Commit graph

8172 commits

Author SHA1 Message Date
David Marcec
45cc022ea9 Implement RG32UI and R32UI
Needed for xenoblade
2018-08-13 22:55:16 +10:00
B3n30
eab35c8235 Core::CoreTiming: add UnscheduleEventThreadsafe 2018-08-13 13:56:41 +02:00
MerryMage
01d199965a audio_renderer: samples_remaining counts frames, not samples 2018-08-13 11:26:50 +01:00
MerryMage
4b44b8b4fb audio_core: Interpolate 2018-08-13 11:26:50 +01:00
MerryMage
56300f2928 audio_core: Implement low-pass filter 2018-08-13 11:26:50 +01:00
bunnei
e67630b51e
Merge pull request #1032 from lioncash/sanitize
vfs: Use sanitized paths within MoveFile() and MoveDirectory()
2018-08-13 01:43:35 -04:00
bunnei
bd14653417
Merge pull request #1031 from lioncash/verbosity
card_image: Simplify return statement of GetSubdirectories()
2018-08-13 01:42:07 -04:00
bunnei
2e89719d3e
Merge pull request #1048 from lioncash/atomic
kernel/object: Tighten object against data races
2018-08-13 01:41:23 -04:00
bunnei
41b77c4e0a maxwell_to_gl: Implement VertexAttribute::Size::Size_8.
- Used by Breath of the Wild.
2018-08-13 01:34:21 -04:00
bunnei
baaafbd5ea
Merge pull request #1047 from bunnei/rgba16-uint
renderer_opengl: Implement RenderTargetFormat::RGBA16_UINT.
2018-08-13 01:32:26 -04:00
Lioncash
3476f5b4d3 kernel/object: Tighten object against data races
Despite being covered by a global mutex, we should still ensure that the
class handles its reference counts properly. This avoids potential
shenanigans when it comes to data races.

Given this is the root object that drives quite a bit of the kernel
object hierarchy, ensuring we always have the correct behavior (and no
races) is a good thing.
2018-08-13 00:16:40 -04:00
bunnei
bdf17fe0cc renderer_opengl: Implement RenderTargetFormat::RGBA16_UINT.
- Used by Breath of the Wild.
2018-08-13 00:06:22 -04:00
bunnei
54ef9302a2
Merge pull request #1045 from bunnei/rg8-unorm
renderer_opengl: Implement RenderTargetFormat::RG8_UNORM.
2018-08-13 00:05:25 -04:00
David Marcec
76fad8410d Registered missing channel devices 2018-08-13 14:03:50 +10:00
David Marcec
92492ee23b Added missing channel devices 2018-08-13 14:00:27 +10:00
bunnei
e56a444da9
Merge pull request #1044 from bunnei/linestrip
maxwell_to_gl: Implement PrimitiveTopology::LineStrip.
2018-08-12 23:21:34 -04:00
bunnei
8fe118bcaa maxwell_to_gl: Implement PrimitiveTopology::LineStrip.
- Used by Breath of the Wild.
2018-08-12 23:09:32 -04:00
bunnei
c56a0e3c34 renderer_opengl: Implement RenderTargetFormat::RG8_UNORM.
- Used by Breath of the Wild.
2018-08-12 23:08:50 -04:00
bunnei
fecffeb0dd
Merge pull request #1043 from Subv/timing
Use an approximated amortized amount of ticks when advancing timing.
2018-08-12 22:31:55 -04:00
bunnei
9608f51cde
Merge pull request #1036 from lioncash/thread
scheduler: Make HaveReadyThreads() a const member function
2018-08-12 22:13:14 -04:00
bunnei
e4ed5bc836
Merge pull request #1042 from Subv/races
Fixed a bunch of race conditions when running in multicore mode.
2018-08-12 22:05:48 -04:00
bunnei
de5d431eec
Merge pull request #1041 from Subv/duplicated_mutex
Kernel/Mutex: Don't duplicate threads in the mutex waiter list.
2018-08-12 22:02:03 -04:00
bunnei
8da753ab81
Merge pull request #1040 from bunnei/xmad
gl_shader_decompiler: Implement XMAD instruction.
2018-08-12 21:56:40 -04:00
Subv
d923766042 CPU/Timing: Use an approximated amortized amount of ticks when advancing timing.
We divide the number of ticks to add by the number of cores (4) to obtain a more or less rough estimate of the actual number of ticks added. This assumes that all 4 cores are doing similar work. Previously we were adding ~4 times the number of ticks, thus making the games think that time was going way too fast.

This lets us bypass certain hangs in some games like Breath of the Wild.

We should modify our CoreTiming to support multiple cores (both running in a single thread, and in multiple host threads).
2018-08-12 20:41:28 -05:00
Subv
a9877c8f65 Kernel/SVC: Don't reschedule the current core when creating a new thread.
The current core may have nothing to do with the core where the new thread was scheduled to run. In case it's the same core, then the following PrepareReshedule call will take care of that.
2018-08-12 20:38:37 -05:00
Subv
2e7802ad7d Core/HLE: Make the 'reschedule_pending' flag atomic.
Another thread may write to this variable while the core in question is in the middle of checking for a reschedule request.
2018-08-12 18:41:12 -05:00
Subv
3a338d9286 CPU/HLE: Lock the HLE mutex before performing a reschedule.
Another thread might be in the middle of an SVC, thus altering the state of the schedulers.
2018-08-12 18:41:11 -05:00
Subv
84b542c386 Kernel/Threads: Lock the HLE mutex when executing the wakeup callback.
Another thread might be in the middle of a reschedule, thus altering the state of the schedulers.
2018-08-12 18:41:11 -05:00
Subv
0135b328ed Kernel/Thread: Always use the threadsafe option when scheduling wakeups.
WakeAfterDelay might be called from any host thread, so err on the side of caution and use the thread-safe CoreTiming::ScheduleEventThreadsafe.

Note that CoreTiming is still far from thread-safe, there may be more things we have to work on for it to be up to par with what we want.
2018-08-12 18:40:56 -05:00
bunnei
a970709d5d
Merge pull request #1039 from lioncash/type
vfs: Make type hierarchy objects classes instead of structs
2018-08-12 18:43:27 -04:00
bunnei
534abf9d97 gl_shader_decompiler: Implement XMAD instruction. 2018-08-12 18:30:24 -04:00
Subv
5224cc49c4 Kernel/Mutex: Don't duplicate threads in the mutex waiter list.
Exit from AddMutexWaiter early if the thread is already waiting for a mutex owned by the owner thread.

This accounts for the possibility of a thread that is waiting on a condition variable being awakened twice in a row.

Also added more validation asserts.

This should fix one of the random crashes in Breath Of The Wild.
2018-08-12 16:35:27 -05:00
Lioncash
b82b093108 vfs: Make VfsFilesystem constructor explicit
Makes it consistent with the other VFS interfaces and prevents implicit
construction.
2018-08-12 16:55:40 -04:00
Lioncash
cf0a7cd1c1 vfs: Make type hierarchy objects classes instead of structs
struct should be used when the data type is very simple or otherwise has
no invariants associated with it. Given these are used to form a
hierarchy, class should be used instead.
2018-08-12 16:55:40 -04:00
bunnei
424e90f0f5
Merge pull request #1025 from ogniK5377/bad-cast
Fixed invalid cast in loader
2018-08-12 16:22:35 -04:00
bunnei
e12a07079e
Merge pull request #1038 from MerryMage/lock-cubeb
cubeb_sink: Protect queue with a mutex
2018-08-12 16:22:11 -04:00
Zach Hilman
35e4a47be0 registration: Various style and documentation improvements
Fix logic in RealVfsFilesystem Create methods
Remove magic numbers
Fix regex errors
2018-08-12 15:55:44 -04:00
MerryMage
fcc5ffdfdd cubeb_sink: Protect queue with a mutex 2018-08-12 20:41:46 +01:00
bunnei
4cafc24a4e
Merge pull request #1035 from ogniK5377/audio-dev-revision-info
GetAudioDeviceServiceWithRevisionInfo (Used by Bloodstained: Curse of the Moon)
2018-08-12 14:56:11 -04:00
bunnei
68c44ca0ee
Merge pull request #1028 from ogniK5377/aoa
Added GetAudioRendererSampleRate, GetAudioRendererSampleCount & GetAudioRendererMixBufferCount
2018-08-12 13:33:08 -04:00
bunnei
e858a72a22
Merge pull request #1034 from lioncash/hid
hid: Stub DisconnectNpad()
2018-08-12 13:15:56 -04:00
bunnei
4db8acd30a
Merge pull request #1030 from bunnei/sdl2-2.0.8
externals: Update to SDL2-2.0.8.
2018-08-12 13:15:04 -04:00
bunnei
b8c1dca62f
Merge pull request #1006 from degasus/stream_buffer
GL renderer: Pick the streambuffer from citra and use them.
2018-08-12 13:14:42 -04:00
Lioncash
e850ff63bc scheduler: Make HaveReadyThreads() a const member function
This function doesn't modify instance state, so the const qualifier can
be added to it.
2018-08-12 12:55:58 -04:00
Lioncash
11470f331a thread_queue_list: Make contains() and get_first() const member functions
These don't directly modify the contained data.
2018-08-12 12:54:14 -04:00
Lioncash
55c73e10a7 thread_queue_list: Convert typedef to a type alias 2018-08-12 12:47:11 -04:00
Markus Wick
0eb39922f6 gl_rasterizer: Use a shared helper to upload from CPU memory. 2018-08-12 16:10:26 +02:00
Markus Wick
0af7e93763 gl_state: Don't track constant buffer mappings. 2018-08-12 16:10:26 +02:00
Markus Wick
6ff7906ddc gl_rasterizer: Use the stream buffer for constant buffers. 2018-08-12 16:10:26 +02:00
Markus Wick
ce722e317b gl_rasterizer: Use the streaming buffer itself for the constant buffer.
Don't emut copies, especially not for data, which is used once. They just end in a huge GPU overhead.
2018-08-12 15:48:59 +02:00