Fix [1] is introducing another problem in that memset is added twice to
the libc makefile when OVERRIDE_LIBC=1 (the C and asm implementations).
Correct by adding memset.c when OVERRIDE_LIBC=0 and memset.S when
OVERRIDE_LIBC=1.
[1] https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/26091
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: Ie4b7e04880d4cd871e7b51cd8ff5bddcf8d0308c
Change [1] introduced the memcpy_s function and added the source file to
lib/libc/libc.mk but omitted to update lib/libc/libc_asm.mk
Arm platforms (and platforms from one partner) use OVERRIDE_LIBC=1
option as a platform default hence consume libc_asm.mk
To prevent this confusion to happen again, create libc_common.mk for the
common set of C files to build.
libc_common.mk is included by both libc.mk and libc_asm.mk
The latter adds asm implementations of libc functions.
[1] https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/21450
Signed-off-by: Olivier Deprez <olivier.deprez@arm.com>
Change-Id: Ibd36ea5c561e35f268048ccbefc8f47485f64bdf
This patch makes printf capable of padding strings with spaces
to their left, allowing right-aligning numeric values and strings.
It uses the same width field for the format specifier as in the
original libc function : %<width><type> (e.g.: %10d pads an
integer value with spaces to the left to complete 10 characters).
Change-Id: Ib7b5519dae17742181352ce58e507a05ba5250d4
Signed-off-by: Juan Pablo Conde <juanpablo.conde@arm.com>
Align entire TF-A to use Arm in copyright header.
Change-Id: Ief9992169efdab61d0da6bd8c5180de7a4bc2244
Signed-off-by: Govindraj Raja <govindraj.raja@arm.com>
Enables printing captial hex chars as well as lowercase
Change-Id: I4dc48c3db97b908f0bb344d7765807967de8cf02
Signed-off-by: Matt Schulte <matsch@google.com>
To support memcpy_s for better security purpose
to avoid overflowing the dest while copy from src.
Signed-off-by: Jit Loon Lim <jit.loon.lim@intel.com>
Change-Id: I63c3ea6a3e99c10d69be6bce04843c14b0a28a4d
Adds %c support for printf and snprintf to print one character. Required
by most recent MbedTLS 3.4.0.
Change-Id: I4d9b2725127a929d58946353324f99ff22b3b28b
Signed-off-by: Maksims Svecovs <maksims.svecovs@arm.com>
This issue was triggered by sparse tool:
lib/libc/putchar.c:9:5: warning:
symbol '__putchar' was not declared. Should it be static?
Instead of setting __putchar as static, just remove the function and
directly use putchar() with a weak attribute.
Signed-off-by: Yann Gautier <yann.gautier@st.com>
Change-Id: Ib35e4ba064f06010851bb860269b08462fe3d3bd
Compiling with -Wimplicit-fallthrough=3 (enabled by -Wextra) produces
many warnings about fallthrough comments either missing or being wrong.
Unify the comments so we comply with -Wextra.
Note that Coverity recommends against using the __attribute__ directive.
Also, zlib does not build with a higher value of -Wimplicit-fallthrough.
Finally, compilers strip comments before expanding macros. As such,
checkpatch's fallthrough annotation (or higher levels of the flag) isn't
really possible.
Signed-off-by: Boyan Karatotev <boyan.karatotev@arm.com>
Change-Id: I060cf4f8dc04c02cbb45cf4ceb69569a8369ccee
Moving putchar() out of libc and adding a weak dummy
implementation in libc.
This is to remove libc's dependencies to the platform
driver.
Signed-off-by: Claus Pedersen <claustbp@google.com>
Change-Id: Ib7fefaec0babb783def614ea23521f482fa4a28a
stddef.h is needed for the definition of size_t
stdio.h is needed for the declaration of putchar
Signed-off-by: Jorge Troncoso <jatron@google.com>
Change-Id: I72dac843dbbfc440cff0f9e9d13669b78a812abc
- Removing platform dependencies from libc modules.
- Replacing panicking with actual error handling.
- Debug macros are included indirectly from assert.h. Removing
"platform_def.h" from assert.h and adding "common/debug.h"
where the macros are used.
- Removing hack for fixing PLAT_LOG_LEVEL_ASSERT to 40.
Instead removing assert with expression, as this
does not provide additional information.
Signed-off-by: Claus Pedersen <claustbp@google.com>
Change-Id: Icc201ea7b63c1277e423c1cfd13fd6816c2bc568
Add long, long long and size_t length specifiers to
snprintf similar to vprintf. This will help capturing
all the UART logs into a logbuffer and makes snprintf
functionally equivalent to vprintf.
Change-Id: Ib9bd20e2b040c9b8755cf7ed7c9b4da555604810
Signed-off-by: Channagoud kadabi <kadabi@google.com>
Replace %d with %u in logs, to avoid warning when
-Wformat-signedness is enabled.
Signed-off-by: Yann Gautier <yann.gautier@st.com>
Change-Id: Id058f6fb0fd25ce5f83b1be41082403fcb205841
The snprintf code uses the uintptr_t type, which is defined in stdint.h.
We do not include this header explicitly, but get the definition
indirectly through some other header doing so.
However this breaks when snprintf is compiled in isolation (for instance
for unit-testing), so let's add this #include to make things right.
Change-Id: I1299767ee482f5cf1af30c4df2e8f7e596969b41
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
In our unsigned_num_print() function we first print the integer into a
local buffer, then put this through alignment and padding and output the
result. For this we use a local buffer, sized by the maximum possible
length of the largest possible number.
However this assumes that the radix is not smaller than 10, which is
indeed the smallest value we pass into this static function at the
moment. To prevent accidents in the future, should we add support for
other radices, add an assert to enforce our assumption.
Unfortunately this cannot be a static assert (CASSERT), since the
compiler is not smart enough to see that the argument is always coming
from a literal.
Change-Id: Ic204462600d9f4c281d899cf9f2c698a0a33a874
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
The number formatting routine in snprintf was trying to be clever with
the buffer handling, but tripped over its own feet: snprintf() users
expect output to be emitted, even if not everything fits into the
buffer. The current code gives up completely when the buffer is too
small.
Fix those issues and simplify the code on the way, by consequently using
the CHECK_AND_PUT_CHAR() macro, which both checks for the buffer size
correctly, but also keeps track of the number of should-be-printed
characters for the return value.
Change-Id: Ifd2b03b9a73f9279abed53081a2d88720ecbdbc1
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
MISRA complained about "0"s not being followed by an "U" (please note
my protest about this!) and about values not being explicitly compared
to 0 (fair enough).
Also use explicit pointer types.
Fix those issues to make the CI happy.
Change-Id: I4d11e49c14f16223a71c78b0fc3e68ba9a1382d3
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
The coding guidelines[1] in TF-A forbid the use of ato*() functions
in favour of strto*(). However, the TF-A libc does not provide an
implementation of strto*(), making this rule impossible to satisfy.
Also made small changes to fit into TF-A project. Added the source
files to the libc makefile
[1] https://trustedfirmware-a.readthedocs.io/en/latest/process/coding-guidelines.html#libc-functions-that-are-banned-or-to-be-used-with-caution
Change-Id: I2e94a0b227ec39f6f4530dc50bb477999d27730f
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
The coding guidelines[1] in TF-A forbid the use of ato*() functions
in favour of strto*(). However, the TF-A libc does not provide an
implementation of strto*(), making this rule impossible to satisfy.
Also made small changes to fit into TF-A project. Added the source
files to the libc makefile
[1] https://trustedfirmware-a.readthedocs.io/en/latest/process/coding-guidelines.html#libc-functions-that-are-banned-or-to-be-used-with-caution
Change-Id: I9cb581574d46de73c3d6917ebf78935fc5ac075a
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
The coding guidelines[1] in TF-A forbid the use of ato*() functions
in favour of strto*(). However, the TF-A libc does not provide an
implementation of strto*(), making this rule impossible to satisfy.
Also made small changes to fit into TF-A project. Added the source
files to the libc makefile
[1] https://trustedfirmware-a.readthedocs.io/en/latest/process/coding-guidelines.html#libc-functions-that-are-banned-or-to-be-used-with-caution
Change-Id: I8c3b92751d1ce226c966f7c81fedd83f0846865e
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
The coding guidelines[1] in TF-A forbid the use of ato*() functions
in favour of strto*(). However, the TF-A libc does not provide an
implementation of strto*(), making this rule impossible to satisfy.
Also made small changes to fit into TF-A project. Added the source
files to the libc makefile
[1] https://trustedfirmware-a.readthedocs.io/en/latest/process/coding-guidelines.html#libc-functions-that-are-banned-or-to-be-used-with-caution
Change-Id: Ica95bf5da722913834fe90bf3fe743aa34e01e80
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
Add macro CHECK_AND_PUT_CHAR to check buffer capacity, save one
character to buffer, and then increase character counter by one in one
single statement, so that 4 similar code pieces can be cleaned.
Signed-off-by: Heyi Guo <guoheyi@linux.alibaba.com>
Change-Id: I2add6b4bd6c24ea3c0d2499a44924e3e8db0f4d1
Enable snprintf()/vsnprintf() in TF-A to print "%" character as C
standard, which may be used in platform porting to print percentage
information.
Signed-off-by: Heyi Guo <guoheyi@linux.alibaba.com>
Change-Id: I9b296372a1002046eabac1df5e8eb99a27efd4a8
Enable printf() in TF-A to print "%" character as C standard, which
may be used in platform porting to print percentage information.
Signed-off-by: Heyi Guo <guoheyi@linux.alibaba.com>
Change-Id: I7af2f1d153548e426f423fce15dc48b0da56c622
And from crash_console_flush.
We ignore the error information return by console_flush in _every_
place where we call it, and casting the return type to void does not
work around the MISRA violation that this causes. Instead, we collect
the error information from the driver (to avoid changing that API), and
don't return it to the caller.
Change-Id: I1e35afe01764d5c8f0efd04f8949d333ffb688c1
Signed-off-by: Jimmy Brisson <jimmy.brisson@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
Made small changes to fit into TF-A project
Change-Id: I991f653a7ace04f9c84bcda78ad8d7114ea18e93
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
It uses the existing implementation of snprintf() function
Change-Id: Ie59418564c2e415222e819cf322c34e9a4d1f336
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
From commit: 21571b1d140ae7bb44e94c0afba2ec61456b275b
Made small changes to fit into TF-A project
Change-Id: I07fd7fe1037857f6b299c35367c104fb51fa5cfa
Signed-off-by: Madhukar Pappireddy <madhukar.pappireddy@arm.com>
Currently our memset() implementation is safe, but slow. The main reason
for that seems to be the single byte writes that it issues, which can
show horrible performance, depending on the implementation of the
load/store subsystem.
Improve the algorithm by trying to issue 64-bit writes. As this only
works with aligned pointers, have a head and a tail section which
covers unaligned pointers, and leave the bulk of the work to the middle
section that does use 64-bit writes.
Put through some unit tests, which exercise all combinations of nasty
input parameters (pointers with various alignments, various odd and even
sizes, corner cases of content to write (-1, 256)).
Change-Id: I28ddd3d388cc4989030f1a70447581985368d5bb
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Trace analysis of FVP_Base_AEMv8A 0.0/6063 model
running in Aarch32 mode with the build options
listed below:
TRUSTED_BOARD_BOOT=1 GENERATE_COT=1
ARM_ROTPK_LOCATION=devel_ecdsa KEY_ALG=ecdsa
ROT_KEY=plat/arm/board/common/rotpk/arm_rotprivk_ecdsa.pem
shows that when auth_signature() gets called
71.99% of CPU execution time is spent in memset() function
written in C using single byte write operations,
see lib\libc\memset.c.
This patch introduces new libc_asm.mk makefile which
replaces C memset() implementation with assembler
version giving the following results:
- for Aarch32 in auth_signature() call memset() CPU time
reduced to 20.56%.
The number of CPU instructions (Inst) executed during
TF-A boot stage before start of BL33 in RELEASE builds
for different versions is presented in the tables below,
where:
- C TF-A: existing TF-A C code;
- C musl: "lightweight code" C "implementation of the
standard library for Linux-based systems"
https://git.musl-libc.org/cgit/musl/tree/src/string/memset.c
- Asm Opt: assemler version from "Arm Optimized Routines"
project
https://github.com/ARM-software/optimized-routines/blob/
master/string/arm/memset.S
- Asm Linux: assembler version from Linux kernel
https://github.com/torvalds/linux/blob/master/arch/arm/lib/memset.S
- Asm TF-A: assembler version from this patch
Aarch32:
+-----------+------+------+--------------+----------+
| Variant | Set | Size | Inst | Ratio |
+-----------+------+------+--------------+----------+
| C TF-A | T32 | 16 | 2122110003 | 1.000000 |
| C musl | T32 | 156 | 1643917668 | 0.774662 |
| Asm Opt | T32 | 84 | 1604810003 | 0.756233 |
| Asm Linux | A32 | 168 | 1566255018 | 0.738065 |
| Asm TF-A | A32 | 160 | 1525865101 | 0.719032 |
+-----------+------+------+--------------+----------+
AArch64:
+-----------+------+------------+----------+
| Variant | Size | Inst | Ratio |
+-----------+------+------------+----------+
| C TF-A | 28 | 2732497518 | 1.000000 |
| C musl | 212 | 1802999999 | 0.659836 |
| Asm TF-A | 140 | 1680260003 | 0.614917 |
+-----------+------+------------+----------+
This patch modifies 'plat\arm\common\arm_common.mk'
by overriding libc.mk makefile with libc_asm.mk and
does not effect other platforms.
Change-Id: Ie89dd0b74ba1079420733a0d76b7366ad0157c2e
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>
The current implementation of snprintf() does not support pointer and
hexadecimal format specifiers, which can be needed, for instance, for
DTB manipulations.
This patch adds that functionality by borrowing some code from the
printf() implementation.
Signed-off-by: Javier Almansa Sobrino <javier.almansasobrino@arm.com>
Change-Id: I2076ea46693a73a04890982bf20e3c633c2767fb
This reverts commit e7d344de01.
This reverts the patch https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/5313 due to a timing issue with the merge. The merge occurred at the same time as the additional comments and thusly were were not seen until the merge was done. This reverts the change and additional patches from Alexei will follow to address the concerns expressed in the orignal patch.
Change-Id: Iae5f6403c93ac13ceeda29463883fcd4c437f2b7
Trace analysis of FVP_Base_AEMv8A model running in
Aarch32 mode with the build options listed below:
TRUSTED_BOARD_BOOT=1 GENERATE_COT=1
ARM_ROTPK_LOCATION=devel_ecdsa KEY_ALG=ecdsa
ROT_KEY=plat/arm/board/common/rotpk/arm_rotprivk_ecdsa.pem
shows that when auth_signature() gets called
71.84% of CPU execution time is spent in memset() function
written in C using single byte write operations,
see lib\libc\memset.c.
This patch replaces C memset() implementation with assembler
version giving the following results:
- for Aarch32 in auth_signature() call memset() CPU time
reduced to 24.84%.
- Number of CPU instructions executed during TF-A
boot stage before start of BL33 in RELEASE builds:
----------------------------------------------
| Arch | C | assembler | % |
----------------------------------------------
| Aarch32 | 2073275460 | 1487400003 | -28.25 |
| Aarch64 | 2056807158 | 1244898303 | -39.47 |
----------------------------------------------
The patch also replaces memset.c with aarch64/memset.S
in plat\nvidia\tegra\platform.mk.
Change-Id: Ifbf085a2f577a25491e2d28446ee95a4ac891597
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>
This function scans a string backwards from the end for the first
instance of a character.
Change-Id: I46b21573ed25a0ff222eac340e1e1fb93b040763
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Sparse warns this:
lib/libc/assert.c:29:6: error: symbol '__assert' redeclared with different type (originally declared at include/lib/libc/assert.h:36) - different modifiers
Add __dead2 to match the header declaration and C definition.
I also changed '__dead2 void' to 'void __dead2' for the consistency
with other parts.
Change-Id: Iefa4f0e787c24fa7e7e499d2e7baf54d4deb49ef
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
The previous implementation could behave incorrectly because of the sign
extension of the char when compared to the int.
Change-Id: I397838b0ec87a6f1af6972d022a8c19a5184b447
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Improved support for W=1 compilation flag by solving missing-prototypes
and old-style-definition warnings.
The libraries are compiling with warnings (which turn into errors with
the Werror flag).
Outside of libraries, some warnings cannot be fixed without heavy
structural changes.
Change-Id: I1668cf99123ac4195c2a6a1d48945f7a64c67f16
Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>
Now that setjmp() and longjmp() are compliant with the standard they can
be moved with the other libc files.
Change-Id: Iea3b91c34eb353ace5e171e72f331602d57774d5
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Enforce full include path for includes. Deprecate old paths.
The following folders inside include/lib have been left unchanged:
- include/lib/cpus/${ARCH}
- include/lib/el3_runtime/${ARCH}
The reason for this change is that having a global namespace for
includes isn't a good idea. It defeats one of the advantages of having
folders and it introduces problems that are sometimes subtle (because
you may not know the header you are actually including if there are two
of them).
For example, this patch had to be created because two headers were
called the same way: e0ea0928d5 ("Fix gpio includes of mt8173 platform
to avoid collision."). More recently, this patch has had similar
problems: 46f9b2c3a2 ("drivers: add tzc380 support").
This problem was introduced in commit 4ecca33988 ("Move include and
source files to logical locations"). At that time, there weren't too
many headers so it wasn't a real issue. However, time has shown that
this creates problems.
Platforms that want to preserve the way they include headers may add the
removed paths to PLAT_INCLUDES, but this is discouraged.
Change-Id: I39dc53ed98f9e297a5966e723d1936d6ccf2fc8f
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Imported from lib/libc/string/strrchr.c from commit:
59fd2fb98e4cc7e9bfc89598e28e21d405fd470c
Change-Id: I898206c6f0372d4d211c149ec0fb9522d0a5b01c
Signed-off-by: Antonio Nino Diaz <antonio.ninodiaz@arm.com>
Ensure case clauses:
* Terminate with an unconditional break, return or goto statement.
* Use conditional break, return or goto statements as long as the end
of the case clause is unreachable; such case clauses must terminate
with assert(0) /* Unreachable */ or an unconditional __dead2 function
call
* Only fallthough when doing otherwise would result in less
readable/maintainable code; such case clauses must terminate with a
/* Fallthrough */ comment to make it clear this is the case and
indicate that a fallthrough is intended.
This reduces the chance of bugs appearing due to unintended flow through a
switch statement
Change-Id: I70fc2d1f4fd679042397dec12fd1982976646168
Signed-off-by: Daniel Boulby <daniel.boulby@arm.com>
