mono.git
8 years ago[coop] Fix the state transition for a pending suspend in reset blocking.
Rodrigo Kumpera [Tue, 1 Sep 2015 23:24:00 +0000 (19:24 -0400)]
[coop] Fix the state transition for a pending suspend in reset blocking.

The problematic scenario is the following: [1]

thread A) [DO_BLOCKING] runnable -> blocking (0)                      -- i.e, do a pinvoke
thread B) [ASYNC_SUSPEND_REQUESTED] blocking -> blocking (+1)         -- thread B tries to suspend thread A
thread A) [ABORT_BLOCKING] blocking -> async_suspend_requested (0)    -- i.e, that pinvoke called into a reverse delegate and needs to go back to runnable state
thread A) [STATE_POLL] async_suspend_requested -> self_suspended (0)  -- thread A self suspend due to the request from B

If polling witness async suspend, it must notify the initiator of a successful suspend.
If the initiator witness a thread in blocking, it bumps the suspend count, assumes it suspended and don't wait for a notification.

This means that thread A would post to the suspend semaphore but thread B (the suspend initiator) will not wait on it.

Given the initiator won't wait in this case, it makes no sense for abort blocking to put the thread in async_suspend_requested.
The solution is to put it on self_suspend_requested, which won't trigger state_poll to notify.

[1] Read it as: thread) [TRANSITION NAME] old state -> new state (suspend count delta) -- explanation

8 years ago[coop] Improve logging and add a pair of suspend related due asserts.
Rodrigo Kumpera [Tue, 1 Sep 2015 23:18:27 +0000 (19:18 -0400)]
[coop] Improve logging and add a pair of suspend related due asserts.

Add asserts to both mono_threads_begin_global_suspend and mono_threads_end_global_suspend to ensure
balanced usage of the suspend semaphore.

The suspend semaphore must end a suspend operation at zero, meaning there must an equal number of
post and wait calls.

This ensure the initiator can't jump the gun when waiting for pending suspends by consuming a
post from a previous suspend/resume op.

8 years ago[coop] Fix debug logging.
Rodrigo Kumpera [Tue, 1 Sep 2015 23:03:49 +0000 (19:03 -0400)]
[coop] Fix debug logging.

8 years ago[runtime] Remove the usage of the GC aware hash table from the tasklets code.
Zoltan Varga [Tue, 1 Sep 2015 22:37:43 +0000 (18:37 -0400)]
[runtime] Remove the usage of the GC aware hash table from the tasklets code.

8 years ago[sgen] Properly synchronize `allocated_blocks`.
Mark Probst [Tue, 1 Sep 2015 22:02:40 +0000 (15:02 -0700)]
[sgen] Properly synchronize `allocated_blocks`.

We were missing a memory barrier here.  This might fix the current
intermittent crashes on wrench.

8 years agoFix IL_SEQPOINT processing; Add names to trampolines
Neale Ferguson [Tue, 1 Sep 2015 21:35:06 +0000 (17:35 -0400)]
Fix IL_SEQPOINT processing; Add names to trampolines

8 years agoMerge pull request #2015 from alexrp/profiler-roots-v2
Alex Rønne Petersen [Tue, 1 Sep 2015 21:06:22 +0000 (23:06 +0200)]
Merge pull request #2015 from alexrp/profiler-roots-v2

[runtime] Add root type and human-readable description for all GC roots.

8 years ago[runtime] Throw an exception instead of asserting when trying to call remoting method...
Zoltan Varga [Tue, 1 Sep 2015 20:16:54 +0000 (16:16 -0400)]
[runtime] Throw an exception instead of asserting when trying to call remoting methods linked away.

8 years ago[bcl] Disable some tests which rely on fp behavior on mobile. (#33510).
Zoltan Varga [Tue, 1 Sep 2015 19:08:57 +0000 (15:08 -0400)]
[bcl] Disable some tests which rely on fp behavior on mobile. (#33510).

8 years ago[Facades] Fix parallel builds for non-general profiles
Miguel de Icaza [Tue, 1 Sep 2015 19:05:24 +0000 (15:05 -0400)]
[Facades] Fix parallel builds for non-general profiles

Joao reported that the facades build broke with 9022df44211c434c95c5e2d1f4e5f5e3979f3dd0
as it tried to compile monodroid assemblies and this attempt to compile System.IO.Pipes
which in turn needs a type not presented on Android: SafePipeHandle

This should fix that

8 years ago[ppdb] Fix support for Assembly.LoadFrom(). Fixes #33550.
Zoltan Varga [Tue, 1 Sep 2015 19:03:41 +0000 (15:03 -0400)]
[ppdb] Fix support for Assembly.LoadFrom(). Fixes #33550.

8 years agoMerge pull request #2012 from joelmartinez/mdoc-memberremoval-fix
Jonathan Pryor [Tue, 1 Sep 2015 18:57:20 +0000 (14:57 -0400)]
Merge pull request #2012 from joelmartinez/mdoc-memberremoval-fix

[mdoc] Fixes an issue with member deletion.

8 years ago[mdoc] Fixes an issue with member and type deletion.
Joel Martinez [Tue, 25 Aug 2015 19:52:13 +0000 (15:52 -0400)]
[mdoc] Fixes an issue with member and type deletion.

Specifically, when updating classic and unified assemblies, members
that were completely removed from one or both assemblies were not
being removed from the XML. This patch updates mdoc's `DeleteMember`
method to properly handle deletion in all scenarios (normal, classic, unified).

Of course, deletion will still not happen if one of two things is true:

- `--delete` is *not* passed into the `mdoc update` call and the member node
  contains existing documentation.
- `--preserve` *is* passed into the `mdoc update` invocation.

In both of those cases, the member will not be deleted, and a message stating
as such will be written as a warning to the output.

8 years agoMerge pull request #2019 from directhex/powerpc-fixes
Alex Rønne Petersen [Tue, 1 Sep 2015 16:51:20 +0000 (18:51 +0200)]
Merge pull request #2019 from directhex/powerpc-fixes

Add move_f_to_i4 and move_i4_to_f to PowerPC

8 years ago[Facades] Use Zoltan's parallel build infrastructure
Miguel de Icaza [Tue, 1 Sep 2015 16:37:19 +0000 (12:37 -0400)]
[Facades] Use Zoltan's parallel build infrastructure

8 years ago[runtime] Add root type and human-readable description for all GC roots.
Alex Rønne Petersen [Mon, 31 Aug 2015 22:28:35 +0000 (00:28 +0200)]
[runtime] Add root type and human-readable description for all GC roots.

8 years agoReverted too long to run test.
Marcos Henrich [Tue, 1 Sep 2015 10:31:24 +0000 (11:31 +0100)]
Reverted too long to run test.

1abf5baf16a8051091a3c30e11118651c334aa6c added a tests that was making
jenkins to timeout.

8 years ago[threadpool-io] Reduce the value of poll_fds_size on socket removal
Ludovic Henry [Tue, 1 Sep 2015 09:13:59 +0000 (11:13 +0200)]
[threadpool-io] Reduce the value of poll_fds_size on socket removal

8 years agoAdd move_f_to_i4 and move_i4_to_f to PowerPC
Lennart Sorensen [Tue, 1 Sep 2015 10:20:15 +0000 (11:20 +0100)]
Add move_f_to_i4 and move_i4_to_f to PowerPC

8 years ago[mcs] Remove not implemented from ambiguous static type lookup. Fixes #33527
Marek Safar [Tue, 1 Sep 2015 09:09:35 +0000 (11:09 +0200)]
[mcs] Remove not implemented from ambiguous static type lookup. Fixes #33527

8 years agoMerge pull request #2016 from tpokorra/patch-1
Marek Safar [Tue, 1 Sep 2015 07:59:53 +0000 (09:59 +0200)]
Merge pull request #2016 from tpokorra/patch-1

fixing typo appropiate in WebHeaderCollection.cs

8 years agofixing typo appropiate in WebHeaderCollection.cs
Timotheus Pokorra [Tue, 1 Sep 2015 07:56:28 +0000 (09:56 +0200)]
fixing typo appropiate in WebHeaderCollection.cs

as mentioned by FastCode in IRC

8 years agoRevert "[llvm] Disable support for nested clauses again, it still causes test failure...
Zoltan Varga [Mon, 31 Aug 2015 22:48:00 +0000 (18:48 -0400)]
Revert "[llvm] Disable support for nested clauses again, it still causes test failures like #32583."

This reverts commit d48d8e93e033f5850d2f6fdb86b1f6c35be789d7.

Revert this as it breaks the build.

8 years ago[llvm] Disable support for nested clauses again, it still causes test failures like...
Zoltan Varga [Mon, 31 Aug 2015 22:35:53 +0000 (18:35 -0400)]
[llvm] Disable support for nested clauses again, it still causes test failures like #32583.

8 years ago[threadpool-io] Remove sparse-array for poll backend
Ludovic Henry [Mon, 31 Aug 2015 14:35:08 +0000 (16:35 +0200)]
[threadpool-io] Remove sparse-array for poll backend

The sparse-array approach led to better insert/remove performance, by allowing a O(1) access, but led to memory consumption issue. The kernel would, for example, have to parse the whole array, even if it's mostly empty, and that can be an issue for large arrays.

8 years agoRemove incorrect warnings
Marek Safar [Mon, 31 Aug 2015 13:22:30 +0000 (15:22 +0200)]
Remove incorrect warnings

8 years ago[mcs] Don't report error for warnings which are not intended to be compiler warnings...
Marek Safar [Mon, 31 Aug 2015 12:14:51 +0000 (14:14 +0200)]
[mcs] Don't report error for warnings which are not intended to be compiler warnings. Fixes #33487

8 years ago[system] WebUtility from reference sources
Marek Safar [Mon, 31 Aug 2015 11:30:35 +0000 (13:30 +0200)]
[system] WebUtility from reference sources

8 years ago[runtime] Removed duplicated code
Marcos Henrich [Sun, 30 Aug 2015 11:22:56 +0000 (12:22 +0100)]
[runtime] Removed duplicated code

8 years ago[mcs] 'future' option is now 'experimental' - clarify in man page and help text.
Alex Rønne Petersen [Sun, 30 Aug 2015 19:18:13 +0000 (21:18 +0200)]
[mcs] 'future' option is now 'experimental' - clarify in man page and help text.

8 years agoRevert "Merge pull request #2008 from dlech/CustomAttributeData"
Marek Safar [Sat, 29 Aug 2015 14:24:06 +0000 (16:24 +0200)]
Revert "Merge pull request #2008 from dlech/CustomAttributeData"

This reverts commit 2775839f66faf67803efd96af7aa075888a1e09b, reversing
changes made to 9cf3500204c70d0cf69584a2c47c2c1591f198b0.

8 years agoMerge pull request #1936 from esdrubal/DotNetRelativeOrAbsolute
Marcos Henrich [Sat, 29 Aug 2015 11:24:02 +0000 (12:24 +0100)]
Merge pull request #1936 from esdrubal/DotNetRelativeOrAbsolute

[System] UriKind.RelativeOrAbsolute workaround.

8 years agoMerge pull request #2008 from dlech/CustomAttributeData
Marcos Henrich [Sat, 29 Aug 2015 09:55:43 +0000 (10:55 +0100)]
Merge pull request #2008 from dlech/CustomAttributeData

Make CustomAttributeData.AttributeType virtual

8 years agoUpdated list of available MONO_DEBUG options.
Marcos Henrich [Sat, 29 Aug 2015 09:44:18 +0000 (10:44 +0100)]
Updated list of available MONO_DEBUG options.

8 years agoMake CustomAttributeData.AttributeType virtual
David Lechner [Sat, 29 Aug 2015 03:07:53 +0000 (22:07 -0500)]
Make CustomAttributeData.AttributeType virtual

According to https://msdn.microsoft.com/en-us/library/system.reflection.customattributedata.attributetype%28v=vs.110%29.aspx, System.Reflection.CustomAttributeData.AttributeType is a virtual property.

8 years agoMerge pull request #2005 from BrzVlad/feature-concurrent-work
Vlad Brezae [Sat, 29 Aug 2015 01:04:48 +0000 (18:04 -0700)]
Merge pull request #2005 from BrzVlad/feature-concurrent-work

Concurrent sgen improvements

8 years ago[sgen] Add missing write barriers for the concurrent collector
Vlad Brezae [Thu, 27 Aug 2015 21:33:38 +0000 (14:33 -0700)]
[sgen] Add missing write barriers for the concurrent collector

8 years ago[sgen] Major allocation should not happen while we sweep
Vlad Brezae [Wed, 26 Aug 2015 20:47:14 +0000 (13:47 -0700)]
[sgen] Major allocation should not happen while we sweep

8 years ago[sgen] Enable cementing for concurrent collector
Vlad Brezae [Mon, 17 Aug 2015 22:50:35 +0000 (15:50 -0700)]
[sgen] Enable cementing for concurrent collector

8 years ago[sgen] Remove the finishing nursery collection within the concurrent collector
Vlad Brezae [Mon, 24 Aug 2015 22:35:08 +0000 (15:35 -0700)]
[sgen] Remove the finishing nursery collection within the concurrent collector

8 years ago[sgen] Clear nursery at gc if we are doing concurrent checks
Vlad Brezae [Tue, 25 Aug 2015 19:38:03 +0000 (12:38 -0700)]
[sgen] Clear nursery at gc if we are doing concurrent checks

8 years ago[sgen] Fix scanning of invalid entries within arrays
Vlad Brezae [Thu, 27 Aug 2015 23:44:50 +0000 (16:44 -0700)]
[sgen] Fix scanning of invalid entries within arrays

8 years agoMerge pull request #2007 from madewokherd/imagelock
Zoltan Varga [Fri, 28 Aug 2015 19:55:08 +0000 (15:55 -0400)]
Merge pull request #2007 from madewokherd/imagelock

Fix an unlock without a corresponding lock.

8 years agoFix an unlock without a corresponding lock.
Vincent Povirk [Fri, 28 Aug 2015 19:33:58 +0000 (14:33 -0500)]
Fix an unlock without a corresponding lock.

8eeb2cef moved the lock into an if statement.

This commit is licensed as MIT/X11.

8 years ago[runtime] Remove some dead code.
Zoltan Varga [Fri, 28 Aug 2015 19:39:25 +0000 (15:39 -0400)]
[runtime] Remove some dead code.

8 years ago[sgen] Flush after writing the protocol entry
Vlad Brezae [Fri, 14 Aug 2015 02:03:17 +0000 (19:03 -0700)]
[sgen] Flush after writing the protocol entry

8 years agoMerge pull request #2004 from esdrubal/singlestacktrace
Marcos Henrich [Fri, 28 Aug 2015 15:26:28 +0000 (16:26 +0100)]
Merge pull request #2004 from esdrubal/singlestacktrace

[corlib] Changes StackTrace.ToString to use Exception format.

8 years ago[corlib] Changes StackTrace.ToString to use Exception format.
Marcos Henrich [Fri, 28 Aug 2015 11:51:13 +0000 (12:51 +0100)]
[corlib] Changes StackTrace.ToString to use Exception format.

StackTrace.ToString format was different from Exception.ToString.
StackTrace.ToString had an identation of 3 spaces instead of 2, outputed
"line " before the line number which Exception.ToString does not, and it
did not print offsets.

This change allows StackTrace.ToString output to be used by the mono-symbolicate tool.

8 years agoTest mono-symbolicate with StackTrace.ToString.
Marcos Henrich [Fri, 28 Aug 2015 14:26:35 +0000 (15:26 +0100)]
Test mono-symbolicate with StackTrace.ToString.

We were testing mono-symbolicate tool with Exception.Tostring stack
traces but now we also test that the tool works with StackTrace.ToString.

8 years agoMerge pull request #2000 from BogdanovKirill/master
Martin Baulig [Fri, 28 Aug 2015 14:06:22 +0000 (16:06 +0200)]
Merge pull request #2000 from BogdanovKirill/master

Bug fix: reading of infinite web response can lead to connection loss

8 years agoRemove stray ifdef
Marek Habersack [Fri, 28 Aug 2015 13:18:47 +0000 (15:18 +0200)]
Remove stray ifdef

8 years agoMerge pull request #2003 from esdrubal/seq_test_fix2
Marcos Henrich [Fri, 28 Aug 2015 12:21:00 +0000 (13:21 +0100)]
Merge pull request #2003 from esdrubal/seq_test_fix2

[runtime] Fix test_op_il_seq_point in amd64.

8 years agoTarget test-local in mono-symbolicate Makefile.
Marcos Henrich [Fri, 28 Aug 2015 11:39:03 +0000 (12:39 +0100)]
Target test-local in mono-symbolicate Makefile.

mono-symbolicate tests were not called by recursive test due to not having the
target test-local in its Makefile.

8 years ago[runtime] Fix test_op_il_seq_point in amd64.
Marcos Henrich [Thu, 20 Aug 2015 10:27:02 +0000 (11:27 +0100)]
[runtime] Fix test_op_il_seq_point in amd64.

In amd64 when possible 32bit instructions and registers are usedi
instead of 64bit ones.

Using MONO_DEBUG=single-imm-size avoids 32bit optimizations thus
maintaining the native code size between compilations.

8 years ago[runtime] Added MONO_DEBUG=single-imm-size.
Marcos Henrich [Wed, 19 Aug 2015 19:05:42 +0000 (20:05 +0100)]
[runtime] Added MONO_DEBUG=single-imm-size.

In amd64 when i8const are use with values within the first 32 bits
instructions and registers used can be optimized to their 32 bit
versions.

Setting MONO_DEBUG=single-imm-size variable disables such optimizations
and guarantee that the the generated native code size is independent
of the i8const values.

8 years ago[mcs] Bump ikvm to support up to 4GB output files
Marek Safar [Fri, 28 Aug 2015 07:30:01 +0000 (09:30 +0200)]
[mcs] Bump ikvm to support up to 4GB output files

8 years ago[gsharedvt] Disable the box+brtrue optimization for gsharedvt types. Fixes #33324.
Zoltan Varga [Thu, 27 Aug 2015 20:15:01 +0000 (16:15 -0400)]
[gsharedvt] Disable the box+brtrue optimization for gsharedvt types. Fixes #33324.

8 years ago[MonoPosixHelper] Fix compilation for Android.
Jonathan Pryor [Thu, 27 Aug 2015 20:05:41 +0000 (16:05 -0400)]
[MonoPosixHelper] Fix compilation for Android.

The Android NDK decalres the `struct timeval` struct in <sys/time.h>.
Update sys-socket.c to #include <sys/time.h> to fix compilation
with the Android NDK.

8 years agoFix the build.
Zoltan Varga [Thu, 27 Aug 2015 19:52:10 +0000 (15:52 -0400)]
Fix the build.

8 years ago[x86] Fix fullaot support.
Zoltan Varga [Thu, 27 Aug 2015 19:46:32 +0000 (15:46 -0400)]
[x86] Fix fullaot support.

8 years ago[coop] Fix the public coop api.
Rodrigo Kumpera [Thu, 27 Aug 2015 19:21:40 +0000 (15:21 -0400)]
[coop] Fix the public coop api.

8 years ago[coop] Blacklist two more wrappers when generating safepoints.
Rodrigo Kumpera [Thu, 27 Aug 2015 19:10:10 +0000 (15:10 -0400)]
[coop] Blacklist two more wrappers when generating safepoints.

A GC Safepoint require the thread to be in GC unsafe mode to run properly.

Both mono_threads_finish_blocking and mono_threads_reset_blocking_start are called
in GC Safe mode, which would cause them to assert in checked-build.

mono_threads_finish_blocking is called after a pinvoke finished, which means the thread will
still be in blocking mode.

mono_threads_reset_blocking_start is called by reverse pinvoke wrappers before calling managed
code, so it will be in blocking mode too.

This looks like something checked-build should catch by setting the polling variable more frequently.

8 years ago[JIT] Prefer gdb over lldb on OSX as the later still produces useless output on my...
Rodrigo Kumpera [Thu, 27 Aug 2015 19:08:38 +0000 (15:08 -0400)]
[JIT] Prefer gdb over lldb on OSX as the later still produces useless output on my system.

8 years agoMerge pull request #1993 from steffen-kiess/posix-sockets-1
Jonathan Pryor [Thu, 27 Aug 2015 18:08:47 +0000 (14:08 -0400)]
Merge pull request #1993 from steffen-kiess/posix-sockets-1

Add socket-related functions and data structures to Mono.Posix

8 years agoMerge pull request #2002 from mono/revert-1991-seq_test_fix
Marcos Henrich [Thu, 27 Aug 2015 16:20:12 +0000 (17:20 +0100)]
Merge pull request #2002 from mono/revert-1991-seq_test_fix

Revert "[runtime] Fix test_op_il_seq_point in amd64."

8 years agoRevert "[runtime] Fix test_op_il_seq_point in amd64."
Marcos Henrich [Thu, 27 Aug 2015 16:03:27 +0000 (17:03 +0100)]
Revert "[runtime] Fix test_op_il_seq_point in amd64."

8 years agoMerge pull request #2001 from esdrubal/get_module_filename
Marcos Henrich [Thu, 27 Aug 2015 16:03:07 +0000 (17:03 +0100)]
Merge pull request #2001 from esdrubal/get_module_filename

Fixes get_module_filename.

8 years agoFixes get_module_filename.
Marcos Henrich [Thu, 27 Aug 2015 14:30:25 +0000 (15:30 +0100)]
Fixes get_module_filename.

8 years agoMerge pull request #1991 from esdrubal/seq_test_fix
Marcos Henrich [Thu, 27 Aug 2015 13:21:54 +0000 (14:21 +0100)]
Merge pull request #1991 from esdrubal/seq_test_fix

[runtime] Fix test_op_il_seq_point in amd64.

8 years ago[coop] Initial drop of the supporting public API.
Rodrigo Kumpera [Wed, 26 Aug 2015 22:22:08 +0000 (18:22 -0400)]
[coop] Initial drop of the supporting public API.

8 years ago[corlib] Fix the mobile test suite.
Rodrigo Kumpera [Wed, 26 Aug 2015 22:06:43 +0000 (18:06 -0400)]
[corlib] Fix the mobile test suite.

8 years ago[threadpool-io] Fixed the Windows backend to call the correct polling function.
triton [Wed, 26 Aug 2015 22:27:05 +0000 (23:27 +0100)]
[threadpool-io] Fixed the Windows backend to call the correct polling function.

8 years ago[sgen] Don't access LOS object size directly because it has to be masked.
Mark Probst [Wed, 26 Aug 2015 20:07:29 +0000 (13:07 -0700)]
[sgen] Don't access LOS object size directly because it has to be masked.

We use the lowest bit in the LOS object size to denote whether the object is
marked/pinned, so we can't directly use the size.  Use the getter function
instead.

8 years ago[sdb] Disallow nested invokes.
Zoltan Varga [Wed, 26 Aug 2015 19:56:42 +0000 (15:56 -0400)]
[sdb] Disallow nested invokes.

8 years agoMerge pull request #1999 from rodrmoya/master
Alex Rønne Petersen [Wed, 26 Aug 2015 19:54:04 +0000 (21:54 +0200)]
Merge pull request #1999 from rodrmoya/master

Fix format description for TYPE_CONTEXT

8 years agoFix format description for TYPE_CONTEXT
Rodrigo Moya [Wed, 26 Aug 2015 16:21:08 +0000 (18:21 +0200)]
Fix format description for TYPE_CONTEXT

8 years ago[llvm] Implement support for GC_SAFE_POINT.
Zoltan Varga [Wed, 26 Aug 2015 18:47:24 +0000 (14:47 -0400)]
[llvm] Implement support for GC_SAFE_POINT.

8 years ago[corlib] Fixes test not compiling with older NUnit.
Marcos Henrich [Wed, 26 Aug 2015 10:02:42 +0000 (11:02 +0100)]
[corlib] Fixes test not compiling with older NUnit.

8 years ago[WebConnectionStream] Fix connection loss for >2GB reads
BogdanovKirill [Tue, 18 Aug 2015 09:32:25 +0000 (14:32 +0500)]
[WebConnectionStream] Fix connection loss for >2GB reads

When HttpWebRequest is used to download infinite (ContentLength not set) response from web server, we can see that after reading about 2 GB (Int32.MaxValue) of data, totalRead variable can be equal to contentLength (which is int32.MaxValue). Then WebConnectionStream thinks that all data is received (see soft condition totalRead >= contentLength), but this is not correct.

This change is released under the MIT license.

8 years agoMerge pull request #1998 from BillSeurer/master
Rodrigo Kumpera [Tue, 25 Aug 2015 21:07:08 +0000 (17:07 -0400)]
Merge pull request #1998 from BillSeurer/master

Fix coding error with ! operator in two places in libtest.c

8 years agoFix coding error with ! operator in two places.
Bill Seurer [Tue, 25 Aug 2015 21:00:04 +0000 (16:00 -0500)]
Fix coding error with ! operator in two places.

8 years agoMerge pull request #1992 from BogdanovKirill/ChunkedReading
Martin Baulig [Tue, 25 Aug 2015 12:52:31 +0000 (14:52 +0200)]
Merge pull request #1992 from BogdanovKirill/ChunkedReading

[ChunkStream] CPU and memory leak fix when writing and reading infinitely

8 years agoMerge pull request #1996 from BillSeurer/master
Zoltan Varga [Mon, 24 Aug 2015 22:22:10 +0000 (18:22 -0400)]
Merge pull request #1996 from BillSeurer/master

[ppc] Add MonoContext and asynch callback support

8 years ago[ppc] Add MonoContext and asynch callback support
Bill Seurer [Mon, 24 Aug 2015 22:10:36 +0000 (17:10 -0500)]
[ppc] Add MonoContext and asynch callback support

8 years ago[runtime] Remove the is_mb_open field from MonoMethod, it doesn't seem to be needed...
Zoltan Varga [Mon, 24 Aug 2015 20:25:44 +0000 (16:25 -0400)]
[runtime] Remove the is_mb_open field from MonoMethod, it doesn't seem to be needed, inflated MethodBuilders are represented by MethodOnTypeBuilderInst objects.

8 years ago[sdb] Remove the code which stopped single stepping when exiting a runtime invoke...
Zoltan Varga [Mon, 24 Aug 2015 20:19:30 +0000 (16:19 -0400)]
[sdb] Remove the code which stopped single stepping when exiting a runtime invoke, it was only enabled on ios/android, and it doesn't seem to be needed anymore.

8 years ago[threadpool-io] Remove unecessary operation change
Ludovic Henry [Thu, 13 Aug 2015 19:10:15 +0000 (16:10 -0300)]
[threadpool-io] Remove unecessary operation change

These 2 operations are going to be treated the same in managed, so the change is useless

8 years ago[threadpool-io] Add log tracing
Ludovic Henry [Thu, 13 Aug 2015 19:09:31 +0000 (16:09 -0300)]
[threadpool-io] Add log tracing

8 years ago[threadpool-io] Rework backend API
Ludovic Henry [Thu, 13 Aug 2015 17:55:28 +0000 (14:55 -0300)]
[threadpool-io] Rework backend API

The main change is the fd will not be removed from the backend when there is no event to poll, but only when we close the socket. This also reduce the number of functions of the backend.

8 years ago[threadpool-io] Move the states table to the selector thread exclusively
Ludovic Henry [Wed, 12 Aug 2015 15:24:12 +0000 (12:24 -0300)]
[threadpool-io] Move the states table to the selector thread exclusively

This greatly simplify the model of the states as we do not need to worry about modifications in different threads.

8 years ago[threadpool-io] Disable AIO by default
Ludovic Henry [Thu, 13 Aug 2015 14:11:59 +0000 (11:11 -0300)]
[threadpool-io] Disable AIO by default

8 years ago[coop] Add some TODOs to the stress test suite.
Rodrigo Kumpera [Mon, 24 Aug 2015 14:51:26 +0000 (10:51 -0400)]
[coop] Add some TODOs to the stress test suite.

8 years agoAdd socket-related functions and data structures to Mono.Posix
Steffen Kieß [Mon, 24 Aug 2015 17:42:54 +0000 (19:42 +0200)]
Add socket-related functions and data structures to Mono.Posix

This commit adds those socket-related functions which do not require the
sockaddr_* structures or the msghdr structure.

8 years agoMerge pull request #1989 from esdrubal/local_tz_from_link
Rodrigo Kumpera [Mon, 24 Aug 2015 16:03:44 +0000 (12:03 -0400)]
Merge pull request #1989 from esdrubal/local_tz_from_link

Local tz id from symbolic link

8 years ago[runtime] Fix test_op_il_seq_point in amd64.
Marcos Henrich [Thu, 20 Aug 2015 10:27:02 +0000 (11:27 +0100)]
[runtime] Fix test_op_il_seq_point in amd64.

In amd64 when possible 32bit instructions and registers are usedi
instead of 64bit ones.

Using MONO_DEBUG=single-imm-size avoids 32bit optimizations thus
maintaining the native code size between compilations.

8 years ago[runtime] Added MONO_DEBUG=single-imm-size.
Marcos Henrich [Wed, 19 Aug 2015 19:05:42 +0000 (20:05 +0100)]
[runtime] Added MONO_DEBUG=single-imm-size.

In amd64 when i8const are use with values within the first 32 bits
instructions and registers used can be optimized to their 32 bit
versions.

Setting MONO_DEBUG=single-imm-size variable disables such optimizations
and guarantee that the the generated native code size is independent
of the i8const values.

8 years ago[runtime] Fix native backtraces with lldb by closing stdin.
Zoltan Varga [Sat, 22 Aug 2015 02:00:41 +0000 (22:00 -0400)]
[runtime] Fix native backtraces with lldb by closing stdin.

8 years agoFix the build.
Zoltan Varga [Fri, 21 Aug 2015 22:04:10 +0000 (18:04 -0400)]
Fix the build.

8 years ago[runtime] Implement support for dynamic methods in stack traces in a different way...
Zoltan Varga [Fri, 21 Aug 2015 21:46:55 +0000 (17:46 -0400)]
[runtime] Implement support for dynamic methods in stack traces in a different way: instead of constructing the stack trace eagerly when it contains a dynamic method, save the list of dynamic methods into the Exception object so they are kept alive.

8 years ago[runtime] Remove eager exception stack construction when it contains dynamic method...
Marek Safar [Fri, 21 Aug 2015 19:36:52 +0000 (21:36 +0200)]
[runtime] Remove eager exception stack construction when it contains dynamic method. Fixes #33064

We used to construct exception stack traces as string at throw site (for dynamic methods) but
that does not work correctly because exception stack trace can be decomposed via
new StackTrace (exception) or merged via ExceptionDispatchInfo both of them need the exception
stack in raw form and not as a string.

Further the format used by local stack builder was not same as format used by
System.Diagnostics.StackTrace.

The orignal reason for the specialization was that dynamic method can be collected
before stack is rendered which leads to stack frame to be reported as

at <unknown method> instead of
at (wrapper dynamic-method)

That should be addressed by gc-link between dynamic method and exception or
stack-frame.