such as code: double value = +0x1.000000000000080000000000p-600;
printf("Hello, World!! %.7g:f64\n", value);
expected output : Hello, World!! 2.40992e-181:f64
but real output : Hello, World!! 2.40992e-B1:f64
the reason: we want printf "18", but the flag of character is 'B'
and 'B' is '0'+18,so printf 'B';
In order to turn longjmp context-switch safe, it's necessary
to disable interrupts before modifying windowbase and windowstart.
Otherwise, after a context switch, windowstart and windowbase
would be different, leading to a wrongly set windowstart bit due to
longjmp writing it based on the windowbase before the context switch.
This corrupts the registers at the next window overflow reaching
that wrongly set bit.
*Background:*
This PR is related to an issue first observed on ESP-IDF
https://github.com/espressif/esp-idf/issues/5229 and it was, then,
checked on NuttX using a test application.
*The test application:*
To check if the problem affects ESP32, ESP32-S2 and ESP32-S3 on
NuttX, it was created an application based on:
https://en.cppreference.com/w/c/program/longjmp
The application creates 16 tasks (`#define NUMBER_OF_TASKS 16`)
that implements the following daemon:
```
static int setjmp_longjmp_daemon(int argc, char *argv[])
{
for (int i = 0; i < NUMBER_OF_TASKS * 2; i++)
{
jmp_buf env;
volatile int count = 0;
if (setjmp(env) != UINT16_MAX)
{
foo(&env, ++count);
}
}
sem_post(&g_sem);
return EXIT_SUCCESS;
}
```
The main function also initializes a semaphore to avoid application
exiting before tasks return successfully:
```
sem_init(&g_sem, 0, -NUMBER_OF_TASKS);
```
Finally, the round-robin interval was lowered to 1ms to raise the
chances of the longjmp being interrupted by a context switch
(`CONFIG_RR_INTERVAL=1).
This setup was able to reproduce the problem prior to this patch
being applied.
When a7 curl calls dns_query via usrsock, it will wait a usrsock
response with the dns lock held. If a7 rptun recv a dns_addnameserver
event from another core at this time, rptun waits the lock held by
curl in callback. Then curl will never get its usrsock response with
rptun blocked, so the deadlock occurs.
This change will break lock when do usrsock request.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
this is fix a ltp test case and update
the error code of pthread_barrierdestroy
ltptcase is : apps/testing/ltp/ltp/testcases/open_posix_testsuite/conformance/interfaces/pthread_barrier_destroy/2-1.c
Signed-off-by: anjiahao <anjiahao@xiaomi.com>
In the past, header file paths were generated by the incdir command
Now they are generated by concatenating environment variables
In this way, when executing makefile, no shell command will be executed,
it will improve the speed of executing makfile
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
In the past, predefined macros were generated by define.sh scripts
Now they are generated by concatenating environment variables
In this way, when executing makefile, no shell command will be executed,
it will improve the speed of executing makfile
Signed-off-by: yinshengkai <yinshengkai@xiaomi.com>
instead of lib_psa_[get|set]stackaddr.c since the later doesn't exist in kernel mode
regression by:
commit b9b032af72
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
Date: Thu Oct 20 03:10:23 2022 +0800
sched/spawn: Support task_spawnattr_[set|get]stacksize in kernel mode
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
modify the return value according to posix standard
------------
SCHED_GET_PRIORITY_MAX(2)
NAME
sched_get_priority_max, sched_get_priority_min - get static priority range
RETURN VALUE
On success, sched_get_priority_max() and sched_get_priority_min() return the
maximum/minimum priority value for the named scheduling policy.
On error, -1 is returned, and errno is set appropriately.
ERRORS
EINVAL The argument policy does not identify a defined scheduling policy.
Signed-off-by: chao an <anchao@xiaomi.com>
testcase:
set TZ CEST
time_t t = time(NULL);
struct tm *ts = localtime(&t);
t = timegm(ts);
t will return (time_t)-1;
Signed-off-by: dongjiuzhu1 <dongjiuzhu1@xiaomi.com>
arm-none-eabi-ld: nuttx/staging/libc.a(lib_impure.o):(.data.impure_data+0x4): undefined reference to `__sf'
arm-none-eabi-ld: nuttx/staging/libc.a(lib_impure.o):(.data.impure_data+0x8): undefined reference to `__sf'
arm-none-eabi-ld: nuttx/staging/libc.a(lib_impure.o):(.data.impure_data+0xc): undefined reference to `__sf'
Toolchain:
Version: GNU Arm Embedded Toolchain to Version 11.3.Rel1(August 8, 2022)
Tar: arm-gnu-toolchain-11.3.rel1-x86_64-arm-none-eabi.tar.xz
Break change:
https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=b0cb9f85ca3626e0e68fd451c3090d253ceb4300
Signed-off-by: chao an <anchao@xiaomi.com>
1.Don't include unwind.h when arch specific backtrace is enable
2.Built arch specific backtrace wrapper only when enable
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
to ensure the generated key is unique from each other since only
root pseduo file system really support st_ino field after:
commit d35fbf534d51f7ef72382d9666a1f19e07c6f00f
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
Date: Fri Sep 16 14:24:55 2022 +0800
fs: Allocate unique serial number for the root pseduo file system node
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
- Add check in roundx() functions for infinite or NaN cases
- Add block to avoid style warnings
- Define long double constants and macros for infinity and nan
- Correct return syntax to match NuttX style
- Make c89 compliant
- Fix definitions of INFINITY_L/NAN_L
* include/nuttx/lib/math.h - match standard naming conventions
- Rename isinf_l to isinfl
- Rename isinf_f to isinff
- Add finite()
- Add finitel()
- Add finitef()
- Define isnanl and isnanf
- Define isfinite() so that it uses the appropriate macro
* libs/libc/math/lib_asinf.c
libs/libc/math/lib_asinl.c
libs/libc/math/lib_roundf.c
libs/libc/math/lib_roundl.c
libs/libc/math/lib_sqrtf.c
libs/libc/math/lib_sqrtl.c
- Use renamed macros
- Use correct NAN_x or INFINIT_x macro on returns
```
ubsan_prologue: ================================================================================
ubsan_prologue: UBSAN: array-index-out-of-bounds in symtab/symtab_allsyms.c:62:37
__ubsan_handle_out_of_bounds: index 1619 is out of range for type 'symtab_s [1]'
ubsan_epilogue: ================================================================================
```
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
* libs/libc/stdlib/lib_strtod.c:
(strtod): Add a note about limitations of this implementation
as compared to POSIX in the function's docstring. Also fix a
typo.
* libs/libc/stdlib/lib_strtof.c:
(strtof): Port the changes made to strtod in PR-6952 (commit
c83985c5ce) and add same note as above to docstring.
* libs/libc/stdlib/lib_strtold.c:
(strtold): Same changes as strtof.
g_dns_servers need init before dns_query, Otherwise sim can not add default
dns server when use usrsock mode to share host network.
Avoid similar problems in the future, so directly initialize g_dns_servers.
Signed-off-by: zhanghongyu <zhanghongyu@xiaomi.com>
Summary:
- I noticed that adb shell failed when sh is spawned.
- Finally, I found that an error happened when executing dup2() action
if the file descriptor has the O_CLOEXEC option.
- This commit fixes this issue by dropping the option in the API.
Impact:
- posix_spawn_file_actions_adddup2() only
Testing:
- adbd with sabre-6quad:netnsh (will be merged later)
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
some toolchain provider remove const qualifier from global_impure_ptr
(e.g. Espressif) but other provider keep it(e.g. ARM or Cadence), so
let's define const to empty macro to pass the compile and more safety.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
but just format the low 32bits if CONFIG_LIBC_LONG_LONG isn't enabled to
avoid to expand the code space to much.
Note: the size will increase 192 bytes on stm32_tiny:nsh.
Before the change:
text data bss dec hex filename
41444 184 1656 43284 a914 nuttx
After the change:
text data bss dec hex filename
41636 184 1656 43476 a9d4 nuttx
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
and remove CONFIG_LIBC_LONG_LONG option to simplify the usage.
note: the size will increase 668
before change:
text data bss dec hex filename
168440 348 4480 173268 2a4d4 nuttx
after change:
text data bss dec hex filename
169108 348 4480 173936 2a770 nuttx
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
* libs/libc/misc/lib_getrandom.c, in function getrandom():
We were assigning the return value of _NX_READ() to nbytes, a
variable of the unsigned type size_t. Note that _NX_READ() resolves
to either read() or nx_read(), both of which return the signed type
ssize_t to indicate either the number of bytes read successfully
(>= 0) or an error (< 0). Then we were testing for a negative
size_t value, a condition that can never occur. The end result is
that if an error occured in _NX_READ(), it would never be detected
and getrandom() would return some large positive value. This bug is
corrected by assigning the return value of _NX_READ() to a new local
variable, ret, of size ssize_t.
The "p" format specifier already prepends the pointer address with
"0x" when printing.
Signed-off-by: Gustavo Henrique Nihei <gustavo.nihei@espressif.com>
Make use of XCHG/BSWAP on x86, REV16 and REV on ARMv6-m and above,
and whatever other optimized instructions on other platforms.
Defines extra CONFIG variables, and removes the unused functions for
endian-swapping. Fixes some oversights in using the macros.
For CONFIG_BUILD_KERNEL using the sched/task/task_exithook implementation
will just not work. It calls user code with kernel privileges which is
a bit of a security issue.
Targets build during the kernel phase did not have their dependencies
specified and thus they were not rebuilt after their dependencies have
changed, for example by changing options in menuconfig.
Signed-off-by: Abdelatif Guettouche <abdelatif.guettouche@espressif.com>
Summary:
- This commit adds R_RISCV_JAL support to arch_elf.c
- The code only checks the immediate value but does not relocate it
because it is done by the compiler.
Impact:
- None
Testing:
- Tested with rv-virt:knsh64 (not merged yet)
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
Do not call strlen() here. Old implementation iterated
over string twice, if searched for position was at the
beginning. This commit changes strrchr() to scan string
only once, regardless of input.
Signed-off-by: Juha Niskanen <juha.niskanen@haltian.com>
to fix the gcov link error:
arm-none-eabi-ld: /home/xiaoxiang/vela/prebuilts/gcc/linux/arm/bin/../lib/gcc/arm-none-eabi/10.3.1/thumb/v7e-m/nofp/libgcov.a(_gcov.o): in function `gcov_error_exit':
libgcov-driver.c:(.text+0x7c): undefined reference to `_impure_ptr'
arm-none-eabi-ld: /home/xiaoxiang/vela/prebuilts/gcc/linux/arm/bin/../lib/gcc/arm-none-eabi/10.3.1/thumb/v7e-m/nofp/libgcov.a(_gcov.o): in function `get_gcov_error_file':
libgcov-driver.c:(.text+0x1e0): undefined reference to `_impure_ptr'
arm-none-eabi-ld: /home/xiaoxiang/vela/prebuilts/gcc/linux/arm/bin/../lib/gcc/arm-none-eabi/10.3.1/thumb/v7e-m/nofp/libgcov.a(_gcov.o): in function `gcov_exit_open_gcda_file.isra.0':
libgcov-driver.c:(.text+0x660): undefined reference to `_impure_ptr'
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
1.Remove CXX_INITIALIZE_MACHO and CXX_INITIALIZE_SINIT
2.Merge cxx_initialize_sinit.c and cxx_initialize_macho.c into cxx_initialize.c
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Summary:
- Cast to unsigned char for strcmp and strncmp
- strcmp and strncmp are described following by opengroup.org
The sign of a non-zero return value shall be determined by the sign
of the difference between the values of the first pair of bytes
(both interpreted as type unsigned char) that differ in the strings
being compared.
https://pubs.opengroup.org/onlinepubs/9699919799/functions/strcmp.htmlhttps://pubs.opengroup.org/onlinepubs/9699919799/functions/strncmp.html
Impact:
- strcmp and strncmp return value
Testing:
- ostest on sabre-6quad:smp w/ qemu
Signed-off-by: Oki Minabe <minabe.oki@gmail.com>
Restore the checks which got completely broken by the following change.
```
commit 15142a8b10
Author: anjiahao <anjiahao@xiaomi.com>
Date: Fri Apr 1 16:20:03 2022 +0800
modlib/modlib_bind:fix unsigned_compare with zero
Signed-off-by: anjiahao <anjiahao@xiaomi.com>
```
