Pieter du Preez [Sun, 14 Feb 2016 10:12:06 +0000 (11:12 +0100)]
Define the newlib version macros in one place: _newlib_version.h.
Currently, the newlib version information needs to be updated in two places:
- newlib/acinclude.m4
- newlib/libc/include/sys/features.h
The goal of this patch is to:
- supply a single location for defining the newlib version
information: newlib/acinclude.m4
- define __NEWLIB__, __NEWLIB_MINOR__ and __NEWLIB_PATCHLEVEL__
This is in line with what gcc does for its version macros. See:
https://gcc.gnu.org/onlinedocs/cpp/Common-Predefined-Macros.html
This patch moves the definition of the _NEWLIB_VERSION, __NEWLIB__
and __NEWLIB_MINOR__ macros from newlib/libc/include/sys/features.h,
to the newly generated newlib/_newlib_version.h file. Additionally,
the __NEWLIB_PATCHLEVEL__ macro was created, for completeness.
In order to stay backwards compatible, newlib/_newlib_version.h gets
included by newlib/newlib.h and newlib/libc/include/sys/features.h.
Note: This patch does _not_ include the modifications to the following
files, as these should all be generated any way.
*Makefile.in,
*aclocal.m4,
*configure
stamp-* files
Signed-off-by: Pieter du Preez <pdupreez@gmail.com>
Corinna Vinschen [Wed, 17 Feb 2016 15:08:59 +0000 (16:08 +0100)]
cyglsa: Drop code handling Windows 2000 and earlier
* cyglsa.c (must_create_logon_sid): Remove.
(LsaApInitializePackage): Drop checking version and setting
must_create_logon_sid accordingly.
(LsaApLogonUserEx): Drop code setting login SID only required on
Windows 2000 and earlier.
Mark Geisert [Wed, 17 Feb 2016 09:55:16 +0000 (10:55 +0100)]
Silent relocation truncations considered harmful
This follows up from my msg re GMP-ECM failing its 'make check' on the
main list https://cygwin.com/ml/cygwin/2016-02/msg00147.html .
There's an error that ought to be reported during dynamic linking if the
linked-to address is too far from the relocation site. However the error
is not reported if __OPTIMIZE__ was #defined when building the Cygwin DLL.
I can't see why optimization settings should affect this.
Corinna Vinschen [Mon, 15 Feb 2016 15:25:10 +0000 (16:25 +0100)]
cygwin_conv_path: Really always preserve trailing slash in conversion to POSIX path
* mount.cc (mount_info::conv_to_posix_path): Account for scenarios
where the path is identical to a mount point in terms of preserving
the trailing slash in the output.
Václav Haisman [Fri, 12 Feb 2016 21:25:59 +0000 (22:25 +0100)]
POSIX barrier implementation, take 3
The attached patch should address all of the review comments.
Modifed change log:
Newlib:
* libc/include/sys/features.h (_POSIX_BARRIERS): Define for Cygwin.
* libc/include/sys/types.h (pthread_barrier_t)
(pthread_barrierattr_t): Do not define for Cygwin.
Yaakov Selkowitz [Fri, 12 Feb 2016 17:17:35 +0000 (11:17 -0600)]
cygwin: fix errors with GCC 5
GCC 5 switched from C89 to C11 by default. This implies a change from
GNU to C99 inline by default, which have very different meanings of
extern inline vs. static inline:
https://gcc.gnu.org/onlinedocs/gcc/Inline.html
Marking these as gnu_inline retains the previous behaviour.
winsup/cygwin/
* exceptions.cc (exception::handle): Change debugging to int to fix
an always-true boolean comparison warning.
* include/cygwin/config.h (__getreent): Mark gnu_inline.
* winbase.h (ilockcmpexch, ilockcmpexch64): Ditto.
Yaakov Selkowitz [Fri, 12 Feb 2016 02:23:10 +0000 (20:23 -0600)]
Remove broken ieeefp.h macros
Any attempt to use isnanf, isinff, or finitef from <ieeefp.h> with
GCC 5 on platforms other than SPU result in a "lvalue required as
unary '&' operand" error.
Yaakov Selkowitz [Fri, 12 Feb 2016 02:16:06 +0000 (20:16 -0600)]
Fix compile with GCC 5 -Werror
newlib/libc/
* stdio64/freopen64.c: Include <string.h> for memset().
* stdlib/quick_exit.c: Include <unistd.h> for _exit().
* string/gnu_basename.c (__gnu_basename): Fix discarded const
qualifier warning.
* stdlib/strtold.c: Include "mprec.h" for _strtorx_r().
printf(3): Handle multibyte decimal point in field size computation
This patch fixes the problem reported in
https://cygwin.com/ml/cygwin/2016-02/msg00014.html
The 2009 changes to handle multibyte decimal point and thousands
separator missed to take the length of a multibyte decimal point into
account when computing the field size.
Jon Turney [Thu, 4 Feb 2016 17:07:44 +0000 (17:07 +0000)]
stdio.h: Use __POSIX_VISIBLE etc. guards for POSIX 1003.1:2001 functions
Current mesa expects stdio.h to prototype fileno() when compiling with gcc
-std=c99 -D_XOPEN_SOURCE
Fix the 'POSIX 1003.1:2001' block in stdio.h to prototype functions if not
__STRICT_ANSI__, or _BSD_SOURCE, _POSIX_C_SOURCE or _XOPEN_SOURCE are defined
appropriately
As far as I can tell, despite being in this block, setbuffer() and setlinebuf()
are BSDisms which aren't in POSIX 1003.1:2001
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Игорь Веневцев [Fri, 29 Jan 2016 15:19:57 +0000 (18:19 +0300)]
Newlib build is broken if configured with nano-malloc and non-reentrant system calls
Non-reentrant system calls version implies both MISSING_SYSCALL_NAMES
and REENTRANT_SYSCALL_PROVIDED macros to be defined.
Being coupled with --enable-newlib-nano-malloc knob it breaks the build:
../../../../../../newlib-2.3.0.20160104/newlib/libc/stdlib/nano-mallocr.c:
In function ‘_mallinfo_r’:
../../../../../../newlib-2.3.0.20160104/newlib/libc/stdlib/nano-mallocr.c:489:35:
error: macro "_sbrk_r" requires 2 arguments, but only 1 given
sbrk_now = _sbrk_r(RCALL 0);
^
../../../../../../newlib-2.3.0.20160104/newlib/libc/stdlib/nano-mallocr.c:489:20:
error: ‘_sbrk_r’ undeclared (first use in this function)
sbrk_now = _sbrk_r(RCALL 0);
^
../../../../../../newlib-2.3.0.20160104/newlib/libc/stdlib/nano-mallocr.c:489:20:
note: each undeclared identifier is reported only once for each
function it appears in
Makefile:1512: recipe for target 'lib_a-nano-mallinfor.o' failed
make[8]: *** [lib_a-nano-mallinfor.o] Error 1
In case of non-reentrant system calls _sbrk_r became a macro with TWO
args (defined in reent.h):
#define _sbrk_r(__reent, __incr) sbrk(__incr)
But in our case only one argument is present. (RCALL 0) is considered
as a single argument despite RCALL itself is a macro:)
So intermediate one-arg macro will be enough to expand args before
final _sbrk_r expansion:
Corinna Vinschen [Thu, 28 Jan 2016 21:05:49 +0000 (22:05 +0100)]
Treat ACLs with extra ACEs for Admins and SYSTEM like a trivial ACL
POSIX.1e requires that chmod changes the MASK rather than the
GROUP_OBJ value if the ACL is non-trivial.
On Windows, especially on home machines, a standard ACL often
consists of entries for the user, maybe the group, and additional
entries for SYSTEM and the Administrators group. A user calling
chmod on a file with bog standard Windows perms usually expects
that chmod changes the GROUP_OBJ perms, but given the rules from
POSIX.1e we can't do that.
However, since we already treat Admins and SYSTEM special in a
ACL (they are not used in MASK computations) we go a step in the
Windows direction to follow user expectations. If an ACL only
consists of the three POSIX permissions, plus entries for Admins
and SYSTEM *only*, then we change the permissions of the GROUP_OBJ
entry *and* the MASK entry.
* fhandler_disk_file.cc (fhandler_disk_file::chmod): Drop unused
code. Add special handling for a "standard" Windows ACL. Add
comment to explain.
* sec_acl.cc (get_posix_access): Allow to return "standard-ness"
of an ACL to the caller. Add preceeding comment to explain a bit.
* security.h (get_posix_access): Align prototype.
* arm/Makefile.in: Add newlib/libc/machine/arm to the include path if
newlib is present.
* arm/arm.h: Include acle-compat.h.
(THUMB_V7_V6M): Rename to ...
(PREFER_THUMB): This. Use ACLE macros __ARM_ARCH_ISA_ARM instead of
__ARM_ARCH_6M__ to decide whether to define it.
(THUMB1_ONLY): Define for Thumb-1 only targets.
(THUMB_V7M_V6M): Rename to ...
(THUMB_VXM): This. Defined based on __ARM_ARCH_ISA_ARM, excluding
ARMv7.
* arm/crt0.S: Use THUMB1_ONLY rather than __ARM_ARCH_6M__,
!__ARM_ARCH_ISA_ARM rather than THUMB_V7M_V6M for fp enabling, and
PREFER_THUMB rather than THUMB_V7_V6M. Rename other occurences of
THUMB_V7M_V6M to THUMB_VXM.
* arm/linux-crt0.c: Likewise.
* arm/redboot-crt0.S: Likewise.
* arm/swi.h: Likewise.
* arm/trap.S: Likewise.
newlib:
* libc/machine/arm/memcpy-stub.c: Use ACLE macros __ARM_ARCH_ISA_THUMB
and __ARM_ARCH_ISA_ARM to check for Thumb-2 only targets rather than
__ARM_ARCH and __ARM_ARCH_PROFILE.
* libc/machine/arm/memcpy.S: Likewise.
* libc/machine/arm/setjmp.S: Likewise for Thumb-1 only target and
include acle-compat.h.
* libc/machine/arm/strcmp.S: Likewise for Thumb-1 and Thumb-2 only
target and include acle-compat.h.
* libc/sys/arm/arm.h: Include acle-compat.h.
(THUMB_V7_V6M): Rename to ...
(PREFER_THUMB): This. Use ACLE macro __ARM_ARCH_ISA_ARM instead of
__ARM_ARCH_6M__ to decide whether to define it.
(THUMB1_ONLY): Define for Thumb-1 only targets.
(THUMB_V7M_V6M): Rename to ...
(THUMB_VXM): This. Defined based on __ARM_ARCH_ISA_ARM, excluding
ARMv7.
* libc/sys/arm/crt0.S: Use PREFER_THUMB rather than THUMB_V7_V6M and
rename THUMB_V7M_V6M into THUMB_VXM.
* libc/sys/arm/swi.h: Likewise.
Corinna Vinschen [Wed, 27 Jan 2016 14:43:15 +0000 (15:43 +0100)]
Don't use LoadLibraryEx(..., LOAD_LIBRARY_SEARCH_SYSTEM32). It hangs
Observed running hexchat under X. For some reason the call to
LoadLibraryEx(..., LOAD_LIBRARY_SEARCH_SYSTEM32) in dll_load
hangs when trying to autoload MsgWaitForMultipleObjectsEx in
select.cc after hexchat forks to run DNS calls. Dropping the
call and just using full paths as in 2.3.1 fixes the issue.
* autoload.cc (dll_load): Drop call to LoadLibraryEx with
LOAD_LIBRARY_SEARCH_SYSTEM32 flag.
* wincap.h (wincaps::has_load_lib_search_flags): Remove.
* wincap.cc (wincaps::has_load_lib_search_flags): Drop handling
this flag.
Corinna Vinschen [Mon, 25 Jan 2016 20:00:14 +0000 (21:00 +0100)]
Make select wait state more readable
Rename "res" to "wait_state" and change its type to
select_stuff::wait_states. Use select_stuff::wait_states
values instead of ints throughout. Rearrange a few comments.
Corinna Vinschen [Thu, 21 Jan 2016 17:40:30 +0000 (18:40 +0100)]
cygpath: Avoid returning SysWOW64
On Cygwin 32 running under WOW64:
When case-correcting the path fetched with -S, the underlying
Windows function fetching the normalized path returns the real
path C:\Windows\SysWOW64 instead of the path redirection
enabled C:\Windows\System32 path. This breaks using the result
of `cygpath -S' to fetch the POSIX path of the network related
files under SYSTEMROOT\drivers\etc. This path is in fact under
the *real* C:\Windows\System32 and only mapped into the 32 bit
C:\Windows\System32 (aka C:\Windows\SysWOW64) via path redirection.
Sounds messy?
This patch checks if we're running under WOW64. If so, it
changes the path returned by GetSystemDirectoryW from "system32"
to "Sysnative". This in turn is changed to "System32" by
NtQueryInformationFile, so we're back to what we need.
Corinna Vinschen [Thu, 21 Jan 2016 17:32:16 +0000 (18:32 +0100)]
cygwin_logon_user: Return non-privileged token as well
If the calling process doesn't have sufficient privileges to
fetch the linked token of an admin-user token, cygwin_logon_user
fails. This patch changes that by returning the original,
unprivileged token of the admin user to allow authentication
and calling setuid for the current process.
Corinna Vinschen [Fri, 15 Jan 2016 14:13:11 +0000 (15:13 +0100)]
get_posix_access: Fix group deny bits leaking into file type attribute bits
* sec_acl.cc (get_posix_access): Fix bracketing in expression
constructing POSIX group permissions so as not leaking deny bits
into POSIX file type bits.
Corinna Vinschen [Tue, 12 Jan 2016 14:23:14 +0000 (15:23 +0100)]
Try loading with safe path using LOAD_LIBRARY_SEARCH_SYSTEM32 first
* autoload.cc (dll_load): Move safe loading from std_dll_init here.
Add code to handle systems supporting LOAD_LIBRARY_SEARCH flags.
Add comments to explain what the code is doing. Fix up comment
preceeding this function.
(std_dll_init): Move safe loading code to dll_load.
* wincap.h (wincaps::has_load_lib_search_flags): New element.
* wincap.cc: Implement above element throughout.
Corinna Vinschen [Mon, 11 Jan 2016 18:10:45 +0000 (19:10 +0100)]
Fix previous fix for generating unique inode numbers for sockets
* fhandler.h (fhandler_base::get_plain_ino): New inline method.
Add comment to explain what it's supposed to be used for.
* fhandler_socket.cc (fhandler_socket::get_proc_fd_name): Create
filename using inode number.
(fhandler_socket::fstat): Generate inode number from filename if
ino is not set (that's the case in a stat(2) call).
* pipe.cc: Throughout, use get_plain_ino when appropriate.
The DLLs always guaranteed to be loaded from the system dir are only
those in the KnownDLLs list. We're using some DLLs not in that list on
all supported OSes, thus we need to make sure to use full paths.
Corinna Vinschen [Mon, 11 Jan 2016 11:35:41 +0000 (12:35 +0100)]
Return unique inode numbers when calling stat/fstat on pipes and IP sockets
* fhandler.h (class fhandler_base): Convert unique_id to int64_t.
(fhandler_base::set_ino): New protected inline method.
(fhandler_base::get_unique_id): Convert to int64_t.
(fhandler_base::set_unique_id): New inline method taking int64_t.
(fhandler_pipe::fstat): Declare.
(fhandler_pipe::init): Take extra parameter.
(fhandler_pipe::create): Ditto.
* fhandler_socket.cc (fhandler_socket::init_events): Set inode number
to serial number.
(fhandler_socket::fstat): Set device to DEV_TCP_MAJOR. Create st_ino
from get_ino.
* include/cygwin/signal.h (struct _sigcommune): Replace
_si_pipe_fhandler with _si_pipe_unique_id.
* pinfo.h (_pinfo::pipe_fhandler): Take unique id instead of HANDLE.
* pinfo.cc (commune_process): Accommodate change to _si_pipe_unique_id.
(_pinfo::commune_request): Ditto.
(_pinfo::pipe_fhandler): Ditto.
* pipe.cc (fhandler_pipe::init): Take unique id as argument and set
inode number and unique_id from there.
(fhandler_pipe::open): Rework to find any matching pipe from unique
id in filename.
(fhandler_pipe::get_proc_fd_name): Create filename using inode number.
(fhandler_pipe::create): Generate and return unique id from process pid
and pipe_unique_id. In outer method, call init with additional unique
id as parameter.
(fhandler_pipe::fstat): New method.
(pipe_worker): Accommodate using 64 bit inode number in filename.
select: Don't timeout without setting descriptor arrays to all zero
* select.cc (copyfd_set): Remove.
(select): Don't copy local wait fd arrays over to returned fd arrays
since bits set there are not accounted for in return value. Zero out
returned fd arrays instead. Always call sel.poll even in case of a
timeout. Always zero out fd array when timing out. Convert while/do
to do/while for clarity. Use dedicated variable as return value to
decouple return value from artificial return code from sel.wait.
Do not treat the command line or environment like paths
* dcrt0.cc (dll_crt0_1), environ.cc (environ_init, getwinenveq,
build_env), strfuncs.cc (sys_wcstombs, sys_wcstombs_alloc),
wchar.c (sys_wcstombs, sys_wcstombs_alloc): avoid mis-conversions
of text that does not, actually, refer to a path or file name
Detailed explanation:
Our WCS -> UTF conversion handles the private Unicode page specially
to allow for otherwise invalid file names. However, this handling makes
no sense for command-lines, nor environment variables, which we would
rather convert verbatim.
As a stop-gap solution, let's just introduce a version of the
sys_wcstombs() function that specifically excludes that file name
conversion magic.
The proper solution is to change sys_wcstombs() to assume that it is not
a path that wants to be converted, and introduce sys_wcstombs_path()
that does, but that is a bigger task which we leave for another patch.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
cygpath: Try to return system directories with correct case
* cygpath.cc (do_sysfolders): Drop lame workaround to fix case of
directory returned by GetSystemDirectoryW. Try to fix case of
any path returned by this function in case it has to return a
POSIX path to support case-sensitivity.
Convert utmp{x}name to int, return useful value. Define _PATH_UTMPX
* syscalls.cc (utmpname): Convert to int. Return 0 if strdup
worked, -1 otherwise.
* include/utmpx.h (_PATH_UTMPX): Define as _PATH_UTMP.
(utmpxname): Declare as int function.
* include/sys/utmp.h (utmpname): Ditto.
* include/cygwin/version.h: Bump API minor version.
Corinna Vinschen [Fri, 25 Dec 2015 20:40:51 +0000 (21:40 +0100)]
Only request WRITE_DAC rights when writing an ACL
* fhandler_disk_file.cc (fhandler_disk_file::fchmod): Only request
query_write_dac rather than query_write_control access when writing
an ACL. Fix a comment.
Corinna Vinschen [Mon, 21 Dec 2015 17:15:57 +0000 (18:15 +0100)]
setfacl(1): Fix handling of -m and -x on a single commandline
* setfacl.cc (delace): New helper function to delete a single ACE.
(delacl): Call delace.
(modacl): Ditto, if entry is supposed to get deleted. Align comments.
Jeff Johnston [Mon, 21 Dec 2015 16:53:14 +0000 (11:53 -0500)]
Fix atexit logic to honor _ATEXIT_DYNAMIC_ALLOC setting.
If small reent is enabled (_REENT_SMALL is defined) then malloc() was
used in __register_exitproc() even if user requested it to be disabled
(_ATEXIT_DYNAMIC_ALLOC is defined). With this fix, function fails when
_ATEXIT_DYNAMIC_ALLOC is defined and whole static storage is already
used.
* libc/stdlib/on_exit_args.{c,h}: New files.
* libc/stdlib/Makefile.am: Add new source file.
* libc/stdlib/Makefile.in: Regenerate.
* libc/stdlib/__atexit.c (__register_exitproc): Initialize
_on_exit_args_ptr field of _GLOBAL_ATEXIT on first run.
* libc/stdlib/on_exit.c: Force linking of static instance of
_on_exit_args.
* libc/stdlib/cxa_atexit.c: Likewise.
Corinna Vinschen [Mon, 21 Dec 2015 12:05:41 +0000 (13:05 +0100)]
setfacl(1): Rewrite support for mask recomputation
* setfacl.cc (modacl): Move recomputing mask into new function.
(check_got_mask): New function checking if mask is in input.
(recompute_mask): New function to recompute mask.
(addmissing): Align mask computation to Linux setfacl.
(setfacl): Call check_got_mask and recompute_mask on Set, Delete and
Modify actions.
(usage): Rename --substitute to --set.
(longopts): Add --set option.
DJ Delorie [Thu, 17 Dec 2015 21:49:08 +0000 (16:49 -0500)]
Build msp430-specific libnosys
The MSP430 debuggers support I/O on hardware through CIO, so
we can use a CIO-enabled library as the "nosys" library (in
addition to the libsim library, which talks to our simulator)
* configure.in: Don't build default libnosys for msp430
* configure: Regenerate.
* msp430/Makefile: Rename libcio to libnosys.
Anton Kolesov [Tue, 15 Dec 2015 17:54:58 +0000 (20:54 +0300)]
ARC: Use new definitions for optional ARC CPU features
GCC for ARC has been updated to provide consistent naming of preprocessor
definitions for different optional architecture features:
* __ARC_BARREL_SHIFTER__ instead of __Xbarrel_shifter for
-mbarrel-shifter
* __ARC_LL64__ instead of __LL64__ for -mll64
* __ARCEM__ instead of __EM__ for -mcpu=arcem
* __ARCHS__ instead of __HS__ for -mcpu=archs
* etc (not used in newlib)
This patch updates assembly routines for ARC to use new definitions instead
of a deprecated ones. To ensure compatibility with older compiler new
definitions are also defined in asm.h if needed, based on deprecated
preprocessor definitions.
*** newlib/ChangeLog ***
2015-12-15 Anton Kolesov <Anton.Kolesov@synopsys.com>
Kevin Buettner [Tue, 15 Dec 2015 03:41:53 +0000 (20:41 -0700)]
rl78: Don't output CR when LF is encountered in write().
The file libgloss/rl78/write.c currently contains code which outputs
\r when \n is seen. The code will then output the \n as well.
This patch removes the bit of code that tests for \n and then outputs
\r.
I made this change to fix some failures in gdb.base/call-ar-st.exp. In
that test, I see two carriage returns followed by a newline. One CR is
output by the libgloss code. The other is output by the terminal driver.
The total list of failures fixed (using the default rl78 multilib) are:
Sebastian Huber [Mon, 14 Dec 2015 09:06:18 +0000 (10:06 +0100)]
FreeBSD compatibility for <sys/select.h>
* libc/include/sys/_sigset.h: New.
* libc/include/sys/select.h: Do not include <sys/types.h> and
<sys/time.h> to avoid cyclic header file dependencies. Include
specialized header files instead.
(sigset_t): Conditionally define.
* libc/include/sys/signal.h (sigset_t): Likewise.
* libc/include/sys/time.h: Include <sys/select.h> if
__BSD_VISIBLE.
* libc/include/sys/types.h: Likewise.
Fetch and store FileAllInformation rather than FileNetworkOpenInformation
* path.h (class path_conv_handle): Use FILE_ALL_INFORMATION instead of
FILE_NETWORK_OPEN_INFORMATION. Use definitions from ntdll.h since it's
included anyway.
(path_conv_handle::fai): Change name from fnoi.
(path_conv::fai): Ditto.
(file_get_fai): Change name from file_get_fnoi. Drop second parameter.
* path.cc (file_get_fai): Ditto. Fetch FileAllInformation rather than
FileNetworkOpenInformation. Convert STATUS_BUFFER_OVERFLOW to
STATUS_SUCCESS. Remove workaround to fetch FileBasicInformation and
FileStandardInformation on filesystems with broken
FileNetworkOpenInformation handling.
(symlink_info::check): Accommodate above changes. In case of using
the NtQueryDirectoryFile fallback, fetch FileIdBothDirectoryInformation
to get inode number as well.
* fhandler_disk_file.cc (fhandler_base::fstat_by_handle): Drop outdated
comment. Accommodate change to using FileAllInformation. Drop
extra function calls to fetch NumberOfLinks and IndexNumber. Set ino
directly from IndexNumber stored in pc.fai(). Drop second argument
from call to fstat_helper.
(fhandler_base::fstat_by_name): Drop second argument from call to
fstat_helper.
(fhandler_base::fstat_helper): Drop second parameter. Accommodate
the fact that we access a FILE_ALL_INFORMATION structure now.
(fhandler_base::open_fs): Set ino directly from IndexNumber stored in
pc.fai().
* fhandler.h (fhandler_base::fstat_helper): Fix declaration accrdingly.
Check for 64 FP instructions availability in libm before using them
Currently, double precision math functions in newlib/libm/machine/arm detect whether neon instructions can be used for double precision computation by checking the architecture version (>= 8) and the availability of floating-point instructions. However, these instructions would not be available if targeting fpv5-sp-d16 as the FPU, which by definition does not have 64bit float instructions. This patch adds a check that __ARM_FP advertises 64bit float instructions.
* libm/machine/arm/s_ceil.c: Also check that 64bit FP instructions are
available in the guard.
* libm/machine/arm/s_floor.c: Likewise.
* libm/machine/arm/s_nearbyint.c: Likewise.
* libm/machine/arm/s_rint.c: Likewise.
* libm/machine/arm/s_round.c: Likewise.
* libm/machine/arm/s_trunc.c: Likewise.
Fix re-initialization of FILE flags and mbstate in freopen
* libc/stdio/freopen.c (_freopen_r): Only reset __SWID bit per SUSv4.
* libc/stdio64/freopen64.c (_freopen64_r): Add missing resetting of
flag values and _mbstate.
Rename FileId to IndexNumber in NT structures to align with upstream
* ntdll.h (FILE_ID_BOTH_DIR_INFORMATION): Rename FileId to IndexNumber
to align with Microsoft naming scheme.
(FILE_INTERNAL_INFORMATION): Ditto.
* fhandler_disk_file.cc: Rename FileId to IndexNumber throughout.
* syscalls.cc: Ditto.
* mount.cc (mount_info::from_fstab_line): Support mounting the current
user's temp folder as /tmp/. This is particularly useful a feature
when Cygwin's own files are write-protected.
* pathnames.xml: document the new usertemp file system type
Detailed explanation:
In the context of Windows, there is a per-user directory for temporary
files, by default specified via the environment variable %TEMP%. Let's
allow to use that directory for our /tmp/ directory.
With this patch, we introduce the special filesystem type "usertemp":
By specifying
none /tmp usertemp binary,posix=0 0 0
in /etc/fstab, the /tmp/ directory gets auto-mounted to the directory
specified by the %TEMP% variable.
This feature comes handy in particularly in scenarios where the
administrator might want to write-protect the entire Cygwin directory
yet still needs to allow users to write into the /tmp/ directory.
This is the case in the context of Git for Windows, where the
Cygwin (MSys2) root directory lives inside C:\Program Files and hence
/tmp/ would not be writable otherwise.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Corinna Vinschen [Wed, 18 Nov 2015 21:13:31 +0000 (22:13 +0100)]
Move fd_stuff from sys/types.h to sys/select.h
* libc/include/sys/types.h: Move definitions of NBBY and howmany to
sys/param.h. Move definitions of select(2) macros to sys/select.h.
* libc/include/sys/param.h: See above.
* libc/include/sys/select.h: Move Cygwin's sys/select.h here.
* include/sys/select.h: Move select(2) macros from newlib's sys/types.h
here. Rename howmany to _howmany to unclutter namespace. Move file to
newlib.
* libc/rexex.cc: Add declaration for cygwin_gethostname.
* poll.cc: Include sys/param.h and locale select.h.
* select.h (cygwin_select): Declare.
* uname.cc: Declare cygwin_gethostname.
* winsup.h: Drop declarations of cygwin_select and cygwin_gethostname.
Always allocate main thread stack from pthread stack area on x86_64.
* dcrt0.cc: Semi-revert commit 12743c2d5d2721f3a80b4d7671a349be03c1f520.
(dll_crt0_0): Drop setting wow64_needs_stack_adjustment on 64 bit.
(_dll_crt0): Split out 64 bit code again and always create new main
thread stack, unless forked off from the non main thread in the parent.
Call create_new_main_thread_stack with parent stack commitsize if
started from the parent's main thread.
Only call child_info_fork::alloc_stack for the latter case on 64 bit.
Slightly rearrange moving rsp and rbp to new stack and document how.
Revert 32 bit wow64 handling to its former self.
* miscfunc.cc (create_new_main_thread_stack): Take a commitsize
parameter and use it if it's not 0. Don't set _main_tls here, it's
done in the caller _dll_crt0 anyway. Return stackbase - 16 bytes,
rather than stacklimit (which was very wrong anyway).
* miscfuncs.h (create_new_main_thread_stack): Accommodate declaration
to aforementioned change.
* wincap.h (wincaps::has_3264_stack_broken): Remove element.
* wincap.cc: Ditto, throughout.
* wow64.cc: Semi-revert to pre-12743c2d5d2721f3a80b4d7671a349be03c1f520
but keep architecture-agnostic type changes intact. Fix formatting.
* wow64.h: Revert to pre-12743c2d5d2721f3a80b4d7671a349be03c1f520.
Allow cygwin_conv_path(3) and cygpath(1) to emit /proc/cygdrive prefixed path
* include/sys/cygwin.h (CCP_PROC_CYGDRIVE): New flag.
* mount.cc (mount_info::cygdrive_posix_path): Take flag values rather
than just a trailing_slash_p bool. Emit /proc/cygdrive path if
CCP_PROC_CYGDRIVE flag is given.
(mount_info::conv_to_posix_path): Take flag values rather than just
a keep_rel_p bool. Rename _p variables. Print flag value as hex in
debug_printf. Call cygdrive_posix_path with flag values.
* mount.h (mount_info::cygdrive_posix_path): Accommodate above change
in declaration.
(mount_info::conv_to_posix_path): Ditto.
* fhandler_process.cc (format_process_exename): Accommodate change to
mount_info::conv_to_posix_path.
* path.cc (cygwin_conv_path): Ditto.
* cygpath.cc (absolute_flag): Initialize to CCP_RELATIVE to simplify
expressions.
(cygdrive_flag): New global flag.
(long_options): Add --proc-cygdrive option.
(options): Add -U option.
(usage): Add description for -U option.
(do_sysfolders): Or cygdrive_flag to cygwin_conv_path call.
(do_pathconv): Simply or absolute_flag to conv_func. Or
cygdrive_flag to conv_func.
(do_options): Initalize absolute_flag to CCP_RELATIVE. Initialize new
cygdrive_flag. Set absolute_flag to CCP_ABSOLUTE on -a. Set
cygdrive_flag to CCP_PROC_CYGDRIVE on -U.