mono.git
8 years agoMerge pull request #2763 from esdrubal/mono-symbolicate-standalone2
Marcos Henrich [Wed, 30 Mar 2016 07:55:15 +0000 (08:55 +0100)]
Merge pull request #2763 from esdrubal/mono-symbolicate-standalone2

Mono-symbolicate standalone

8 years agoMerge pull request #2787 from esdrubal/endpointaddress10
Marcos Henrich [Wed, 30 Mar 2016 07:37:40 +0000 (08:37 +0100)]
Merge pull request #2787 from esdrubal/endpointaddress10

[System.ServiceModel] Fixes reflection call to undefined method

8 years agoMerge pull request #2781 from alexanderkyte/inflated_method_header_leak
Rodrigo Kumpera [Wed, 30 Mar 2016 00:15:24 +0000 (17:15 -0700)]
Merge pull request #2781 from alexanderkyte/inflated_method_header_leak

[runtime] Fix leaking of inflated method headers

8 years agoMerge pull request #1718 from madewokherd/sgenthreadcleanup
monojenkins [Tue, 29 Mar 2016 20:01:05 +0000 (21:01 +0100)]
Merge pull request #1718 from madewokherd/sgenthreadcleanup

[sgen] Clean up thread pool on shutdown.

Keeping threads around can cause processes to keep running when they shouldn't, particularly on Windows when the entry point returns.

(The reason this is important for me is that Wine provides _CorExeMain, which is an entry point that does return, unlike normal compiled C programs which appear to call exit() after main() returns. A Windows process doesn't exit until ALL threads have exited, so a thread that never exits means a hung process.)

It doesn't seem like we have other GC-specific cleanup code, so I don't know if this is a good approach. Feedback on that question would be appreciated.

8 years ago[tests] Force tests to print thread dumps when they time out.
Zoltan Varga [Tue, 29 Mar 2016 18:32:00 +0000 (14:32 -0400)]
[tests] Force tests to print thread dumps when they time out.

8 years agoMerge pull request #2801 from BrzVlad/feature-conc-cementing
Vlad Brezae [Tue, 29 Mar 2016 17:45:10 +0000 (00:45 +0700)]
Merge pull request #2801 from BrzVlad/feature-conc-cementing

[sgen] Cementing for the finishing pause

8 years ago[corlib] Remove local variables inadvertently overriding class variables.
Rolf Bjarne Kvinge [Tue, 29 Mar 2016 17:04:58 +0000 (19:04 +0200)]
[corlib] Remove local variables inadvertently overriding class variables.

8 years ago[corlib] Add Console.[Input|Output]Encoding to the mobile profiles.
Rolf Bjarne Kvinge [Tue, 29 Mar 2016 16:24:19 +0000 (18:24 +0200)]
[corlib] Add Console.[Input|Output]Encoding to the mobile profiles.

8 years agoMerge pull request #2794 from xmcclure/win-aot-basic
monojenkins [Tue, 29 Mar 2016 16:31:13 +0000 (17:31 +0100)]
Merge pull request #2794 from xmcclure/win-aot-basic

Make AOT compilation on 64 bit windows produce executables

This PR contains three commits (see commit descriptions) which:

- Fix 64-bit debug windows builds in MSVC 2015
- Make it possible to use the tool-prefix aot option with mono/mini make fullaotcheck (necessary if you are in cygwin and wish to invoke `x86_64-w64-mingw32-gcc`)
- Change mini so that `--aot=full` successfully emits 64-bit windows executables.

Building 32 bit executables will generally fail because we do not follow cdecl. The 64 bit executables do not run properly when executed.

8 years agoMerge pull request #2767 from lambdageek/dev/monoerror-mono_lookup_dynamic_token
monojenkins [Tue, 29 Mar 2016 15:15:20 +0000 (16:15 +0100)]
Merge pull request #2767 from lambdageek/dev/monoerror-mono_lookup_dynamic_token

[runtime] MonoError-ize mono_loookup_dynamic_token{,_class}

Push `MonoError` through `mono_lookup_dynamic_token_class` and `mono_lookup_dynamic_token`

Also simplify the implementations of those functions so that they call `mono_reflection_lookup_dynamic_token` directly, rather than via a function pointer.  The old implementation was a holdover from when the runtime was implemented as separate libraries.

8 years ago[runtime] Use mono_object_isinst_checked for ppc and ia64
Alexander Köplinger [Tue, 29 Mar 2016 13:02:20 +0000 (15:02 +0200)]
[runtime] Use mono_object_isinst_checked for ppc and ia64

The usages were only partially changed in the move from mono_object_isinst in fbe121079035450251d4169b71b64196c9f5657c and caused build errors.

8 years agoBump referencesource
Marek Safar [Tue, 29 Mar 2016 11:24:32 +0000 (13:24 +0200)]
Bump referencesource

8 years agoMerge pull request #896 from echampet/webresource
Marek Safar [Tue, 29 Mar 2016 07:33:39 +0000 (09:33 +0200)]
Merge pull request #896 from echampet/webresource

[Fix] ReusableHashAlgorithm doesn't set the Key of the ValidationAlgorit...

8 years ago[bcl] Avoid printing debug messages to stdout in the default trace listener when...
Zoltan Varga [Tue, 29 Mar 2016 01:46:23 +0000 (21:46 -0400)]
[bcl] Avoid printing debug messages to stdout in the default trace listener when a debugger is connected. Fixes #13538.

8 years ago[amd64] Avoid passing the pc/pc offset to the throw corlib exception trampoline in...
Zoltan Varga [Tue, 29 Mar 2016 00:44:40 +0000 (20:44 -0400)]
[amd64] Avoid passing the pc/pc offset to the throw corlib exception trampoline in llvm mode, instead simply use the return address pushed by the call itself. In JIT mode, this is needed because the throw code is generated out of line so its not inside the try range, but in llvm mode, the try-range encompasses the call.

8 years ago[jit] Pass the type token index instead of the full type token to the throw corlib...
Zoltan Varga [Mon, 28 Mar 2016 22:33:53 +0000 (18:33 -0400)]
[jit] Pass the type token index instead of the full type token to the throw corlib exceptions functions on most platforms to reduce the size of the generated code.

8 years agoMerge pull request #2798 from BrzVlad/fix-sgen-timestamp
monojenkins [Mon, 28 Mar 2016 20:12:07 +0000 (21:12 +0100)]
Merge pull request #2798 from BrzVlad/fix-sgen-timestamp

[sgen] Fix sgen binary protocol timestamp

TV_ELAPSED was cast to long instead of gint64, overflowing on some platforms.

8 years ago[bcl] Avoid 'sed: net_4_x_System.ServiceModel_test.dll.excludes: No such file or...
Zoltan Varga [Mon, 28 Mar 2016 20:02:05 +0000 (16:02 -0400)]
[bcl] Avoid 'sed: net_4_x_System.ServiceModel_test.dll.excludes: No such file or directory' warnings during the build.

8 years ago[arm] In the llvm throw corlib exception trampoline, use the normal caller ip instead...
Zoltan Varga [Mon, 28 Mar 2016 19:50:35 +0000 (15:50 -0400)]
[arm] In the llvm throw corlib exception trampoline, use the normal caller ip instead of the argument passed in by the caller since that might not point directly to the call instruction. Fixes #39420.

8 years ago[runtime] Make an argument name more descriptive.
Zoltan Varga [Mon, 28 Mar 2016 19:48:01 +0000 (15:48 -0400)]
[runtime] Make an argument name more descriptive.

8 years agoMerge pull request #2797 from kumpera/android_krait_bug
Rodrigo Kumpera [Mon, 28 Mar 2016 19:08:43 +0000 (12:08 -0700)]
Merge pull request #2797 from kumpera/android_krait_bug

[android] Fix Krait CPU bug on Android.

8 years agoMerge pull request #2785 from alexanderkyte/cached_info_leak
monojenkins [Mon, 28 Mar 2016 18:56:03 +0000 (19:56 +0100)]
Merge pull request #2785 from alexanderkyte/cached_info_leak

[runtime] Free cached_info tables on cleanup

8 years ago[sgen] Cement objects for the finishing pause
Vlad Brezae [Tue, 9 Feb 2016 09:29:56 +0000 (11:29 +0200)]
[sgen] Cement objects for the finishing pause

With the concurrent collector, we used to reset cementing at the start of the finishing pause (we can uncement since we process all the remsets when the world is stopped, through the mod union entries). This processing means cardtable scan in the finishing pause. In order to reduce pause time and overhead, we extend cementing for the finishing pause. Now we don't uncement objects (during sgen_cement_reset) that are still pinned in the roots at the start of the concurrent collection. We won't mark mod_union entries for these objects and they will survive the collection in the nursery. They can become uncemented at the start of the next major collection, given that they are no longer pinned in the roots.

8 years ago[sgen] Don't scan the roots on the worker
Vlad Brezae [Tue, 9 Feb 2016 09:31:04 +0000 (11:31 +0200)]
[sgen] Don't scan the roots on the worker

We used to enqueue the jobs on the worker only for us to wait for them to finish in the same pause.

8 years ago[runtime] MonoError-ize mono_lookup_dynamic_token
Aleksey Kliger [Wed, 16 Mar 2016 00:24:12 +0000 (17:24 -0700)]
[runtime] MonoError-ize mono_lookup_dynamic_token

8 years ago[runtime] MonoError-ize mono_lookup_dynamic_token_class
Aleksey Kliger [Wed, 16 Mar 2016 00:03:35 +0000 (17:03 -0700)]
[runtime] MonoError-ize mono_lookup_dynamic_token_class

It just calls mono_reflection_lookup_dynamic_token directly.

8 years ago[runtime] Remove MonoLookupDynamicToken indirection
Aleksey Kliger [Tue, 15 Mar 2016 23:14:07 +0000 (16:14 -0700)]
[runtime] Remove MonoLookupDynamicToken indirection

Call mono_reflection_lookup_dynamic_token directly from
mono_lookup_dynamic_token.

Delete mono_install_lookup_dynamic_token () and the
MonoLookupDynamicToken typedef.

8 years ago[reflection] MonoError-ize mono_reflection_lookup_dynamic_token
Aleksey Kliger [Tue, 15 Mar 2016 22:46:43 +0000 (15:46 -0700)]
[reflection] MonoError-ize mono_reflection_lookup_dynamic_token

Since it gets installed as a callback, update the MonoLookupDynamicToken typedef.

8 years ago[SRE] Fix compile warning in encode_cattr_value
Aleksey Kliger [Mon, 28 Mar 2016 16:30:16 +0000 (12:30 -0400)]
[SRE] Fix compile warning in encode_cattr_value

Landed mono_object_isinst_checked PR in a way that broke encode_cattr_value.

8 years ago[sgen] Fix sgen binary protocol timestamp
Vlad Brezae [Fri, 18 Mar 2016 09:05:14 +0000 (16:05 +0700)]
[sgen] Fix sgen binary protocol timestamp

TV_ELAPSED was cast to long instead of gint64, overflowing on some platforms.

8 years agoMerge pull request #2621 from BrzVlad/feature-conc-preclean
Vlad Brezae [Mon, 28 Mar 2016 16:26:08 +0000 (23:26 +0700)]
Merge pull request #2621 from BrzVlad/feature-conc-preclean

[sgen] Concurrent precleaning of card table

8 years agoMerge pull request #2795 from lambdageek/dev/marshal-no-raise
Aleksey Kliger (λgeek) [Mon, 28 Mar 2016 15:27:36 +0000 (11:27 -0400)]
Merge pull request #2795 from lambdageek/dev/marshal-no-raise

[marshal] Don't raise an exception, just assert

8 years agoMerge pull request #2770 from lambdageek/dev/monoerror-object_isinst_mbyref
monojenkins [Mon, 28 Mar 2016 14:45:22 +0000 (15:45 +0100)]
Merge pull request #2770 from lambdageek/dev/monoerror-object_isinst_mbyref

[runtime] MonoError-ize mono_object_isinst{,_mbyref}

Also mark them external only.  Runtime should use mono_object_isinst_mbyref_checked and mono_object_isinst_checked

8 years ago[aot] Emit a .subsections_via_symbols directive on arm.
Zoltan Varga [Fri, 25 Mar 2016 22:35:10 +0000 (18:35 -0400)]
[aot] Emit a .subsections_via_symbols directive on arm.

8 years ago[aot] Fix a memory leak.
Zoltan Varga [Sun, 27 Mar 2016 22:49:57 +0000 (18:49 -0400)]
[aot] Fix a memory leak.

8 years ago[runtime] Avoid setting klass->instance_size if its already set, this can happen...
Zoltan Varga [Sun, 27 Mar 2016 22:43:29 +0000 (18:43 -0400)]
[runtime] Avoid setting klass->instance_size if its already set, this can happen when loading cached class info from an AOT image. This could lead to random crashes/asserts since klass->instance_size is directly modified while it is computed.

8 years ago[runtime] Fix a memory leak in ves_icall_Type_GetNestedTypes ().
Zoltan Varga [Sun, 27 Mar 2016 22:41:11 +0000 (18:41 -0400)]
[runtime] Fix a memory leak in ves_icall_Type_GetNestedTypes ().

8 years agoMerge pull request #2779 from akoeplinger/fix-socket-closed
Alexander Köplinger [Sun, 27 Mar 2016 15:33:48 +0000 (17:33 +0200)]
Merge pull request #2779 from akoeplinger/fix-socket-closed

[System] Use Monitor.Wait/Pulse in SafeSocketHandle instead of Thread.Sleep

8 years ago[System] Use Monitor.Wait/Pulse in SafeSocketHandle instead of Thread.Sleep
Alexander Köplinger [Mon, 21 Mar 2016 15:27:19 +0000 (16:27 +0100)]
[System] Use Monitor.Wait/Pulse in SafeSocketHandle instead of Thread.Sleep

We are quite frequently seeing an error on the ARM workers on Jenkins:

```
Test Case Failures:
1) MonoTests.System.Net.Sockets.SocketTest.SendAsyncFile : System.Exception : Could not abort registered blocking threads before closing socket.
at System.Net.Sockets.SafeSocketHandle.ReleaseHandle () [0x000be] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/System/System.Net.Sockets/SafeSocketHandle.cs:77
at System.Runtime.InteropServices.SafeHandle.DangerousReleaseInternal (Boolean dispose) <0x42cb0050 + 0x0020b> in <filename unknown>:0
at System.Runtime.InteropServices.SafeHandle.InternalDispose () <0x446ba758 + 0x00023> in <filename unknown>:0
at System.Runtime.InteropServices.SafeHandle.Dispose (Boolean disposing) [0x00000] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/external/referencesource/mscorlib/system/runtime/remoting/callcontext.cs:804
at System.Runtime.InteropServices.SafeHandle.Dispose () [0x00000] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/external/referencesource/mscorlib/system/runtime/remoting/callcontext.cs:795
at System.Net.Sockets.Socket.Dispose (Boolean disposing) [0x00047] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/System/System.Net.Sockets/Socket.cs:3321
at System.Net.Sockets.Socket.Dispose () [0x00000] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/System/System.Net.Sockets/Socket.cs:3327
at System.Net.Sockets.Socket.Close () [0x00007] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/System/System.Net.Sockets/Socket.cs:3256
at MonoTests.System.Net.Sockets.SocketTest.SendAsyncFile () [0x000e7] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/System/Test/System.Net.Sockets/SocketTest.cs:4331
at (wrapper managed-to-native) System.Reflection.MonoMethod:InternalInvoke (System.Reflection.MonoMethod,object,object[],System.Exception&)
at System.Reflection.MonoMethod.Invoke (System.Object obj, BindingFlags invokeAttr, System.Reflection.Binder binder, System.Object[] parameters, System.Globalization.CultureInfo culture) [0x0003d] in /media/ssd/jenkins/workspace/test-mono-mainline/label/debian-armhf/mcs/class/corlib/System.Reflection/StrongNameKeyPair.cs:103
```

Let's see if implementing the suggestion from https://github.com/mono/mono/pull/1612/files#r26393224
fixes this. I couldn't reproduce the test failure even on the same ARM machine so this is a bit of
an opportunistic change, but it feels like the more "correct" solution anyway.

I decided to inline AbortRegisteredThreads as at the only place we're calling it we already know
blocking_threads can't be null and acquired the lock.

8 years agoMerge pull request #2789 from lambdageek/dev/monoerror-threading
monojenkins [Sat, 26 Mar 2016 01:41:10 +0000 (01:41 +0000)]
Merge pull request #2789 from lambdageek/dev/monoerror-threading

[threadpool-ms] MonoError-ize

1. Convert icalls that throw `NotImplementedException` to do it via mono_error_set_pending_exception
2. Add MonoError outargs to `mono_threadpool_ms_enqueue_work_item` and `mono_threadpool_ms_begin_invoke`

8 years ago[runtime] MonoError-ize mono_object_isinst
Aleksey Kliger [Thu, 17 Mar 2016 21:41:20 +0000 (14:41 -0700)]
[runtime] MonoError-ize mono_object_isinst

Mark it external only. Runtime should use mono_object_isinst_checked.

8 years ago[threadpool-ms] MonoError-ize mono_threadpool_ms_begin_invoke
Aleksey Kliger [Wed, 23 Mar 2016 18:06:01 +0000 (14:06 -0400)]
[threadpool-ms] MonoError-ize mono_threadpool_ms_begin_invoke

8 years agoMerge pull request #2748 from lambdageek/dev/monoerror-mono_value_box
monojenkins [Fri, 25 Mar 2016 20:55:38 +0000 (20:55 +0000)]
Merge pull request #2748 from lambdageek/dev/monoerror-mono_value_box

[runtime] Make mono_value_box use MonoError

Mark it external only.  Runtime should use mono_value_box_checked.

8 years ago[runtime] MonoError-ize mono_object_isinst_mbyref
Aleksey Kliger [Wed, 16 Mar 2016 23:42:30 +0000 (16:42 -0700)]
[runtime] MonoError-ize mono_object_isinst_mbyref

Also mark it external only.  Runtime should use mono_object_isinst_mbyref_checked

8 years ago[MonoError] change mono_error_set_pending_exception to return a bool
Aleksey Kliger [Fri, 25 Mar 2016 19:23:19 +0000 (15:23 -0400)]
[MonoError] change mono_error_set_pending_exception to return a bool

Return TRUE if pending exception was set, FALSE otherwise.

The idea is that you can write:
  if (mono_error_set_pending_exception (error)) {
    /* cleanup */
    return;
  }

8 years ago[android] Fix Krait CPU bug on Android.
Rodrigo Kumpera [Mon, 16 Nov 2015 05:22:56 +0000 (00:22 -0500)]
[android] Fix Krait CPU bug on Android.

Krait cpus don't ignore thumb ITT state in ARM mode (it's unspecified CPU behavior).
The linux kernel then didn't bother resetting that state when landing an ARM signal handler.

The end result was that Krait cpus would skip the first few instructions of the signal handler
and things would explode in the most amazing possible ways.

Patch by Mark Probst <mark.probst@gmail.com>.

8 years ago[marshal] Don't raise an exception, just assert
Aleksey Kliger [Fri, 25 Mar 2016 17:29:03 +0000 (13:29 -0400)]
[marshal] Don't raise an exception, just assert

Don't raise deep in the marshalling code, just assert.

(Raising means we'd have to convert the code to use MonoError and then
thread it out through all of the marshalling code resulting in a lot of
code churn)

8 years ago[runtime] MonoError-ize mono_nullable_box
Aleksey Kliger [Thu, 10 Mar 2016 19:23:43 +0000 (14:23 -0500)]
[runtime] MonoError-ize mono_nullable_box

8 years ago[runtime] Make mono_value_box use MonoError
Aleksey Kliger [Thu, 10 Mar 2016 15:35:04 +0000 (10:35 -0500)]
[runtime] Make mono_value_box use MonoError

Mark it external only.  Runtime should use mono_value_box_checked.

8 years ago[threadpool-ms] MonoError-ize mono_threadpool_ms_enqueue_work_item
Aleksey Kliger [Wed, 23 Mar 2016 16:56:01 +0000 (12:56 -0400)]
[threadpool-ms] MonoError-ize mono_threadpool_ms_enqueue_work_item

Take a MonoError outarg; return boolean on success/failure.

8 years ago[threadpool-ms] MonoError-ize some not implemented icalls.
Aleksey Kliger [Wed, 23 Mar 2016 16:49:51 +0000 (12:49 -0400)]
[threadpool-ms] MonoError-ize some not implemented icalls.

1. ves_icall_System_Threading_ThreadPool_ReportThreadStatus
2. ves_icall_System_Threading_ThreadPool_PostQueuedCompletionStatus

8 years agoMerge pull request #2791 from lambdageek/dev/monoerror-cleanup-deadcode
Aleksey Kliger (λgeek) [Fri, 25 Mar 2016 15:33:37 +0000 (11:33 -0400)]
Merge pull request #2791 from lambdageek/dev/monoerror-cleanup-deadcode

[runtime] Remove dead code and don't throw in embedding API functions

8 years ago[sgen] Add option for enabling/disabling precleaning
Vlad Brezae [Tue, 8 Mar 2016 02:41:20 +0000 (04:41 +0200)]
[sgen] Add option for enabling/disabling precleaning

8 years ago[sgen] Preclean mod union entries before the finishing pause
Vlad Brezae [Tue, 19 Jan 2016 09:21:39 +0000 (11:21 +0200)]
[sgen] Preclean mod union entries before the finishing pause

When the concurrent mark ends, we have another concurrent phase where we scan the cardtable trying to clean as many cards as possible. The way we achieve this is by first copying all the mod union cards (for each block/los_obj) to a separate temporary card_table. When scanning we will only scan these cards. Therefore we will first preclean these cards in the actual mod_union card_tables, followed by a full memory barrier (which assures us that, when scanning, we will observe the changes that led to the corresponding card marks) and then by a normal mod_union scan, which can remark some of these cards, without any races with the preclean phase.

8 years ago[sgen] Move assert to the right place
Vlad Brezae [Tue, 8 Mar 2016 02:42:41 +0000 (04:42 +0200)]
[sgen] Move assert to the right place

8 years ago[runtime] Enclose define parameters in parentheses
Vlad Brezae [Mon, 7 Mar 2016 16:36:28 +0000 (18:36 +0200)]
[runtime] Enclose define parameters in parentheses

8 years ago[sgen] Add separate scan function for handling a single pointer entry
Vlad Brezae [Sun, 7 Feb 2016 23:08:24 +0000 (01:08 +0200)]
[sgen] Add separate scan function for handling a single pointer entry

This way we avoid duplicating cardtable marking code, making sure we have consistent code.

For example, when scanning los objects in the mod union scan step, we didn't check whether the object is cemented or not, leading to additional unnecessary global remsets for the next nursery collection.

8 years ago[sgen] Move scan vtype with the other object ops
Vlad Brezae [Mon, 25 Jan 2016 01:17:55 +0000 (03:17 +0200)]
[sgen] Move scan vtype with the other object ops

8 years ago[sgen] Simplify defines for object ops
Vlad Brezae [Mon, 25 Jan 2016 01:02:41 +0000 (03:02 +0200)]
[sgen] Simplify defines for object ops

8 years ago[sgen] Allow read only access on block list without taking GC_LOCK
Vlad Brezae [Fri, 22 Jan 2016 18:16:01 +0000 (20:16 +0200)]
[sgen] Allow read only access on block list without taking GC_LOCK

Still, block list should not be accessed while sweep is running.

8 years ago[sgen] Allow read only access of los object list without taking GC_LOCK
Vlad Brezae [Fri, 22 Jan 2016 16:13:49 +0000 (18:13 +0200)]
[sgen] Allow read only access of los object list without taking GC_LOCK

Still, object lists should not be accessed while sweep is running.

8 years ago[sgen] Extract the handle array code for reuse
Vlad Brezae [Fri, 22 Jan 2016 00:08:06 +0000 (02:08 +0200)]
[sgen] Extract the handle array code for reuse

We add a new data structure which contains a bucket of entries, exponentially increasing in sizes. The structure is customizable through a few callbacks.

Setting is_slot_set_func enables us to also lookup empty slots, instead of only adding entries at the end of the list.

Setting set_slot_func enables the use of a customized function for setting the slot. If this function is not set, the slots and array fields will be set directly, therefore the structure won't be thread safe.

8 years ago[sgen] Don't prefetch unused memory when scanning the card table
Vlad Brezae [Tue, 19 Jan 2016 17:43:04 +0000 (19:43 +0200)]
[sgen] Don't prefetch unused memory when scanning the card table

When doing the mod union scan, we use the block's cardtable (instead of the global one), which we also cannot prefetch, since that would mean dereferencing the block info.

8 years agoMerge pull request #2793 from radical/fix-mcs-args
Alexander Köplinger [Fri, 25 Mar 2016 00:05:23 +0000 (01:05 +0100)]
Merge pull request #2793 from radical/fix-mcs-args

[mcs] Accept and ignore command line args supported by csc that we don't

8 years ago[runtime] mark mono_domain_unload external only
Aleksey Kliger [Wed, 23 Mar 2016 21:13:16 +0000 (17:13 -0400)]
[runtime] mark mono_domain_unload external only

Runtime should use mono_domain_try_unload

8 years ago[runtime] Remove CHECK_MONOTYPE macro
Aleksey Kliger [Wed, 23 Mar 2016 21:02:04 +0000 (17:02 -0400)]
[runtime] Remove CHECK_MONOTYPE macro

It's dead code.

8 years ago[runtime] Delete mono_assert and mono_assert_not_reached
Aleksey Kliger [Wed, 23 Mar 2016 20:53:32 +0000 (16:53 -0400)]
[runtime] Delete mono_assert and mono_assert_not_reached

Runtime should use g_assert and g_assert_not_reached or set a MonoError,
as needed.

Also remove REAL_PRINT_REG dead macros from mini-x86.c and mini-amd64.c

8 years ago[runtime] Change mono_object_castclass_mbyref not to throw
Aleksey Kliger [Wed, 23 Mar 2016 20:21:52 +0000 (16:21 -0400)]
[runtime] Change mono_object_castclass_mbyref not to throw

Mark the function external only. (There are no uses inside the runtime).

Change behavior to return NULL if unable to cast, instead of throwing.

8 years agoMerge pull request #2790 from lambdageek/dev/monoerror-corlib-execeptions
Aleksey Kliger (λgeek) [Thu, 24 Mar 2016 23:19:49 +0000 (19:19 -0400)]
Merge pull request #2790 from lambdageek/dev/monoerror-corlib-execeptions

[MonoError] Helpers for some common corlib exceptions

8 years agoMake --aot=full able to emit executables on 64 bit Windows
Andi McClure [Thu, 24 Mar 2016 21:11:48 +0000 (14:11 -0700)]
Make --aot=full able to emit executables on 64 bit Windows

- COFF does not use .local or .size
- COFF does use .def to mark functions and .data to mark data
- The seh_handler functions should not be called on full aot builds (Zoltan says they are unnecessary in full aot, and they attempt to jit)

The executables do not yet run without crashing.

8 years agoSupport tool-prefix in mono/mini fullaotcheck
Andi McClure [Mon, 28 Mar 2016 21:30:56 +0000 (17:30 -0400)]
Support tool-prefix in mono/mini fullaotcheck

* Support MONO_FULLAOT_ADDITIONAL_ARGS env var for make fullaotcheck, appended to --aot parameters
* To support above, permit "blank" parameters (ie trailing commas on parameter list) to --aot=
* --aot tool-prefix parameter now supported for LD as well as AS

8 years agoFix debug Windows builds in MSVC 2015
Andi McClure [Mon, 28 Mar 2016 21:26:09 +0000 (17:26 -0400)]
Fix debug Windows builds in MSVC 2015

We were sometimes passing invalid input to isalnum(). On the debug version of MSVCRT this would result in an assert.

8 years ago[arm] Fix support for soft breakpoints.
Zoltan Varga [Thu, 24 Mar 2016 20:04:59 +0000 (20:04 +0000)]
[arm] Fix support for soft breakpoints.

8 years ago[mcs] Accept and ignore command line args supported by csc that we don't
Ankit Jain [Thu, 24 Mar 2016 19:21:58 +0000 (15:21 -0400)]
[mcs] Accept and ignore command line args supported by csc that we don't

support:

/analyzer
/appconfig
/baseaddress
/deterministic
/errorendlocation
/errorlog
/features
/link
/moduleassemblyname
/pathmap
/pdb
/preferreduilang
/publicsign
/reportanalyzer
/ruleset
/sqmsessionguid

8 years ago[runtime] Make the appdomain-unload.exe test more verbose.
Zoltan Varga [Thu, 24 Mar 2016 16:49:47 +0000 (12:49 -0400)]
[runtime] Make the appdomain-unload.exe test more verbose.

8 years ago[aot] Reenable the emission of line number info using GAS .file/.loc directives,...
Zoltan Varga [Thu, 24 Mar 2016 16:06:33 +0000 (12:06 -0400)]
[aot] Reenable the emission of line number info using GAS .file/.loc directives, recent versions of clang seems to support it.

8 years ago[jit] Fix the dwarf info structure when there is no line number table.
Zoltan Varga [Thu, 24 Mar 2016 15:45:56 +0000 (11:45 -0400)]
[jit] Fix the dwarf info structure when there is no line number table.

8 years ago[jit] Remove the 'appending' support from the dwarf writer, its no longer used.
Zoltan Varga [Thu, 24 Mar 2016 15:28:33 +0000 (11:28 -0400)]
[jit] Remove the 'appending' support from the dwarf writer, its no longer used.

8 years ago[jit] Get rid of the create_delegate_trampoline_data () helper function which only...
Zoltan Varga [Thu, 24 Mar 2016 01:01:05 +0000 (21:01 -0400)]
[jit] Get rid of the create_delegate_trampoline_data () helper function which only has one caller.

8 years agoMerge pull request #2768 from lambdageek/dev/monoerror-cominterop
Aleksey Kliger (λgeek) [Thu, 24 Mar 2016 15:19:42 +0000 (11:19 -0400)]
Merge pull request #2768 from lambdageek/dev/monoerror-cominterop

[com] MonoError-ize cominterop.c

8 years agoMerge pull request #2663 from esdrubal/islocal
Alexis Christoforides [Thu, 24 Mar 2016 14:45:48 +0000 (10:45 -0400)]
Merge pull request #2663 from esdrubal/islocal

[System] Fixes HttpListenerRequest.IsLocal.

8 years agoCreate xammac_net_4_5 variant of System.Runtime.Remoting
Chris Hamons [Wed, 23 Mar 2016 21:45:29 +0000 (16:45 -0500)]
Create xammac_net_4_5 variant of System.Runtime.Remoting

8 years agoTrigger a rebuild of Mono.Security
Alexander Köplinger [Thu, 24 Mar 2016 02:11:47 +0000 (03:11 +0100)]
Trigger a rebuild of Mono.Security

8 years ago[Mono.Security]: MonoTlsProvider.InvokeSystemCertificateValidator() now takes 'ref...
Martin Baulig [Wed, 23 Mar 2016 23:42:26 +0000 (19:42 -0400)]
[Mono.Security]: MonoTlsProvider.InvokeSystemCertificateValidator() now takes 'ref X509Chain'.

8 years ago[System]: Add private 'X500DistinguishedName' API to store the canonical encoding...
Martin Baulig [Tue, 22 Mar 2016 03:43:59 +0000 (23:43 -0400)]
[System]: Add private 'X500DistinguishedName' API to store the canonical encoding and use it for fast comparison.

8 years ago[System]: SystemCertificateValidator.CheckUsage(): lazily convert to X509Certificate2...
Martin Baulig [Wed, 16 Mar 2016 04:20:39 +0000 (00:20 -0400)]
[System]: SystemCertificateValidator.CheckUsage(): lazily convert to X509Certificate2 if needed.

8 years ago[System]: Make 'X509Certificate2' and 'X509Chain' lazily-initialized.
Martin Baulig [Tue, 15 Mar 2016 23:36:48 +0000 (19:36 -0400)]
[System]: Make 'X509Certificate2' and 'X509Chain' lazily-initialized.

Add new internal 'X509Certificate2Impl' and 'X509ChainImpl' classes similar to
'X509CertificateImpl' in corlib and only initialize them when actually needed.

8 years ago[corlib]: Cleanup the internal 'X509CertificateImpl' API.
Martin Baulig [Tue, 15 Mar 2016 23:33:11 +0000 (19:33 -0400)]
[corlib]: Cleanup the internal 'X509CertificateImpl' API.

8 years ago[ppdb] Disable the metadata verifier for ppdb tables for now. Fixes #39631.
Zoltan Varga [Wed, 23 Mar 2016 21:11:28 +0000 (17:11 -0400)]
[ppdb] Disable the metadata verifier for ppdb tables for now. Fixes #39631.

8 years ago[MonoError] Use mono_error_set_execution_engine helper
Aleksey Kliger [Wed, 23 Mar 2016 19:10:53 +0000 (15:10 -0400)]
[MonoError] Use mono_error_set_execution_engine helper

8 years ago[MonoError] Use mono_error_set_not_supported helper
Aleksey Kliger [Wed, 23 Mar 2016 19:11:16 +0000 (15:11 -0400)]
[MonoError] Use mono_error_set_not_supported helper

8 years ago[MonoError] Use mono_error_set_not_implemented helper
Aleksey Kliger [Wed, 23 Mar 2016 19:02:44 +0000 (15:02 -0400)]
[MonoError] Use mono_error_set_not_implemented helper

8 years ago[MonoError] add helpers for some corlib exceptions
Aleksey Kliger [Wed, 23 Mar 2016 18:56:34 +0000 (14:56 -0400)]
[MonoError] add helpers for some corlib exceptions

System.ExecutionEngineException
System.NotSupportedException
System.NotImplementedException

8 years ago[runtime] Make memory allocation for method headers consistent, stop leaking
Alexander Kyte [Mon, 21 Mar 2016 23:05:31 +0000 (19:05 -0400)]
[runtime] Make memory allocation for method headers consistent, stop leaking

8 years ago[man] Remove documentation for the nonexistent MONO_ENABLE_SHM environment variable.
Zoltan Varga [Wed, 23 Mar 2016 18:10:16 +0000 (14:10 -0400)]
[man] Remove documentation for the nonexistent MONO_ENABLE_SHM environment variable.

8 years agoMerge pull request #2784 from mhutch/msbuild-fixes
Alexander Köplinger [Wed, 23 Mar 2016 15:43:24 +0000 (16:43 +0100)]
Merge pull request #2784 from mhutch/msbuild-fixes

Fix regressions in Microsoft.Build.dll

8 years ago[cil-stringreplacer] Use Cecil as reference dependency
Marek Safar [Wed, 23 Mar 2016 14:44:25 +0000 (15:44 +0100)]
[cil-stringreplacer] Use Cecil as reference dependency

8 years ago[cil-stringreplacer] Update debug symbol
Marek Safar [Wed, 23 Mar 2016 12:08:40 +0000 (13:08 +0100)]
[cil-stringreplacer] Update debug symbol

8 years agoMerge pull request #2783 from alexanderkyte/strdup_memory_leaks
monojenkins [Tue, 22 Mar 2016 22:35:21 +0000 (22:35 +0000)]
Merge pull request #2783 from alexanderkyte/strdup_memory_leaks

[runtime] Fix memory leaks related to short-lived strings

8 years agoMerge pull request #2775 from lambdageek/dev/monoerror-reflection-icalls
monojenkins [Tue, 22 Mar 2016 22:35:16 +0000 (22:35 +0000)]
Merge pull request #2775 from lambdageek/dev/monoerror-reflection-icalls

[SRE] MonoError-ize some icalls

* Cleanup some memory that would leak on failure; improve a few error messages
* Add MonoError to `mono_reflection_call_is_assignable_to` and `fieldbuilder_to_mono_class_field`.
* Update several SRE icalls to follow the pattern:
  ```c
    void
    the_icall (args... )
    {
       MonoError error;
       (void) the_real_work_function (args..., &error);
       mono_error_set_pending_exception (&error);
    }

    static gboolean
    the_real_work_function (args..., MonoError *error)
    {
       ... /* return TRUE on success, FALSE and set error on failure */ ...
    }
  ```
  Do this to:
        1. `mono_reflection_register_with_runtime`
        2. `mono_reflection_setup_internal_class`
        3. `mono_reflection_create_internal_class`
        4. `mono_reflection_generic_class_initialize`
        5. `mono_reflection_event_builder_get_event_info`
        6. `mono_reflection_initialize_generic_parameter`
        7. `mono_reflection_create_dynamic_method`