summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* cygwin: workaround GCC 6 changesYaakov Selkowitz2017-09-123-2/+6
| | | | | | | | | | | | | GCC 6 includes a number of new warnings which cause Cygwin to either not compile, or not work properly even if said warnings are ignored: https://cygwin.com/ml/cygwin-developers/2017-09/msg00000.html https://gcc.gnu.org/gcc-6/porting_to.html For now, we use the flags necessary to revert to GCC 5 behaviour until we can fix the code properly. Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
* Revert "cygwin: only expose /dev/con{in,out,sole} when started from a ↵Yaakov Selkowitz2017-09-111-1/+1
| | | | | | | | | | | | Windows console" This caused serious regressions when running from a cmd window: https://cygwin.com/ml/cygwin/2017-09/msg00114.html This reverts commit b706c6b479422d31f0124b92c21b4cb536bbddff. Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
* Change time_t to 64-bit by defaultSebastian Huber2017-09-077-3/+60
| | | | | | | | | In order to avoid the year 2038 problem, define time_t to a signed integer with at least 64-bits. The type for time_t can be forced to long with the --enable-newlib-long-time_t configure option or with the _USE_LONG_TIME_T system configuration define. Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* Remove harmful casts in gmtime_r()Sebastian Huber2017-09-071-2/+2
| | | | | | | | In case time_t is long, then the cast to long is a nop. In case time_t is __int_least64_t, then the cast to long may truncate the value before the division. Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* Let RTEMS provide clock()Sebastian Huber2017-09-071-1/+1
| | | | | | | | | | | Newlib uses _times_r() in clock(). The problem is that the _times_r() clock frequency is defined by sysconf(_SC_CLK_TCK). The clock frequency of clock() is the constant CLOCKS_PER_SEC. FreeBSD uses getrusage() for clock(). Since RTEMS has only one process, the implementation can be simplified. Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* stdio.h: Don't define unlocked macros using static inline on C++cygwin-2_9_0-releaseCorinna Vinschen2017-09-041-1/+1
| | | | | | | | | | | | In C++, the usage of static inline functions for getchar_unlocked and putchar_unlocked may result in error messages like error: ‘_putchar_unlocked’ was not declared in this scope Fix this by not using the _getchar_unlocked and _putchar_unlocked macros in C++. Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* cygwin: Document crypt_rYaakov Selkowitz2017-09-021-0/+1
| | | | Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
* Remove some dangerous advice from the FAQAchim Gratz2017-09-021-4/+8
|
* include: fix ffs, fls guardsYaakov Selkowitz2017-08-301-2/+4
| | | | Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
* newlib: string/Makefile.am (CHEWOUT_FILES): Add strnstr.defCorinna Vinschen2017-08-302-2/+2
| | | | | | Regenerate strings/Makefile.in Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* newlib: strnstr: drop traditional synopisisCorinna Vinschen2017-08-301-7/+0
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* Add man page entry for strnstr.c.Sichen Zhao2017-08-302-0/+43
|
* Modify strnstr.c.Sichen Zhao2017-08-301-55/+10
|
* cygwin: add strptime %F fix to release notesCorinna Vinschen2017-08-291-0/+3
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* winsup/cygwin/libc/strptime.cc(__strptime) fix %F widthBrian Inglis2017-08-291-2/+4
|
* cygwin: document %s support in strptimeCorinna Vinschen2017-08-292-0/+6
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* Fix compile error due to new strnstr()Sebastian Huber2017-08-281-43/+0
| | | | | | | | | | | | | | | Remove local strnstr() implementation to fix compile error: newlib/libc/iconv/lib/aliasesi.c:53:8: error: conflicting types for 'strnstr' _DEFUN(strnstr, (haystack, needle, length), ^ In file included from newlib/libc/iconv/lib/aliasesi.c:29:0: newlib/libc/include/string.h:125:10: note: previous declaration of 'strnstr' was here char *strnstr(const char *, const char *, size_t) __pure; ^~~~~~~ Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* cygwin: export strnstrCorinna Vinschen2017-08-254-3/+5
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* newlib: rebuild string/Makefile.inCorinna Vinschen2017-08-251-27/+49
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* Port strnstr.c to newlib.Sichen Zhao2017-08-252-0/+4
|
* Import strnstr.c from FreeBSD.Sichen Zhao2017-08-251-0/+65
|
* Revert "headers: avoid bareword attributes" for clangEric Blake2017-08-251-14/+16
| | | | | | | | | | | | | | | This reverts most of commit 979d467ff6e39ee5c52cf1aac7a6c9c63058141c. We cannot avoid some bareword attributes until clang is fixed to properly support __-decorated attributes; see this bug: https://bugs.llvm.org/show_bug.cgi?id=34319 The macros in question expand to the empty string under gcc, so only compilation under clang is affected, and since clang has the bug, the obvious solution is to roll back the changes, and document the issue. Signed-off-by: Eric Blake <eblake@redhat.com>
* RTEMS: Use __uint64_t for _CLOCK_T_Sebastian Huber2017-08-251-0/+3
| | | | | | | | This addresses: https://devel.rtems.org/ticket/2135 Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* Make _CLOCK_T_ system configurableSebastian Huber2017-08-251-0/+3
| | | | | | | Let systems optionally provide the _CLOCK_T_ type via <machine/_types.h>. Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* newlib/libc/time/strptime.c(strptime_l) add %F %s support for strptimeBrian Inglis2017-08-251-2/+32
|
* winsup/cygwin/libc/strptime.cc(__strptime) add %s support to strptimeBrian Inglis2017-08-251-0/+20
|
* move ILP32 sanity check on heap base code under ARM_RDI_MONITORAlexander Fedotov-B556132017-08-241-4/+5
|
* cygwin: only expose /dev/con{in,out,sole} when started from a Windows consoleCorinna Vinschen2017-08-231-1/+1
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* Add myself to RISC-V Port MaintainerKito Cheng2017-08-211-0/+1
|
* Change license to FreeBSD License for RISC-VKito Cheng2017-08-2116-16/+16
| | | | | | - For prevent confuse about what BSD license variant we used, 2- or 3-clause license, we change the license to FreeBSD license to make it unambiguously refers to the 2-clause license.
* Document renameat2Ken Brown2017-08-193-0/+10
|
* cygwin: Implement renameat2Ken Brown2017-08-195-7/+54
| | | | | | Define the RENAME_NOREPLACE flag in <cygwin/fs.h> as defined on Linux in <linux/fs.h>. The other RENAME_* flags defined on Linux are not supported.
* Add RISC-V port for libglossnewlib-snapshot-20170818Kito Cheng2017-08-179-0/+4340
| | | | | | | | Contributor list: - Andrew Waterman <andrew@sifive.com> - Palmer Dabbelt <palmer@dabbelt.com> - Kito Cheng <kito.cheng@gmail.com> - Alex Suykov <alex.suykov@gmail.com>
* Add RISC-V port for libmKito Cheng2017-08-1720-1/+7315
| | | | | | Contributor list: - Michael Neilly <mneilly@yahoo.com> - Kito Cheng <kito.cheng@gmail.com>
* headers: avoid bareword attributesEric Blake2017-08-172-15/+15
| | | | | | | | | Always use the __-decorated form of an attribute name in public headers, as the bareword form is in the user's namespace, and we don't want compilation to break just because the user defines the bareword to mean something else. Signed-off-by: Eric Blake <eblake@redhat.com>
* headers: avoid bareword attributesEric Blake2017-08-174-5/+5
| | | | | | | | | Always use the __-decorated form of an attribute name in public headers, as the bareword form is in the user's namespace, and we don't want compilation to break just because the user defines the bareword to mean something else. Signed-off-by: Eric Blake <eblake@redhat.com>
* Fix crt0 init fini codeSzabolcs Nagy2017-08-171-4/+0
| | | | | __USES_INITFINI__ ifdef was incorrectly copied from arm (it's an arm backend thing in gcc, not meaningful on aarch64)
* Add RISC-V port for newlibKito Cheng2017-08-1625-2/+7198
| | | | | | | | Contributor list: - Andrew Waterman <andrew@sifive.com> - Palmer Dabbelt <palmer@dabbelt.com> - Kito Cheng <kito.cheng@gmail.com> - Scott Beamer <sbeamer@eecs.berkeley.edu>
* fix typo in AArch64 crt0Alexander Fedotov-B556132017-08-151-1/+1
|
* use stack from linker script when nosysAlexander Fedotov-B556132017-08-141-1/+18
|
* newlib: regenerate libc/stdlib/Makefile.amCorinna Vinschen2017-08-111-9/+36
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* Don't fetch command line options without semi-hostingYao Qi2017-08-091-0/+6
| | | | | | | | | | | | | | | | | Nowadays, the code fetching command line options via semi-hosting are unconditionally pulled in, so that the semi-hosting code is still there even I compile with option --specs=nosys.specs. gdb ./aarch64-none-elf/libgloss/aarch64/crt0.o (gdb) disassemble _start 0x0000000000000050 <+80>: ldr x1, 0x128 <_cpu_init_hook+48> 0x0000000000000054 <+84>: mov w0, #0x15 0x0000000000000058 <+88>: hlt #0xf000 This patch fixes this problem by wrapping the code by ARM_RDI_MONITOR. When semi-hosting is not used, set command line options to NULL.
* Fix crt0 overwriting.Tamar Christina2017-08-094-1/+10
| | | | | | | | | | | On AArch64 we currently always link in crt0 regardless of if another one is being provided by something else, like rdimon.a. This was never an issue before as nosys was not supported on AArch64. This updates the specs to supply a different crt0 when a semihosting call is required. Signed-off-by: Tamar Christina <tamar.christina@arm.com>
* Define sigsetjmp/siglongjmp only if __POSIX_VISIBLEKen Brown2017-08-071-2/+2
|
* Proper locking for getchar() and putchar()Sebastian Huber2017-08-071-7/+32
| | | | | | | | | | | | | Add internal inline functions _getchar_unlocked() and _putchar_unlocked() if __CUSTOM_FILE_IO__ is not defined. These functions get _REENT only once. Use them for getchar_unlocked() and putchar_unlocked(). Define getchar() and putchar() to these unlocked internal functions if __SINGLE_THREAD__ is defined, otherwise use the external functions to use proper locking of the FILE object. Assumes that __SINGLE_THREAD__ is not defined if __CYGWIN__ is defined. Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
* cygwin: pthread timed locks: actually timeout on timeoutCorinna Vinschen2017-08-031-4/+26
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* cygwin: pthread_rwlock_rdlock: don't set errno, just return error codeCorinna Vinschen2017-08-031-1/+1
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* cygwin: Implement pthread_rwlock_timedrdlock, pthread_rwlock_timedwrlockCorinna Vinschen2017-08-037-11/+84
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* cygwin: Bump DLL version to 2.9.0Corinna Vinschen2017-08-034-4/+20
| | | | Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
* cygwin: Implement pthread_mutex_timedlockCorinna Vinschen2017-08-035-4/+35
| | | | | | | - pthread_mutex::lock now takes a PLARGE_INTEGER timeout pointer and uses that in the call to cygwait. Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
generated by cgit v1.2.3 (git 2.25.1) at 2025-07-16 16:11:35 +0000