for core stuff:
just remove unique ptrs that dont need any pointer stability at all (afterall its an allocation within an allocation so yeah)
for fibers:
Main reasoning behind this is because virtualBuffer<> is stupidly fucking expensive and it also clutters my fstat view
ALSO mmap is a syscall, syscalls are bad for performance or whatever
ALSO std::vector<> is better suited for handling this kind of "fixed size thing where its like big but not THAT big" (512 KiB isn't going to kill your memory usage for each fiber...)
for core.cpp stuff
- inlines stuff into std::optional<> as opposed to std::unique_ptr<> (because yknow, we are making the Impl from an unique_ptr, allocating within an allocation is unnecessary)
- reorganizes the structures a bit so padding doesnt screw us up (it's not perfect but eh saves a measly 44 bytes)
- removes unused/dead code
- uses std::vector<> instead of std::deque<>
no perf impact expected, maybe some initialisation boost but very minimal impact nonethless
lto gets rid of most calls anyways - the heavy issue is with shared_ptr and the cache coherency from the atomics... but i clumped them together because well, they kinda do not suffer from cache coherency - hopefully not a mistake
this balloons the size of Impl to about 1.67 MB - which is fine because we throw it in the stack anyways
REST OF INTERFACES: most of them ballooned in size as well, but overhead is ok since its an allocation within an alloc, no stack is used (when it comes to storing these i mean)
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3306
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Raises the size of each page entry to 32 bytes, however, it merges them into a single structure
THEORETICALLY this is better since the access pattern observed corresponds to the program wanting backing_addr/pointers/blocks immediately after one another.
This may improve performance at the cost of some extra memory.
Another implementation would be to structure only backing_addr/blocks within the same virtual buffer.
Alas spamming virtual buffers is evil since each of them is a cache trasher (imagine jumping from wildly different block to wildly different block immediately).
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3215
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
So when libc starts it has to start at an entry point located into crt0, now most OSes will do "enough" setup to allow mprotect() and mmap() to be called in static ctors (remember they're called BEFORE main)
By some stupid miracle, NetBSD doesn't; this means that using those functions on NetBSD will result in spurious results
The reason why is still unknown to me, but this is also combined with the fact that allocating a big chunk of memory for the JIT will make NetBSD refuse to mprotect()/mmap() it in low memory situations (even when space is available); so I take the same approach as with solaris
Also I now make it so fastmem handlers are NOT registered for OSes that disabled fastmem, this is because they pollute sigsegv and makes debugging stupidier
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3092
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Uh - the other pr that had "Fix dragonfly" got commits lost and I didn't notice... oops; cherry picked them back through :)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2860
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Still had the issues with libusb, but that should get solved with the other PRs anyways
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2805
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
The debug toggle and the CPU accuracy options are mutually exclusive, if debug toggle on => cpu accuracy is ignored, if debug toggle off => cpu accuracy is used. So just add it to cpu accuracy and avoid the extra hassle.
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Co-authored-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2640
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Compilation and CMake fixes for both Windows on ARM and clang-cl, meaning Windows can now be built on both MSVC and clang on both amd64 and aarch64.
Compiling on clang is *dramatically* faster so this should be useful for CI.
Co-authored-by: crueter <crueter@eden-emu.dev>
Co-authored-by: crueter <crueter@crueter.xyz>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/348
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
This reverts commit c9a3baab5d.
this commit caused issues in ender magnolia or something, need to make
sure I didn't mess up the revert
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/382
Reviewed-by: Shinmegumi <shinmegumi@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@outlook.com>
This adds an option to enable or disable Host MMU Emulation [Android/PC], brings better config per-game handling with Disable Buffer Reorder, disables Flush Debug Lines by Log, option which was enabled by default on Android/PC taxing performance and translates to all supported languages the recent changes.
Leaves room for NCE improvements in the foreseable future.
Co-authored-by: crueter <crueter@eden-emu.dev>
Co-authored-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/324
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
Rationale: Throwing when running out of memory just creates sad paths for no reason (and at that point, just abort immediately). We are using MAP_NORESERVE, if there isn't enough memory a crash will follow anyways.
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/316
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
revert [jit] Increase x86_64 default code size to full 2GiB hugepage (#318)
Abuses the existence of transparent huge pages on Unix. 4*2 = 8GiB virtual memory used total by JIT. May reduce native host TLB trees.
--------------
WIP: Wasn't meant to be merged, it's going to be refined to be added later when more data/ testing have been made about this approach.
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/318
Reviewed-by: Shinmegumi <shinmegumi@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/337
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Co-committed-by: CamilleLaVey <camillelavey99@gmail.com>
revert [jit] Disable fastmem (by default) on FreeBSD, Solaris and OpenBSD due to subpar timings of SIGSEGV (#319)
According to MaranBR, this should have never been merged and should have been closed instead as they iterated on it in 324.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/319
Reviewed-by: Shinmegumi <shinmegumi@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/331
Co-authored-by: Shinmegumi <shinmegumi@eden-emu.dev>
Co-committed-by: Shinmegumi <shinmegumi@eden-emu.dev>
Abuses the existence of transparent huge pages on Unix. 4*2 = 8GiB virtual memory used total by JIT. May reduce native host TLB trees.
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/318
Reviewed-by: Shinmegumi <shinmegumi@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Some header files, specifically for OSX and Musl libc define PAGE_SIZE to be a number
This is great except in yuzu we're using PAGE_SIZE as a variable
Specific example
`static constexpr u64 PAGE_SIZE = u64(1) << PAGE_BITS;`
PAGE_SIZE PAGE_BITS PAGE_MASK are all similar variables.
Simply deleted the underscores, and then added YUZU_ prefix
Might be worth noting that there are multiple uses in different classes/namespaces
This list may not be exhaustive
Core::Memory 12 bits (4096)
QueryCacheBase 12 bits
ShaderCache 14 bits (16384)
TextureCache 20 bits (1048576, or 1MB)
Fixes#8779
