Thank you.
BTW i realized that Chris hasn't got an assignment from me yet.
I will send one in the next days.
Regards,
Thomas
On Wed, 5 Jun 2002, Robert Collins wrote:
> Thanks, applied.
>
> > -Original Message-
> > From: Thomas Pfaff [mailto:[EMAIL PROTECTED]]
unctions (from the standard) are not
> cancellable.
>
> Rob
>
> > -Original Message-
> > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED]] On Behalf Of Robert Collins
> > Sent: Monday, 10 June 2002 12:24 PM
> > To: 'Thomas Pfaf
Hi Rob,
i had a minor problem with your latest code:
You decided to change the mutex pointer into an object. This will break
the verifyable_object_isvalid call in pthread::create.
I have attached a small patch, a cancellation patch will follow.
Thomas
2002-06-12 Thomas Pfaff <[EM
On Wed, 12 Jun 2002, Robert Collins wrote:
>
>
> > -Original Message-
> > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED]] On Behalf Of Thomas Pfaff
> > Sent: Wednesday, 12 June 2002 9:17 PM
> > To: [EMAIL PROTECTED]
> > Subject: R
I am sorry, but i recognized that my patch was incomplete. The diff
included only threads.cc.
I have attached a new one.
Thomas
2002-06-12 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (pthread::cleanup_stack): Renamed cleanup_handlers to
cleanup_stack.
* thr
This patch provides a pthread cancellation implementation.
Rob, please mail me if you can't read the attachment.
Thomas
2002-06-25 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pthread.h (PTHREAD_CANCELED): Defined a reasonable
value.
* pthread.cc (pthread_
Changelog
2002-06-27 Thomas Pfaff <[EMAIL PROTECTED]>
*thread.cc (pthread::create): Added trace printf to get
CreateThread LastError.
diff -urp src.old/winsup/cygwin/thread.cc src/winsup/cygwin/thread.cc
--- src.old/winsup/cygwin/thread.cc Mon Jun 24 03:12:3
On Mon, 16 Sep 2002, Robert Collins wrote:
> On Sat, 2002-08-17 at 06:18, Thomas Pfaff wrote:
> >
> > Rob suggested to break the pthread_fork patch into smaller chunks. Ths is
> > part 1 providing a fork save key value handling.
> > The patch will add a linked list
you will not get
an EDEADLOCK on locking twice, it WILL deadlock instead).
With my previous 2 patches and this one i was able to build and run a
threaded perl that has passed all tests (to be true it failed 3 but these
were really not pthread related).
2002-08-15 Thomas Pfaff <[EMAIL PROTEC
On Tue, 17 Sep 2002, Robert Collins wrote:
> On Sat, 2002-08-17 at 06:32, Thomas Pfaff wrote:
> >
> > Some small fixes in the pthread key handling.
> > @@ -1020,16 +1020,27 @@ pthread_key::~pthread_key ()
> > int
> > pthread_key::set (const void *value
On Tue, 17 Sep 2002, Robert Collins wrote:
> On Tue, 2002-09-17 at 19:34, Thomas Pfaff wrote:
> >
> > This patch contains a new mutex implementation.
> >
> > The advantages are:
> >
> > - Same code on Win9x and NT. Actual are critical sections used on NT
On Fri, 20 Sep 2002, Robert Collins wrote:
> On Tue, 2002-09-17 at 19:34, Thomas Pfaff wrote:
>
> Thomas, the patch is incomplete.
>
> pthread_cond::TimedWait needs updating as well...
Yup, but it seems that this was broken on NT before i made my changes,
because it was never
On Fri, 20 Sep 2002, Robert Collins wrote:
> On Sat, 2002-08-17 at 06:55, Thomas Pfaff wrote:
> >
> > Pthread key destructor handling revised. IMHO it does not make sense to
> > handle two lists with keys, one with all keys, one with its destructors.
> > The destruct
On Sun, 22 Sep 2002, Robert Collins wrote:
> On Sat, 2002-09-21 at 01:47, Christopher Faylor wrote:
> > I haven't been following very closely. Is the reason why we are not using
> > critical sections that TryEnterCriticalSection isn't available anywhere?
> > If so, then we can probably fix tha
This patch contains an update of my mutex implementation that should apply
to current.
2002-09-23 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pthread.h: Change default mutex type to errorcheck.
* thread.cc (pthread_cond::TimedWait): Fix mutex unlock call.
(pthread
See
http://www.opengroup.org/onlinepubs/007904975/functions/pthread_key_create.html
I do not think that we should support more than one iterations at the
moment. This seems to be a rather new addition to the pthread
specification.
2002-09-23 Thomas Pfaff <[EMAIL PROTEC
2002-09-23 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (MTinterface::fixup_after_fork): Reset threadcount to
1 after fork.
diff -urp src.old/winsup/cygwin/thread.cc src/winsup/cygwin/thread.cc
--- src.old/winsup/cygwin/thread.cc Mon Sep 23 14:26:30 2002
+++ src/
On Sun, 22 Sep 2002, Robert Collins wrote:
> On Sat, 2002-09-21 at 01:47, Christopher Faylor wrote:
> > I haven't been following very closely. Is the reason why we are not using
> > critical sections that TryEnterCriticalSection isn't available anywhere?
> > If so, then we can probably fix tha
On Mon, 23 Sep 2002, Robert Collins wrote:
> On Mon, 2002-09-23 at 22:36, Thomas Pfaff wrote:
> >
> > 2002-09-23 Thomas Pfaff <[EMAIL PROTECTED]>
> >
> > * thread.cc (MTinterface::fixup_after_fork): Reset threadcount to
> > 1 after fork.
>
&
This is part 1 of 3 patches for the MTinterface.
It provides a more general implementation of Robs initMainThread that
could be used also to assign thread handles for threads that are created
outside pthread (via CreateThread).
2002-09-24 Thomas Pfaff <[EMAIL PROTECTED]>
* thr
This is patch 2 of 3 for MTinterface.
It will convert the TLS handling for reents and thread self pointer to use
the fork save pthread keys. This is important for forks from threads other
than the main thread.
2002-09-24 Thomas Pfaff <[EMAIL PROTECTED]>
* init.cc (dll_entry):
This is patch 3 of 3.
It will change thread cleanup to use a destructor function.
2002-09-24 Thomas Pfaff <[EMAIL PROTECTED]>
* init.cc (dll_entry): Run pthread key destructors on thread
detach.
* thread.cc (pthread::destructor): Implement.
(pthread
Check for valid pthread object to avoid deletion of Nullpthread.
2002-09-25 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (pthread::destructor): Check for valid pthread object
prior to delete.
diff -urp src.old/winsup/cygwin/thread.cc src/winsup/cygwin/thread.cc
--- s
This patch should fix the segfault in pthread_mutex::init by changing the
test order for a valid object and checking for valid initializer object
first..
Thomas
2002-10-17 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (verifyable_object_isvalid): Test for static object
Stripped the patch a little. Changelog remains unchanged.
Thomas
2002-10-17 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (verifyable_object_isvalid): Test for static object
first.
(pthread_mutex::init): Add test for valid initializer object.
On Thu, 17 Oc
On Thu, 17 Oct 2002, Robert Collins wrote:
> On Thu, 2002-10-17 at 18:11, Thomas Pfaff wrote:
> >
> > This patch should fix the segfault in pthread_mutex::init by changing the
> > test order for a valid object and checking for valid initializer object
> > fir
Rob,
the location for the static test is wrong. I have moved it 2 lines down.
Thomas
2002-10-18 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (verifyable_object_isvalid): Test for a valid object
pointer before testing for static ptr.
--- thread.cc.org Thu Oct 17
f
>
> http://cygwin.com/ml/cygwin/2002-10/msg01974.html
>
It seems that the fromlen in WSARecvFrom must be NULL if from is NULL and
not a pointer to 0.
Actually msg->msg_name is NULL and msg->msg_namelen is 0 which will result
in WSAEFAULT.
Thomas
2002-11-05 Thomas Pfa
read_exit). And the newly created
object never gets freed.
To avoid these errors i have made changes that will create the mainthread
object dynamic and store the reents and thread self pointer via fork safe
keys.
Thomas
2002-11-05 Thomas Pfaff <[EMAIL PROTECTED]>
* dcrt
On Tue, 5 Nov 2002, Robert Collins wrote:
> On Wed, 2002-11-06 at 00:14, Thomas Pfaff wrote:
> >
> > I have discovered some problems with the current MTinterface
> > implementation. Here are 2 test cases:
>
> > Even if the handles would be valid the pthread_join
On Tue, 5 Nov 2002, Robert Collins wrote:
> Overall this looks good. What happens to non-cygwinapi created threads
> now though? You mention you don't agree with the code, but I can't see
> (from a brief look) how you correct it.
>
> BTW: I'm currently packing to move house, so don't expect much
the nothrow new operators, i think
that the cleanest way is to define own operators. As a side effect the
cygwin1 dll can be build with an sjlj gcc and the dll is
significantly smaller.
Thomas
2002-11-26 Thomas Pfaff <[EMAIL PROTECTED]>
* cxx.cc: New file. Implement new, new[],
mailing list
called "DLLs and Exceptions" discussing the pros and cons of sjlj and
dwarf2.
Whatever exception type is preferred by the cygwin gcc package i do not
think that the cygwin1.dll should ever throw an exception.
2002-11-26 Thomas Pfaff <[EMAIL PROTECTED]>
*
Rob has added a native mutex around mutex initialization to avoid a
race. This is required in pthread_cond_init too.
I have moved the nativeMutex class outside of pthread_mutex and added
similar calls to pthread_cond to avoid that race.
Thomas
2002-11-29 Thomas Pfaff <[EMAIL PROTEC
This is my patch from 2002-09-17 ported to the recent cygwin src. (See
http://cygwin.com/ml/cygwin-patches/2002-q3/msg00436.html).
pthread_cond_init patch should be applied first.
2002-12-04 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pthread.h: Add define for errorchecking m
Changed spaces to tabs in Changelog. Sorry.
2002-12-04 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pthread.h: Add define for errorchecking mutexes.
Change default mutex type.
* thread.cc (pthread_cond::TimedWait): Update mutex unlock
calls.
(pthread
The patch has been made a little smaller and better readable by not
changing the functions order in thread.cc
2002-12-10 Thomas Pfaff <[EMAIL PROTECTED]>
* pthread.cc (pthread_cond_init): Use new pthread_cond::init.
* thread.cc: Some white spaces cleanups.
-win32 maintainer.
2002-12-11 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pthread.h: Add define for errorchecking mutexes.
Change default mutex type.
* thread.cc (pthread_cond::TimedWait): Update mutex unlock
calls.
(pthread_mutex::pthread_mutex
I have attached a bunch of mutex testcases borrowed from the
pthreads-win32 project.
Thomas
On Wed, 11 Dec 2002, Thomas Pfaff wrote:
>
>
> Rob,
>
> thank you that you are willing to add my patches.
>
> FYI, the logic of the mutex stuff is running in pthreads-win32 ab
While Chris is waiting on Rob (or vice versa) i got a new pthread related
patch which will make pthread_join and sem_wait real cancellation points.
TODO: Add testcancel to other cancelation points.
Thomas
2002-12-20 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (WAIT_CANCELED
This patch will make sleep and usleep a pthread cancellation point.
2003-01-10 Thomas Pfaff <[EMAIL PROTECTED]>
* signal.cc (sleep): Add pthread_testcancel call.
Wait for signal and cancellation event.
(usleep): Ditto.
diff -urp src.old/winsup/cygwin/signal.
This patch will make handle_sigsuspend (used by pause, sigpause
and sigsuspend) a pthread cancellation point.
2003-01-10 Thomas Pfaff <[EMAIL PROTECTED]>
* exceptions.cc (handle_sigsuspend): Add pthread_testcancel call.
Wait for signal and cancellation
Attached are test cases for the cancellation patches.
Thomas
pthreads-test.tar.gz
Description: GNU Zip compressed data
Attached is a patch and a test case for wait4 (used by wait, wait3 and
waitpid).
2003-01-13 Thomas Pfaff <[EMAIL PROTECTED]>
* wait.cc: Include thread.h
(wait4): Add pthread_testcancel call.
Wait for child process and cancellation event.
* thread.cc: Updat
Sorry, no testcase for that patch (it is really to simple).
I will commit my outstanding patches today.
Thomas
2003-01-14 Thomas Pfaff <[EMAIL PROTECTED]>
* syscalls.cc (system): Add pthread_testcancel call.
* thread.cc: Update list of cancellation points.
dif
On Tue, 14 Jan 2003, Robert Collins wrote:
> On Tue, 2003-01-14 at 21:50, Thomas Pfaff wrote:
> > Sorry, no testcase for that patch (it is really to simple).
>
> I think it really is worth adding test cases - even for simple things.
>
> It prevents regressions, which i
This patch will make sure that the signal handlers that are saved in the
system call are restored even if the thread got cancelled. Since
spawn_guts uses waitpid when mode is _P_WAIT spawn_guts is a cancellation
point.
Attached is the patch and a new test case.
2003-01-15 Thomas Paff <[EMAIL P
On Wed, 15 Jan 2003, Robert Collins wrote:
> On Wed, 2003-01-15 at 22:23, Thomas Pfaff wrote:
> > This patch will make sure that the signal handlers that are saved in the
> > system call are restored even if the thread got cancelled. Since
> > spawn_guts uses waitpid
ds_win32.
As a side effect thread.cc will shrink significantly.
2003-01-23 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (pthread_mutex::canBeUnlocked): New static method.
(pthread_cond::ExitingWait): Remove.
(pthread_cond::mutex): Ditto.
(pthread
On Thu, 23 Jan 2003, Robert Collins wrote:
> On Thu, 2003-01-23 at 23:21, Thomas Pfaff wrote:
>
>
> > 4. The spec requires that the mutex which is used with the condition
> > shall be locked by the calling thread. The current code does not check
> > this and will ad
Required for the rwlock patch.
2003-02-27 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (MTinterface::fixup_after_fork): Initialize mainthread
prior to pthread objects.
diff -urp src.old/winsup/cygwin/thread.cc src/winsup/cygwin/thread.cc
--- src.old/winsup/cygwin/thr
This patch adds support for PTHREAD_MUTEX_NORMAL (fast and
deadlocking) mutexes and slightly modifies the lock counter logic.
The counter now start at 0.
Attached are also additional testcases for the MUTEX_NORMAL type.
2002-03-27 Thomas Pfaff <[EMAIL PROTECTED]>
* include/pth
This is a slightly modified version of my still pending patch. The
modifications were made to support rwlocks.
2003-02-27 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (pthread_cond::ExitingWait): Remove.
(pthread_cond::mutex): Ditto.
(pthread_cond::cond_access):
This patch adds support for pthread_rwlocks. Attached are testcases from
the pthreads-win32 project.
The rwlock implementation follows the opengroup spec that it will
prefer writers among readers and check for deadlock conditions.
2003-02-27 Thomas Pfaff <[EMAIL PROTEC
Hi Corinna,
this is a slightly modified version of my proposed solution from
yesterday. Not tested exhaustiv but seems to work pretty well.
Thomas
2003-03-27 Thomas Pfaff <[EMAIL PROTECTED]>
* fhandler_socket.cc (fhandler_socket::connect): Add support for
an interru
Reorganize the list handling of the pthreads objects by using the List
template class and remove a lot of duplicate code.
2002-02-28 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (class List): Move inline code inside of class
declaration.
(pthread
This patch removes all wrapper functions in pthread.cc that only add an
additional function call. Export the functions in thread.cc instead.
2003-02-28 Thomas Pfaff <[EMAIL PROTECTED]>
* pthread.cc (pthread_attr_init): Remove
(pthread_attr_destroy):
On Fri, 28 Feb 2003, Robert Collins wrote:
> On Sat, 2003-03-01 at 00:53, Thomas Pfaff wrote:
> > This patch removes all wrapper functions in pthread.cc that only add an
> > additional function call. Export the functions in thread.cc instead.
>
> Please apply.
>
I
On Thu, 27 Feb 2003, Robert Collins wrote:
> On Thu, 2003-02-27 at 23:26, Thomas Pfaff wrote:
> > This patch adds support for PTHREAD_MUTEX_NORMAL (fast and
> > deadlocking) mutexes and slightly modifies the lock counter logic.
> > The counter now start at 0.
>
&g
On Thu, 13 Mar 2003, Cygwin (Robert Collins) wrote:
> This:
>
>if (1 == InterlockedIncrement ((long *)&lock_counter))
>
> is not safe. You can only check for equal to 0, less than 0, and greater
> than 0 with InterlockedIncrement | Decrement.
>
The xadd based inline interlocked functions in
On Wed, 19 Mar 2003, Robert Collins wrote:
> On Mon, 2003-03-17 at 19:17, Thomas Pfaff wrote:
> > On Thu, 13 Mar 2003, Cygwin (Robert Collins) wrote:
> >
> > > This:
> > >
> > >if (1 == InterlockedIncrement ((long *)&lock_counter))
> > &g
> template
> T& for_each(L const &start,, L const& finish, T& visitor)
> {
> for (L current (start); current != finish; ++current)
> visitor(*current);
> return visitor;
> }
>
>
> On Fri, 2003-02-28 at 22:10, Thomas Pfaff wrote:
&
On Wed, 19 Mar 2003, Robert Collins wrote:
> On Wed, 2003-03-19 at 21:15, Thomas Pfaff wrote:
>
> > > > > Secondly, IIRC lock_counter should be long, so the (long *) typecasting
> > > > > isn't needed.
> > > >
> > > > IM
2003-03-19 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (pthread_cond::Wait): Remove typecasts for unsigned
long values when calling Interlocked functions. Use new UL functions
instead.
(pthread_mutex::_Lock): Ditto.
(pthread_mutex::_TryLock):
2003-03-19 Thomas Pfaff <[EMAIL PROTECTED]>
* pthread.cc (pthread_equal): Replacement for pthread_equal in
thread.cc.
* thread.cc: Rename pthread_equal to pthread::equal throughout.
(pthread_equal): Remove.
* thread.h (pthread::equal): New static
2002-03-19 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.h (class List): Move inline code inside class
declaration.
(List::forEach): Change callback parameter to member function
pointer.
(pthread_keys::fixup_before_fork): Change to inline. Use
On Wed, 19 Mar 2003, Robert Collins wrote:
> On Thu, 2003-03-20 at 00:54, Thomas Pfaff wrote:
> > 2003-03-19 Thomas Pfaff <[EMAIL PROTECTED]>
> >
> > * pthread.cc (pthread_equal): Replacement for pthread_equal in
> > thread.cc.
> > * thr
On Wed, 19 Mar 2003, Robert Collins wrote:
> I'm happy with changing the method naming format, but is it GNU
> standard? Thats a requirement for the cygwin project.
>
> http://www.gnu.org/prep/standards_26.html#SEC26 says that
> "For example, you should use names like ignore_space_change_flag; d
Christopher Faylor wrote:
I assume that these patches are to clean up some warnings. Do they
make sense?
Yes.
I have recognized these warnings yesterday too and will apply the patch.
Thomas
2003-03-27 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc: Change 1==foo equations to foo==1 throughout.
diff -urp src.old/winsup/cygwin/thread.cc src/winsup/cygwin/thread.cc
--- src.old/winsup/cygwin/thread.cc 2003-03-27 12:56:03.0 +0100
+++ src/winsup/cygwin/thread.c
EINPROGRESS. Attached patch fix
this to return EALREADY on the second and all following connects. As a
side effect this will change interrupted blocking connects to return
EALREADY too (and not block again) as we have discussed before.
I think that this more opengroup alike.
Thomas
2003-06-06 Thomas
Christopher Faylor wrote:
On Fri, May 16, 2003 at 02:12:16PM +0200, Thomas Pfaff wrote:
Thomas Pfaff wrote:
While single threaded apps should keep run without problems (_impure_ptr
is still used for the mainthread) multithreaded apps should be recompiled
to get the full power of the thread safe
Christopher Faylor wrote:
On Fri, Jun 13, 2003 at 08:20:47AM +0200, Thomas Pfaff wrote:
Unfortunately stdin, stdout and stderr were defined with _REENT
directly. Sigh.
I've checked in this patch. I'll ask again if it makes sense
to fix this problem in newlib. Does it?
It is a littl
Christopher Faylor wrote:
Thomas, I made the change below to stop a SEGV on thread exit as evinced
by the threadidafterfork test in the testsuite.
The problem is that this code overwrites impure_ptr with the contents of
the thread which called fork, which is not the correct thing to do since
_impur
This patch suspends all threads on SIGSTOP and resumes them on SIGCONT.
The corresponding functions in the pthread class are already committed.
Thomas
2003-10-31 Thomas Pfaff <[EMAIL PROTECTED]>
* exceptions.cc (sig_handle_tty_stop): Suspend all
threads on SIGSTOP,
debbugging loop is
implemented this way.
Thomas
003-10-31 Thomas Pfaff <[EMAIL PROTECTED]>
* exceptions.cc (try_to_debug): Suspend/resume all threads when
a debugger is attaching.
(handle_exceptions): Give debugger CPU time to attach.
? suspend_all_on_stop
This time with attachment.
This patch suspends all threads on SIGSTOP and resumes them on SIGCONT.
The corresponding functions in the pthread class are already committed.
Thomas
2003-10-31 Thomas Pfaff <[EMAIL PROTECTED]>
* exceptions.cc (sig_handle_tty_stop): Suspe
Christopher Faylor wrote:
On Fri, Oct 31, 2003 at 10:17:37PM +0100, Thomas Pfaff wrote:
This time with attachment.
This patch suspends all threads on SIGSTOP and resumes them on SIGCONT.
The corresponding functions in the pthread class are already committed.
Thomas
2003-10-31 Thomas Pfaff
Christopher Faylor wrote:
On Fri, Oct 31, 2003 at 10:11:46PM +0100, Thomas Pfaff wrote:
This patch allows a debugger to attach when an exception occurs in a
thread other than the mainthread.
I am not happy about the wait in handle_exceptions, but it works on my
machine. I think that a waitloop
Christopher Faylor wrote:
On Fri, Oct 31, 2003 at 10:11:46PM +0100, Thomas Pfaff wrote:
This patch allows a debugger to attach when an exception occurs in a
thread other than the mainthread.
I am not happy about the wait in handle_exceptions, but it works on my
machine. I think that a waitloop
Attached patch fixes the memory leak reported by Arash Partow by
initializing stdio during startup and setting __sdidinit from thread
local clib appropriately.
Thomas
2003-11-10 Thomas Pfaff <[EMAIL PROTECTED]>
* dcrt0.cc: Add prototype for __sinit.
(dll_crt0_1): Init
Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 12:23:35PM +0100, Thomas Pfaff wrote:
Attached patch fixes the memory leak reported by Arash Partow by
initializing stdio during startup and setting __sdidinit from thread
local clib appropriately.
Thomas
2003-11-10 Thomas Pfaff <[EM
Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 03:03:06PM +0100, Thomas Pfaff wrote:
Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 12:23:35PM +0100, Thomas Pfaff wrote:
Attached patch fixes the memory leak reported by Arash Partow by
initializing stdio during startup and setting
Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 10:09:52AM -0500, Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 03:03:06PM +0100, Thomas Pfaff wrote:
Christopher Faylor wrote:
On Mon, Nov 10, 2003 at 12:23:35PM +0100, Thomas Pfaff wrote:
Attached patch fixes the memory leak reported by
Christopher Faylor wrote:
Actually, on poking around a little, I wonder if we should be calling
_reclaim_reent to get back all of the stuff allocated in the REENT
structure?
I think you are right.
Here is my patch to thread.cc that i will apply if there are no objections.
diff -urp src.org/thread.c
Chris,
while the cancel event creation looks good now i would make sure that
the process is created only when the handles are valid.
Thomas
2003-15-15 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (pthread::init_main_thread): Make sure that the
main thread has valid h
On Mon, 15 Dec 2003, Thomas Pfaff wrote:
> Chris,
>
> while the cancel event creation looks good now i would make sure that
> the process is created only when the handles are valid.
>
> Thomas
>
> 2003-15-15 Thomas Pfaff <[EMAIL PROTECTED]>
2003-12-15
my patch for newlib will be accepted.
Regards,
Thomas
2004-01-12 Thomas Pfaff <[EMAIL PROTECTED]>
* include/sys/_types.h: New file.
* include/sys/lock.h: Ditto.
* include/sys/stdio.h: Ditto.
* thread.cc: Include sys/lock.h
(__cygwin_lock_init_rec
Please ignore this patch, i will generate a new one soon.
Thomas
Thomas Pfaff wrote:
This patch adds support for thread safe stdio.
It will add 3 new header files three which supersedes newlib ones.
One of these (_types.h) is just a copy of newlibs _types.h with a
modified _lock_t. It is not
) ==
sizeof(void*).
2004-01-19 Thomas Pfaff <[EMAIL PROTECTED]>
* include/sys/_types.h: New file.
* include/sys/lock.h: Ditto.
* include/sys/stdio.h: Ditto.
* thread.cc: Include sys/lock.h
(__cygwin_lock_init_recursive): New fu
This is an update of my previous patch. It adds support for newlibs
__LOCK_INIT macro.
Thomas
2004-01-22 Thomas Pfaff <[EMAIL PROTECTED]>
* include/sys/_types.h: New file.
* include/sys/lock.h: Ditto.
* include/sys/stdio.h: Ditto.
* thread.cc: Include sys/
Corinna Vinschen wrote:
On Jan 22 21:48, Thomas Pfaff wrote:
This is an update of my previous patch. It adds support for newlibs
__LOCK_INIT macro.
Thomas
2004-01-22 Thomas Pfaff <[EMAIL PROTECTED]>
* include/sys/_types.h: New file.
I'm not quite sure if that's the
;
The __sinit call must be done after malloc is initialized, otherwise the
mutex creation will fail.
Thomas
2004-02-11 Thomas Pfaff <[EMAIL PROTECTED]>
* include/cygwin/_types.h: New file.
* include/sys/lock.h: Ditto.
* include/sys/stdio.h: Ditto.
* dc
earlier does no harm, AFAICS. The
> rest of the Cygwin patch should obviously wait.
> Igor
>
Igor,
include/cygwin/_types.h is part of the patch, see Changelog:
2004-02-11 Thomas Pfaff <[EMAIL PROTECTED]>
* include/cygwin/_types.h: New file.
The patch for newlibs _
Christopher Faylor wrote:
On Wed, Feb 11, 2004 at 11:08:57AM +0100, Thomas Pfaff wrote:
The __sinit call must be done after malloc is initialized, otherwise the
mutex creation will fail.
I am not comfortable with this part of the patch. I moved the __sinit
call where I did for a reason. It
initializers can be used for example to enable thread safe stdio
in newlib without moving __sinit. I really want to get this fixed.
Thomas
P.S.: I will be on winter vacation next week, no reply during that
time.
2004-02-20 Thomas Pfaff <[EMAIL PROTECTED]>
* winsup.api/pthread/mutex8e.
Christopher Faylor wrote:
This looks like it simplifies mutex handling a lot. I haven't investigated
in great detail but since you've provided some test cases, feel free to
check in.
I tried but it seems that my ssh login is no longer valid, i get:
cvs -z9 -d :ext:[EMAIL PROTECTED]:/cvs/src co src
This time i am using the non portable mutex initializers, therefore
moving __sinit is no longer needed. And i added calls to newlibs
__fp_lock_all and __fp_unlock_all at fork.
2004-03-14 Thomas Pfaff <[EMAIL PROTECTED]>
* include/cygwin/_types.h: New file.
* include/sys/
Christopher Faylor wrote:
On Fri, Mar 26, 2004 at 01:50:08AM -0500, Christopher Faylor wrote:
On Sun, Mar 14, 2004 at 08:28:02PM +0100, Thomas Pfaff wrote:
This time i am using the non portable mutex initializers, therefore
moving __sinit is no longer needed. And i added calls to newlibs
MT_INTERFACE->fixup_before_fork () should be done as the last step in
pthread::atforkprepare.
I am sorrry if the Changelog contains spaces, but i have limited internet
access at the moment (only Webmail).
2004-03-29 Thomas Pfaff <[EMAIL PROTECTED]>
* thread.cc (pthread::atfo
1 - 100 of 116 matches
Mail list logo