libutils: mempool based raw malloc functions

Instead of the old stack like internal memory allocator, use the raw
malloc functions instead for more efficient memory usage.

CFG_WITH_STATS is enabled

libutils: mempool based raw malloc functions

Instead of the old stack like internal memory allocator, use the raw
malloc functions instead for more efficient memory usage.

CFG_WITH_STATS is enabled automatically if
CFG_MEMPOOL_REPORT_LAST_OFFSET is enabled to secure a new dependency in
the code.

Acked-by: Sahil Malhotra <sahil.malhotra@nxp.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: export raw malloc functions

Exports raw_{memalign,malloc,free,calloc,realloc}() and also adds
raw_malloc_get_ctx_size(), raw_malloc_init_ctx(),
raw_malloc_add_pool() and raw_malloc_get_sta

libutils: export raw malloc functions

Exports raw_{memalign,malloc,free,calloc,realloc}() and also adds
raw_malloc_get_ctx_size(), raw_malloc_init_ctx(),
raw_malloc_add_pool() and raw_malloc_get_stats().

This allows using the malloc functions to allocate with a independent
memory pool.

Acked-by: Sahil Malhotra <sahil.malhotra@nxp.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: provide __sprintf_chk implementation

While building optee_test CXX test-cases natively on aarch64, OP-TEE
build relies on toolchain provided by buildroot. The buildroot toolchain
is built

libutils: provide __sprintf_chk implementation

While building optee_test CXX test-cases natively on aarch64, OP-TEE
build relies on toolchain provided by buildroot. The buildroot toolchain
is built with flag: -fstack-protector-strong which requires
__sprintf_chk symbol provided by standard glibc. For OP-TEE we use a
customized libc which leads to below error:

...
CC out/init.o
CC out/os_test.o
CC out/ta_entry.o
CXX out/cxx_tests.o
CC out/user_ta_header.o
CPP out/ta.lds
LD out/5b9e0e40-2636-11e1-ad9e-0002a5d5c51b.elf
/home/sumit/optee_br/build/../toolchains/aarch64/bin/aarch64-linux-ld.bfd: /home/sumit/optee_br/toolchains/aarch64/bin/../lib/gcc/aarch64-buildroot-linux-gnu/10.3.0/../../../../aarch64-buildroot-linux-gnu/lib/../lib64/libstdc++.a(cp-demangle.o): in function d_append_num': cp-demangle.c:(.text+0x830): undefined reference to __sprintf_chk'
...

Fix this issue by providing __sprintf_chk implementation.

Signed-off-by: Sumit Garg <sumit.garg@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

lib: libutils: Use ret for longjmp with Clang and BTI

longjmp uses br instead of ret to jump to the target. Thus the target
location should have the right BTI launchpad to handle this.
clang has a b

lib: libutils: Use ret for longjmp with Clang and BTI

longjmp uses br instead of ret to jump to the target. Thus the target
location should have the right BTI launchpad to handle this.
clang has a bug [1] and doesn't add the BTI after setjmp causing
exception when BTI is enabled. This works well with gcc [2]
and can be tested when compiling xtests with WITH_CXX_TESTS=n.
To avoid the exception, use ret instead of br with clang and BTI.

[1] - https://bugs.llvm.org/show_bug.cgi?id=49544
[2] - https://gcc.gnu.org/legacy-ml/gcc-patches/2018-11/msg02472.html

Signed-off-by: Ruchika Gupta <ruchika.gupta@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>

show more ...

Generate ELF Note for BTI in all arm64 asm files

Add program property note section in the assembly files to
ensure that when linking them, program property note section
is generated in the final ELF

Generate ELF Note for BTI in all arm64 asm files

Add program property note section in the assembly files to
ensure that when linking them, program property note section
is generated in the final ELF.

Signed-off-by: Ruchika Gupta <ruchika.gupta@linaro.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

Add BTI launch pads in aarch64 assembly files

Compiler adds BTI launchpads only in C source files. For
assembly files, BTI launchpad is also required at locations
where "br" is used and at the start

Add BTI launch pads in aarch64 assembly files

Compiler adds BTI launchpads only in C source files. For
assembly files, BTI launchpad is also required at locations
where "br" is used and at the start of the functions. This
needs to be added manually.

Signed-off-by: Ruchika Gupta <ruchika.gupta@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: qsort.c: fix undefined pointer subtraction

Clang 13.0.0 produces the following warning:

lib/libutils/isoc/qsort.c:81:8: warning: performing pointer subtraction with a null pointer has un

libutils: qsort.c: fix undefined pointer subtraction

Clang 13.0.0 produces the following warning:

lib/libutils/isoc/qsort.c:81:8: warning: performing pointer subtraction with a null pointer has undefined behavior [-Wnull-pointer-subtraction]
loop: SWAPINIT(a, es);
^~~~~~~~~~~~~~~
lib/libutils/isoc/qsort.c:50:47: note: expanded from macro 'SWAPINIT'
#define SWAPINIT(a, es) swaptype = ((char *)a - (char *)0) % sizeof(long)
^ ~~~~~~~~~

Replace the subtraction with a simple cast to uintptr_t.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: ta: provide malloc(), calloc() and realloc() when debug is on

When enabling malloc debug for TAs (CFG_TEE_TA_MALLOC_DEBUG=y), the
standard malloc entry points malloc(), calloc() and reallo

libutils: ta: provide malloc(), calloc() and realloc() when debug is on

When enabling malloc debug for TAs (CFG_TEE_TA_MALLOC_DEBUG=y), the
standard malloc entry points malloc(), calloc() and realloc() are
redirected by C macros to instrumented variants: mdbg_malloc(),
mdbg_calloc() and mdbg_realloc(). In addition, the 'normal' symbols are
not exported by libutils. That is a problem because a TA might still
reference them. For example the C++ code in optee_test requires libstdc++
which relies on malloc() etc.:

build (master)$ make -j10 CFG_TEE_TA_MALLOC_DEBUG=y CFG_TEE_TA_LOG_LEVEL=2
...
/home/jerome/work/optee_repo_qemu/build/../toolchains/aarch32/bin/arm-linux-gnueabihf-ld.bfd: /home/jerome/work/toolchains-gcc10.2/aarch32/bin/../lib/gcc/arm-none-linux-gnueabihf/10.2.1/../../../../arm-none-linux-gnueabihf/lib/libstdc++.a(eh_alloc.o): in function `__cxa_allocate_exception':
/tmp/dgboter/bbs/build03--cen7x86_64/buildbot/cen7x86_64--arm-none-linux-gnueabihf/build/src/gcc/libstdc++-v3/libsupc++/eh_alloc.cc:284: undefined reference to `malloc'
...

Fix the issue by defining the standard malloc() functions in libutils,
calling the debug variants.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

libutils: bget_malloc: fix test in pool min size

Requires at least 1 kB for the initial malloc memory pool. The rational
is that the initial pool min size is not straightforward to compute as
it dep

libutils: bget_malloc: fix test in pool min size

Requires at least 1 kB for the initial malloc memory pool. The rational
is that the initial pool min size is not straightforward to compute as
it depends on the internals of the BGET implementation. However, with
a requirement of at least 1 kB in initial memory pool we'll have good
margin while still being reasonable.

Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: fix range check in gen_malloc_add_pool()

Prior to this patch was the length of supplied buffer not checked
accurately for very small buffer. This could result in an unexpected
assert():
E/

libutils: fix range check in gen_malloc_add_pool()

Prior to this patch was the length of supplied buffer not checked
accurately for very small buffer. This could result in an unexpected
assert():
E/TC:0 0 assertion 'start < end' failed at lib/libutils/isoc/bget_malloc.c:775 <gen_malloc_add_pool>
E/TC:0 0 Panic at core/kernel/assert.c:28 <_assert_break>

So fix this with a proper test.

Reviewed-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: sys/queue.h: add SLIST_FOREACH_SAFE()

Import macro SLIST_FOREACH_SAFE from FreeBSD.

Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissie

libutils: sys/queue.h: add SLIST_FOREACH_SAFE()

Import macro SLIST_FOREACH_SAFE from FreeBSD.

Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>
Acked-by: Jerome Forissier <jerome@forissier.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: sys/queue.h: add LIST_FOREACH_SAFE()

Import macro LIST_FOREACH_SAFE from FreeBSD.

Signed-off-by: Ruchika Gupta <ruchika.gupta@linaro.org>
Reviewed-by: Rouven Czerwinski <r.czerwinski@peng

libutils: sys/queue.h: add LIST_FOREACH_SAFE()

Import macro LIST_FOREACH_SAFE from FreeBSD.

Signed-off-by: Ruchika Gupta <ruchika.gupta@linaro.org>
Reviewed-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

lib: libutils: ext/isoc: sub.mk: make sources path platform generic

The path to platform specific code is hard-coded. This commit
changes it to use defined variable. This is helpful in case
of porti

lib: libutils: ext/isoc: sub.mk: make sources path platform generic

The path to platform specific code is hard-coded. This commit
changes it to use defined variable. This is helpful in case
of porting OP-TEE OS to a new architecture such we make
maximum reuse of existing sources.

Signed-off-by: Marouene Boubakri <marouene.boubakri@nxp.com>
Acked-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Jerome Forissier <jerome@forissier.org>

show more ...

Reintroduce memalign() and friends

memalign() and friends where removed with the
commit 8cd8a6296974 ("Remove memalign()").

At the time memalign() was unused and a bit buggy. This new memalign()
is

Reintroduce memalign() and friends

memalign() and friends where removed with the
commit 8cd8a6296974 ("Remove memalign()").

At the time memalign() was unused and a bit buggy. This new memalign()
is believed to work correctly due to extensive testing. Recently
memalign() has been needed by certain drivers so it makes sense to add
it again.

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutil: add hdr_size parameter to bget

Adds a hdr_size parameter to bget(), bgetz() and bgetr(). hdr_size must
be a multiple of BGET_HDR_QUANTUM. If hdr_size is larger than 0 the
buffer will be all

libutil: add hdr_size parameter to bget

Adds a hdr_size parameter to bget(), bgetz() and bgetr(). hdr_size must
be a multiple of BGET_HDR_QUANTUM. If hdr_size is larger than 0 the
buffer will be allocated so that the alignment constraints are fulfilled
after advancing hdr_size bytes into the returned buffer.

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutil: add alignment parameter to bget

Adds alignment parameter to bget(), bgetz() and bgetr(). If alignment is
larger then 0 the returned buffer is guaranteed to have an address which
is a multip

libutil: add alignment parameter to bget

Adds alignment parameter to bget(), bgetz() and bgetr(). If alignment is
larger then 0 the returned buffer is guaranteed to have an address which
is a multiple of this value.

The algorithm is basically unchanged, in the way that the memory is
still allocated from the end of a free memory block. The difference is
in the core implementation in bget() where now alignment of the returned
memory is taken into account. If only allocating with the minimum
alignment the memory blocks are expected to be allocated in the same
pattern.

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

Introduce CFG_TA_BGET_TEST

Introduces CFG_TA_BGET_TEST which compiles the integrated bget test
suite together with the rest of bget. When enabled, the test entry point
is bget_main_test() in libutil

Introduce CFG_TA_BGET_TEST

Introduces CFG_TA_BGET_TEST which compiles the integrated bget test
suite together with the rest of bget. When enabled, the test entry point
is bget_main_test() in libutils.

Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: strtoul.c: adjust license text to upstream

Adjust the license text in strtoul.c according to upstream commit "Use
remove-advertising-clause script to edit BSD licenses" [1].
With this, the

libutils: strtoul.c: adjust license text to upstream

Adjust the license text in strtoul.c according to upstream commit "Use
remove-advertising-clause script to edit BSD licenses" [1].
With this, the file also matches its SPDX License identifier correctly.

Link: [1]: https://sourceware.org/git/?p=newlib-cygwin.git;a=commit;h=9042d0ce65533a26fc3264206db5828d5692332c
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Acked-by: Jerome Forissier <jerome@forissier.org>
[jf: minor edits to commit message]
Signed-off-by: Jerome Forissier <jerome@forissier.org>

show more ...

libutils: fix memset(), memcpy(), memmove() with -O3

When libutils is built with optimization -O3 (make CFG_CC_OPT_LEVEL=3),
memset() may cause infinite recursion. This bug was observed with GCC
8.3

libutils: fix memset(), memcpy(), memmove() with -O3

When libutils is built with optimization -O3 (make CFG_CC_OPT_LEVEL=3),
memset() may cause infinite recursion. This bug was observed with GCC
8.3 with QEMUv8. The reason is, at this optimization level the compiler
may decide to replace loops with calls to memset(), which is obviously
not valid when the loop is in the memset function itself. This behavior
can be turned off with -fno-tree-loop-distribute-patterns.

Therefore, add the appropriate build flag to memset.c. Note that a
similar fix was introduced in upstream newlib [1] but since it depends
on the newlib build script to detect compiler support, the patch is not
directly applicable. Instead $(call cc-option,...) is used here.

Although I have not observed any issue with memcpy() and memmove(),
upstream did apply the compiler flag to these functions too [1], [2],
which seems quite reasonable so do the same here.

Link: [1] https://github.com/bminor/newlib/commit/10e4d79a191f07999bc94b21535fba5d7c04f419
Link: [2] https://github.com/bminor/newlib/commit/82dfae9ab0734b9946321590162dc6021057fec1
Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: memset(): add -O2 only if optimization is -Os

Commit 5b2aaa117eee ("libutee: optimize memcpy() for speed") added -O2
to memset.c to exclude the slower implementation which would be selecte

libutils: memset(): add -O2 only if optimization is -Os

Commit 5b2aaa117eee ("libutee: optimize memcpy() for speed") added -O2
to memset.c to exclude the slower implementation which would be selected
when level is -Os (in which case __OPTIMIZE_SIZE__ is defined).

Since the optimization level can now be selected globally with
CFG_CC_OPT_LEVEL, -O2 should not be forced unconditionally or it could
actually reduce the desired level (when CFG_CC_OPT_LEVEL=3).

Therefore, add -O2 only if the global optimization is -Os.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

arm32: fold UNWIND(.fnstart/.fnend) into the FUNC macros

This change applies to arm32 assembler sources.

Instead of using UNWIND(.fnstart) after FUNC or LOCAL_FUNC and
UNWIND(.fnend) before END_FUN

arm32: fold UNWIND(.fnstart/.fnend) into the FUNC macros

This change applies to arm32 assembler sources.

Instead of using UNWIND(.fnstart) after FUNC or LOCAL_FUNC and
UNWIND(.fnend) before END_FUNC, let's fold these statements into the
FUNC macros.

The .fnstart/.fnend directives mark the start and end of a function
with an unwind table entry (.ARM.exidx) and therefore a function
without them has no entry and cannot be unwound. This means that a
stack dump (on abort or panic) would stop when reaching such a
function.

As a result of this patch, a small number of functions now have an
entry in the unwind table when they had none before (the functions
which were using FUNC or LOCAL_FUNC but had no .fnstart/.fnend). It was
almost always a bug and this pacth only increases the size of the
.ARM.exidx section by a few bytes (tested on QEMU).

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

arm32: move the UNWIND() macro to <asm.S>

All the users of the UNWIND() macro include <asm.S> already, which is
therefore a good place to define this macro. Let's move it from
<kernel/unwind.h> to <

arm32: move the UNWIND() macro to <asm.S>

All the users of the UNWIND() macro include <asm.S> already, which is
therefore a good place to define this macro. Let's move it from
<kernel/unwind.h> to <asm.S>, remove a couple of duplicates in
assembler files, and drop the useless includes.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

core: remove stack dump macros and multiple log levels

Of the various xPRINT_STACK() macros (x in {E,I,D,F}), only
EPRINT_STACK() is used. Let's simplify the code by removing the macros
altogether a

core: remove stack dump macros and multiple log levels

Of the various xPRINT_STACK() macros (x in {E,I,D,F}), only
EPRINT_STACK() is used. Let's simplify the code by removing the macros
altogether and calling print_kernel_stack() instead. Since only the
TRACE_ERROR is used, the 'level' argument to print_kernel_stack(),
print_stack_arm32() and print_stack_arm64() is removed too.

In addition to simplifying the code, these changes will allow the
consolidation of the stack unwinding code between core and ldelf.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
Reviewed-by: Etienne Carriere <etienne.carriere@linaro.org>

show more ...

libutils: add simplified fputc(), fputs(), fwrite(), write()

Preparing for C++ support in TAs.

Adds a few <stdio.h> functions to libutils:

fputc() fputs() fwrite() write()

The proposed implem

libutils: add simplified fputc(), fputs(), fwrite(), write()

Preparing for C++ support in TAs.

Adds a few <stdio.h> functions to libutils:

fputc() fputs() fwrite() write()

The proposed implementations are limited in the sense that they only
accept writing to stdout or stderr. The output goes directly to the
secure console and no difference is made between stdout and stderr.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...

libutils: add sprintf()

Preparing for C++ support in TAs.

Implement sprintf() in libutils.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Acked-by: Etienne Carriere <etienne.carriere@linar

libutils: add sprintf()

Preparing for C++ support in TAs.

Implement sprintf() in libutils.

Signed-off-by: Jerome Forissier <jerome@forissier.org>
Acked-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>

show more ...