It's better to save one argument by returning pid directly.
This change also follow the convention of task_create.
BTW, it is reasonable to adjust the function prototype a
little bit from both implementation and consistency since
task_spawn is NuttX specific API.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
OS functions in syscall.csv are ordered alphabetically. However, two recently added functions are not in the correct location. This PR simply corrects that ordering.
The ordering of one entry was also corrected in libs/libc/libc.csv. Same issue.
This change is only cosmetic.
Verified only by CI
Gregory Nutt is the copyright holder for those files and he has submitted the
SGA as a result we can migrate the licenses to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Gregory Nutt is the copyright holder for those files and he has submitted the
SGA as a result we can migrate the licenses to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Drop to user-space in kernel/protected build with up_pthread_exit,
now all pthread_cleanup functions executed in user mode.
* A new syscall SYS_pthread_exit added
* A new tcb flag TCB_FLAG_CANCEL_DOING added
* up_pthread_exit implemented for riscv/arm arch
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
Pros:
* Reduce code differences
* Smaller allocations for !CONFIG_ARCH_USE_MODULE_TEXT
Cons:
* Likely to use more memory for !CONFIG_ARCH_USE_MODULE_TEXT in total
Tested with:
* sim:module on macOS
* esp32-devkit:nsh + CONFIG_MODULE on qemu
* lm3s6965-ek:qemu-protected + CONFIG_EXAMPLES_SOTEST on qemu
arch: Allocate the space from the beginning in up_stack_frame
and modify the affected portion:
1.Correct the stack dump and check
2.Allocate tls_info_s by up_stack_frame too
3.Move the stack fork allocation from arch to sched
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
I've seen a module with 16 bytes .rodata alignment for xmm operations.
It was getting SEGV on sim/Linux because of the alignment issue.
The same module binary seems working fine after applying this patch.
Also, tested on sim/macOS and esp32 on qemu,
using a module with an artificially large alignment. (64 bytes)
Fix comment in libs/libc/tls/tls_getinfo.c: The TLS data must lie at the beginning of the allocated stack memory for both push-up and push-down stacks.
Ken Pettit has submitted the ICLA and we can migrate the licenses
to Apache.
Sebastien Lorquet has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Or genromfs failed to generate the right romfs.img
Change-Id: Icbcc2e89da1ede644b994d33e70f1a48662c412b
Signed-off-by: liuhaitao <liuhaitao@xiaomi.com>
Per the Linux man page, "By default, getopt() permutes the contents of argv as it scans, so that eventually all the nonoptions are at the end." This behavior, however, is not implemented in the NuttX getopt() logic.
The Linux man page requires that the getopt_long() and getopt_long_only() functions accept arguments to options in a form like:
--option=argument
This PR adds that missing functionality.
This change effects only getopt_long() and getopt_long_only()
Tested on a simulator NSH configuration with a modified version of the getopt() test in apps/testing/ostest.
If an unrecognized long option is encountered, we must skip over that argv[] entry or getopt_long() will seriously misbehave.
Affects getopt_long() and getopt_long_only()
Problem found and fix verified with an updated version of the OS test.
Found a place in getopt_common() where the option string is used before it is checked if it is NULL. This can happen because the short option string is optional for getopt_long() and getopt_long_only()
If optstring is NULL, that would be an ERROR for getopt(), but not for the getopt_long() versions.
Should effect only the getopt() APIs
Tested on the simulator using apps/testing/ostest.
Several licenses were missed in the initial work
David Sidrane has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
1. Null pointer dereference:
- for (ndx = 0; longopts[ndx].name[0] != '\0'; ndx++)
+ for (ndx = 0; longopts[ndx].name != NULL; ndx++)
2. Handle single character long options. An option like -x could be either a short option or a long option (under getopt_long_only()). This case was not being handled correctly.
3. Add missing support for optional arguments to short options (indicated with two "::"
This effects all members of the getopt() family of APIs.
Tested on the simulator using extensions to apps/testing/ostest.
Add implementations of getopt_long() and getopt_long_only()
There could be impacts to getop() since that implementation is shared withe new getopts_long() and getops_long_only() implementation.
Tested using a modified version of sim:nsh
Since 7a046358d9 the top-level COPYING
file has been deleted and replaced by DISCLAIMER, LICENSE, and NOTICE
files. However, some references to the old COPYING file remained in
Kconfig help text and documentation.
Documentation/contributing/coding_style.rst:
Documentation/introduction/about.rst:
boards/arm/lpc17xx_40xx/olimex-lpc1766stk/README.txt:
boards/arm/sam34/arduino-due/README.txt:
boards/arm/sam34/sam4l-xplained/README.txt:
boards/arm/sama5/giant-board/README.md:
boards/arm/sama5/sama5d2-xult/README.txt:
boards/arm/sama5/sama5d4-ek/README.txt:
boards/arm/samd2l2/samd20-xplained/README.txt:
boards/arm/samd2l2/samd21-xplained/README.txt:
boards/arm/samd2l2/saml21-xplained/README.txt:
boards/arm/stm32/hymini-stm32v/README.txt:
boards/arm/stm32/stm3210e-eval/README.txt:
fs/fat/Kconfig:
libs/libc/string/Kconfig:
* Updates stale references to the old top-level COPYING file to
either LICENSE or NOTICE (or both), as appropriate in each
instance.
Mark Schulte has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Pierre-Noel Bouteville has submitted the ICLA and we can migrate the licenses
to Apache.
David Sidrane has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Uros Platise has submitted the ICLA and we can migrate the licenses
to Apache.
David Sidrane has submitted the ICLA and we can migrate the licenses
to Apache.
Bob Feretich has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Johannes Schock has submitted the ICLA and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
getopt() in the FLAT build environment is not thread safe. This is because global variables that are process-specific in Unix are truly global in the FLAT build. Moving the getopt() variables into TLS resolves this issue.
No side-effects are expected other than to getopt()
Tested with sim:nsh
OpenGroup specification was updated regarding the return value for
pthread_once, which after Issue 7 states that "the [EINVAL] error for an
uninitialized pthread_once_t object is removed; this condition results
in undefined behavior".
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Several files were missed in the past
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
David Sidrane has submitted the ICL and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
David S. Alessio has submitted the ICL and we can migrate the licenses
to Apache.
Gregory Nutt has submitted the SGA and we can migrate the licenses
to Apache.
Signed-off-by: Alin Jerpelea <alin.jerpelea@sony.com>
Stream buffer flags has a dependency with allocated buffer,
and it cannot clear without buffer operation.
Change clearerr to keep buffer flags to prevent unexpected
buffer operation.
When CONFIG_SYMTAB_ORDEREDBYNAME is selected most code will use
the ordered search function. When it is not selected no code will
use the ordered search function. This change merges the two
functions and varies its behaviour based on the config setting,
such that all callers can simply call the one search function
and get the best behaviour.
An additional configuration option allows leading underscores to
be stripped from symbols being relocated in loaded objects. This
allows toolchains which prefix C symbol with underscores to make
loadable ELF objects.
Implement si_send/sendto/recvfrom with si_sendmsg/recvmsg, instead of
the other way round.
Change-Id: I7b858556996e0862df22807a6edf6d7cfe6518fc
Signed-off-by: Peter Bee <bijunda1@xiaomi.com>
Change the copyright header of assorted source files in libs/libc to the
ASF version. I was listed as the author in these files. I did also
check the heritage of those files and besides me they have only been
changed by the following people, all of whom did sign a CLA to the best
of my knowledge:
Alin Jerpelea
Haitao Liu
Gregory Nutt
Yamamoto Takashi
Xiang Xiao
Signed-off-by: Michael Jung <mijung@gmx.net>
Summary:
- The code was added in Mar 2018 to stabilize the SMP kernel
- I confirmed that the code is no longer needed now.
Impact:
- SMP only
Testing:
- Tested with ostest the following configs
- esp32-devkitc:smp (QEMU), sabre-6quad:smp (QEMU)
- maix-bit:smp (QEMU), sim:smp
- spresense:smp
- Tested with nxplayer and stress test with spresense:wifi_smp
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
ld error caused by multiple definition:
ld: arm-none-eabi/lib/thumb/v8-m.main+fp/hard/libsupc++.a(atexit_arm.o): in function `__aeabi_atexit':
atexit_arm.cc:(.text.__aeabi_atexit+0x0): multiple definition of `__aeabi_atexit';
nuttx/staging/libxx.a(libxx_eabi_atexit.o):nuttx/libs/libxx/libxx_eabi_atexit.cxx:75: first defined here
Signed-off-by: chao.an <anchao@xiaomi.com>
Summary:
- This commit fixes compile errors in libxx_new.cxx and libxx_newa.cxx
Impact:
- None
Testing:
- Build with spresense:nsh with CONFIG_DEBUG_ERROR=y
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
all the characters placed before the space on the ASCII table
and the 0x7F character (DEL) are control characters.
Signed-off-by: chao.an <anchao@xiaomi.com>
This might be specific to gcc or Linux
as I didn't see it with clang + macOS.
libcxx/src/locale.cpp: In member function 'void std::__1::locale::__install_ctor(const std::__1::locale&, std::__1::locale::facet*, long int)':
libcxx/src/locale.cpp:576:62: warning: declaration of 'id' shadows a member of 'std::__1::locale' [-Wshadow]
576 | locale::__install_ctor(const locale& other, facet* f, long id)
| ^
In file included from /Users/yamamoto/git/nuttx/nuttx/include/libcxx/locale:181,
from libcxx/src/locale.cpp:16:
/Users/yamamoto/git/nuttx/nuttx/include/libcxx/__locale:203:32: note: shadowed declaration is here
203 | class _LIBCPP_TYPE_VIS locale::id
| ^~
This might be specific to gcc or Linux
as I didn't see it with clang + macOS.
libcxx/src/barrier.cpp: In constructor 'std::__1::__barrier_algorithm_base::__barrier_algorithm_base(ptrdiff_t&)':
libcxx/src/barrier.cpp:35:9: warning: declaration of '__expected' shadows a member of 'std::__1::__barrier_algorithm_base' [-Wshadow]
35 | : __expected(__expected)
| ^
libcxx/src/barrier.cpp:29:24: note: shadowed declaration is here
29 | ptrdiff_t& __expected;
| ^~~~~~~~~~
libcxx/src/barrier.cpp: In constructor 'std::__1::__barrier_algorithm_base::__barrier_algorithm_base(ptrdiff_t&)':
libcxx/src/barrier.cpp:44:5: warning: declaration of '__expected' shadows a member of 'std::__1::__barrier_algorithm_base' [-Wshadow]
44 | }
| ^
libcxx/src/barrier.cpp:29:24: note: shadowed declaration is here
29 | ptrdiff_t& __expected;
| ^~~~~~~~~~
libcxx/src/barrier.cpp: In constructor 'std::__1::__barrier_algorithm_base::__barrier_algorithm_base(ptrdiff_t&)':
libcxx/src/barrier.cpp:44:5: warning: declaration of '__expected' shadows a member of 'std::__1::__barrier_algorithm_base' [-Wshadow]
44 | }
| ^
libcxx/src/barrier.cpp:29:24: note: shadowed declaration is here
29 | ptrdiff_t& __expected;
| ^~~~~~~~~~
CXX: libcxx/src/new.cpp
In file included from libcxx/src/new.cpp:19:
/usr/include/c++/9/cxxabi.h: In member function 'virtual bool __cxxabiv1::__pbase_type_info::__pointer_catch(const __cxxabiv1::__pbase_type_info*, void**, unsigned int) const':
/usr/include/c++/9/cxxabi.h:308:23: error: 'const class std::type_info' has no member named '__do_catch'
308 | return __pointee->__do_catch (thrown_type->__pointee, thr_obj, outer + 2);
| ^~~~~~~~~~
The length of hostbuffer come from Kconfig, it isn't safe to
allocate this variable on the stack.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I75213688153a1cba738544de2a51fa3247626dd3
wget is missing from some system (like macOS and Windows native),
it's better to use curl to simplify build environment.
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
The following errors are intentionally left.
They are a part of tables which are not trivial to fix.
libs/libc/machine/arm/armv6-m/arch_elf.c:228:94: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:230:89: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:238:94: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:240:89: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:401:94: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:403:91: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:411:94: error: Long line found
libs/libc/machine/arm/armv6-m/arch_elf.c:413:91: error: Long line found
For commands without "-" arguments, ex:"ls", we should always let
optind = 1 after getopt is called in order to get what follows
correctly.
Change-Id: Iac3cfbadd27fb96e47070c4e3198229306299b6b
Signed-off-by: dongjiuzhu <dongjiuzhu1@xiaomi.com>
modify the return value according to posix standard
Change-Id: I6e32a8f7a5cac85fe7a395a8115710fdff61d985
Signed-off-by: dongjiuzhu <dongjiuzhu1@xiaomi.com>
The following errors are intentionally left.
They are a part of tables which are not trivial to fix.
libs/libc/machine/arm/armv7-m/arch_elf.c:230:94: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:232:89: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:240:94: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:242:89: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:403:94: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:405:91: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:413:94: error: Long line found
libs/libc/machine/arm/armv7-m/arch_elf.c:415:91: error: Long line found
Also implement getpagesize() based on sysconf(_SC_PAGESIZE).
Change-Id: I812eb8f34ed602f7bc12c4cafafcebc0d98fd136
Signed-off-by: liuhaitao <liuhaitao@xiaomi.com>
The previous implementation of strtoul(l) is flawed. The range check
assumed that when overflow happens, the truncated value is smaller than
the original value. As a counter example, passing "10000000000" to
strtol will not trigger ERANGE, but return a truncated value. This patch
adds more accurate range checks.
Change-Id: I239e034e390b4974157ed6efa17110f2e74904cf
Signed-off-by: Peter Bee <bijunda1@xiaomi.com>
since it is wrong to close the builtin stream and specially note
https://pubs.opengroup.org/onlinepubs/9699919799/functions/fclose.html:
Since after the call to fclose() any use of stream results in
undefined behavior, fclose() should not be used on stdin, stdout,
or stderr except immediately before process termination (see XBD
Process Termination), so as to avoid triggering undefined behavior
in other standard interfaces that rely on these streams. If there
are any atexit() handlers registered by the application, such a
call to fclose() should not occur until the last handler is
finishing. Once fclose() has been used to close stdin, stdout, or
stderr, there is no standard way to reopen any of these streams.
and it is also unnecessary because the stream always get flushed.
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Follow the POSIX description.
SIGTSTP should be sent when the Ctrl-Z characters is encountered, not SIGSTP.
Testing:
Built with hifive1-revb:nsh (CONFIG_SERIAL_TERMIOS=y, CONFIG_SIG_DEFAULT=y and CONFIG_TTY_SIGTSTP=y)
- Turn some macros into functions
- Implement some type-agnostic functions.
(Just use __builtin_xxx)
- Add some missing function prototypes
(Just prototypes, not actually implemented in this commit)
to avoid FILE pointer change after this commit:
commit b0797263ca
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
Date: Thu Aug 13 18:17:29 2020 +0800
libc/stdio: Allocate file_struct dynamically
1.Reduce the default size of task_group_s(~512B each task)
2.Scale better between simple and complex application
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Endian-ness issues are mind boggling sometimes. I have been confused by the logic in inet_ntop() a few times so I thought I would add some comments so that I (and others) will understand the endian-ness issues in the future. No change to logic, only comments changed.
Move sched/task/task/task_gettid.c to libs/libc/unistd/lib_gettid.c. gettid() is a dumb wrapper around getpid(). It is wasteful of resources to support TWO systme calls, one for getpid() and one for gettid(). Instead, move gettid() in the C library where it calls the single sysgtem call, getpid(). Much cleaner.
base/time/time_exploded_posix.cc:190:14: error: ‘struct tm’ has no member named ‘tm_zone’; did you mean ‘tm_mon’?
190 | timestruct.tm_zone = nullptr; // not a POSIX field, so mktime/timegm ignore
| ^~~~~~~
| tm_mon
Change-Id: I9f93e63b50c0692a7a2bfc47abd9d07aa2c8e8db
Signed-off-by: chao.an <anchao@xiaomi.com>
Modify reason:
When build Nuttx SIM, in x86_64 system:
Compile with gcc option '-m64' (default):
sizeof(double_t) = 8
sizeof(double) = 8
Compile with gcc option '-mx32':
sizeof(double_t) = 8
sizeof(double) = 8
Compile with gcc option '-m32':
sizeof(double_t) = 12 // long double
sizeof(double) = 8
When use '-m32', and print sth. like this:
printf("%f\n", (double)3.0);
SIM will print out: nan
This is because sizeof(double_t) is not equal with double.
Resolve:
replace all double_t to double in libs/libc/stdio.
As a user of '-m32', you should know double_t is one type
long double, and len is 12. And you use use '%lf' to print.
like:
printf("%lf\n", (double_t)3.0);
Currently we don't support '%lf'.
Change-Id: I9b9d11853140d5296dd80416c8ed6a260a9d2d9c
Signed-off-by: ligd <liguiding1@xiaomi.com>
In some cases, when NuttX configuration changes and this makes the
object list used to build one of the .a libraries change as well,
since the command used to build it is "ar crs" and this simply appends
the list of object files, the library could still include object
files from prior builds. This commit modifies the ARCHIVE macro to
erase the .a file if it already exists.
Since in some cases this behavior was actually expected (object
files from a subdirectory were appended to a library created one
level above) I added a ARCHIVE_ADD which works as ARCHIVE did.
This change should greatly improve behavior of building after
configuration changes.
1.Reduce the default size of task_group_s(~512B each task)
2.Scale better between simple and complex application
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: Ia872137504fddcf64d89c48d6f0593d76d582710
and enable on sim as a demo. Here is the paper:
ftp://gcc.gnu.org/pub/gcc/summit/2003/Stackguard.pdf
Signed-off-by: qiaowei <qiaowei@xiaomi.com>
Change-Id: I1926936328dad54eee16b322dcaad0b42f9c4a62
1.Expose the notification through fb_vtable_s::updatearea
2.Incorporate old nx_notify_rectangle into the new updatearea callback
3.Migrate the calle of nx_notify_rectangle to fb_vtable_s::updatearea
Change-Id: Ia3d1f73e8757b2d381586d76ec6adc16c810018d
Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
drivers/net/ftmac100.c, libs/libc/stdlib/lib_aligned_alloc.c
A continuation of PRs #1507, #1510, and #1512. See Issue #1481 for additional information.
Found by clang-check:
modlib/modlib_sections.c:93:3: warning: Value stored to 'buffer' is never read
buffer = loadinfo->iobuffer;
^ ~~~~~~~~~~~~~~~~~~
1 warning generated.
simplify and symmetry the implementation in KERNEL/PROTECTED build
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: Iefdeea5f6ef6348c774b2ca9f7e45fe89c0c22dd
Make.dep file should be updated by .config changed after first make.
There are 2 cases affected for this problem:
1) Add source files by config symbol
2) Include header files in #ifdef directive
These 2 cases may not be included in Make.dep and this may prevent the
differential build from working correctly.
since libc++ declare these function in ctime by:
using ::localtime[_r];
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: Ic0bb68b44c0cab838ab7cc34baee2aaa3ca8a9b5
since exit will be only callable from userspace and change
the 1st argument from "const uint8_t *" to "const char *"
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I86487d57210ab63109148232da71dbc4d60a563b
because the logic:
1.only work on arm platform
2.couple with elf format
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I25dc95b5fc7b24196e2e71fdcf82d71d621ee2d3
since sim arch has to use other instance provided by host glibc
also initialize __dso_handle to self as glibc:
https://github.com/bminor/glibc/blob/master/csu/dso_handle.c#L21
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I7531ae58fc4dbe4600bcb2c2c3a6cac021378bc1
otherwise the user is shocked that using cin/cout/cerr will crash the system
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I453427261f3e2a6e60f7dd2398f7d3bd1043a0d5
1.Remove CONFIG_HAVE_INLINE macro
2.Change the ANSI C function to normal function
3.Other simple non ANSI function to macro
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
to avoid the similar code spread around each application
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I8967d647eaf2ecae47f29f83e7fa322ef1b42a02
since double_t move from sys/types.h to math.h now and remove
math.h inclusion too because lib_dtoa_engine.h already include
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: I3497a73908301d999cf1cfc4a66552a7ca4868c6
As it's stated in the standards.
The original code look at the first Xs. It can end up with
an unexpected behavior, if a template contains multiple series of Xs.
E.g. /tmp/XXXXXX/XXXXXX
This backs out a part of PR 1179 which has a very serious error: If C buffered I/O is available, then printf() and vprintf MUST use it. Otherwise, the ordering of the I/O will be screwed up. They must not use direct file descriptor I/O UNLESS C buffered I/O is disabled.
and update the related stuff in libs/libc/libc.csv
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Change-Id: Id695a7f07bf18a7b4e526297f9131c75ddb79d30
Functions within the OS must never set the errno value. fs_fdopen() was setting the errno value. Now, after some parameter changes, it reports errors via a negated errno integer return value as do most all other internal OS functions.
It's better to have a default working for many cases.
Usually DNS servers are not optimized for embedded clients.
Users can fine tune for their environment anyway.
1. Move pthread-specific data files from sched/pthread/ to libs/libc/pthread.
2. Remove pthread-specific data functions from syscalls.
3. Implement tls_alloc() and tls_free() with system calls.
4. Reimplement pthread_key_create() and pthread_key_free() using tls_alloc() and tls_free().
5. Reimplement pthread_set_specific() and pthread_get_specicif() using tls_set_value() and tls_get_value()
- Remove per-thread errno from the TCB structure (pterrno)
- Remove get_errno() and set_errno() as functions. The macros are still available as stubs and will be needed in the future if we need to access the errno from a different address environment (KERNEL mode).
- Add errno value to the tls_info_s structure definitions
- Move sched/errno to libs/libc/errno. Replace old TCB access to the errno with TLS access to the errno.
Move the logic to get TLS information from an inline function to a normal function. For the unaligned case, it is probably too large to be inlined.
Also fixes some minor things from review of previous commits.
CONFIG_TLS_ALIGNED will select the (legacy) aligned stack implementation of TLS. If CONFIG_TLS_ALIGNED is not defined, then the new, implementation of TLS using an unaligned stack will be enabled.
The new OS interface, sched_get_stackinfo() combines two pthread-specific interfaces into a single generic interface. The existing pthread_get_stackaddr_np() and pthread_get_stacksize_np() are moved from sched/pthread to libs/libc/pthread.
There are two motivations for this change: First, it reduces the number of system calls. Secondly, it adds a common hook that is going to used for a future implementation of TLS.
