| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
-Wundef warns if an undefined identifier is evaluated in an #if
directive. This would be valid as they are replaced with 0. However, it
is often an early warning sign and not intentional. To allow for
enabling -Wundef even outside system directories (where compilers
ignore such problems unless -Wsystem-headers is enabled) this patch
adds the required defined() checks.
glibc also has been supporting this for 10 years now:
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=f248238cf43bd751db29e6f151d6da7645337ff5
I have not exhaustively tested this but you can see the effect with
something like the following (+ using -stdc= and/or -D...):
echo | gcc -include newlib/libc/include/sys/features.h -Inewlib/libc/include/ -E - -Wundef
Signed-off-by: Stefan Tauner <stefan.tauner@gmx.at>
|
|
The commit a2a8bc771f2f has a problem that __atexit which includes
args for atexit()'ed function may be touched by another thread
since the mutex is unlocked while calling the function.
With this path, the args, etc. are copyed to local variable to
prevent this problem.
Fixes: a2a8bc771f2f ("newlib: Unlock the mutex while calling atexit()'ed functions")
Suggested-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>, Sebastian Huber <sebastian.huber@embedded-brains.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
|
|
The atexit()'ed function may deadlock if it waits for another thread
which calls atexit() in the current __call_atexit.c code. This patch
unlock __atexit_recursive_mutex while calling atexit()'ed functions
to avoid the deadlock mentioned above. glibc and Darwin do the same,
so it sounds reasonable.
Addresses: https://cygwin.com/pipermail/cygwin/2025-October/258930.html
Reported-by: Tomohiro Kashiwada <tomohiro-kashiwada@ezweb.ne.jp>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
|
|
The documentation states that _GNU_SOURCE enables "all of the above
plus GNU extensions.". Furthermore the documentation states that
_DEFAULT_SOURCE would enable POSIX-1.2008 with BSD and SVr4 extensions.
The code did different things though:
* For _GNU_SOURCE it only enabled POSIX.1-2008 but it should have been the
highest available one, POSIX.1-2024.
* For _DEFAULT_SOURCE it enabled POSIX.1-2024.
Let's use the highest available POSIX standard possible and fix
documentation where necessary.
|
|
POSIX.1-2024 added strlcpy, strlcat, and memmem to strings.h. Previously
these functions were only available as parts of BSD or GNU extensions.
Fix the header so that the symbol are visible for either the right
extensions, or for POSIX.1-2024.
|
|
This patch adds a missing function declaration in system.c file. This helps avoid declaration compiler
warning or error.
Tested by building newlib for m68k-elf with GCC.
|
|
The RTEMS section defined macros for features RTEMS does not have.
|
|
The setjmp() function needs this attribute to help GCC avoid false
positives for the -Wclobbered warning. The -Wclobbered warning is
part of -Wextra.
|
|
* libc/sys/h8300hms/crt0.S: General cleanup: declare stack section
in a proper way, merge H8/300H and H8/300S/SX parts of code into
one, add support for normal mode.
|
|
* libc/machine/h8300/strcmp.S (_strcmp): First extend 8 bit
character codes to 16 bit integers and then compare them.
|
|
This helper is only used within strtodg.c; keep it internal to avoid
exporting a generic symbol name and reduce namespace pollution.
* libc/stdlib/strtodg.c (increment): Make static.
|
|
The ARM fegetexceptflag() implementation was accidentally replaced with
fesetexceptflag() during refactoring in b7a6e02dc6a5289bfa489c0e7b6539abd281e2c6.
Restore the original implementation.
Tested on Cortex-M7.
|
|
The implementation of strtorQ is imported from FreeBSD's gdtoa library (By David
M. Gay) with some adaptations to fit with newlib.
`strtorQ.c` enables `strtold` to perform correct parsing on targets where long
double uses the IEEE754 binary128 format (113-bit mantissa), such as AArch64.
Without this patch, strtold would wrongly fallback to `strtorx` which will
parse into a 80-bit long double and give invalid result.
* libc/stdlib/strtorQ.c: New file, adapted from FreeBSD gdtoa.
* libc/stdlib/strtold.c (_strtold_impl): New helper selecting strtorx
for 80-bit and strtorQ for 128-bit long double.
(_strtold_r, strtold_l, strtold): Use _strtold_impl.
* libc/stdlib/mprec.h (_strtorQ_l): Declare.
* libc/stdlib/Makefile.inc (libc_a_SOURCES): Add strtorQ.c.
* newlib/Makefile.in: Regenerate with automake
Signed-off-by: Zakaria Fadli <fadli@adacore.com>
|
|
The vector ilogb routines, including the ones inlined into fmod, had a bug
in which the conditional masks were not properly applied, causing the value of
one lane to be affected by conditional choices of another lane. The problem
was not immediately obviously because all values were calculated correctly when
no lane contained a subnormal input.
The problem is fixed by proper use of VECTOR_COND_MOVE and VECTOR_WHILE.
|
|
The _arc4random_getentropy_fail() function does not return. Mark it as
such.
|
|
The end condition on this loop, unlike all the other similar loops, is
"i >= 0", which is a problem because "i <<= 1" can go negative and then zero if
you continue shifting, and so back to true, again. This isn't a problem for
the loop in the scalar implementation, but it means we need to mask the shift
in the vector implementation.
This fixes GCC PR#121392.
|
|
|
|
Linux and POSIX-1.2024 define operations to create OFD (Open
File Description) locks.
OFD locks are like POSIX record locks, but bound to the file descriptor,
not bound to the process. As such, they are also inherited via fork(2)
and dup(2), just like BSD flock(2) locks.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
|
|
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
|
|
This patch is returning back support for AArch64 ILP32 ABI that was
removed in de479a54e22e8fcb6262639a8e67fe8b00a27c37 commit but is needed
to ensure source code compatibility with GCC 14.
The change in newlib/libc/machine/aarch64/asmdefs.h makes it
out-of-the-sync with the current upstream implementation in
https://github.com/ARM-software/optimized-routines repository.
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
This was what I was using locally before Radek Bartoň
<radek.barton@microsoft.com> had his version of the patch. Revert in
favor of his final version.
Revert 70c5505766ad4ae62e4d045835ed2a6b928d5760
|
|
ilp32 support was removed prematurely. It is still in GCC 15 which is
the latest GCC release.
From: <radek.barton@microsoft.com>
Date: Thu, 5 Jun 2025 11:32:08 +0200
Subject: [PATCH] newlib: libc: update asmdefs.h compatible with Cygwin AArch64
This patch synchronizes newlib/libc/machine/aarch64/asmdefs.h header with
version from https://github.com/ARM-software/optimized-routines/commit/4352245388a55a836f3ac9ac5907022c24ab8e4c
commit that added support for AArch64 Cygwin.
This version of the header removed PTR_ARG and SIZE_ARG macros as ILP32 was
deprecated which introduced changes in many .S files so the patch contains
removal of usages of those macros.
On top of that, setjmp.S and rawmemchr.S were refactored to use
ENTRY/ENTRY_ALIGN and END macros. `
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
UTF-16"
This reverts commit b374973d14ac7969b10ba719feedc709f6971c0d.
Turns out this patch breaks mbrtowc. Example:
--- SNIP ---
void mb(unsigned char c)
{
wchar_t wc;
int ret = mbrtowc(&wc, &c, 1, 0);
printf("%02X -> %04X : %d\n", c, wc, ret);
}
void main ()
{
setlocale (LC_CTYPE, "");
mb(0xF0);
mb(0x9F);
mb(0x98);
mb(0x8E);
}
--- SNAP ---
Output before commit b374973d14ac:
F0 -> 0000 : -2
9F -> 0000 : -2
98 -> D83D : 1
8E -> DE0E : 1
Output after commit b374973d14ac:
F0 -> 0000 : -2
9F -> 0000 : -2
98 -> 0000 : -2
8E -> D83D : 3
By using mbrtowc(), the high surrogate is only emitted after byte 4, and
there's no way to recover the low surrogate. The byte count is also incorrect.
Conclusion: We have to emit the high surrogate already after byte 3
to be able to emit the low surrogate after byte 4.
Reported-by: Thomas Wolff <towo@towo.net>
Addresses: https://cygwin.com/pipermail/cygwin/2025-July/258513.html
Fixes: b374973d14ac ("mbrtowc: fix handling invalid UTF-8 4 byte sequences if wchar_t == UTF-16")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
|
|
The macro was not protecting properly its first parameter, which caused
it to silently return an invalid value when passing a composed
parameter, ie.
__builtin_is_aligned(src | dst | len, 32);
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
|
|
This reverts commit bb2c338520b9da9fd3ac067438e9aeeffc0b122a.
Accidental push.
|
|
This reverts commit 3b97a5ec67a5a52c130158bb143949cd842de305.
|
|
This patch introduces dummy implementation of fegetprec and fsetprec for Cygwin
build as those symbols are being exported by cygwin1.dll and AArch64 do not support
changing floating point precision at runtime.
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
The commit cited below introduced changes which included
conditionals strictly on __mips__. This ignored tailoring
for any environment on MIPS except that targeted by the
author of the change. This patch just fixes this code for
RTEMS.
commit 467a2bdf17ad376dafada9f1734784f4611fa6fd
Author: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
Date: Wed Jun 11 10:11:33 2025 +0200
mips: Implement MIPS HAL and UHI
Implement abstract interface for MIPS, including unified hosting
interface (UHI).
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
This commit changes the return type of the read() and write() syscalls for
nvptx to ssize_t. This would allow large files to be handled properly by
these syscalls in situations where the read/write buffer length exceeds
INT_MAX, for example. This also makes the syscall signatures fully complaint
with their current POSIX specifications.
We additionally define two macros: '_READ_WRITE_RETURN_TYPE' as _ssize_t and
'_READ_WRITE_BUFSIZE_TYPE' as __size_t in libc/include/sys/config.h under
__nvptx__ for consistency.
Signed-off-by: Arijit Kumar Das <arijitkdgit.official@gmail.com>
|
|
This patch synchronizes newlib/libc/machine/aarch64/asmdefs.h header with
version from https://github.com/ARM-software/optimized-routines/commit/4352245388a55a836f3ac9ac5907022c24ab8e4c
commit that added support for AArch64 Cygwin.
This version of the header removed PTR_ARG and SIZE_ARG macros as ILP32 was
deprecated which introduced changes in many .S files so the patch contains
removal of usages of those macros.
On top of that, setjmp.S and rawmemchr.S were refactored to use
ENTRY/ENTRY_ALIGN and END macros. `
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
With the commit 656df313e08a, if a thread acquires sfp lock after
another thread calls fclose() and fp lock is acquired, the first
thread falls into deadlock if it tries to acquire fp lock. This can
happen if the first thread calls __sfp_lock_all() while the second
thread calls fclose().
This patch reverts the changes for newlib/libc/stdio/fclose.c in
the commit 656df313e08a.
Addresses: https://cygwin.com/pipermail/cygwin/2025-June/258323.html
Fixes: 656df313e08a ("* libc/stdio/fclose.c: Only use sfp lock to guard non-atomic changes of flags and fp lock.")
Reported-by: Takashi Yano <takashi.yano@nifty.ne.jp>
Reviewed-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Takashi Yano <takashi.yano@nifty.ne.jp>
|
|
For _GNU_VISIBLE, provide the CLOCK_TAI clock identifier for the
International Atomic Time. Use the value specified by glibc and Linux.
Add _BSD_VISIBLE given FreeBSD also provides this clock identifier.
|
|
Regenerate the configuration files since a file was replaced
in the RISC-V port, and one other file was renamed.
Reviewed-by: Christian Herber <christian.herber@oss.nxp.com>
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
Redirect to memcpy() if the memory areas of source and destination
do not overlap. Only redirect if length is > SZREG in order to
reduce overhead on very short copies.
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
If misaligned accesses are slow or prohibited, either
source or destination address are unaligned and the number
of bytes to be copied is > SZREG*2, align the source address
to xlen. This speeds up the function in the case where at
least one address is unaligned, since now one word (or doubleword
for rv64) is loaded at a time, therefore reducing the amount of
memory accesses necessary.
We still need to store back individual bytes since the
destination address might (still) be unaligned after aligning
the source.
The threshold of SZREG*2 was chosen to keep the negative effect
on shorter copies caused by the additional overhead from aligning
the source low.
This change also affects the case where both adresses are xlen-
aligned, the memory areas overlap destructively, and length is not
a multiple of SZREG. In the destructive-overlap case, the copying
needs to be done in reversed order. Therefore the length is added
to the addresses first, which causes them to become unaligned.
Reviewed-by: Christian Herber <christian.herber@oss.nxp.com>
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
Add loop-unrolling for the case where both source and destination
address are aligned in the case of a destructive overlap, and
increase the unroll factor from 4 to 9 for the word-by-word
copy loop in the non-destructive case.
This matches the loop-unrolling done in memcpy() and increases
performance for lenghts >= SZREG*9 while almost not at all
degrading performance for shorter lengths.
Reviewed-by: Christian Herber <christian.herber@oss.nxp.com>
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
Replace macros with static inline functions or RISC-V specifc
macros in order to keep consistency between all functions in the port.
Change data types to fixed-width and/or RISC-V specific types.
Reviewed-by: Christian Herber <christian.herber@oss.nxp.com>
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
Copy the common implementation of memmove() to the RISC-V port.
Rename memmove.S to memmove-asm.S to keep naming of files
consistent between functions. Update Makefile.inc with the changed
filenames.
Reviewed-by: Christian Herber <christian.herber@oss.nxp.com>
Signed-off-by: m fally <marlene.fally@gmail.com>
|
|
According to the POSIX documentation, "when the new process image is
executed, any file descriptor (from this new set) which has its
FD_CLOEXEC flag set shall be closed (see posix_spawn())." The "new set"
is after processing the file actions, so if addopen had the O_CLOEXEC
flag set the descriptor should be closed after the exec.
The adddup2 docs, by contrast, specify that the flag should be cleared,
even if dup2 wouldn't have done so due to the specified file descriptors
being equal. Add a comment to that effect.
Addresses: https://sourceware.org/pipermail/newlib/2025/021968.html
Fixes: c7c1a1ca1b ("2013-10-01 Petr Hosek <phosek@chromium.org>")
Signed-off-by: Jeremy Drake <cygwin@jdrake.com>
|
|
strcmp gives incorrect result for little endian targets under
the following conditions:
1. Length of 1st string is 1 less than a multiple of 4 (i.e len%4=3)
2. First string is a prefix of the second string
3. The first differing character in the second string is extended
ASCII (that is > 127)
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
Improve `strcmp` by using `ext` instruction, if available.
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
Fix prefetching in core loop to avoid exceeding the operated upon
memory region. Revert accidentally changed prefetch-hint back to
streaming mode. Refactor various bits and provide pre-processor
checks to allow parameters to be overridden from compiler command
line.
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
Implement abstract interface for MIPS, including unified hosting
interface (UHI).
Signed-off-by: Jovan Dmitrović <jovan.dmitrovic@htecgroup.com>
|
|
For aarch64 on ELF targets, the library does not export
fe{enable,disable,get}except as symbols from the library, relying on
static inline functions to provide suitable definitions if required.
But for Cygwin we need to create real definitions to satisfy the DLL
export script.
So arrange for real definitions of these functions when building on Cygwin.
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
A __WCTOMB call on the terminating NUL may emit more than a NUL byte.
This is the case if the string ends with a lone UTF-16 high surrogate.
Fixes: 2a3a02a68764 ("Add SUSV2 support for calculating size if output buffer is NULL")
Signed-off-by: Christian Franke <christian.franke@t-online.de>
|
|
for aarch64-pc-cygwin
Signed-off-by: Radek Bartoň <radek.barton@microsoft.com>
|
|
GCC 13 does not define the __riscv_misaligned_* builtin defines. They are
supported by GCC 14 or later. Test for __riscv_misaligned_fast to select an
always correct memcpy() implementation for GCC 13.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
|
|
Addresses: https://cygwin.com/pipermail/cygwin/2025-June/258378.html
Fixes: 6ff28fc3b121 ("Allow CESU-8 surrogate value encoding")
Signed-off-by: Christian Franke <christian.franke@t-online.de>
|
|
When commit 28186e81d947 split _mbtowc_r into per-codeset functions, the
code generating wchar_t from UTF-8 input was slightly rearranged.
Unfortunately the new code introduced a bug:
On systems with wchar_t being UTF-16, 4 byte sequences have to be
converted to surrogate pairs. The low surrogate pair can be fully
created from the first 3 bytes of the sequence. However, the surrogates
should only be created if it's clear that the 4th byte is valid, and the
entire 4 byte string represents a valid UTF-8 sequence.
The code change in 28186e81d947 neglected just that: In contrast to the
original code, it now created the low surrogate after having read the
first 3 bytes of the sequence, without checking validity of the 4th byte.
This patch moves the test sequence to check the 4th byte in front of the
code generating the low surrogate. Make sure to return the value 3 (3
bytes digested) rather than the content of the local variable i, which
is already set to 4 at this point.
Reported-by: Christian Franke <Christian.Franke@t-online.de>
Addresses: https://cygwin.com/pipermail/cygwin/2025-June/258358.html
Fixes: 28186e81d947 ("* libc/ctype/iswalpha.c: Handle all wchar_t as unicode on _MB_CAPABLE systems.")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
|
