Anton Kolesov [Fri, 13 May 2016 15:09:43 +0000 (18:09 +0300)]
arc: Rework default exception handlers for ARC EM and HS
Initially crt0.S used a special function, declared as weak as a default
exception handler in interrupt vector table. To let user override individual
handlers, this function had multiple names - one for each IVT entry, which,
however, was terribly confusing for the debugger and user - because it
wasn't clear which symbol will be used as a function name in debugger.
Defining multiple separate functions - one for each handler, would resolve
the mess, but would increase code size of crt0.o.
To clean this up, this patch defines exception handlers as weak symbols as
well, but those are defined as just symbols, not functions, hence there
would be less confusion over what is what. At the same time, users still can
redefine exception handlers symbol by creating functions with respective
names.
libgloss/
2016-05-24 Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc/crt0.S: Convert memory_error and friends to non-function
symbols.
Corinna Vinschen [Wed, 25 May 2016 10:13:03 +0000 (12:13 +0200)]
TZ: Replace unreliable isupper calls on wchars
In case the TZ variable is empty, Cygwin fetches timezone info from
Windows. Extracting the timezone short name uses isupper on wide chars.
Replace with explicit check for A <= character <= Z to be independent
of undefined behaviour.
Yaakov Selkowitz [Fri, 20 May 2016 21:55:50 +0000 (16:55 -0500)]
Feature test macros: ctermid, cuserid
The proper location for these functions has always been <stdio.h>, however
XPG4 and SUSv2 did mandate a duplicate declaration in <unistd.h>. cuserid
was dropped in SUSv3 (it was marked legacy since XPG4) and the ctermid
declaration in <unistd.h> was made optional and obsolete in SUSv4.
Corinna Vinschen [Fri, 20 May 2016 19:48:10 +0000 (21:48 +0200)]
mmap: Fix size restriction of maps due to using 32 bit size type
Throughout mmap, size-related variables and parameters are still using
DWORD as type, which disallows mapping ranges > 4Gigs. Fix this by
using SIZE_T throughout for those vars and parameters.
Also, drop unused off parameter from 1st variant of mmap_record::map_pages.
Corinna Vinschen [Fri, 20 May 2016 15:45:24 +0000 (17:45 +0200)]
Fix thread priority handling
So far pthread::postcreate() only sets the thread priority at all, only
if the inherit-scheduler attribute is PTHREAD_EXPLICIT_SCHED. This
completely ignores the PTHREAD_INHERIT_SCHED case, since in contrast
to POSIX, a thread does not inherit its priority from the creating
thread, but always starts with THREAD_PRIORITY_NORMAL.
pthread_getschedparam() only returns what's stored in the thread attributes,
not the actual thread priority.
Corinna Vinschen [Fri, 20 May 2016 15:38:22 +0000 (17:38 +0200)]
Rewrite scheduler functions getting and setting process and thread priority
So far the scheduler priority handling is not POSIX compatible.
The priorities use a range of -14 up to +15, which means it's not clear
if the POSIX-required return value of -1 in case of an error is *really*
an error or just the valid priority value -1. Even more confusing, -14
is the *max* value and 15 is the *min* value. Last but not least this
range doesn't match the POSIX requirement of at least 32 priority values.
This patch cleans up scheduler priority handling and moves the valid
priority range to 1 (min) - 32 (max). It also adds a function
sched_get_thread_priority() which will help to make thread priority
more POSIX-like.
John Hood [Wed, 18 May 2016 23:14:17 +0000 (19:14 -0400)]
Improve and simplify select().
* select.h: Eliminate redundant select_stuff::select_loop state.
* select.cc (select): Eliminate redundant
select_stuff::select_loop state. Eliminate redundant code for
zero timeout. Do not return early on early timer return.
(select_stuff::wait): Eliminate redundant
select_stuff::select_loop state.
John Hood [Wed, 18 May 2016 23:14:16 +0000 (19:14 -0400)]
Use high-resolution timebases for select().
* select.h: Change prototype for select_stuff::wait() for larger
microsecond timeouts.
* select.cc (pselect): Convert from old cygwin_select().
Implement microsecond timeouts.
(cygwin_select): Rewrite as a wrapper on pselect().
(select): Implement microsecond timeouts.
(select_stuff::wait): Implement microsecond timeouts with a timer
object.
Anton Kolesov [Tue, 17 May 2016 11:08:32 +0000 (14:08 +0300)]
Sync toplevel configure with upstream GCC.
This fixes a problem, where libgloss wouldn't be built for ARC processors,
even though it is actually supported. The original patch that introduced
support for ARC in libgloss hasn't been submitted to GCC maillist, hence
when top-level configure has been synced with GCC - this libgloss patch has
been effectively reverted.
Libgloss support for ARC has been accepted in GCC:
https://gcc.gnu.org/ml/gcc-patches/2016-05/msg01148.html
ChangeLog
2016-05-17 Anton Kolesov <Anton.Kolesov@synopsys.com>
* configure.ac: Sync with upstream GCC.
* configure: Regenerate.
Signed-off-by: Anton Kolesov <Anton.Kolesov@synopsys.com>
Wilco Dijkstra [Thu, 12 May 2016 16:16:58 +0000 (16:16 +0000)]
Add rawmemchr
Marcus Shawcroft wrote:
> This patch appears to have been munged by the mail system, can you
> repost as an attachment please.
Sure, I've attached the patch.
Wilco
Add a simple rawmemchr implementation. Use strlen for rawmemchr(s, '\0') as it is the
fastest way to search for '\0', and use memchr with an infinite size for other cases.
This is 3x faster for large sizes.
The reason is that the load order from sys/select.h includes sys/_types.h
before sys/config.h has been included from anywhere else. sys/_types.h
defines _fpos64_t only if __LARGE64_FILES is defined, but it never is in
this scenario. So sys/_types.h has to make sure to get the configuration
info by itself.
Jeff Johnston [Wed, 11 May 2016 21:18:48 +0000 (17:18 -0400)]
Fix strlen using Thumb-2 with -Os -marm
2016-04-18 Thomas Preud'homme <thomas.preudhomme@arm.com>
* libc/machine/arm/strlen-stub.c: Check capabilities of architecture
to decide which Thumb implementation to use and fall back to C
implementation for architecture not supporting Thumb mode.
* libc/machine/arm/strlen.S: Likewise.
Jeff Johnston [Tue, 3 May 2016 18:21:17 +0000 (14:21 -0400)]
Add _Thread_queue_Queue::_owner for RTEMS
Add _Thread_queue_Queue::_owner which will be used for the upcomming
priority inheritance implementation and an O(m) independence-preserving
protocol (OMIP) implementation.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Jeff Johnston [Tue, 3 May 2016 18:17:47 +0000 (14:17 -0400)]
Move _Thread_queue_Queue::_Lock for RTEMS
Move _Thread_queue_Queue::_Lock to begin of the structure. On RTEMS,
the presence of a lock component in the thread queue structures actually
depends on the build-time RTEMS_SMP configuration option. A move of
this part to the begin of the structure allows an implementation re-use
for the other parts.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Jeff Johnston [Mon, 2 May 2016 20:11:26 +0000 (16:11 -0400)]
Always assign return value to passed pointer in time function.
If the passed t pointer is not a null pointer, always assign the return
value to the object it points to, regardless of whether the return value
is an error.
This is what the GNU C Library does, and this is also the expected
behavior according to the latest draft of the C programming language
standard (C11 ISO/IEC 9899:201x WG14 N1570, dated 2011-04-12):
Jeff Johnston [Mon, 2 May 2016 16:04:40 +0000 (12:04 -0400)]
Fix support ARC processors without barrel-shifter
crt0.S for ARC used to use instruction "asr.f lp_count, r3, 2" for all cores
except ARC601. However instructions which shift more than 1 bit are
optional, so this crt0.S didn't worked for all ARC cores.
Luckily this is a shift just by 2 bits on all occassions, so fix is trivial
- use two single-bit shifts.
libgloss/ChangeLog
2016-04-29 Anton Kolesov <anton.kolesov@synopsys.com>
* arc/crt0.S: Fix support for processors without barrel-shifter.
Signed-off-by: Anton Kolesov <Anton.Kolesov@synopsys.com>
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
* __ARCEM__ instead of __EM__ for ARC EM cores
* __ARCHS__ instead of __HS__ for ARC HS cores
* etc (not used in libgloss)
This patch updates crt0.S for ARC to use new definitions instead of a
deprecated ones. To ensure compatibility with older compiler new definitions
are also defined in crt0.S if needed, based on presence of deprecated
preprocessor definitions.
libgloss/ChangeLog
2016-04-29 Anton Kolesov <Anton.Kolesov@synopsys.com>
* arc/crt0.S: Use new GCC defines to detect processor features.
Sebastian Huber [Thu, 21 Apr 2016 06:21:44 +0000 (08:21 +0200)]
Resurrect <machine/types.h> for <sys/types.h>
Resurrect <machine/_user_types.h> for use in <sys/types.h>. Newlib
targets may provide an own version of <machine/types.h> in their machine
directory to add custom user types for <sys/types.h>. Check the
_SYS_TYPES_H header guard to prevent a direct include of
<machine/types.h>, since the <machine/types.h> file is a Newlib
speciality.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Don't test pthread objects for being already initialized at init time
For all pthread init functions, POSIX says
Results are undefined if pthread_FOO_init() is called specifying an
already initialized pthread_FOO object.
So far our pthread init functions tested the incoming object if it's
already an initialized object and, if so, returned EBUSY. That's ok
*iff* the object was already initialized. However, as the example in
https://cygwin.com/ml/cygwin/2016-04/msg00473.html shows, an uninitialized
pthread object could also accidentally look like an initialized object
and then returning EBUSY is not ok.
Consequentially, all those tests are dangerous. Per POSIX, an application
has to know what its doing when calling any of the pthread init functions
anyway, and re-initializing the object is just as well as undefined
behaviour as is returning EBUSY on already initialized objects.
* thread.cc (pthread_attr_init): Drop check for already initialized
object.
(pthread_condattr_init): Ditto.
(pthread_rwlockattr_init): Ditto.
(pthread_mutexattr_init): Ditto.
Disallow S_ISGID on directories without default ACL entries
We can't handle the S_ISGID bit if the child didn't inherit a NULL SID
ACE with the S_ISGID bit set. On directories without default ACL
entries we would have to add an inheritable NULL SID ACE and nothing else.
This in turn results in permission problems when calling set_file_sd
from set_created_file_access. That's fixable, but it would only work
for files created from Cygwin while files created from native Windows
tools end up with really ugly permissions.
This patch only makes sure that the S_ISGID bit is reset for a directory
if it has no inheritable ACEs. Still having the 's' bit shown in ls or
getfacl output would be misleading. So, calling `setfacl -k' on a dir
also removes the S_ISGID bit now.
* sec_acl.cc (set_posix_access): Drop S_ISGID bit on directories
without inheritable ACEs. Explain why.
So far we tweaked ACL_GROUP_OBJ and ACL_MASK values the same way when
creating a file. We now do what POSIX requires, namely just change
ACL_MASK if it's present, otherwise ACL_GROUP_OBJ. Note that we only
do this at creation time. Chmod still tweaks both to create less
surprising results for the unsuspecting user.
Additionally make sure to take umask only into account if no ACL_MASK
value is present. That has been missed so far.
* sec_acl.cc (set_posix_access): Perform check for non-existant
default ACEs earlier. Ignore umask also if ACL_MASK is present.
Only set owner_eq_group if we're actually handling a user entry.
Mention chmod in a comment.
* security.cc (set_created_file_access): Perform group/mask
permission setting as required by POSIX 1003.1e.
Commit f75114fc was supposed to drop NULL SIDs in case the permissions
are simple enough not to require mask values or special POSIX bits
(S_ISVTX, etc). The check was incorrect. This patch is supposed to
fix the problem.
* sec_acl.cc (set_posix_access): Fix condition under which we
write a NULL SID.
Joel Sherrill [Wed, 13 Apr 2016 20:50:25 +0000 (15:50 -0500)]
libc/sys/rtems/crt0.c: Add <sys/lock.h> symbols required to link autoconf probes
The dummy crt0.c provided by the RTEMS target provides stubs of
symbols which would normally be provided by RTEMS. This patch adds
stubs for posix_memalign() as well as the synchronization methods
prototyped in <sys/lock.h>.
Sebastian Huber [Fri, 15 Apr 2016 11:44:28 +0000 (13:44 +0200)]
Provide __size_t via <sys/_types.h>
Various FreeBSD source and header files need a typedef __size_t via
<sys/_types.h>. Unfortunately the GCC provided <stddef.h> uses
#if (defined (__FreeBSD__) && (__FreeBSD__ >= 5)) \
|| defined(__DragonFly__) \
|| defined(__FreeBSD_kernel__)
/* __size_t is a typedef on FreeBSD 5, must not trash it. */
#elif defined (__VMS__)
/* __size_t is also a typedef on VMS. */
#else
#define __size_t
#endif
and therefore defines __size_t on Newlib targets which would trash a
__size_t typedef. Include <stddef.h> before <sys/_types.h> in
<sys/types.h> and undefine __size_t in <sys/_types.h> as a workaround.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Sebastian Huber [Fri, 15 Apr 2016 11:44:26 +0000 (13:44 +0200)]
Add __va_list to <sys/_types.h>
Add __va_list to <sys/_types.h> for BSD compatibility. In FreeBSD this
typedef is provided by the various architecture-specific
<machine/_types.h> in a copy and paste manner.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Sebastian Huber [Fri, 15 Apr 2016 11:44:12 +0000 (13:44 +0200)]
Use __machine_*_t_defined for internal types
Newlib defines defaults for internal types via <sys/_types.h> and uses
<machine/_types.h> to let targets define their own type if necessary.
Previously for example
#ifndef __dev_t_defined
typedef short __dev_t;
#endif
However, the __*_t_defined pattern conflicts with the glibc type guard
pattern for user types, e.g. dev_t in this example. Introduce a
__machine_*_t_defined pattern for internal types (defined by
<machine/_types.h>, used by <sys/_types.h>). For example
#ifndef __machine_dev_t_defined
typedef short __dev_t;
#endif
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Feature test macros overhaul: string.h and strings.h overlaps
strings.h is the header mandated for these functions in POSIX.1 prior to
2008 (when most of these were removed). The declarations in string.h are
only for BSD compatibility. But when both headers are included, avoid
duplicate declarations.
Get rid of some special cases for Cygwin in sys/types.h
Remove off_t typedef from cygwin/types.h thus relying on sys/types.h.
Introduce winsup/cygwin/machine/_types.h and move some types shared
with newlib into it. Get rid of their definition in cygwin/types.h.
Add same handling for __key_t/key_t as for the other types.
Sebastian Huber [Wed, 13 Apr 2016 11:10:15 +0000 (13:10 +0200)]
Eliminate use of Newlib-specific <machine/types.h>
This change solves a glibc/BSD compatibility problem.
glibc and BSD use double underscore types for internal types. The Linux
port of Newlib uses some glibc provided internal type definitions which
are not protected by guard defines, e.g. __off_t. To avoid a conflict
Newlib uses single underscore types for some internal types, e.g.
_off_t. However, for BSD compatibility we have to define the internal
types with double underscore names in <sys/_types.h>.
The header file <machine/types.h> is Newlib-specific. It was used
instead of <sys/_types.h> to provide the internal type definitions
_CLOCK_T, _TIME_T_, _CLOCKID_T_, _TIMER_T_, and __suseconds_t. Move
these definitions to <sys/_types.h> (there exist two instances of this
file, one for Linux and one for all other targets). This makes the
_HAVE_SYSTYPES configuration define obsolete (could possibly break the
__RDOS__ target). Use the standard <sys/_types.h> include throughout.
Move __loff_t defintion to default (non-Linux) <sys/_types.h>. Define
it via _off64_t to avoid a dependency on the compiler.
Provide the __off_t definition via default (non-Linux) <sys/_types.h>
based on _off_t for all systems except Cygwin. For Cygwin use _off64_t.
Define off_t via __off_t.
Provide the __pid_t definition via default (non-Linux) <sys/_types.h>.
This prevents a potential __pid_t and pid_t incompatibility. Add BSD
guard defines for pid_t.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
strxfrm/wcsxfrm: Always return length of the transformed string
Cygwin's strxfrm/wcsfrm treated a too short output buffer as an error
condition and always returned the size value provided as third parameter.
This is not as it's documented in POSIX.1-2008. Rather, the only error
condition is an invalid input string(*).
Other than that, the functions are supposed to return the length of the
resulting sort key, even if the output buffer is too small. In the latter
case the content of the output array is unspecified, but it's the job
of the application to check that the return value is greater or equal to
the provided buffer size.
(*) We have to make an exception in Cygwin: strxfrm has to call the
UNICODE function LCMapStringW for reasons outlined in a source comment.
If the incoming multibyte string is so large that we fail to malloc
the space required to convert it to a wchar_t string, we have to
ser errno as well since we have nothing to call LCMapStringW with.
* nlsfuncs.cc (wcsxfrm): Fix expression computing offset of
trailing wchar_t NUL. Compute correct return value even if
output buffer is too small.
(strxfrm): Handle failing malloc. Compute correct return value
even if output buffer is too small.
Sebastian Huber [Thu, 7 Apr 2016 10:58:43 +0000 (12:58 +0200)]
Move kernel dependent parts of <sys/time.h>
Move the kernel dependent parts of <sys/time.h> to new system-specific
header file <machine/_time.h>. Provide an empty default implementation.
Add a specialized implementation for RTEMS.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Sebastian Huber [Thu, 7 Apr 2016 10:58:44 +0000 (12:58 +0200)]
Drop <unistd.h> include from RTEMS <sys/param.h>
This include is not present in default Newlib, glibc and FreeBSD
<sys/param.h>. With it there is now a conflict with <sys/libkern.h>
introduced by ecf453f9635fb278cff4d4bae21a1e249313b817.
Signed-off-by: Sebastian Huber <sebastian.huber@embedded-brains.de>
Be truthful about reporting whether readahead is available
In 7346568 (Make requested console reports work, 2016-03-16), code was
introduced to report the current cursor position. It works by using a
pointer that either points to the next byte in the readahead buffer, or
to a NUL byte if the buffer is depleted, or the pointer is NULL.
These conditions are heeded in the fhandler_console::read() method, but
the condition that the pointer can point at the end of the readahead
buffer was not handled properly in the get_cons_readahead_valid()
method.
This poses a problem e.g. in Git for Windows (which uses a slightly
modified MSYS2 runtime which is in turn a slightly modified Cygwin
runtime) when vim queries the cursor position and immediately goes on to
read console input, erroneously thinking that the readahead buffer is
valid when it is already depleted instead. This condition results in an
apparent freeze that can be helped only by pressing keys repeatedly.
The full Git for Windows bug report is here:
https://github.com/git-for-windows/git/issues/711
Let's just teach the get_cons_readahead_valid() method to handle a
depleted readahead buffer correctly.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
math: guard BSD finite/isinf/isnan functions properly in math.h
Now that we have properly functioning feature test macros, the BSD
floating-point classification functions can go into math.h instead of
the non-standard ieeefp.h, and not under the C99 guard:
Note that the isinf and isnan functions (but not the variants) conflict
with functions by the same name in C++11, hence they (and only they)
need to be hidden:
As a Cygwin-specific header, there is no need to guard functions based on
capability macros. Instead, guard several blocks based on additions or
removals in later versions of POSIX.1, along with a few which are only
XSI or GNU extensions.
cygwin: make POSIX/XSI version macros dependent on feature test macros
Each version of SUS specifies a different value for _POSIX_VERSION,
_POSIX2_VERSION, and _XOPEN_VERSION. glibc also changes the value
of the other _POSIX2_ variables but not the _POSIX_* variables.
_POSIX_TIMERS should be set to a version number, not just 1.
The _POSIX_V7_* macros were missing, which was not noticed because
the V6 values were aliased in sysconf (<unistd.h>).