__STDINT_EXP() is provided by newlib but not by stdint-gcc.h. stdint-gcc.h
is used when the GCC argument -ffreestanding is used and this results in this
file not compiling.
Modern gcc's generate additional DWARF 5 debug sections, which were
still missing in our Cygwin loader script. With ld from binutils 2.37,
this results in diagnostic output when linking the Cygwin DLL...
ld: cygwin0.dll:/4: section below image base
ld: cygwin0.dll:/20: section below image base
ld: cygwin0.dll:/36: section below image base
...and the section addresses given to these sections (.debug_loclists,
.debug_rnglists, debug_line_str) will be wrong.
Fix this by adding the missing DWARF 5 sections to our linker script
template cygwin.sc.in. Add a comment in terms of the deprecated
DWARF 4 section .debug_types.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
This was used before switching to automake to allow easy tweaking
of optimization and debugging settings from the command line during
testing. Reenable.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
trying to use aggregate initialization syntax on a member of a
nameless union member failes in g++ 11.2.
Workaround this by using explicit initialization.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The GCC diagnostic ignored "-Wstringop-overflow" pragma doesn't work
as expected anymore. Use the still working expression.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
The register keyword was already deprecated with C++11, but
with C++17 it has been entirely removed.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
This patch to the libc/machine/nvptx port of newlib implements an
approximation of "clock" and provides some additional stub routines.
These changes not only reduce the number of (link) failures in the GCC
testsuite when targeting nvptx-none, but also allow the NIST scimark4
benchmark to compile and run without modification.
newlib already contains support for backends to provide their own
clock implementations via -DCLOCK_PROVIDED. That functionality is
used here to return an approximate elapsed time based on the NVidia
GPU's clock64 cycle counter. Although not great, this is better than
the current behaviour of link error from the unresolved symbol
_times_r.
The other part of the patch is to add a small number of stub functions
to nvptx's misc.c. Adding isatty, for example, resolves linking
problems in libc from the dependency in __smakebuf_r, and the sync
stub, for example, fixes the failure with GCC's
testsuite/gfortran.dg/ISO_Fortran_binding_14.f90 [which simply tests
that gfortran can call a/any C function].
newlib/
configure.host: Add -DCLOCK_PROVIDED to newlib_cflags on nvptx*.
newlib/libc/machine/nvptx
Makefile.am: Add clock.c to lib_a_SOURCES.
clock.c: New source file to implement/approximate clock().
misc.c: Add stubs for fstat, isatty, open, sync and unlink.
Revert mx parameter and mutex lock while operating the list.
Mutex was removed with 94d24160 informing that:
'Use InterlockedCompareExchangePointer to ensure race safeness
without using a mutex.'
But it does not.
Calling pthread_mutex_init and pthread_mutex_destroy from two or
more threads occasionally leads to hang in pthread_mutex_destroy.
To not change the behaviour of other cases where List_insert was called,
List_insert_nolock is added.
_strtod_l as well as the gethex function both fetch the decimal point
from the current LC_NUMERIC locale info. This pulls in _C_numeric_locale
unconditionally even on targets not supporting locales at all.
Another problem is that strtod.c and gdtoa-gethex.c are ELIX 1, while
locale information in general isn't. This leads to potential build
breakage on bare metal targets.
Fix this by setting the decimal point to "." on all targets not
defining __HAVE_LOCALE_INFO__.
While at it, const'ify the entire local decimal point info in the
affected functions.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
As outlined in the previous patch, the non-atomicity of iterating
over a directory in the NT namespace via NtQueryDirectoryObject
one entry each, results in potential duplication of directory entries.
Fix this for fhandler_procsys::readdir as well by fetching the entire
dir inside fhandler_procsys::opendir, storing it in a buffer, and just
return buffer content from fhandler_procsys::readdir.
Fixes: 43f65cdd7d ("fhandler_procsys.cc: New file.")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Due to reports on the Cygwin mailing list[1][2], it was uncovered
that a NtOpenDirectoryObject/NtQueryDirectoryObject/NtClose sequence
with NtQueryDirectoryObject iterating over the directory entries,
one entry per invocation, is not running atomically. If new entries
are inserted into the queried directory, other entries may be moved
around and then accidentally show up twice while iterating.
Change (almost) all NtQueryDirectoryObject invocations so that it gets
a really big buffer (64K) and ideally fetches all entries at once.
This appears to work atomically.
"Almost" all, because fhandler_procsys::readdir can't be easily changed.
[1] https://cygwin.com/pipermail/cygwin/2021-July/248998.html
[2] https://cygwin.com/pipermail/cygwin/2021-August/249124.html
Fixes: e9c8cb3193 ("(format_proc_partitions): Revamp loop over existing harddisks by scanning the NT native \Device object directory and looking for Harddisk entries.")
Fixes: a998dd7055 ("Implement advisory file locking.")
Fixes: 3b7cd74bfd ("(winpids::enum_processes): Fetch Cygwin processes from listing of shared cygwin object dir in the native NT namespace.")
Fixes: 0d6f2b0117 ("syscalls.cc (sync_worker): Rewrite using native NT functions.")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Commit 3434d35a64 fixed a problem when
accessing block devices via their /proc/sys/Device entries. This
changed the way stat info is generated for these devices, resulting
in identical inode numbers for all block devices under /proc/sys/Device.
This patch fixes that by faking a device number for these devices, just as
before.
Fixes: 3434d35a64 ("Cygwin: Fix access to block devices below /proc/sys.")
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
svfwscanf replaces getwc and ungetwc_r. The comments in the code talk
about avoiding file operations, but they also need to bypass the
mbtowc calls as svfwscanf operates on wchar_t, not multibyte data,
which is a more important reason here; they would not work correctly
otherwise.
The ungetwc replacement has code which uses the 3 byte FILE _ubuf
field, but if wchar_t is 32-bits, this field is not large enough to
hold even one wchar_t value. Building in this mode generates warnings
about array overflow:
In file included from ../../newlib/libc/stdio/svfiwscanf.c:35:
../../newlib/libc/stdio/vfwscanf.c: In function '_sungetwc_r.isra':
../../newlib/libc/stdio/vfwscanf.c:316:12: warning: array subscript 4294967295 is above array bounds of 'unsigned char[3]' [-Warray-bounds]
316 | fp->_p = &fp->_ubuf[sizeof (fp->_ubuf) - sizeof (wchar_t)];
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../../newlib/libc/stdio/stdio.h:46,
from ../../newlib/libc/stdio/vfwscanf.c:82,
from ../../newlib/libc/stdio/svfiwscanf.c:35:
../../newlib/libc/include/sys/reent.h:216:17: note: while referencing '_ubuf'
216 | unsigned char _ubuf[3]; /* guarantee an ungetc() buffer */
| ^~~~~
However, the vfwscanf code *never* ungets data before the start of the
scanning operation, and *always* ungets data which matches the input
at that point, so the code always hits the block which backs up over
the input data and never hits the block which uses the _ubuf field.
In addition, the svfwscanf code will always start with the unget
buffer empty, so the ungetwc replacement never needs to support an
unget buffer at all.
Simplify the code by removing support for everything other than
backing up over the input data, leaving the check to make sure it
doesn't get underflowed in case the vfscanf code has a bug in it.
Signed-off-by: Keith Packard <keithp@keithp.com>
Added function prototypes to newlib/libc/include/pthread.h
for the following Issue 8 Standard APIs:
pthread_cond_clockwait()
pthread_mutex_clocklock()
pthread_rwlock_clockrdlock()
pthread_rwlock_clockwrlock()
The current implementation does not reliably initialize t0 once.
Additionally the initialization requires two calls to _gettimeofday().
Let's sacrifice a byte to keep the initialization status
and reduce the maximum number of calls to _gettimeofday().
This has caused issues in an application that invokes clock().
The problematic situation is as follows:
1) The program calls clock() which calls _times().
2) _gettimeofday(&t0, 0) puts 0 in t0.tv_usec (because less than 1 us has
elapsed since the beginning of time).
3) _gettimeofday(&t, 0) puts 1 in t.tv_usec (since now more than 1 us has
elapsed since the beginning of time).
4) That call to clock() returns 1 (the value from step 3 minus the value in
step 2).
5) The program does a second call to clock().
6) The code above still sees 0 in t0 so it tries to update t0 again and
_gettimeofday(&t0, 0) puts 1 in t0.tv_usec.
7) The _gettimeofday(&t, 0) puts 1 in t.tv_usec (since less than 1us has
elapsed since step 3).
8) clock() returns 0 (step 7 minus step 6) and indicates that time is
moving backwards.
Signed-off-by: Christoph Muellner <cmuellner@gcc.gnu.org>
Make sure to cast to ulong all DWORD values displayed with format "%lu".
More instances are fixed here than in either my earlier unused patch or
Corinna's patch. I decided to use typedef..ulong for more compact code.
Address jturney's reported small issues:
- Remove explicit external ref for cygwin_internal() as it is already
provided by <sys/cygwin.h>.
- Leave intact ref for cygwin_dll_path[] as it is required by function(s)
in path.cc that profiler uses. Added comment to that effect.
- Delete existing main() wrapper. Rename main2() to main(). This because
profiler is now a Cygwin program and doesn't need to dynamically load
cygwin1.dll.
- Documentation issues will be addressed in a separate xml patch.
(I would have linked message-ids of Corinna's and Jon's messages for
proper theading but I no longer have their original emails and the mail
archives don't show msgids any more.)
The doc for gmondump says 1 or more FILENAME are expected, but 0 is
handled. That's an oversight. Make invocation with 0 FILENAMEs print a
one-line help message.
Reword the beginning of profiler's description doc to clarify target's
child processes are run but only optionally profiled.
A recent patch introduced new code for sig2str/str2sig.
This code does not properly exclude code that requires
SIGRTMIN/SIGRTMAX to be defined and triggers the following
compile error:
newlib/libc/signal/sig2str.c:199:8: error: 'SIGRTMIN' undeclared
newlib/libc/signal/sig2str.c:200:29: error: 'SIGRTMAX' undeclared
Let's add the missing guards.
Fixes: 2b50ec0cd2 ("libc: Fix compilation for new sig2str/str2sig implementation")
Signed-off-by: Christoph Muellner <cmuellner@gcc.gnu.org>
Added implementations for sig2str() and str2sig() in libc/signal
in order to improve POSIX compliance. Added fucntion prototypes
in libc/include/sys/signal.h.
If a service is supported as TCP and UDP service, GetAddrInfo does not
return two entries, one for TCP, one for UDP, as on Linux. Rather, it
just returns a single entry with ai_socktype and ai_protocol set to 0.
If the service only exists as TCP or UDP service, then ai->ai_socktype
is set, but ai_protocol isn't.
Fortunately we copy over the result from Windows into local storage
anyway, so this patch adds code to fix up the fields neglected by
Windows. In case ai_socktype as well as ai_protocol are 0, duplicate
the entry with valid values for ai_socktype and ai_protocol.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
If an interface is disconnected, net.cc:get_ifs tries to fetch IPv4
addresses from the registry. If it fails, it currently returns
pointers to sockaddr structs with zero address. Return a NULL pointer
instead, to signal the caller of getifaddrs that we do not have a
valid struct sockaddr.
Partially addresses: https://cygwin.com/pipermail/cygwin/2021-July/248970.html
riscv64-unknown-elf-g++-11.1.0 regression suite reports the following
failures for
$ make check-gcc-c++ RUNTESTFLAGS='dg.exp=Wstringop-overflow-6.C'
```
FAIL: g++.dg/warn/Wstringop-overflow-6.C -std=gnu++14 (test for excess errors)
FAIL: g++.dg/warn/Wstringop-overflow-6.C -std=gnu++17 (test for excess errors)
FAIL: g++.dg/warn/Wstringop-overflow-6.C -std=gnu++2a (test for excess errors)
UNSUPPORTED: g++.dg/warn/Wstringop-overflow-6.C -std=gnu++98
```
The "excess errors" being
```
output is In file included from /home/maxim/prj/riscv-upstream/install/riscv64-unknown-elf/include/wchar.h:6,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/cwchar:44,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/bits/postypes.h:40,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/iosfwd:40,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/ios:38,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/ostream:38,
from /home/maxim/prj/riscv-upstream/build/gcc-stage2/riscv64-unknown-elf/libstdc++-v3/include/iostream:39,
from /home/maxim/prj/riscv-upstream/gcc-11.1.0/gcc/testsuite/g++.dg/warn/Wstringop-overflow-6.C:6:
/home/maxim/prj/riscv-upstream/install/riscv64-unknown-elf/include/sys/reent.h:685:11: warning: unnecessary parentheses in declaration of '_sig_func' [-Wparentheses]
```
It appears to be the case that NtQueryTimer can return a negative time
remaining for an unsignalled timer. The value appears to be less than
the timer resolution.
Signed-off-by: David Allsopp <david.allsopp@metastack.com>
DWORD has different types on 32 and 64 bit. Use a common cast to
unsigned long to use %lu format for DWORD values throughout.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
cc Aldy Hernandez <aldyh@redhat.com> and Andrew MacLeod <amacleod@redhat.com>,
they are author of new VRP analysis for GCC, just to make sure I didn't
mis-understanding or mis-interpreting anything on GCC site.
GCC 11 have better value range analysis, that give GCC more confidence
to perform more aggressive optimization, but it cause scalbn/scalbnf get
wrong result.
Using scalbn to demostrate what happened on GCC 11, see comments with VRP
prefix:
```c
double scalbn (double x, int n)
{
/* VRP RESULT: n = [-INF, +INF] */
__int32_t k,hx,lx;
...
k = (hx&0x7ff00000)>>20;
/* VRP RESULT: k = [0, 2047] */
if (k==0) {
/* VRP RESULT: k = 0 */
...
k = ((hx&0x7ff00000)>>20) - 54;
if (n< -50000) return tiny*x; /*underflow*/
/* VRP RESULT: k = -54 */
}
/* VRP RESULT: k = [-54, 2047] */
if (k==0x7ff) return x+x; /* NaN or Inf */
/* VRP RESULT: k = [-54, 2046] */
k = k+n;
if (k > 0x7fe) return huge*copysign(huge,x); /* overflow */
/* VRP RESULT: k = [-INF, 2046] */
/* VRP RESULT: n = [-INF, 2100],
because k + n <= 0x7fe is false, so:
1. -INF < [-54, 2046] + n <= 0x7fe(2046) < INF
2. -INF < [-54, 2046] + n <= 2046 < INF
3. -INF < n <= 2046 - [-54, 2046] < INF
4. -INF < n <= [0, 2100] < INF
5. n = [-INF, 2100] */
if (k > 0) /* normal result */
{SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20)); return x;}
if (k <= -54) {
/* VRP OPT: Evaluate n > 50000 as true...*/
if (n > 50000) /* in case integer overflow in n+k */
return huge*copysign(huge,x); /*overflow*/
else return tiny*copysign(tiny,x); /*underflow*/
}
k += 54; /* subnormal result */
SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20));
return x*twom54;
}
```
However give the input n = INT32_MAX, k = k+n will overflow, and then we
expect got `huge*copysign(huge,x)`, but new VRP optimization think
`n > 50000` is never be true, so optimize that into `tiny*copysign(tiny,x)`.
so the solution here is to moving the overflow handle logic before `k = k + n`.
These are updates to wire into the build tree the new tools profiler and
gmondump, and to supply documentation for the tools.
The documentation for profiler and ssp now mention each other but do not
discuss their similarities or differences. That will be handled in a
future update to the "Profiling Cygwin Programs" section of the Cygwin
User's Guide, to be supplied.
This new tool was formerly part of 'profiler' but was spun out thanks to
Jon T's reasonable review comment. Gmondump is more of a debugging tool
than something users might have need for. Users would more likely use
gprof to make use of symbolic info like function names and source line
numbers.
The new tool formerly known as cygmon is renamed to 'profiler'. For the
name I considered 'ipsampler' and could not think of any others. I'm open
to a different name if any is suggested.
I decided that a discussion of the pros and cons of this profiler vs the
existing ssp should probably be in the "Profiling Cygwin Programs" section
of the Cygwin User's Guide rather than in the help for either. That
material will be supplied at some point.
CONTEXT buffers are made child-specific and thus thread-specific since
there is one profiler thread for each child program being profiled.
The SetThreadPriority() warning comment has been expanded.
chmod() works on Cygwin so the "//XXX ineffective" comment is gone.
I decided to make the "sample all executable sections" and "sample
dynamically generated code" suggestions simply expanded comments for now.
The profiler program is now a Cygwin exe rather than a native exe.
The Linux man page for cfsetspeed(3) specifies that the speed argument
must be one of the constants Bnnn (e.g., B9600) defined in termios.h.
But Linux in fact allows the speed to be the numerical baud rate
(e.g., 9600). For consistency with Linux, we now do the same.
Addresses: https://cygwin.com/pipermail/cygwin/2021-July/248887.html
- GCC will set __FLT_EVAL_METHOD__ to 16 if __fp16 supported, e.g.
cortex-a55/aarch64.
- $ aarch64-unknown-elf-gcc -v 2>&1 |grep version
gcc version 9.2.0 (GCC)
- $ aarch64-unknown-elf-gcc -E -dM -mcpu=cortex-a55 - < /dev/null |grep FLT_EVAL_METHOD
#define __FLT_EVAL_METHOD__ 16
#define __FLT_EVAL_METHOD_TS_18661_3__ 16
#define __FLT_EVAL_METHOD_C99__ 16
- The behavior of __FLT_EVAL_METHOD__ == 16 is same as
__FLT_EVAL_METHOD__ == 0 except for float16_t, but newlib didn't
support float16_t.
ISO/IEC TS 18661-3:
http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2405.pdf
V2 Changes:
- List Howland, Craig D as co-author since he provide the draft of comment
in math.h.
Co-authored-by: "Howland, Craig D" <howland@LGSInnovations.com>
The default PSAPI_VERSION is controlled by WIN32_WINNT, which we set to
0x0a00 when building utils since 48a76190 (and is the default in w32api
>= 9.0.0)
In order for the built executables to run on Windows Vista, we must also
define PSAPI_VERSION as 1 (otherwise '#define GetModuleFileNameExA
K32GetModuleFileNameExA' causes a 'The procedure entry point
K32GetModuleFilenameExA could not be located in the dynamic link library
kernel32.dll' error at run time).
Also drop uneeded psapi.h from dlfcn.cc (31ddf45d), resource.cc
(34a6eeab) and ps.cc (1def2148).
Use <cmdsynopsis> element markup in utils docbook documentation, rather
than some preformatted text inside <screen>.
(This didn't happen as part of 646745cb, when we first started using
refentry elements to make it possible to generate manpages)
This helps produce better looking manpages:
* uses bold (for command names) and italic (for replaceable items)
* different output formats inconsistently treat tabs inside <screen>
(so we have to be careful to not use them in that preformatted text)
Also clean up various issues:
* Replace '[OPTIONS]' with a real synopsis of the options
* Consistently use 'ITEM...' rather than 'ITEM1 [ITEM2...]' for an item
which should appear 1 or more times (cygcheck -f, getfacl, kill)
* Consistently document the '-h | -V' invocation form
* Since replaceable items are now marked up so they have some formatting
indicating they are replaceable, we can drop wrapping them in angle
brackets, as is done in some places
* Add missing '-W' and '-p PID' options to ps synopsis
* Adjust cygpath synopsis to show that only one 'System information'
option is allowed, possibly modified by -A
Future work:
* Sync up the actual help emitted by the util, where it's been improved
* Also don't use <screen> for formatting 'OPTIONS' section of manpage
* pldd inconsistently uses '-?' rather than '-h'!
Joel Sherrill
Corinna Vinschen
Roger Sayle
Aleksand Malikov
Keith Packard
Matt Joyce
Christoph Muellner
Mark Geisert
Christoph Muellner
Jon Turney
Ken Brown
Maxim Blinov
David Allsopp
Alex White
Kito Cheng