bunnei
e791da9791
core: hle: kernel: KPageTable: Various improvements to MapPages and UnmapPages.
2022-01-22 20:51:34 -08:00
bunnei
07add23251
core: hle: kernel: KPageTable: MapProcessCode: Various cleanup.
2022-01-22 20:51:34 -08:00
bunnei
ee25e0a40b
core: hle: kernel: KPageTable: ReserveTransferMemory: Various cleanup.
2022-01-22 20:51:34 -08:00
bunnei
0cee5e1af8
core: hle: kernel: KPageTable: ResetTransferMemory: Various cleanup.
2022-01-22 20:51:34 -08:00
bunnei
ffcaf5af90
core: hle: kernel: KPageTable: SetMemoryAttribute: Various cleanup.
2022-01-22 20:51:34 -08:00
bunnei
2935c9d8de
core: hle: kernel: KPageTable: Assert valid address on GetPhysicalAddr.
2022-01-22 01:33:26 -08:00
bunnei
264bb5abf7
core: hle: kernel: KPageTable: Operate: Assert lock ownership.
2022-01-22 01:33:26 -08:00
bunnei
0137f2e6e1
core: hle: kernel: KPageTable: SetHeapSize: Cleanup & take physical memory lock.
2022-01-22 01:33:26 -08:00
bunnei
6d8e498f76
core: hle: kernel: Refactor Un/MapPhysicalMemory to remove unnecessary methods.
2022-01-22 01:33:26 -08:00
bunnei
b8b1b58f36
core: hle: kernel: Rename Un/Map to Un/MapMeory.
2022-01-22 01:33:26 -08:00
bunnei
615fb40416
hle: kernel: KThread: Ensure host (dummy) threads block on locking.
...
- But do not enter the priority queue, as otherwise they will be scheduled.
- Allows dummy threads to use guest synchronization primitives.
2022-01-21 17:12:06 -08:00
bunnei
f6815086a1
hle: kernel: Remove redundant tracking of dummy threads.
...
- These are already tracked by kernel's registered_objects member.
2022-01-20 17:08:00 -08:00
bunnei
91ff6d4cb3
hle: kernel: KThread: DummyThread can be waited, ensure wait_queue is not nullptr.
2022-01-20 17:08:00 -08:00
bunnei
46a620f9d7
hle: kernel: KThread: Decrease DummyThread priority to ensure it is never scheduled.
2022-01-20 17:08:00 -08:00
bunnei
0b37e7cb39
hle: kernel: service_thread: Ensure dummy thread is closed & destroyed on thread exit.
2022-01-20 17:08:00 -08:00
bunnei
384e24d3e9
hle: kernel: KServerSession: Remove hack for CompleteSyncRequest.
...
- This does not appear to be necessary anymore.
2022-01-20 17:08:00 -08:00
bunnei
ad53dc22fd
hle: kernel: KServerSession: Simplify CompleteSyncRequest EndWait.
...
- Considering is_thread_waiting is never set, so we can remove IsThreadWaiting.
- KThread::EndWait will take the scheduler lock, so we can remove the redundant lock.
2022-01-20 17:08:00 -08:00
bunnei
5ffec69dc7
hle: kernel: KThread: Ensure dummy threads never call EndWait.
...
- These are only used by host threads for locking and will never have a wait_queue.
2022-01-20 17:08:00 -08:00
bunnei
11a380c3da
hle: kernel: KScheduler: Ensure dummy threads are never scheduled.
...
- These are only used by host threads for locking.
2022-01-20 17:08:00 -08:00
bunnei
f6cbb14dce
hle: kernel: KThread: Rename thread_type_for_debugging -> thread_type.
...
- This will be used to ensure that we do not schedule dummy threads.
2022-01-20 17:08:00 -08:00
bunnei
b5e83bcc7b
Merge pull request #7701 from bunnei/clear-mem-pages
...
Kernel Memory Updates (Part 3): Clear KMemoryManager pages & other fixes
2022-01-18 21:20:42 -08:00
bunnei
101d86897b
Merge pull request #7712 from bunnei/fix-thread-exit
...
Accurately implement thread exit
2022-01-17 18:08:24 -08:00
Valeri
84786dde00
hle: remove no-op code
...
Found by static analysis with PVS-Studio. Nobody seems to really know what was it doing there.
2022-01-17 13:51:12 +03:00
bunnei
b54cbc985e
hle: kernel: k_memory_manager: Clear pages on allocation & free.
...
- Heap pages should be zero'd.
- Also explicitly passed along heap allocation option.
2022-01-14 21:16:33 -08:00
bunnei
f499c8177e
core: hle: kernel: KThread: Integrate with KWorkerTask and implement DoWorkerTaskImpl.
...
- This is used to terminate a thread asynchronously after it has been exited.
- This fixes a crash that can occur in Pokemon Sword/Shield because a thread is incorrectly closed on svcExitThread, then, the thread is destroyed on svcCloseHandle while it is still scheduled.
- Instead, we now wait for the thread to no longer be scheduled on all cores before destroying it from KWorkerTaskManager, which is accurate to HOS behavior.
2022-01-14 16:44:14 -08:00
bunnei
d8b3f665db
core: hle: kernel: KProcess: Integrate with KWorkerTask and add unimplemented DoWorkerTaskImpl.
2022-01-14 16:44:14 -08:00
bunnei
03884b7ea6
core: hle: kernel: KThread: Replace Suspend with UpdateState & various updates.
...
- This makes our implementations of these more closely match HOS.
2022-01-14 16:44:14 -08:00
bunnei
c905044e1b
core: hle: kernel: Instantiate a kernel instance of KWorkerTaskManager.
2022-01-14 16:44:14 -08:00
bunnei
bf32fcc817
core: hle: kernel: Add KWorkerTask and KWorkerTaskManager.
...
- These primitives are used to dispatch asynchronous kernel tasks from KThread and KProcess.
2022-01-14 16:43:59 -08:00
bunnei
cc112f971e
hle: kernel: Fix service_threads access to be thread safe V2.
...
- PR #7699 attempted to fix CreateServiceThread and ReleaseServiceThread to be thread safe, but inadvertently introduced a possible dead-lock.
- With this PR, we use a worker thread to manage the service thread list, allowing it only to be accessed by a single thread, and guaranteeing threads will not destroy themselves.
- Fixes a rare crash in Pokemon Sword/Shield, I've now run this game for ~12 hours non-stop and am quite confident this is a good solution for this issue.
2022-01-14 16:02:57 -08:00
Mai M
b2d45a4072
Merge pull request #7699 from bunnei/fix-service-thread-race
...
hle: kernel: Fix service_threads access to be thread safe.
2022-01-14 00:46:16 -05:00
bunnei
2147240e47
hle: kernel: Fix service_threads access to be thread safe.
...
- CreateServiceThread and ReleaseServiceThread can be accessed by different threads, uses a lock to make this thread safe.
- Fixes a rare crash in Pokemon Sword/Shield that can occur when a new service thread is being created while an old one is being destroyed.
2022-01-13 21:26:10 -08:00
bunnei
49a0e4330e
hle: kernel: k_page_table: Update SetProcessMemoryPermission.
2022-01-11 16:28:11 -08:00
bunnei
6123b6ea45
hle: kernel: k_page_table: ReadAndWrite -> UserReadWrite.
2022-01-11 16:28:11 -08:00
bunnei
081669c334
hle: kernel: k_page_table: Rename *ProcessCodeMemory -> *CodeMemory.
2022-01-11 16:28:11 -08:00
bunnei
af4696657c
core: hle: kernel: svc: Updates to SetMemoryAttribute and SetMemoryPermission.
2022-01-08 12:18:14 -08:00
bunnei
b9a313057e
core: hle: kernel: k_page_table: Update CheckMemoryState.
2022-01-08 03:20:57 -08:00
bunnei
3a89723d97
core: hle: kernel: Implement thread pinning.
...
- We largely had the mechanics in place for thread pinning, this change hooks these up.
- Validated with tests https://github.com/Atmosphere-NX/Atmosphere/blob/master/tests/TestSvc/source/test_thread_pinning.cpp .
2021-12-30 15:50:45 -08:00
bunnei
091463a429
core: hle: kernel: Updated implementation of svcSetHeapSize.
...
- Updates our svcSetHeapSize with latest HOS, furthermore allowing heap size to properly be extended/shrunk.
- Validated with tests https://github.com/Atmosphere-NX/Atmosphere/blob/master/tests/TestSvc/source/test_set_heap_size.cpp .
2021-12-28 01:25:20 -08:00
bunnei
9a0648ff0a
Merge pull request #7621 from bunnei/set-mem-perm
...
core: hle: kernel: Implement SetMemoryPermission.
2021-12-27 23:33:11 -08:00
bunnei
4e7a6639d2
core: hle: kernel: Implement SetMemoryPermission.
...
- Not seen in any games yet, but validated with kernel tests.
2021-12-23 01:10:36 -08:00
bunnei
a0c7d93b84
core: hle: kernel: KThread: X18 should be a cryptographically random number.
...
- This was added with firmware 11.0.0 (https://switchbrew.org/wiki/11.0.0 ).
- X18 is OR'd by kernel with 1, to make sure it is odd.
2021-12-23 00:03:39 -08:00
bunnei
49e3c073a5
hle: kernel: svc: GetInfo: Fix error checking with IdleTickCount.
...
- Enforce tha the supplied handle is invalid, not valid.
- This gets Witcher 3 booting.
2021-12-21 22:41:23 -08:00
bunnei
ee6d40d414
Merge pull request #7597 from bunnei/remove-global-lock
...
core: hle: Remove global HLE lock.
2021-12-20 14:24:50 -08:00
ameerj
55650c5b75
kernel: Manually destroy the current process during shut down
...
Avoids a memory leak.
2021-12-19 01:38:25 -05:00
bunnei
c73841500a
core: hle: Remove global HLE lock.
...
- This was added early on as a hack to protect against some concurrency issues.
- It's not clear that this serves any purpose anymore, and if it does, individual components should be fixed rather than using a global recursive mutex.
2021-12-17 16:05:51 -08:00
Valeri
04301e1a8a
Remove erroneous #pragma once
2021-12-13 16:49:01 +03:00
bunnei
280c779898
Merge pull request #7462 from bunnei/kernel-improve-scheduling
...
Kernel: Improve threading & scheduling V3
2021-12-12 22:43:25 -08:00
itsmeft24
e05c86aa3c
Update k_code_memory.h
2021-12-07 16:58:23 -05:00
itsmeft24
d197246880
make KCodeMemory::GetSourceAddress const
...
Co-authored-by: Mai M. <mathew1800@gmail.com>
2021-12-07 07:58:33 -05:00