Use QObject->deleteLater() instead of the delete operator to destroy
child widgets of the layout. This prevents crashes caused by pending
events trying to access the now-destroyed widget.
The player_index field in question is ultimately what gets used to determine which ranks get displayed in the leaderboards chart, and because this was missing the chart was simply displaying the top four results no matter what.
The cause of the leaderboard spam was primarily this call where if there was an attempt to get leaderboard info and there wasn't already, there would be a fetch request. This is bad for many reasons: some games have hundreds of boards that will be fetched at startup, if there's simply no data to populate that board, this will just continue to fetch every time the dialog needs to update. To mitigate this, I simply don't load leaderboard information until there are events for that leaderboard - less information for the player, sadly, but heavily cuts down on the number of leaderboard fetches.
Now that we have some test data, it wasn't showing up in the leaderboards tab; this fixes it to ensure (1) that the right ID is being passed to UpdateRow and (2) the map of leaderboard entries is being populated correctly.
Currently we're showing OSD messages for unknown patches in known INI
files, but not for unknown patches in unknown INI files. I don't think
this distinction makes much sense to the user. If there's a patch the
user can't use, they probably want to be aware of that fact.
0c14b0c8a7 made Dolphin load a file from
the Sys folder the first time AchievementManager::GetInstance() is
called. Because Android calls AchievementManager::GetInstance() from
setBackgroundExecutionAllowedNative, this had two negative consequences
on Android:
1. The first time setBackgroundExecutionAllowedNative gets called is
often before directory initialization is done. Getting the path of
the Sys folder before directory initialization is done causes a crash.
2. setBackgroundExecutionAllowedNative is called from the GUI thread,
and we don't want file I/O on the GUI thread for performance reasons.
This change makes us load the data from the Sys folder the first time
the data is needed instead. This also saves us from having to load the
data at all when hardcore mode is inactive.
We mustn't use m_system when it is nullptr. This was causing Dolphin to
crash on Android whenever an activity was recreated or resumed while
emulation is running, which is super common.
Prototype of a system to whitelist known game patches that are allowed to be used while RetroAchievements Hardcore mode is active. ApprovedInis.txt contains known hashes for the ini files as they appear in the repo, and can be compared to the local versions of these files to ensure they have not been edited locally by the player. ApprovedInis.txt is hashed and verified similarly first, with its hash residing as a const string within AchievementManager.h, ensuring ApprovedInis and the hashes within cannot be modified without editing Dolphin's source code and recompiling completely.
The challenge popups have proven to be excessive and are no longer useful thanks to the achievements hotkey. Instead, those events will ask for an immediate RP-level update to the achievements dialog, which will among other things re-sort the dialog to show challenges on top faster.
This way, by pressing Continue on top of a breakpoint, the emulation will actually continue (like on Cached Interpreter and JIT), instead of doing nothing.
Before:
1. In theory there could be multiple, but in practice they were (manually) cleared before creating one
2. (Some of) the conditions to clear one were either to reach it, to create a new one (due to the point above), or to step. This created weird behavior: let's say you Step Over a `bl` (thus creating a temporary breakpoint on `pc+4`), and you reached a regular breakpoint inside the `bl`. The temporary one would still be there: if you resumed, the emulation would still stop there, as a sort of Step Out. But, if before resuming, you made a Step, then it wouldn't do that.
3. The breakpoint widget had no idea concept of them, and will treat them as regular breakpoints. Also, they'll be shown only when the widget is updated in some other way, leading to more confusion.
4. Because only one breakpoint could exist per address, the creation of a temporary breakpoint on a top of a regular one would delete it and inherit its properties (e.g. being log-only). This could happen, for instance, if you Stepped Over a `bl` specifically, and pc+4 had a regular breakpoint.
Now there can only be one temporary breakpoint, which is automatically cleared whenever emulation is paused. So, removing some manual clearing from 1., and removing the weird behavior of 2. As it is stored in a separate variable, it won't be seen at all depending on the function used (fixing 3., and removing some checks in other places), and it won't replace a regular breakpoint, instead simply having priority (fixing 4.).
Now it actually does what it says on the name, instead of creating a breapoint and doing nothing else (not even updating the widget).
Also, it now can't be selected if emulation isn't running.
Closes https://bugs.dolphin-emu.org/issues/13532
Change misleading names.
Fix function usage: Intepreter and Step Out will not check breakpoints in their own wrong way anymore (e.g. breaking on log-only breakpoints).
There are two pieces of functionality to be added here. One, we want to disallow pausing too frequently, as it may be used as an artificial slowdown. This is handled within the client, which can tell us if a pause is allowed. Two, we want to call rc_client_idle on a periodic basis so the connection with the server can be maintained even while the emulator is paused.
When the immediate value is zero, we can do a negation. On ARM64 the NEG
/NEGS instructions are just an alias for SUB/SUBS with a hardcoded WZR.
Before:
```
ldr w22, [x29, #0x28]
mov w21, #0x0 ; =0
subs w22, w21, w22
```
After:
```
ldr w22, [x29, #0x28]
negs w22, w22
```
This reverts commit 72cf2bdb87.
SYSCONF settings are getting cleared when they shouldn't be. Let's
revert the change until I get proper time to figure out why it's broken.
Whenever a request to update the Rich Presence comes in, typically every ten seconds, the Achievement Progress Widget will update the sort order of the achievements and all of their measured values.
bugfix: SetQWidgetWindowDecorations(this); not called before show() for Windows darkmode titlebars.
The actual call to (QWidget) show() needed to come sooner. Show() was originally left alone, but with other checks needing to move with (QWidget) show(), this function became less useful. Show() was originally created to fix the render widget appearing behind the main window, but that appears to work fine in this iteration.
The measured_progress C string for achievements to display potentially contains junk data after the null terminator, which was rendering in the QString in the dialog. This trims those junk characters.
This causes Dual Core to lock up during the boot sequence, because it tries to wait for a not-yet-running GPU thread.
Fixes https://bugs.dolphin-emu.org/issues/13559
rc_client calls the provided memory peeker asynchronously in the callback for starting a session, to validate/invalidate the memory used for achievements. Dolphin cannot access memory from any thread but host or CPU so this access has a small chance of being invalid. This commit adds a MemoryVerifier that the AchievementManager will use to perform this, before changing the peek method back to the original MemoryPeeker for normal operation.
Some pieces of code are calling IsRunning because there's some
particular action that only makes sense when emulation is running, for
instance showing the state of the emulated CPU. IsRunning is appropriate
to use for this. Then there are pieces of code that are calling
IsRunning because there's some particular thing they must avoid doing
e.g. when the CPU thread is running or IOS is running. IsRunning isn't
quite appropriate for this. Such code should also be checking for the
states Starting and Stopping. Keep in mind that:
* When the state is Starting, the state can asynchronously change to
Running at any time.
* When we try to stop the core, the state gets set to Stopping before we
take any action to actually stop things.
This commit adds a new method Core::IsUninitialized, and changes all
callers of IsRunning and GetState that look to me like they should be
changed.
Core::GetState reads from four different pieces of state: s_is_stopping,
s_hardware_initialized, s_is_booting, and CPUManager::IsStepping.
I'm keeping that last one as is for now because there's code in Dolphin
that sets it directly, but we can unify the other three to make things
easier to reason about.
This commit also gets rid of s_is_started. This was previously used in
Core::IsRunningAndStarted to ensure true wouldn't be returned until the
CPU thread was started, but it wasn't used in Core::GetState, so
Core::GetState would happily return State::Running after we had
initialized the hardware but before we had initialized the CPU thread.
As far as I know, there are no callers that have any real need to know
whether the boot process is currently initializing the hardware or the
CPU thread. Perhaps once upon a time there was a desire to make the
apploader debuggable, but a long time has passed without anyone stepping
up to implement it, and the way CBoot::RunApploader is implemented makes
it rather difficult. So this commit makes all the functions in Core.cpp
consider the core to still be starting until the CPU thread is started.
When AchievementProgress::UpdateData(false) is called, it will now empty itself and reinsert all existing boxes, re-sorted into their current buckets, and call UpdateProgress on them all.
Rerendering the entire Achievements dialog every EmulationStateChanged signal is far too often when it turns out that signal fires multiple times to confirm game close, for example. This change results in only the settings changing on EmulationStateChanged, and having the Hardcore mode toggle (which DOES require redrawing the entire dialog) emit its own signal alongside EmulationStateChanged.
AchievementBox now has UpdateData and UpdateProgress, which is called from UpdateData, but may be called elsewhere to update just the progress measurement of the achievement.
CPU Clock Override slider now increments 1% in the UI, with the new lower limit
being 1% instead of 6%.
Prior implementation made it impossible to set exactly 150% in the GUI.
147% -> 152%. Now users can set exact clock % without needing to edit INIs.
rc_client provides basic sorting buckets as a possible option when retrieving the list of achievements or leaderboards; this enables them and labels them in the dialog.
Since rcheevos headers are included in AchievementManager.h, and everyone that depends on Core can include that, we must also pass on the include directory and defines to those dependencies
Storing the log type names in a map results in them getting re-sorted by
their keys, which doesn't quite give us the sorting we want. In
particular, the Achievements category ended up being sorted at R (for
RetroAchivements) instead of at A. Every use of the map is just
iterating through it, so there's no real reason why it has to be a map
anyway.
If an icon is displayed on screen before it downloads, it was displaying a default icon but it would fail to load the actual icon even after it was downloaded. This fixes that.
Add a two second timer to Achievement Progress Indicators to wait until two seconds after the previous message (when it should have decayed out automatically) before posting any new ones.
Prior to this change, attempting to decrease the speed limit below 100% in hardmode would display the new attempted speed and then warn that the speed can not be decreased below 100%; this disables that first message under those conditions.
AchievementManager maintains a unique pointer to a copy of the current volume so it can asynchronously hash that volume. It is not needed otherwise, so I can release that pointer when hashing is complete. This change fixes a bug whereby changing discs in a game and then changing to a different game would result in the loaded volume pointer still being loaded with and hashing to the previous game.
This lets us reduce the number of USE_RETRO_ACHIEVEMENTS ifdefs in the
code base, reducing visual clutter. In particular, needing an ifdef for
each call to IsHardcodeModeActive was annoying to me. This also reduces
the risk that someone writes code that accidentally fails to compile
with USE_RETRO_ACHIEVEMENTS disabled.
We could cut down on ifdefs even further by making HardcodeWarningWidget
always exist, but that would result in non-trivial code ending up in the
binary even with USE_RETRO_ACHIEVEMENTS disabled, so I'm leaving it out
of this PR. It's not a lot of code though, so I might end up revisiting
it at some point.
Enable emulator hotkeys and controller input (when that option is
enabled) when a TAS Input window has focus, as if it was the render
window instead. This allows TASers to use frame advance and the like
without having to switch the focused window or disabling Hotkeys Require
Window Focus which also picks up keypresses while other apps are active.
Cursor updates are disabled when the TAS Input window has focus, as
otherwise the Wii IR widget (and anything else controlled by the mouse)
becomes unusable. The cursor continues to work normally when the render
window has focus.
Using shifts and bit tests makes the code unnecessarily annoying to
reason about. I'm replacing it with subtracting from 3 to translate the
bit order from the PowerPC format to the usual format.
BI contains both the field and the flag (5 bits total), so we need to
shift away the 2 flag bits to get the 3 field bits. (Same as the
CRBA/CRBB handling in the code just below the BI code.)
The names attached to the BadgeStatus object are obsolete and unneeded and are removed from everything that uses them. All BadgeStatus references are updated to just Badge.
The defaults get loaded in by Dolphin at emulator start, and are used if the badge that would normally be displayed has not for whatever reason been downloaded yet. Badges attached to this PR are placeholders (MayIMilae is designing permanent badges) and reside in Sys\Load\RetroAchievements.
Achievement badges/icons are refactored into the type CustomTextureData::ArraySlice::Level as that is the data type images loaded from the filesystem will be. This includes everything that uses the badges in the Qt UI and OnScreenDisplay, and similarly removes the OSD::Icon type because Level already contains that information.
Was informed by the RetroAchievements team that this isn't an option in most emulators, and as the next commits will be to enable default icons, there will always be something to display.
On Windows:
wsi.render_window being set will set/save the initial geometry, which will cause sizing bugs until it's set again by the user resizing/repositioning.
If Rich Presence and Discord Presence are enabled in Achievement Settings, the string generated by rcheevos as the player's current Rich Presence will be sent to the Status field in the Discord Presence object. This will be updated whenever Rich Presence updates.
The "welcome message" that shows up to players when a game starts up and Achievements are active will now either tell the player upon load that there's no support for achievements, or will wait until the first attempt to load the game's badge either succeeds or fails and then display a full message with title and progress and status. This is in part facilitated by a boolean field for when to display a welcome message, set to true upon loading and to false upon a message being displayed.
If achievements were disabled but a player token is in settings, prior to this change the Achievement Manager dialog would show a box with no player name and score zero, which is unnecessary.
Due to an oversight in our CMakeLists, pkg-config would attempt to find *minizip* 3.0.0 (which doesn't exist) instead of *minizip-ng* 3.0.0, or at least it was on my Manjaro Linux machine. This has been fixed. The new submodule is using minizip-ng 3.0.4, the same version that was being used before.
NetBSD doesn't put packages in /usr/local like /CMakeLists.txt thought.
The `#ifdef __NetBSD__` around iconv was actually breaking compilation
on NetBSD when using the system libiconv (there's also a GNU iconv
package)
A C program included from C++ source broke on NetBSD specifically, work
around it.
This doesn't fix compilation on NetBSD, which is currently broken, but
is closer to correct.
967280f140 broke linking against
libLLVM.so because it used the outdated way to link against LLVM from
CMake. This causes a compilation failure on systems that don't have the
LLVM static libraries, such as Arch Linux. On systems that have the
static libraries, it'll use them and increase binary sizes massively.
Switch to the newer llvm_config CMake macro from LLVM.
Previously the Achievements option would only show up if achievements were already enabled, requiring users to manually create a config file in the file system; this now makes it visible no matter what.
Bugfix for hardcore-disabled items being disabled when hardcore was true but achievement integration was false, which should mean hardcore is effectively disabled. Now everything checks the IsHardcoreModeActive method in AchievementManager which processes the setting AND the game state to determine if hardcore mode is actually active.
Spectator Mode is a new mode added by rc_client that allows for achievement and leaderboard functionality, but does not submit this data to the site, partially allowing for offline achievements. It effectively replaces the former settings for disabling achievements, leaderboards, and RP, which are now always active internally as long as the client is active.
The client can take care of itself and handle its own hardcore status when it toggles, so I can tell the settings widget to contact the manager directly to set it.
Also, gradually reorganizing the settings dialog over the next handful of commits.
The client can handle media changes natively so disabling can take place internally. This code uses the same external calls to load data, but will call either BeginLoad or BeginChangeMedia based on whether any media is already loaded.
Due to the client's handling of media changes (it simply disables hardcore if an unknown media is detected) the existing functionality for "disabling" the achievements is no longer necessary and can be deleted.
Two portions of this need updating.
Anything related to points and unlock counts and scoring uses game_summary now instead of the TallyScore method. Unfortunately this comes with the drawback that I cannot easily at this time access the number of points/unlocks from the other hardcore mode, so things like the second progress bar have been deleted.
Rich presence, which no longer needs to be stored, but can be calculated at request. As the AchievementHeader can now update just the Rich Presence, DoFrame can now simply call a header update with .rp=true and the current Rich Presence will be calculated immediately.
As the two items above are the last remaining things to use a number of the components in AchievementManager, this also deletes: Request V1 (V2 is renamed accordingly), ResponseType, PointSpread, TallyScore, UnlockStatus, and the RP generation and ping methods.
UpdateData in AchievementsWindow now only updates the components being requested, massively improving the window's performance. The parameter is UpdatedItems in AchievementManager, which tracks which portions of the system have been updated for every update callback.
Similarly to the Progress widget (though without the separate object for each box, because each Leaderboard unit is just three text fields stacked vertically), AchievementLeaderboardWidget.UpdateData will now accept three options: destroy all and rebuild, update all, or update a set of rows.
As part of this, AchievementManager::GetLeaderboardsInfo has been refactored to GetLeaderboardInfo to return a single leaderboard for the ID passed in.
AchievementProgressWidget maintains in memory a map of AchievementBox pointers so that UpdateData can operate on them individually. UpdateData is overhauled for three options: UpdateData(true) will destroy the entire list and re-create it from scratch as before, to be used if the game or player changes/closes/logs out. UpdateData(false) will loop through the map and call UpdateData on every achievement box, to be used for certain settings changes such as enabling badges or disabling hardcore mode. UpdateData(set<IDs>) will call UpdateData on only the IDs in the set, to be used when achievements are unlocked.
AchievementBox is an extension of QGroupBox that contains the data for a single achievement, initialized with the achievement data and able to reference AchievementManager to update itself.
While state loading is not allowed in the hardcore mode that most players will use, it is allowed in softcore mode; more importantly, if something fails to unlock or unlocks when it shouldn't in either mode the player can create a save that retains the current achievement state.
This is not a 1 to 1 relationship with how the events look primarily because currently achievement
progress messages are in OnScreenDisplay, which currently vanishes messages automatically.
As this covers the last remaining runtime-based event from the old event handler, that handler has been deleted and the new event handler has been renamed to take its place.
Up to four leaderboards are displayed in a window in the bottom right of the screen (vertically above challenge icons, if there are any). As per RetroAchievements standards, the markers only display the current leaderboard values with no further context necessary.
The active leaderboard data (leaderboards currently being attempted, which get displayed on screen) is now tracked. When a leaderboard is started its value is added to a vector (sorted by start frame). There are a separate set of client events specifically to handle leaderboard trackers, that are used to populate and manage this vector. The top portion of this vector (by RetroAchievement standards, the first four items) is exposed to be displayed on screen.
Also deletes the old runtime-based Achievement Triggered event from the old handler, and the methods used by it to publish to the server and reactivate/deactivate achievements in the runtime.
This change was primarily made to refactor the badge fetching to use the client instead of the runtime, but in the process I also refactored the code to cut down on complexity and duplication. Now the FetchBadge method is passed a function that generates the badge name; this is used to ensure that once the badge is loaded that it is still the desired badge to avoid race conditions.
HashGame has become LoadGame, similar structure with the file loaders but using the client instead. LoadGameCallback has been created to handle the results. The old LoadGameSync has been deleted as have
several hash and load methods that it called.
Deletes AchievementManager::Login, renames LoginAsync to Login, and replaces the one synchronous call in the AchievementSettingsWidget with the async call. There is a minor usability regression in that the UI currently does not notify the user when a login has failed; this will be addressed in a later change (possibly in a different PR).
On Windows 11, when playing windowed in a separate window/widget from the main emulator window, we don't want the window to have rounded corners, as it prevents the corner pixels from being visible
Also make the `Decrypt` method private.
As far as I can tell, the only motivation for exposing the `SetBytes`
and `Reset` methods is to allow `CBoot::SetupWiiMemory` to use the same
`SettingsHandler` instance to read settings data and then write it back.
It seems cleaner to just use two separate instances, and require a given
`SettingsHandler` instance to be used for either writing data to a
buffer or reading data from a buffer, but not both.
A natural next step is to split the `SettingsHandler` class into two
classes, one for writing data and one for reading data. I've deferred
that change for a future PR.
This is a JitArm64 version of 219610d8a0.
Due to limitations on how far you can jump with a single AArch64 branch
instruction, going above the former limit of 128 MiB of code (counting
nearcode and farcode combined) requires a bit of restructuring. With the
restructuring in place, the limit now is 256 MiB. See the new large
comment in Jit.h for a description of the new memory layout.
To ensure memory safety, callers of GetPointer have to perform a bounds
check. But how is this bounds check supposed to be performed?
GetPointerForRange contained one implementation of a bounds check, but
it was cumbersome, and it also isn't obvious why it's correct.
To make doing the right thing easier, this commit changes GetPointer to
return a span that tells the caller how many bytes it's allowed to
access.
This fourth part of my series of patches to get rid of unsafe uses of
GetPointer takes care of the "easy" cases in VideoCommon. Three uses of
GetPointer now remain in Dolphin: VertexLoaderManager, TextureInfo, and
the software renderer's TextureSampler.
All relevant games other than Pikmin 1 Wii seem to always set the two
dwords to zero, so previously they were ignored during command dispatch
and now we still ignore them but in the right place.
For some reason Linux is surprisingly slow at closing file descriptors
of event devices. This commit improves GUI startup times on my computer
by about 1.5 seconds.
When writing the software FMA code, I didn't realize that we can't
overwrite d if d is the same register as one of the inputs and
HandleNaNs is going to be called. This fixes that.
On all platforms, this would result in out of bounds accesses when getting the component sizes (which uses stuff from VertexLoader_Position.h/VertexLoader_TextCoord.h/VertexLoader_Normal.h). On platforms other than x64 and ARM64, this would also be out of bounds accesses when getting function pointers for the non-JIT vertex loader (in VertexLoader_Position.cpp etc.). Usually both of these would get data from other entries in the same multi-dimensional array, but the last few entries would be truly out of bounds. This does mean that an out of bounds function pointer can be called on platforms that don't have a JIT vertex loader, but it is limited to invalid component formats with values 5/6/7 due to the size of the bitfield the formats come from, so it seems unlikely that this could be exploited in practice.
This issue affects a few games; Def Jam: Fight for New York (https://bugs.dolphin-emu.org/issues/12719) and Fifa Street are known to be affected.
I have not done any hardware testing for this PR specifically, though I *think* I previously determined that at least a value of 5 behaves the same as float (4). That's what I implemented in any case. I did previously determine that both Def Jam: Fight for New York and Fifa Street use an invalid normal format, but don't actually have lighting enabled when that normal vector is used, so it doesn't change rendering in practice.
The color component format also has two invalid values, but VertexLoader_Color.h/.cpp do check for those invalid ones and return a default value instead of doing an out of bounds access.
IOS::HLE::IOCtlVRequest::Dump sometimes tries to call GetPointerForRange
with an address of 0 and a size of 0. Address 0 is valid, but we were
mistakenly also trying to check that address 3FFFFFFF is valid, which it
isn't.
Fixes https://bugs.dolphin-emu.org/issues/13514.
Normally, the asserts added in 34b0a6ea90 are only triggered when
something actually went wrong in Dolphin. But there is one exception:
In FallBackToInterpreter, we flush all registers regardless of whether
they're discarded. This is fine as long as none of the discarded
registers are inputs to the instruction that the interpreter will run.
To avoid false positive asserts, this change adds a parameter to Flush
that controls whether to skip the asserts for discarded registers.
Additionally, an assert for discarded registers is added to
Arm64FPRCache::Flush. (Previously JitArm64 asserted for GPRs (and CRs)
only, whereas Jit64 asserted both for GPRs and FPRs. I most likely
didn't think of FPRs when writing 34b0a6ea90.)
Move CheatManager's child widgets into scroll areas to allow making the
window smaller than the default.
In CheatSearchWidget, enable word wrapping for the label describing the
address space and search type to help it fit better inside a narrower
window.
Typically when someone uses GetPointer, it's because they want to read
from a range of memory. GetPointer is unsafe to use for this. While it
does check that the passed-in address is valid, it doesn't know the size
of the range that will be accessed, so it can't check that the end
address is valid. The safer alternative GetPointerForRange should be
used instead.
Note that there is still the problem of many callers not checking for
nullptr.
This is part 2 of a series of changes removing the use of GetPointer
throughout the code base. After this, VideoCommon is the one major part
of Dolphin that remains.
Typically when someone uses GetPointer, it's because they want to read
from a range of memory. GetPointer is unsafe to use for this. While it
does check that the passed-in address is valid, it doesn't know the size
of the range that will be accessed, so it can't check that the end
address is valid. The safer alternative GetPointerForRange should be
used instead.
Note that there is still the problem of many callers not checking for
nullptr.
This is the first part of a series of changes that will remove the usage
of GetPointer in different parts of the code base. This commit gets rid
of every GetPointer call from our IOS code except for a particularly
tricky one in BluetoothEmuDevice.
RCOpArg::ExtractWithByteOffset is only used in one place: a special case
of rlwinmx. ExtractWithByteOffset first stores the value of the
specified register into m_ppc_state (unless it's already there), and
then returns an offset into m_ppc_state. Our use of this function has
two undesirable properties (except in the trivial case `offset == 0`):
1. ExtractWithByteOffset calls StoreFromRegister without going through
any of the usual functions. This violated an assumption I made when
working on my constant propagation PR and led to a hard-to-find bug.
2. If the specified register is in a host register and is dirty,
ExtractWithByteOffset will store its value to m_ppc_state even when
it's unnecessary. In particular, this always happens when rlwinmx
uses the same register as input and output, since rlwinmx always
allocates a host register for the output and marks it as dirty.
Since ExtractWithByteOffset is only used in one place, I figure we might
as well inline it there. This commit does that, and also alters
rlwinmx's logic so the special case code is only triggered when the
input is already in m_ppc_state.
Input in `m_ppc_state`, before (11 bytes):
mov esi, dword ptr [rbp-104]
mov dword ptr [rbp-104], esi
movzx esi, byte ptr [rbp-101]
Input in `m_ppc_state`, after (5 bytes):
movzx esi, byte ptr [rbp-101]
Input in host register, before (8 bytes):
mov dword ptr [rbp-104], esi
movzx esi, byte ptr [rbp-101]
Input in host register, after (3 bytes):
shr edi, 0x18
There were three distinct mechanisms for signaling symbol changes in DolphinQt: `Host::NotifyMapLoaded`, `MenuBar::NotifySymbolsUpdated`, and `CodeViewWidget::SymbolsChanged`. The behavior of these signals has been consolidated into the new `Host::PPCSymbolsUpdated` signal, which can be emitted from anywhere in DolphinQt to properly update symbols everywhere in DolphinQt.
Not sure if the behavior I'm implementing here is what real hardware
does, but since this is a buffer overflow, I'd like to get it fixed
quickly. Hardware verification can happen later.
https://bugs.dolphin-emu.org/issues/13506
Having to look up macros that are defined elsewhere makes the code
harder to reason about. The macros don't remove enough repetition to
justify their existence in my opinion.
With this, I intend to make it clearer that Auto, Force 4:3, Force 16:9
and Custom are really the same thing, just with the aspect ratio of the
simulated TV being selected in a different way. I also extended the
introduction in a way I feel will clarify things but which you are
welcome to bikeshed :)
I was thinking of this during the review of 41b19e262f, but wanted to
put it in a separate PR as to avoid blocking it on bikeshedding.
I'm a bit unsure what to do about the word "analog" in "analog TV". I
felt that repeating it for each of these options would be too
repetitive. I suppose there's a reason why we used the word originally,
but digital TVs do give you basically the same aspect ratio for GC/Wii
games as analog TVs. (Of course, whether it's 4:3-like or 16:9-like
depends on what aspect ratio you set in the TV's settings, but that's
the case for widescreen CRTs too.)
When the divisor is a constant value, we can emit more efficient code.
For powers of two, we can use bit shifts. For other values, we can
instead use a multiplication by magic constant method.
- Example 1 - Division by 16 (power of two)
Before:
mov w24, #0x10 ; =16
udiv w27, w25, w24
After:
lsr w27, w25, #4
- Example 2 - Division by 10 (fast)
Before:
mov w25, #0xa ; =10
udiv w27, w26, w25
After:
mov w27, #0xcccd ; =52429
movk w27, #0xcccc, lsl #16
umull x27, w26, w27
lsr x27, x27, #35
- Example 3 - Division by 127 (slow)
Before:
mov w26, #0x7f ; =127
udiv w27, w27, w26
After:
mov w26, #0x408 ; =1032
movk w26, #0x8102, lsl #16
umaddl x27, w27, w26, x26
lsr x27, x27, #38
This implements the GameCube modem adapter. This implementation is stable but not perfect; it drops frames if the receive FIFO length is exceeded. This is probably due to the unimplemented interrupt mentioned in the comments. If the tapserver end of the connection is aware of this limitation, it's easily circumvented by lowering the MTU of the link, but ideally this wouldn't be necessary.
This has been tested with a couple of different versions of Phantasy Star Online, including Episodes 1 & 2 Trial Edition. The Trial Edition is the only version of the game that supports the Modem Adapter and not the Broadband Adapter, which is what made this commit necessary in the first place.
This expands the tapserver BBA interface to be available on all platforms. tapserver itself is still macOS-only, but newserv (the PSO server) is not, and it can directly accept local and remote tapserver connections as well. This makes the tapserver interface potentially useful on all platforms.
Testing found that spamming toggles for Enable Leaderboards etc risked leaderboards being deleted while the runtime was in the process of recalculating them; this should clear up those conflicts.
Cubeb logs a message at CUBEB_LOG_NORMAL verbosity every time you start
or stop a stream which can get a bit annoying when using frame advance
at Dolphin's default verbosity.
Window icon was missing from QDialog lacking a parent.
Giving the QDialog a parent revealed I had failed to make it properly non-modal, necessitating further changes.
Settings save less often, now only upon destruction.
Construction of BranchWatchDialog is now deferred.
Not only was the extra call to the update callback in the AchievementEventHandler method unnecessary, it was getting called on events that don't even need to be tracked here, causing a lot of lag when it turned out one achievement was repeatedly spamming Achievement Reset Events as a shortcut.
Failing to do this was causing challenge icons to carry over into the next game if a game was closed while the icons were active, even if the next game to run was a completely different game entirely with completely different badges.
RetroAchievements plans to use the user_agent in unlock requests to determine which software version was used to play the game, and can filter older software versions out. As such, I have been given the go-ahead to remove the hardcoded line that forces hardcore to always be false.
This takes care of making the image clearer in some edge cases where the game was already running at near perfect
4:3 with no stretching, and the VI aspect ratio didn't match the XFB by one pixel, making the image stretched and blurry.
-Video: Fix `FindClosestIntegerResolution() using the window aspect ratio and not the draw aspect ratio, causing it to prefer
stretching over black bars in cases when it wasn't desirable.
Two minor updates to improve the Achievement Manager's handling of a player's completion rate.
One, UnlockStatus and the unlock map now track achievement category, such that TallyScore does not count unofficial achievements in counts/points.
Two, the determinations for mastery/completion are now improved to check (1) that the achievement triggering this is CORE (not UNOFFICIAL) and (2) that it has not already been unlocked at this level on the site, which should be sufficient to determine that the unlocking of this particular achievement completes/masters the game.
* Fix irregularly shaped corners
* Remove extra space for BalloonTips with no message or no title
* When the target tip location is not on a screen, put the tooltip on
the mouse's screen instead of the primary screen
* Fix description getting cut off when the title was too long
* Expose border width as a parameter
* Fix spacing and sizing issues with larger border widths
Most obviously, there is no longer a warning message to the player in the achievement window that achievements are disabled if a game is not currently running.
Use "bzip2" instead of "bzip" in optparse's compression choices for the
convert command. This is both more accurate and matches what the
ParseCompressionTypeString function expects.
The mismatch between the two parsing functions prevented compression
using bzip2 because either ParseCompressionTypeString or optparse would
generate an error when using "bzip" or "bzip2" respectively.
Fixes https://bugs.dolphin-emu.org/issues/13427
This fixes an issue introduced by 3b0444be6b
where the m_accelerator would not be initialized when loading a savestate if
the current UCode mismatched the UCode in the savestate, leading to a crash.
When performing a default compilation with recent GCC & glibc,
the use of -Werror=nonnull causes a build error.
The error is given as IOFile::ClearError() can call std::clearerr()
with a null file, which can trigger a null-pointer dereference in libc.
Change the std::clearerr() call to be conditional on a file being open.
Unlike ADD (immediate), MOV (register) treats SP as ZR. Therefore the
ADDI2R optimization that was added in 67791d227c can't optimize ADD to
MOV when exactly one of the registers is SP.
There currently isn't any code in Dolphin that calls ADDI2R with
parameters that would trigger this case.
Adds a setting field under the hood to retain which folder the player last saved/loaded a state to/from, so that the dialog box to select a state to save/load reopens at that folder.
Because the last commit made us use separate folders for GCPad and
GCKey profiles, we should also use separate game INI keys for them.
Otherwise setting e.g. PadProfile1 in a game INI will make both GCPad
and GCKey try to load it, typically with one of them succeeding and the
other one showing a panic alert due to the profile not existing in its
folder.
Better do this breaking change for GCKeys in the same PR as the other
breaking change rather than later.
After reading the previous commit, you might think "hold on, what's the
difference between GetProfileName and GetProfileDirectoryName"? These
two are being used for the exact same thing - figuring out where
profiles are stored - yet they return different values for certain
controllers like GC keyboards! As far as I can tell, the existing code
has been broken for GC keyboards since they were introduced a decade
ago. The GUI (and more recently, also InputCycler) would write and read
profiles in one location, and our code for loading profiles specified in
a game INI file would read profiles in another location.
This commit gets rid of the set of values used by the game INI code in
favor of the other set. This does breaking existing setups where a
GCKey profile has been configured in a game INI, but I think the number
of working such setups is vanishingly small. The alternative would make
existing GCKey profiles go missing from the profile dropdown in the GUI,
which I think would be more disruptive. The alternative would also force
new GCKey profiles into the same directory as GCPad profiles.
This commit also fixes a regression from d6c0f8e749. The Android GUI was
using GetProfileName to figure out what key to use in the game INI,
which made it use incorrect game INI entries for GameCube controller
profiles but not Wii Remote profiles. Now the Android GUI uses
GetProfileKey for this, fixing the problem.
By having getters for this information, other code that needs access to
the same information can call the getters instead of duplicating the
information.
Core::RunAsCPUThread is obsoleted by CPUThreadGuard reference already passed into the function. The nonsense lambda in CheatSearchWidget is from changes in fdb7328c73.
Instead, we make the event take a reference to the system and then pass
it in when the event is triggered.
This does introduce two other accessors, but these are much easier to
refactor out over time, and without modification to the existing event
interface.
Makes the hash specialization a little less noisy. Also we mark it
noexcept, since hashes shouldn't be throwing exceptions (and this can be
optimized on).
Allows using keys that aren't directly std::string as the key. This lets
us use std::string_view for the incoming path name, making it more
flexible with other string types.
The Host URL setting in the RetroAchievements config will, if set, be used as the host URL for all server requests for achievements. This allows for an easy switch to the RetroAchievements staging server for testing.
std::set's lower_bound() is optimized better than the generic
implementation of std::lower_bound()
std::lower_bound() works best on random access iterators, where the
number of comparisons can be logarithmic. However, since std::set's
iterators are bidirectional iterators, the comparisons will actually be
linear in practice when using std::lower_bound().
So, we can use std::set's version which is guaranteed to be logarithmic.
Makes these implementations more thread-safe by design. These likely
won't be run on any other thread, but we may as well just use the
re-entrant variant if it's available.
Given we have fixed allocation orders, we can just directly return a
span instead of a pointer and a size via an out parameter.
Makes it a little more convenient, since we get both pieces of info at
once, and also have the ability to iterate directly off the span out of
the box.
This reverts the parts of commit c8c9928eb1 that made translatability
worse rather than better. Changing "Error in column %2" to "%1 in column
%2" not only means that the translators have to check the i18n comments
to know what word hides behind %1, but there's also the problem that
the translator might need to translate "Error" in this context
differently from the standalone string "Error". Having to copy-paste
some HTML tags may be annoying for translators, but it's a far less
serious problem.
Use 1 of the same type as the stored value when shifting left. This
prevents undefined behavior caused by shifting an int more than 31 bits.
Previously iterator incrementation could either hang or prematurely
report it had reached the end of the bitset.
This fixes an issue where the game specific graphics backend would be saved as the global setting after playing a game.
This also now displays the currently running graphics backend when looking in the graphics configuration window.
Keep a shared_ptr to NetKDTimeDevice inside NetKDRequestDevice.
This allows the KDDownload task to finish its work without potentially
trying to dereference nullptr, which can potentially come from either
GetIOS() or GetDeviceByName() if EmulationKernel's destructor has
started running.
Explicitly shut down work queues in NetKDRequestDevice's destructor to
prevent their threads from accessing members after they've been freed.
This crash would occur sporadically if NetKDRequestDevice's periodic
download or mail checks happened to overlap with emulation shutdown in
the wrong way.
An example sequence of events that could cause the crash:
* m_scheduler_timer_thread queues a periodic Download event in
m_scheduler_work_queue, then waits for m_shutdown_event.
* A request to stop emulation results in s_ios being reset by the CPU
thread. This triggers NetKDRequestDevice's destructor which sets
m_shutdown_event and joins m_scheduler_timer_thread.
* m_scheduler_timer_thread wakes from m_shutdown_event and returns from
its thread function, ending the thread.
* The CPU thread resumes execution at the end of NetKDRequestDevice's
destructor and begins destroying NetKDRequestDevice's members in
reverse declaration order.
* m_http is declared after m_scheduler_work_queue and is therefore
destroyed earlier.
* m_scheduler_work_queue's destructor calls its Shutdown function, which
by default finishes the work items in the queue.
* The queued Download event calls KDDownload which calls m_http.Get()
which calls Fetch() which passes garbage data from the freed m_curl
into curl_easy_setopt().
* Curl promptly crashes.
Shutting down the work queues manually in the destructor prevents the
above because m_http and the other members don't get freed until after
the queue threads finish.
So somehow I forgot that AArch64 uses three-operand encoding...
Fixes a regression from 6303416201 which manifested in various ways,
such as incorrect rendering of the Wind Waker title screen.
It was reported that some games (Zelda Wind Waker and Zelda Twilight Princess but others may also exhibit the issue) have graphical issues with the max pixel samplers set to 16 on some Android devices (ex: Pixel6); since this was increased for a performance heavy feature (custom shaders) just disable it for now. In the future, this could be handled more elegantly
Given how many member functions make use of the system instance,
it's likely just better to pass the system instance in on construction.
Makes the interface a little less noisy to use.
We can get rid of the global system accessors by requiring passing in
relevant state that the function needs and making callsites do the work.
This *does* add a global accessor to the PPCAnalyzer, however, this already
has global accessors present elsewhere within its code, so they can be removed
all at once in a follow up change.
This specific issue was already addressed by https://github.com/dolphin-emu/dolphin/pull/11635
though I felt like there was something more we could do, and wasn't too happy with the
likelihood of devices update calls being skipped (due to `m_devices_population_mutex` being locked).
Resolves a deprecation warning on macOS. Also, we can convert
the function to just return a std::string instead of using a
static internal buffer, which is gross and thread-unsafe.
Eliminates a deprecation warnings on macOS. While we're in the
same area, we can remove the call to GetPointer() and instead
use CopyToEmu() to copy the string data back to the emulated memory.
Jit.cpp : Potentially uninitialized local pointer variable 'host_address_after_return' used in a DEBUG_ASSERT on line 470.
dolphin-emu.sln : A copy-paste error.
The codegen for the functions themselves, not for the emitted code.
This seems to save 32 bytes per function. We also get rid of the oddity
we had before where ANDI2R would do masking for 32-bit operations but
the other functions wouldn't.
The alias for __m128i is typically something like:
typedef long long __m128i __attribute__((__vector_size__(16), __may_alias__));
and the part that ends up not getting preserved is the __may_alias__
attribute specifier.
So, in order to preserve that, we can just use a wrapper struct, so the
data type itself isn't being passed through the template.
Previously we were always taking the buffer by value, even if it wasn't
being stored anywhere and only read from.
We can use a std::span for the same thing.
Two fixes: verify that there's an API token before attempting to automatically log in, and don't attempt to hash a game and load it unless achievements are enabled and the player is logged in. This prevents multiple API requests that we already know will fail, one of which will display an unnecessary error message to the player.
Fixes a build failure with clang 17.
The destructor needs to be in the cpp file, since we have a forward
declared std::unique_ptr type as part of the class. So technically the
default inline destructor could invoke without seeing the full data type
definition.
Renames some variables to avoid shadowing warnings on gcc.
Also gets rid of a FilereaderState struct, since one is already defined
in the declaration of the AchievementManager class.
These aren't necessarily cheap to copy, since a control qualifier will
have around 3 std::strings inside of it, so passing by value can churn
allocations a little bit.
std::function is internally allowed to allocate, and these functions
aren't being stored anywhere (only called), so we can freely get rid
of some minor overhead here by passing by reference.
This change also creates aliases for the functions, so that there isn't
a lot of visual noise when reading the function signatures.
We don't need to enforce the use of std::string instances with
AddSetting(). We can accept views and only construct one string,
rather than three temporaries.
Internal details: The large region is split into individual same-sized blocks of memory. On creation, we allocate a single block of memory that will always remain zero, and map that into the entire memory region. Then, the first time any of these blocks is written to, we swap the mapped zero block out with a newly allocated block of memory. On clear, we swap back to the zero block and deallocate the data blocks. That way we only actually allocate one zero block as well as a handful of real data blocks where the JitCache actually writes to.
The active challenges, aka the primed achievements, are displayed on screen as a series of icons in the bottom right corner of the screen via OnScreenUI.
Emulation state changed signals also update the wiimote connection. The signal here is only needed for wiimote connects/disconnects.
Can fix erroneous debugger behavior during booting, as dolphin will sometimes incorrectly report the state as paused, which leads the debugger widgets to run when they shouldn't.
When an achievement is "primed", a challenge is active, for example completing a portion of the game in under a time limit or without taking damage or using certain items. This is stored in a map in the Achievement Manager (and removed when the achievement is unprimed) so a later commit can display it on screen.
The Disabled state sits between Game Closed and completely Shutdown - stronger than the former, as it refuses to let a game be opened again until AchievementManager is restored (which only happens upon a fresh core boot) but it isn't completely shut down and will still allow the player to be logged in and access the achievement settings and their (global) achievement header.
This change splits LoadGameAsync into three methods: two HashGame methods to accept either a string filepath or a volume, and a common LoadGameSync that both HashGames call to actually process the code. In the process, some minor cleanup, and the hash resolution now takes place on the work queue asynchronously.
Normally we only flush registers right at the end of each PPC
instruction. However, for PPC instructions that use a lot of registers
one at a time, it's beneficial to do this flushing work in the middle
of the instruction instead, reducing the risk of register starvation
and improving pipelining.
This should have been done when rebasing 6cc4f593e5 after the merge of
3a00ff625e. There are no correctness implications as far as I know,
only very minor performance implications.
This worked correctly on the JIT vertex loaders, and for the equivalent case with texture coordinates. I'm not aware of any games this affects (but libogc does mention a semi-related scenario at 6bc0317c7d/gc/ogc/gx.h (L1855-L1857).)
Jimmie Johnson's Anything with an Engine is known to use texture coordinate 7 (and only texture coordinate 7) in some cases. There are a lot of possible edge-cases, so this test brute-forces all combinations with coordinates 0, 1, and 2.
This adds the actual switch to turn on Hardcore Mode to the settings tab of the Achievements dialog. It is accompanied by a large tooltip warning explaining what it does and when it can be enabled.
The switch is only enabled to be turned on when no game is running, so that games are started in hardcore mode and can only be loaded via the console's memory card, as in the original hardware. Hardcore may be turned off while a game is running, but cannot be turned back on until the game is disabled.
The toggle trigger for hardcore mode also automatically disables the settings that are not allowed during hardcore mode.
Finally, the original flag in AchievementSettingsWidget to set whether things are enabled in hardcore mode (primarily Leaderboards) is replaced with the actual Hardcore Mode setting.
Play Input Recording would potentially unlock achievements without any player input and needs to be disabled. If a recording is already playing, hardcore mode cannot be enabled.
The player getting a better view of their surroundings than the game would normally allow could possibly give the player an advantage over the original hardware, so Freelook is disabled in hardcore mode. To do this, I disable the config flag for Freelook when it is accessed, to make sure that it is disabled whether it was enabled before or after hardcore mode was enabled.
Memory patches would be an easy way to manipulate the memory needed to calculate achievement logic, so they must be disabled. Riivolution patches that do not affect memory are allowed, as they will be hashed with the game file.
Debug Mode gives players direct read and write access to memory, which could be used to completely manipulate RetroAchievements logic and therefore is not allowed in hardcore mode.
Slowing down the emulator would artificially improve reaction times and is therefore disallowed in hardcore mode. Speeds faster than 1x are allowed, but any speed below 1x is scaled back up to 1x.
Frame advancing is easily exploitable for slowing down a game and artificially improving reaction times and is not allowed in RetroAchievements hardcore mode.
While saving states is allowed (especially for the purpose of debugging), RetroAchievements does not allow loading saved states when hardcore mode is on.
This widget will be used in several places to notify the player that a feature has been disabled because hardcore mode is on. It includes a button to open the Achievement Settings so that Hardcore Mode may be turned off. Also included is the framework required to open AchievementsWindow specifically on the Settings tab.
Hardcore Mode is a RetroAchievements feature for enabling as close to original hardware as possible, to keep a fair, challenging, and competitive playing field for achievements (which get tallied differently and emphasized more in hardcore) and leaderboards (where it is mandatory) at the cost of several common emulator features that provide advantages, such as state loading and slower emulation speeds.
This commit just adds the flag to the AchievementSettings, with more to come.
Dolphin would crash when running with a fastmem arena but without
fastmem. This regression was caused by merging 9192128c50 without
adapting it after the merge of 0606433404.
This has the same effect in the end, but in my opinion, doing it this
way makes it more clear for the reader why we can read from ppcState at
JIT time, something that makes no sense for everything else in ppcState.
By making the JIT cache check if the current state of MMCR0 and MMRC1
matches the state they had at the time the JIT block was compiled, we
solve a correctness issue (marked in a comment as a speed hack).
Not known to affect any games.
This must have broken in a rebase of one of my recently merged PRs.
Dolphin still worked correctly with this bug, for two reasons:
1. Most AArch64 users are not on Windows, and therefore normally do have
the entry points map.
2. When the bug was triggered, Dolphin would fall back to the slower
path rather than crashing.
Instead of shifting left by 1, we can first shift right by 2 and then
left by 3. This is both faster and smaller, because we get the right
shift for free with the masking and the left shift for free with the
address calculation. It also happens to match the pseudocode more
closely, which is always nice for readability.
This slightly improves instruction-level parallelism in Jit64's slow
dispatcher by shifting the PC left instead of the MSR.
In the past, this also enabled an optimization in JitArm64's fast path
where we could use LDP to load normalEntry and msrBits in one
instruction, but this was superseded by fd9c970.
Just to make the code easier to understand at a glance. I especially
found it a bit annoying to reason about whether callee-saved registers
like W28 were being used because we needed a callee-saved register or
just for no reason in particular.
X8 and up is what compilers normally use when they're not register
starved.
AArch64's handling of NaNs in arithmetic instructions matches PowerPC's
as long as no more than one of the operands is NaN. If we know that all
inputs except the last input are non-NaN, we can therefore skip checking
the last input. This is an optimization that in principle only works for
non-SIMD operations, but ps_sumX effectively is non-SIMD as far as the
arithmetic part of it is concerned, so we can use it there too.
FL_EVIL is only used for blocking instructions from being reordered.
There are three types of instructions which have FL_EVIL set:
1. CR operations. The previous commits improved our CR analysis
and removed FL_EVIL from these instructions.
2. Load/store operations. These are always blocked from reordering
due to always having canCauseException set.
3. isync. I don't know if we actually need to prevent reordering
around this one, since as far as I know we only do reorderings
that are guaranteed to not change the behavior of the program.
But just in case, I've renamed FL_EVIL to FL_NO_REORDER instead of
removing it entirely, so that it can be used for this instruction.
Other than the CR instructions, which we now analyze properly,
all the covered instructions are not integer operations and also
have either FL_ENDBLOCK or FL_EVIL set, so there are two other
checks in CanSwapAdjacentOps that will reject them.
This brings the analysis done for condition registers
more in line with the analysis done for GPRs and FPRs.
This gets rid of the old wantsCR member, which wasn't actually
used anyway. In case someone wants it again in the future, they
can compute the bitwise inverse of crDiscardable.
Instead of materializing the quiet bit in a register and ORing the NaN
with it, we can perform an arithmetic operation on the NaN. This is a
cycle or two slower on some CPUs in cases where generating the quiet bit
pipelined well, but this is farcode that rarely runs, so instruction
fetch latency is the bigger concern. And for non-SIMD cases, we also
save a register.
By using MOVI2R+MOVI2R+CSEL in the zero case instead of doing bitwise
operations on the output of the other MOVI2R+MOVI2R+CSEL, we avoid using
BFI, an instruction that takes two cycles on most CPUs. The instruction
count is the same and the pipelining should be at least equally good.
This is a little trick I came up with that lets us restructure our float
classification code so we can exit earlier when the float is normal,
which is the case more often than not.
First we shift left by 1 to get rid of the sign bit, and then we count
the number of leading sign bits. If the result is less than 10 (for
doubles) or 7 (for floats), the float is normal. This is because, if the
float isn't normal, the exponent is either all zeroes or all ones.
Not sure if this was causing correctness issues – it depends on whether
the HLE code was actually reading the discarded registers – but it was
at least causing annoying assert messages in one piece of homebrew.
Nowadays, basically everything except for controller config is handled
by the new config system. Instead of enumerating the systems that are,
let's enumerate the systems that aren't.
I've intentionally not included Config::System::Session in the new list.
While it isn't intended to be saved, it is a setting that's fully
handled by the new config system. See
https://github.com/dolphin-emu/dolphin/pull/9804#discussion_r648949686.
This is used as a base pointer inside CustomPipelineAction, so this
should probably really have a virtual destructor to ensure derived
objects are torn down properly.
The point of farcode is to provide a separate location for code that
rarely runs, so that it doesn't pollute the icache. Taking a conditional
branch is something that happens very often, so the code for that
shouldn't be in farcode.
Bug: https://bugs.dolphin-emu.org/issues/13404
On macOS 13.6 / Intel HD 5000, Dolphin crashes with this message:
> -[MTLIGAccelDevice setShouldMaximizeConcurrentCompilation:]: unrecognized selector
This should be available on all macOS 13.3+ systems – but when using OCLP drivers,
some devices use an older version of Metal.framework, which doesn't expose the selector.
This concerns Intel Ivy Bridge, Haswell and Nvidia Kepler when using OCLP on macOS 13.3
or newer.
(See
34676702f4/docs/PATCHEXPLAIN.md?plain=1#L354C1-L354C83)
As the behavior is an optional optimization anyway, perform a dynamic
detection to avoid crashing if the feature is not available.
Some state changes are meant to be near instantanoues, before switching to something else. By reporting ithe instant switch, the UI will flicker between states (pause/play button) and the debugger will unnecessarily update. Skipping the callback avoids these issues.
This was implemented to prevent UI flickering due to the state rapidly switching between pause/play. Recently, it has been causing issues with debugger windows, which update during frame advance.
This way, the number of loop iterations is equal to the number of set
bits in the bitset rather than the number of bits in the bitset,
which is a good improvement because usually very few bits are set.
This caused us to update the indirect texture information in shaders more often than we needed to, which probably doesn't matter in practice since it's only used in ubershaders and copyyscale and stride are generally only updated before EFB/XFB copies, which generally will have other changes afterwards.
As far as I can tell, it has nothing to do with the mipmap/half_scale functionality, but does change based on the width of the destination texture (and the destination texture is half the width if half_scale is set). The comment that was there (which dates back to the initial megacommit) seems to not have accounted for the width aspect; it was first used as an actual stride in bbbe898839 (the first commit that used it at all).
The move assignment operator for a class is implicitly deleted when the
class has a non-static reference data member, which is true of
WiiSocket's m_socket_manager member.
Explicitly declaring the operator as default generates a
-Wdefaulted-function-deleted warning on Clang.
Delete the move constructor as well for consistency.
Fix -WSwitch warning about unhandled enum value SDL_NUM_LOG_PRIORITIES.
log_level is initialized to LNOTICE right before the switch statement so
this doesn't cause any behavior changes.
Use RunOnCPUThread instead of RunAsCPUThread in BeginRecordingInput.
Most OpenGL functions require an OpenGL context to have been created on
that thread before calling the function; when that isn't the case they
return invalid results which can cause crashes when passed into other
functions.
Dolphin creates the OpenGL context in the EmuThread which then becomes
either the CPU-GPU thread or the Video thread for single and dual core
respectively. OpenGL functions must therefore be called from that
thread.
Movie::BeginRecordingInput is called from the Host thread and runs a
block of code which ultimately creates a savestate, which in turn embeds
the framebuffer which requires calling various OpenGL functions.
In single core the use of RunAsCPUThread leads to this all happening on
the Host thread, eventually leading to invalid OpenGL calls and a crash.
In Dual core the crash is avoided because VideoBackendBase::DoState uses
the AsyncRequests::DO_SAVE_STATE event which causes VideoCommon_DoState
and its subsequent OpenGL calls to safely run on the Video thread.
This commit uses RunOnCPUThread instead of RunAsCPUThread, which causes
the subsequent code to run on the CPU-GPU thread in single core which
has the valid OpenGL context and so doesn't crash.
This makes it so that if you just want to reload the current style (eg. on program start, or in response to a system event), you don't need to know the name of the currently selected user style. It's also more consistent with the way the 'userstyle/enabled' flag works.
Admiral H. Curtiss