Commit Graph

83 Commits

Author SHA1 Message Date
Lioncash
de9c5fd375 ConfigManager: Convert GPUDeterminismMode into an enum class
Makes the values strongly-typed and gets more identifiers out of the
global namespace.

We are forced to use anything that is not "None" to mean none, because
X11 is garbage in that it has:

\#define None 0L

Because clearly no one else will ever want to use that identifier for
anything in their own code (and is why you should prefix literally
any and all preprocessor macros you expose to library users in public
headers).
2018-06-15 14:45:30 -04:00
Lioncash
50a476c371 Assert: Uppercase assertion macros
Macros should be all upper-cased. This is also kind of a wart that's
been sticking out for quite a while now (we avoid prefixing
underscores).
2018-03-14 22:03:12 -04:00
Stenzek
c790077c13 VideoBackend: Remove PeekMessages method
The video thread and backend no longer create any windows, therefore
there will never be any messages dispatched to their thread.
2018-01-27 13:53:55 +10:00
Lioncash
f6e3a39c0e CommandProcessor: Remove unnecessary include
Gets rid of some indirect inclusion.
2017-09-02 13:34:21 -04:00
Shawn Hoffman
ed8f293b4f Change "blocking" BlockingLoop::Stop to give up and die after a timeout.
This fixes the global-static fifo object causing infinite hangs in some
cases. Notably, failure to initialize a graphics backend would result in
BlockingLoop::Prepare being called but never executing Run(), leaving the
object in a bad state.
2017-06-25 20:47:30 -07:00
Shawn Hoffman
e6c15e993b quiet some sign-conversion-in-parameter warnings 2017-06-07 20:09:43 -07:00
Lioncash
41fbc659bf Fifo: Remove unnecessary integral casts
FIFO_SIZE is a u32, so %u can just be used instead.
2017-03-28 10:35:26 -04:00
Lioncash
62db55dee2 Fifo: const correctness
PushFifoAuxBuffer only memcpys data using ptr as the source pointer, so
it can be a pointer to const data because of that.
2017-03-26 23:08:33 -04:00
JosJuice
045de7dd26 Remove special condition for auto dual core determinism 2017-02-01 15:43:47 +01:00
degasus
3787af9ee9 CommandProcessor: Limit scope of ugly SCPFifoStruct.
It's only used as an interface between two classes. So no need to declare
it in the backend export header.
2017-01-27 19:04:56 +01:00
Jules Blok
8203ea929b BlockingLoop: Yield to UI message pump while waiting. 2016-11-11 13:37:02 +01:00
degasus
22681ad975 SyncGPU: Fix savestate. 2016-10-11 18:29:53 +02:00
anthony
b427ead0cc Remove Frameskip 2016-10-08 11:49:51 -05:00
degasus
12f050bb8e Fifo: Fix SyncGPU. 2016-10-03 10:38:16 +02:00
degasus
735da0ed69 Fifo: Use SyncGPU timings for single core. 2016-09-17 16:25:13 +02:00
EmptyChaos
aa16282516 Core: Change CoreTiming event key from int to EventType*
Replace 'int' keys with something that carries type information.
Performance is neutral.
2016-09-03 14:55:44 +10:00
Markus Wick
52af0770c3 Merge pull request #4141 from lioncash/vtx
VertexManagerBase: Get rid of static state
2016-08-23 07:33:25 +02:00
Markus Wick
10d20d47aa Merge pull request #4132 from lioncash/enum-fifo
Fifo: Make SyncGPUReason an enum class
2016-08-22 16:49:40 +02:00
Lioncash
1392efa91d VertexManagerBase: Get rid of static behavior 2016-08-21 23:30:38 -04:00
JosJuice
31c530c7b3 Merge pull request #3386 from lioncash/memory
Common: Namespace MemoryUtil
2016-08-19 11:04:45 +02:00
Lioncash
5635d4b709 Fifo: Make SyncGPUReason an enum class 2016-08-18 22:35:58 -04:00
Léo Lam
dca22e08eb Use Common::Flag and Common::Event when possible
Replaces old and simple usages of std::atomic<bool> with Common::Flag
(which was introduced after the initial usage), so it's clear that
the variable is a flag and because Common::Flag is well tested.

This also replaces the ready logic in WiimoteReal with Common::Event
since it was basically just unnecessarily reimplementing Common::Event.
2016-08-10 16:08:15 +02:00
Lioncash
e01c143379 Common: namespace MemoryUtil 2016-08-07 13:03:07 -04:00
EmptyChaos
5adb29b6cb Fifo: Call AllowSleep instead of Wakeup when pausing.
BlockingLoop was being woken when the emulator was being paused
which would cause the loop to busy spin. Issue 9692.
2016-07-19 15:20:17 +10:00
Pierre Bourdon
3570c7f03a Reformat all the things. Have fun with merge conflicts. 2016-06-24 10:43:46 +02:00
Scott Mansell
5e50c37c13 Alternative fix: promote cycleslate to an s64 everywhere.
Also changed a few functions to be static.
2016-04-10 02:13:06 +12:00
Lioncash
32ce2be2bf Fifo: Make g_use_deterministic_gpu_thread a TU-local variable 2016-01-25 05:24:03 -05:00
Lioncash
5ebd1e215b Fifo: Make g_bSkipCurrentFrame a TU-local variable
This is only ever queried, making it a global isn't necessary.
2016-01-25 05:23:14 -05:00
Pierre Bourdon
0bd649248f Merge pull request #3530 from degasus/syncgpufix
Fifo: Fix SyncGPU.
2016-01-25 10:46:08 +01:00
degasus
4295751662 Fifo: More comments for SyncGPU functions. 2016-01-24 11:06:01 +01:00
degasus
cf4478dc92 Fifo: Fix SyncGPU.
CBoot::BootUp() did call CoreTiming::Advance which itself blocks on the GPU,
but the GPU thread wasn't started already. This commit moves the SyncGPU
initialization into the Fifo.cpp file and call it after BootUp().
2016-01-24 11:06:01 +01:00
Lioncash
e187c55bdd OpcodeDecoder: Add namespace 2016-01-24 01:31:36 -05:00
Lioncash
d9fec92628 VideoCommon: Header cleanup
Also remedies places where the video backends and core rely on things
being indirectly included.
2016-01-17 20:11:45 -05:00
degasus
5f244abf28 Fifo: Create a "Fifo" namespace. 2016-01-12 23:28:26 +01:00
Lioncash
a8ab156a29 Fifo: Convert define into constant
Also moves it to the cpp file where it's used.
2015-12-25 16:07:04 -05:00
Tillmann Karras
c52c73f762 VideoCommon: VertexManager -> VertexManagerBase
It may be a bit weird to see calls to static functions in
VertexManagerBase now, but at least it's easier to see what's going on.
2015-11-02 11:53:54 +01:00
Lioncash
19459e827f Partially revert "General: Toss out PRI macro usage" 2015-09-11 09:49:00 -04:00
degasus
c375111076 Options: merge SCoreStartupParameter into SConfig 2015-06-12 19:07:45 +02:00
degasus
d31bed8b79 Fifo: Rewrite SyncGpu
The new implementation has 3 options:
 SyncGpuMaxDistance
 SyncGpuMinDistance
 SyncGpuOverclock

The MaxDistance controlls how many CPU cycles the CPU is allowed to be in front
of the GPU. Too low values will slow down extremly, too high values are as
unsynchronized and half of the games will crash.
The -MinDistance (negative) set how many cycles the GPU is allowed to be in
front of the CPU. As we are used to emulate an infinitiv fast GPU, this may be
set to any high (negative) number.

The last parameter is to hack a faster (>1.0) or slower(<1.0) GPU. As we don't
emulate GPU timing very well (eg skip the timings of the pixel stage completely),
an overclock factor of ~0.5 is often much more accurate than 1.0
2015-06-08 23:16:24 +02:00
degasus
bfa61105d5 Common: Update BlockingLoop to only use one atomic.
This merges two atomic<bool> into one atomic<int>.
We did move the bit from one bool to another, now we can use operator--.
2015-05-30 12:58:09 +02:00
degasus
02a3a063c3 Fifo: Extract syncing loop
It's now a new helper function within common.
2015-05-30 12:58:09 +02:00
Lioncash
1ba3b4e7ac CommandProcessor: Replace volatile usages with atomics
Also remove said variables from being globals.
2015-05-27 03:53:29 -04:00
Lioncash
4cb55cc427 Fifo: Replace usages of volatile with atomics 2015-05-27 00:42:56 -04:00
Tillmann Karras
30ebb2459e Set copyright year to when a file was created 2015-05-25 13:22:31 +02:00
Tillmann Karras
cefcb0ace9 Update license headers to GPLv2+ 2015-05-25 13:22:31 +02:00
comex
06dd0ba3b4 Exit ReadDataFromFifoOnCPU, PushFifoAuxBuffer early if shutting down (GpuRunningState=false)
This was causing a race condition where the "absurdly large aux buffer"
panic alert would be triggered in the last bit of fifo processing on the
CPU thread in deterministic mode (i.e. netplay).  SyncGPU is supposed to
move the auxiliary queue data to the beginning of the containing buffer
so we don't have to deal with wraparound; if GpuRunningState is false,
however, it just returns, because it's set to false by another thread -
thus it doesn't know whether RunGpuLoop is still executing (in which
case it can't just reset the pointers, because it may still be using the
buffer) or not (in which case the condition variable it normally waits
for to avoid the previous problem will never be signaled).  However,
SyncGPU's caller PushFifoAuxBuffer wasn't aware of this, so if the
buffer was filling at just the right time, it'd stay full and that
function would complain that it was about to overflow it.  Similar
problem with ReadDataFromFifoOnCPU afaik.  Fix this by returning early
from those as well; other callers of SyncGPU should be safe.  A
*slightly* cleaner alternative would be giving the CPU thread a way to
tell when RunGpuLoop has actually exited, but whatever, this works.
2015-04-21 22:33:29 -04:00
degasus
74795b4553 Fifo: rewrite Fifo_PauseAndLock
This lock isn't required any more as our FlushGpu garanty to block until the GPU is idle
2015-04-06 12:35:35 +02:00
degasus
b1ffd32f5f Fifo: only touch the SIMD state once in the single core loop 2015-04-06 12:35:35 +02:00
degasus
d2c62b1744 Fifo: only sleep once within every ms of emulated time 2015-04-06 12:35:35 +02:00
degasus
b020ae1c5d Fifo: rewrite sync on idle skipping hack
Now it's done without a busy loop
2015-04-06 12:35:35 +02:00