*
* The following are externs symbols from the linker. This enables
* the dynamic k_mem_domain and k_mem_partition creation and alignment
* to the section produced in the linker.
*/
extern char _app_smem_start[];
extern char _app_smem_end[];
extern char _app_smem_size[];
extern char _app_smem_rom_start[];
#ifdef CONFIG_APPLICATION_MEMORY
/* Memory owned by the application. Start and end will be aligned for memory

View file

@ -23,6 +23,7 @@
#define _BSS_SECTION_NAME bss
#define _NOINIT_SECTION_NAME noinit
#define _APP_SMEM_SECTION_NAME app_smem
#define _APP_DATA_SECTION_NAME app_datas
#define _APP_BSS_SECTION_NAME app_bss
#define _APP_NOINIT_SECTION_NAME app_noinit

View file

@ -400,5 +400,31 @@ static inline s64_t arithmetic_shift_right(s64_t value, u8_t shift)
#define MACRO_MAP_13(macro, a, ...) macro(a)MACRO_MAP_12(macro, __VA_ARGS__,)
#define MACRO_MAP_14(macro, a, ...) macro(a)MACRO_MAP_13(macro, __VA_ARGS__,)
#define MACRO_MAP_15(macro, a, ...) macro(a)MACRO_MAP_14(macro, __VA_ARGS__,)
/*
* The following provides variadic preprocessor macro support to
* help eliminate multiple, repetitive function/macro calls. This
* allows up to 10 "arguments" in addition to _call .
* Note - derived from work on:
* https://codecraft.co/2014/11/25/variadic-macros-tricks/
*/
#define _GET_ARG(_1, _2, _3, _4, _5, _6, _7, _8, _9, _10, N, ...) N
#define _for_0(_call, ...)
#define _for_1(_call, x) _call(x)
#define _for_2(_call, x, ...) _call(x) _for_1(_call, ##__VA_ARGS__)
#define _for_3(_call, x, ...) _call(x) _for_2(_call, ##__VA_ARGS__)
#define _for_4(_call, x, ...) _call(x) _for_3(_call, ##__VA_ARGS__)
#define _for_5(_call, x, ...) _call(x) _for_4(_call, ##__VA_ARGS__)
#define _for_6(_call, x, ...) _call(x) _for_5(_call, ##__VA_ARGS__)
#define _for_7(_call, x, ...) _call(x) _for_6(_call, ##__VA_ARGS__)
#define _for_8(_call, x, ...) _call(x) _for_7(_call, ##__VA_ARGS__)
#define _for_9(_call, x, ...) _call(x) _for_8(_call, ##__VA_ARGS__)
#define _for_10(_call, x, ...) _call(x) _for_9(_call, ##__VA_ARGS__)
#define FOR_EACH(x, ...) \
_GET_ARG(__VA_ARGS__, \
_for_10, _for_9, _for_8, _for_7, _for_6, _for_5, \
_for_4, _for_3, _for_2, _for_1, _for_0)(x, ##__VA_ARGS__)
#endif /* _UTIL__H_ */

View file

@ -187,6 +187,10 @@ void _data_copy(void)
memcpy(&__ccm_data_start, &__ccm_data_rom_start,
((u32_t) &__ccm_data_end - (u32_t) &__ccm_data_start));
#endif
#ifdef CONFIG_APP_SHARED_MEM
memcpy(&_app_smem_start, &_app_smem_rom_start,
((u32_t) &_app_smem_end - (u32_t) &_app_smem_start));
#endif
#ifdef CONFIG_APPLICATION_MEMORY
memcpy(&__app_data_ram_start, &__app_data_rom_start,
((u32_t) &__app_data_ram_end - (u32_t) &__app_data_ram_start));

142
scripts/gen_app_smem.py Executable file
View file

@ -0,0 +1,142 @@
#!/usr/bin/env python3
import sys
import argparse
import os
import re
import string
from elftools.elf.elffile import ELFFile
partitions=[]
variables=[]
app_smem = "/*\n * The following is a dynamically created linker section.\n" \
" * Please do not modify or delete.\n */\n"
dRegion_size = { "0x20":32 , \
"0x40":64, \
"0x80":128, \
"0x100":256, \
"0x200":512, \
"0x400":1024, \
"0x800":2048, \
"0x1000":4096, \
"0x2000":8192, \
"0x4000":16384, \
"0x8000":32768, \
"0x10000":65536, \
"0x20000":131072, \
"0x40000":262144, \
"0x80000":524288, \
"0x100000":1048576, \
"0x200000":2097152, \
"0x400000":4194304, \
"0x800000":8388608, \
"0x1000000":16777216, \
"0x2000000":33554432, \
"0x4000000":67108864, \
"0x8000000":134217728, \
"0x10000000":268435456, \
"0x20000000":536870912, \
"0x40000000":1073741824, \
"0x80000000":2147483648, \
"0x100000000":4294967296}
def build_linker_section(filename, regionSizes):
with open(filename, 'w') as f:
f.write(app_smem)
for part in partitions:
psize = "0"
if (part.endswith("b")):
continue
for partition, size in regionSizes:
if (partition == str(part).strip("b'")):
f.write("\t\t. = ALIGN(" + size + ");\n");
psize = size
f.write("\t\t" + str(part).strip("b'") + " = .;\n")
f.write("\t\t*(SORT(" + str(part).strip("b'") + "*))\n")
f.write("\t\t. = ALIGN(_app_data_align);\n")
f.write("\t\t" + str(part).strip("b'") + "b_end = .;\n")
f.write("\t\t. = ALIGN(" + psize + ");\n")
def find_variables(filename):
flag = 0
with open(filename, 'rb') as f:
objF = ELFFile( f)
if (not objF):
print("Error parsing file: ",filename)
os.exit(1)
sec = [ x for x in objF.iter_sections()]
for s in sec:
if ("smem" in s.name and not ".rel" in s.name):
variables.append( s)
def build_partitions():
global partitions
s = set()
for var in variables:
s.add(var.name)
for sec in sorted(s):
partitions.append(sec)
def calc_sec_size():
d = dict()
if not variables:
return None
for v in variables:
if( v.name in d):
d[v.name] += v.header.sh_size
else:
d[v.name] = v.header.sh_size
return d
def genRegionDef():
res = []
dSecs = calc_sec_size()
if(dSecs is None):
print("\n***\nError: Failed to identify smem regions in the object files\n***\n")
sys.exit(1)
ordered_defs = []
for key, value in sorted(dRegion_size.items(), key=lambda o: (o[1],o[0]) ):
ordered_defs.append( ( value, key))
ltSecs = sorted(dSecs.items())
szltSecs = len(ltSecs)
if( szltSecs % 2 == 1):
szltSecs += 1
for r in range(int(szltSecs /2 )):
i = r*2
sz = ltSecs[i][1] + ltSecs[1+i][1]
for d in ordered_defs:
if( sz < d[0]):
res.append( ( ltSecs[i][0], d[1]))
break
return res
def parse_args():
global args
parser = argparse.ArgumentParser(
description=__doc__,
formatter_class=argparse.RawDescriptionHelpFormatter)
parser.add_argument("-d", "--directory", required=True,
help="Root build directory")
parser.add_argument("-o", "--output", required=True,
help="Output ld file")
args = parser.parse_args()
def main():
parse_args()
startIndex = args.directory
fileOutput = args.output
for dirpath, dirs, files in os.walk(startIndex):
for filename in files:
if (filename.endswith(".obj") or filename.endswith(".OBJ")):
fullname = os.path.join(dirpath, filename)
find_variables(fullname)
build_partitions()
build_linker_section(fileOutput, genRegionDef())
if __name__ == '__main__':
main()

View file

@ -641,7 +641,7 @@ class SizeCalculator:
"kobject_data", "mmu_tables", "app_pad", "priv_stacks",
"ccm_data", "usb_descriptor", "usb_data", "usb_bos_desc",
'log_backends_sections', 'log_dynamic_sections',
'log_const_sections']
'log_const_sections',"app_smem"]
# These get copied into RAM only on non-XIP
ro_sections = ["text", "ctors", "init_array", "reset", "object_access",
"rodata", "devconfig", "net_l2", "vector", "_bt_settings_area"]

View file

@ -1,3 +1,4 @@
add_subdirectory_ifdef(CONFIG_APP_SHARED_MEM app_memory)
add_subdirectory(debug)
add_subdirectory(logging)
add_subdirectory_ifdef(CONFIG_BT bluetooth)

View file

@ -32,3 +32,5 @@ source "subsys/random/Kconfig"
source "subsys/storage/Kconfig"
source "subsys/settings/Kconfig"
source "subsys/app_memory/Kconfig"

View file

@ -0,0 +1,2 @@
zephyr_library()
zephyr_library_sources(app_memdomain.c)

View file

@ -0,0 +1,9 @@
menu "General Kernel Options"
config APP_SHARED_MEM
bool
prompt "Application shared memory with app_memory"
default n
help
This is a wrapper around app_memory to simplify usage.
endmenu

View file

@ -0,0 +1,81 @@
#include <app_memory/app_memdomain.h>
#include <misc/dlist.h>
#include <stdarg.h>
#include <string.h>
/*
* Initializes a double linked-list for the calculation of
* memory subsections.
*/
sys_dlist_t app_mem_list = SYS_DLIST_STATIC_INIT(&app_mem_list);
/*
* The following zeroizes each "bss" part of each subsection
* as per the entries in the list.
*/
void app_bss_zero(void)
{
sys_dnode_t *node, *next_node;
SYS_DLIST_FOR_EACH_NODE_SAFE(&app_mem_list, node, next_node)
{
struct app_region *region =
CONTAINER_OF(node, struct app_region, lnode);
memset(region->bmem_start, 0, region->bmem_size);
}
}
/*
* The following calculates the size of each subsection and adds
* the computed sizes to the region structures. These calculations
* are needed both for zeroizing "bss" parts of the partitions and
* for the creation of the k_mem_partition.
*/
void app_calc_size(void)
{
sys_dnode_t *node, *next_node;
SYS_DLIST_FOR_EACH_NODE_SAFE(&app_mem_list, node, next_node)
{
if (sys_dlist_is_tail(&app_mem_list, node)) {
struct app_region *region =
CONTAINER_OF(node, struct app_region, lnode);
region->bmem_size =
_app_smem_end -
(char *)region->bmem_start;
region->dmem_size =
(char *)region->bmem_start -
(char *)region->dmem_start;
region->smem_size =
region->bmem_size + region->dmem_size;
region->partition[0].size =
region->dmem_size + region->bmem_size;
} else {
struct app_region *region =
CONTAINER_OF(node, struct app_region, lnode);
struct app_region *nRegion =
CONTAINER_OF(next_node, struct app_region,
lnode);
region->bmem_size =
(char *)nRegion->dmem_start -
(char *)region->bmem_start;
region->dmem_size =
(char *)region->bmem_start -
(char *)region->dmem_start;
region->smem_size =
region->bmem_size + region->dmem_size;
region->partition[0].size =
region->dmem_size + region->bmem_size;
}
}
}
/*
* "Initializes" by calculating subsection sizes and then
* zeroizing "bss" regions.
*/
void appmem_init_app_memory(void)
{
app_calc_size();
app_bss_zero();
}

View file

@ -1,3 +1,4 @@
CONFIG_ZTEST=y
CONFIG_USERSPACE=y
CONFIG_APPLICATION_MEMORY=y
CONFIG_APPLICATION_MEMORY=n
CONFIG_APP_SHARED_MEM=y

View file

@ -12,6 +12,8 @@
#include <kernel_structs.h>
#include <string.h>
#include <stdlib.h>
#include <app_memory/app_memdomain.h>
#include <misc/util.h>
#if defined(CONFIG_ARC)
#include <arch/arc/v2/mpu/arc_core_mpu.h>
@ -26,8 +28,24 @@ K_SEM_DEFINE(uthread_end_sem, 0, 1);
K_SEM_DEFINE(test_revoke_sem, 0, 1);
K_SEM_DEFINE(expect_fault_sem, 0, 1);
static volatile bool give_uthread_end_sem;
static volatile bool expect_fault;
/*
* Create partitions. part0 is for all variables to run
* ztest and this test suite. part1 and part2 are for
* subsequent test specifically for this new implementation.
*/
FOR_EACH(appmem_partition, part0, part1, part2);
/*
* Create memory domains. dom0 is for the ztest and this
* test suite, specifically. dom1 is for a specific test
* in this test suite.
*/
FOR_EACH(appmem_domain, dom0, dom1);
_app_dmem(part0) static volatile bool give_uthread_end_sem;
_app_dmem(part0) bool mem_access_check;
_app_bmem(part0) static volatile bool expect_fault;
#if defined(CONFIG_X86)
#define REASON_HW_EXCEPTION _NANO_ERR_CPU_EXCEPTION
@ -41,7 +59,7 @@ static volatile bool expect_fault;
#else
#error "Not implemented for this architecture"
#endif
static volatile unsigned int expected_reason;
_app_bmem(part0) static volatile unsigned int expected_reason;
/*
* We need something that can act as a memory barrier
@ -243,15 +261,16 @@ static void write_kernel_data(void)
/*
* volatile to avoid compiler mischief.
*/
volatile int *priv_stack_ptr;
_app_dmem(part0) volatile int *priv_stack_ptr;
#if defined(CONFIG_X86)
/*
* We can't inline this in the code or make it static
* or local without triggering a warning on -Warray-bounds.
*/
size_t size = MMU_PAGE_SIZE;
_app_dmem(part0) size_t size = MMU_PAGE_SIZE;
#elif defined(CONFIG_ARC)
int32_t size = (0 - CONFIG_PRIVILEGED_STACK_SIZE - STACK_GUARD_SIZE);
_app_dmem(part0) s32_t size = (0 - CONFIG_PRIVILEGED_STACK_SIZE -
STACK_GUARD_SIZE);
#endif
static void read_priv_stack(void)
@ -297,7 +316,7 @@ static void write_priv_stack(void)
}
static struct k_sem sem;
_app_bmem(part0) static struct k_sem sem;
static void pass_user_object(void)
{
@ -462,7 +481,7 @@ static void user_mode_enter(void)
/* Define and initialize pipe. */
K_PIPE_DEFINE(kpipe, PIPE_LEN, BYTES_TO_READ_WRITE);
static size_t bytes_written_read;
_app_bmem(part0) static size_t bytes_written_read;
static void write_kobject_user_pipe(void)
{
@ -496,6 +515,56 @@ static void read_kobject_user_pipe(void)
"did not fault");
}
/* Removed test for access_non_app_memory
* due to the APPLICATION_MEMORY variable
* defaulting to y, when enabled the
* section app_bss is made available to
* all threads breaking the test
*/
/* Create bool in part1 partitions */
_app_dmem(part1) bool thread_bool;
static void shared_mem_thread(void)
{
/*
* Try to access thread_bool_1 in denied memory
* domain.
*/
expect_fault = true;
expected_reason = REASON_HW_EXCEPTION;
BARRIER();
thread_bool = false;
zassert_unreachable("Thread accessed global in other "
"memory domain\n");
}
static void access_other_memdomain(void)
{
/*
* Following tests the ability for a thread to access data
* in a domain that it is denied.
*/
/* initialize domain dom1 with partition part2 */
appmem_init_domain_dom1(part2);
/* add partition part0 for test globals */
appmem_add_part_dom1(part0);
/* remove current thread from domain dom0 */
appmem_rm_thread_dom0(k_current_get());
/* initialize domain with current thread*/
appmem_add_thread_dom1(k_current_get());
/* Create user mode thread */
k_thread_create(&uthread_thread, uthread_stack, STACKSIZE,
(k_thread_entry_t)shared_mem_thread, NULL,
NULL, NULL, -1, K_USER | K_INHERIT_PERMS, K_NO_WAIT);
k_thread_abort(k_current_get());
}
#if defined(CONFIG_ARM)
extern u8_t *_k_priv_stack_find(void *obj);
extern k_thread_stack_t ztest_thread_stack[];
@ -503,6 +572,20 @@ extern k_thread_stack_t ztest_thread_stack[];
void test_main(void)
{
/* partitions must be initialized first */
FOR_EACH(appmem_init_part, part0, part1, part2);
/*
* Next, the app_memory must be initialized in order to
* calculate size of the dynamically created subsections.
*/
appmem_init_app_memory();
/* Domain is initialized with partition part0 */
appmem_init_domain_dom0(part0);
/* Next, the partition must be added to the domain */
appmem_add_part_dom0(part1);
/* Finally, the current thread is added to domain */
appmem_add_thread_dom0(k_current_get());
#if defined(CONFIG_ARM)
priv_stack_ptr = (int *)_k_priv_stack_find(ztest_thread_stack);
#endif
@ -533,7 +616,9 @@ void test_main(void)
ztest_user_unit_test(access_after_revoke),
ztest_unit_test(user_mode_enter),
ztest_user_unit_test(write_kobject_user_pipe),
ztest_user_unit_test(read_kobject_user_pipe)
ztest_user_unit_test(read_kobject_user_pipe),
ztest_user_unit_test(read_kobject_user_pipe),
ztest_unit_test(access_other_memdomain)
);
ztest_run_test_suite(userspace);
}

View file

@ -2,3 +2,4 @@ tests:
kernel.memory_protection.userspace:
filter: CONFIG_ARCH_HAS_USERSPACE
tags: core security userspace ignore_faults
extra_sections: app_smem

View file

@ -13,6 +13,8 @@
#ifndef __ZTEST_TEST_H__
#define __ZTEST_TEST_H__
#include <app_memory/app_memdomain.h>
struct unit_test {
const char *name;
void (*test)(void);
@ -139,8 +141,17 @@ static inline void unit_test_noop(void)
*
* @param name Name of the testing suite
*/
/* definitions for use with testing application shared memory */
#ifdef CONFIG_APP_SHARED_MEM
#define APPDMEMP0 _app_dmem(part0)
#define APPBMEMP0 _app_bmem(part0)
#else
#define APPDMEMP0
#define APPBMEMP0
#endif
#define ztest_test_suite(name, ...) \
static struct unit_test _##name[] = { \
APPDMEMP0 static struct unit_test _##name[] = { \
__VA_ARGS__, { 0 } \
}
/**

View file

@ -6,19 +6,21 @@
#include <ztest.h>
#include <stdio.h>
#include <app_memory/app_memdomain.h>
#ifdef KERNEL
__kernel static struct k_thread ztest_thread;
#endif
enum {
/* APPDMEMP0 and APPBMEMP0 are used for the application shared memory test */
APPDMEMP0 enum {
TEST_PHASE_SETUP,
TEST_PHASE_TEST,
TEST_PHASE_TEARDOWN,
TEST_PHASE_FRAMEWORK
} phase = TEST_PHASE_FRAMEWORK;
static int test_status;
APPBMEMP0 static int test_status;
static int cleanup_test(struct unit_test *test)
{
@ -150,8 +152,9 @@ out:
K_THREAD_STACK_DEFINE(ztest_thread_stack, CONFIG_ZTEST_STACKSIZE +
CONFIG_TEST_EXTRA_STACKSIZE);
/* APPBMEMP0 is used for the application shared memory test */
APPBMEMP0 static int test_result;
static int test_result;
__kernel static struct k_sem test_end_signal;
void ztest_test_fail(void)