RPCSX/rpcsx
1
0
Fork 0
mirror of https://github.com/RPCSX/rpcsx.git synced 2025-12-06 07:12:14 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity
14108 commits 5 branches 1 tag 150 MiB
Commit graph

618 commits

Author SHA1 Message Date
Nekotekina d6daa0d05b Fix cpu_flag::temp, make sure it removes cpu_flag::wait 2020-10-28 14:00:09 +03:00
Nekotekina c491b73f3a SPU: improve accurate DMA 
Remove vm::reservation_lock from it.
Use lock bits to prevent memory clobbering in GETLLAR.
Improve u128 for MSVC since it's used for bitlocking.
Improve 128 bit atomics for the same reason.
Improve vm::reservation_op and friends.
 2020-10-28 03:47:41 +03:00
Nekotekina c50233cc92 atomics.cpp: add support for waiting on 128-bit atomics 
Complementarily.
Also refactored to make waiting mask non-template arg.
 2020-10-28 03:47:41 +03:00
Nekotekina 4966f6de73 vm: improve range_lock and shareable cache (Non-TSX) 
Allocate "personal" range lock variable for each spu_thread.
Switch from reservation_lock to range lock for all stores.
Detect actual memory mirrors in shareable cache setup logic.
 2020-10-27 17:56:19 +03:00
Nekotekina f1e66085cd Fixup for cpu_flag::temp 
Wrong check_state() result was triggering assertion.
 2020-10-26 01:18:26 +03:00
Nekotekina 130a0ef20e Implement cpu_flag::temp flag 
Accompanies wait flag, indicating that it was set in limited conditions.
Such condition don't allow thread to terminate after its removal.
 2020-10-25 21:48:20 +03:00
kd-11 18ca3ed449 rsx: Block-level reservation access 2020-10-25 20:21:04 +03:00
Eladash 4ea7628204 SPU: Fix LS capture entry point 2020-10-25 16:39:40 +03:00
Nekotekina 2b52b4a749 SPU: use normal notify() thread function 
Using raw_notify() everywhere was overkill.
 2020-10-24 14:16:32 +03:00
Eladash 49610f52f5 SPU: Save LS capture executable in one segment 2020-10-24 14:13:19 +03:00
Eladash b56bc7e087 SPU: cleanup channels logging 2020-10-23 13:13:04 +03:00
Nekotekina dc8252bb9f Remove XABORT in PPU/SPU transactions. 
It's expensive for unknown reason. Simply XEND is usually much cheaper.
Add some minor improvements. Use g_sudo_addr.
 2020-10-20 09:10:21 +03:00
Nekotekina 72d1ac22aa SPU: report too many PUTLLC attempts (TSX) 
Mirrored to PPU STCX code and PUTLLUC (STORE128).
 2020-10-19 19:41:28 +03:00
Nekotekina 8ce5392390 TSX: add prefetchw instruction in transaction code 2020-10-19 19:41:28 +03:00
Nekotekina 311682b341 SPU: fix GETLLAR regression 
Misplaced mov_rdata
 2020-10-19 19:41:28 +03:00
Nekotekina 120849c734 Implement perf stat counter for PPU/SPU reservation ops 
Adds Emu/perf_meter.hpp header file.
Uses RDTSC for speed.
Prints stats at exit.
 2020-10-19 19:41:28 +03:00
Nekotekina adf50b7c4b Implement cpu_thread::if_suspended 
Use it for opportunistic guaranteed GETLLAR execution (TSX-FA).
 2020-10-18 20:10:48 +03:00
Nekotekina f5c575961f Implement priorities for cpu_thread::suspend_all tasks 
Give PUTLLUC increased priority.
 2020-10-18 20:10:48 +03:00
Eladash 402e8b12a6 SPU: Touch unmapoed memory in reservation mismatch 2020-10-18 11:42:54 +03:00
Nekotekina d0057c92e4 Fix spu_putlluc_tx (insignificant) 2020-10-17 21:27:19 +03:00
Nekotekina 583ed61712 SPU: return some give-up behaviour for PUTLLC (TSX) 
Despite using concept of "shared" lock, allow only first to proceed.
This is similar how conditional stores for PPU are implemented.
 2020-10-16 12:14:42 +03:00
Nekotekina facde63460 PPU: fix ppu_stcx_accurate_tx 
Don't destroy xmm6/xmm7 state on exit.
Improve addr arg handling (simplify).
 2020-10-15 19:24:00 +03:00
Nekotekina 494953997e PPU/SPU: give up on conditional stores if locking fails 
Restores Non-TSX behaviour partially.
 2020-10-15 17:18:49 +03:00
Nekotekina 1b89ad00e7 SPU: restore some LR event setting logic after #9048 2020-10-15 17:18:49 +03:00
Nekotekina 3bddba0c7a SPU: fix spu_getllar_tx 
Was not executing.
 2020-10-14 02:53:29 +03:00
Nekotekina 97cd641da9 TSX: reimplement spu_getllar_tx 
Only used as a backup method of reading reservation data.
Increase long GETLLAR reporting threshold.
 2020-10-13 21:10:04 +03:00
Nekotekina 91db4b724c SPU: fix PUTLLC (TSX-FA) 
Some forgotten checks may affect performance.
 2020-10-13 17:46:03 +03:00
Nekotekina dcff8c2637 Fix remaining vm::reservation_lock usages (for now) 
Optimization can be restored later.
 2020-10-13 12:04:59 +03:00
Nekotekina dc39a9b84f SPU: Report 'GETLLAR took too long' 
Also move similar code in PPU.
 2020-10-13 00:12:11 +03:00
Nekotekina a806be8bc4 SPU: Implement S1/S2 (SNR) events (closes #8789) 
Add TSX path in push_snr()
Add locks bits in ch_events
 2020-10-12 21:41:57 +03:00
Eladash 95c1443e30
SPU: Validate reservation in GET commands (Accurate DMA) (#9062) 2020-10-12 15:20:06 +03:00
Nekotekina 5bd5a382c0 PPU: fix LDARX/LWARX in accurate mode (closes #9058) 
Fixup after #9048
Use SSE intrinsics in mov_rdata.
 2020-10-11 19:52:10 +03:00
Nekotekina 050c3e1d6b Rewrite cpu_thread::suspend_all 
Now it's a function of higher order.
Make only one thread do the hard work of thread pausing.
 2020-10-10 13:58:48 +03:00
Nekotekina 346a1d4433 vm: rewrite reservation bits 
Implement classic unique/shared locking concept.
Implement vm::reservation_light_op.
 2020-10-10 13:58:48 +03:00
Eladash 865464f607 SPU Local Storage capture 2020-10-08 19:05:14 +03:00
Eladash 493e57837b
SPU: Fix extremely rare bug of GETLLAR (#9011) 2020-10-03 09:31:28 +01:00
Eladash 56cebd99c2
SPU: Simplify logging of MFC commands (#9004) 2020-10-01 19:52:39 +03:00
Eladash f4ca6f02a1 PPU: Implement support for 128-byte reservations coherency 2020-09-28 22:34:42 +03:00
Eladash 09cddc84be SPU/PPU: Implement Atomic Cache Line Stores 2020-09-27 20:09:21 +03:00
Eladash bfa78870cb SPU: Fix unregistered channels in RCHCNT 
Shouldn't throw exception on realhw.
 2020-09-22 19:47:47 +03:00
Eladash ad37259ccc SPU: Implement many missing channel counts 2020-09-22 19:47:47 +03:00
Eladash c436ef0c6f
SPU: Implement channels 70, 71, add naming for channel 69 (#8932) 2020-09-19 13:08:35 +01:00
eladash 36ac68b436 SPU: Implement events channel count, minor interrupts fixes 2020-09-18 21:57:24 +03:00
Eladash 3206378ae6 sys_spu: Fix overexecution of cpu_return() 2020-09-12 22:11:40 +03:00
Eladash 7ce790f369 SPU: Use ASM for AVX2 coompilation instead of intrinsics 2020-09-12 18:49:49 +03:00
Eladash 9ff0b460a2 SPU: Make PUTLLUC LR event accurate 2020-09-11 09:02:18 +02:00
Eladash 4f0125a0e9 SPU: Remove "Accurate PUTLLUC" setting (always accurate) 2020-09-11 09:02:18 +02:00
Eladash 1e4655aef6
SPU: Remove STOP 0x0 hack (#8873) 2020-09-09 11:36:04 +01:00
Eladash 5060c779da
SPU: Use unaligned instructions in mov_rdata_avx (MSVC) (#8851) 2020-09-07 21:32:44 +01:00
Eladash abc715bc5c
SPU: Make PUT transfers use SEQ-CST ordering on Accurate DMA (#8844) 2020-09-06 12:09:14 +01:00
Eladash 2688081656 SPU: Use unaligned AVX instructions for cmp_data_avx 2020-09-06 07:51:12 +03:00
Eladash e4abd3dc5a SPU: Do not ignore pending PUT tranfers just becase GET may not be cache-line atomic 
This is not the proper way to emulate non-atomic GET tranfers, as it makes it seems as if PUT atomic tranfers arent atomic. (TODO)
Incomplete GET cache line accesses still do not verify data though.
 2020-09-05 22:23:55 +03:00
Eladash c7a185d4e7 SPU: Fix not acuiring reservation locks on DMA with more than one cache line (Accurate DMA) 2020-09-05 22:23:55 +03:00
Eladash 4ffc58a8ce SPU: Cleanup for Accurate PUTLLUC 
Should no longer affect GET commands because Accurate DMA is available for this functionality.
 2020-09-04 10:20:44 +02:00
Eladash c5c9ea1b21 SPU: Make GET's full and aligned cache line accesses atomic with Accurate DMA 2020-09-04 10:20:44 +02:00
Eladash 73d23eb6e6
SPU: Implement Accurate DMA (#8822) 2020-09-02 23:58:29 +02:00
Nekotekina ebc4a0188a Restore some code 2020-08-28 01:54:39 +03:00
Eladash 47b545282e
SPU: Fix events ACK, minor optimizations (#8771) 2020-08-27 21:36:54 +01:00
Eladash 841b8fad38 SPU: Fix timer events 2020-08-24 01:57:32 +03:00
Eladash 0f8ca1f7c5
SPU: Implement RSX accurate reservations on TSX (#8721) 2020-08-13 00:00:37 +01:00
Eladash e52dd9dc6f
SPU: Implement SYS_SPU_THREAD_OPTION_DEC_SYNC_TB_ENABLE (#8657) 2020-07-30 14:01:25 +01:00
Eladash 82068cf802
SPU: Fix spu_thread::cpu_stop() missed executions (#8656) 2020-07-30 10:07:18 +01:00
Eladash a029a94c73 SPU: Use waitable atomics for SPU channels interface 2020-07-23 13:45:58 +03:00
Eladash f8d2d8ca11 SPU/Windows: Fix LS memory mirrors 
This is a workaround but this is because of how utils::shm works on Windows path.
 2020-07-19 17:58:49 +03:00
Eladash c37bc3c55c SPU: Make spu_thread::offset private 2020-07-19 17:58:49 +03:00
Eladash af1ceb1151 SPU LLVM: LS Memory Mirrors (Optimize loads/stores) 2020-07-18 02:01:33 +03:00
Eladash 235d12aa6b SPU MFC: Never clear tag status in WrTagUpdate 2020-07-10 02:52:02 +03:00
Eladash 5d1fc546a8 SPU MFC: Fix MFC_WrTagUpdate channel count 
Always report available, in realhw this is just a hint if the previous tag update hasnt been checked yet by the MFC, avoiding blocking writes and allowing the SPU to execute some code while it processes the previous update request.
Except for MFC_TAG_UPDATE_IMMEDIATE, where it also waits for MFC to process it.
 2020-07-10 02:52:02 +03:00
Eladash 84470c34db SPU: Disable PUTLLC NOP transfers detection on TSX path 2020-07-09 03:17:35 +01:00
Eladash f8dbfa1d1e SPU: Implement GETLLAR polling detection 2020-07-09 03:17:35 +01:00
Eladash dc25a3fa2a PPU debugger: Show stack address of each function 2020-07-06 18:58:16 +02:00
Eladash 2c93fecd8b SPU: Use named constants for MFC tag updates 2020-06-27 20:42:41 +01:00
Eladash f29589e5cf SPU debugger: Add atomic status and tag update channels information 2020-06-27 07:04:37 +01:00
Eladash 731d4330fe
v128: A few optimizations (#8432) 2020-06-15 17:24:04 +03:00
Eladash 5777a1d426 SPU: Implement EBUSY error on non-empty mailbox (sys_spu_thread_send_event/sys_event_flag_set_bit) 
Write into inbound mailbox under mutex.
 2020-06-15 17:08:57 +03:00
Eladash c15b5f1eca SPU: Move check_state() outside of mutex scope 
Can result in a deadlock in some cases, cpu flags are checked after this function as well anyways.
 2020-06-15 17:08:57 +03:00
sampletext32 437f374bae Fix some checks 2020-06-04 19:48:08 +03:00
Nekotekina abf9a08ee3 Fix warnings 2020-05-27 18:41:17 +03:00
Eladash 7c3166a0c6 SPU MFC: Fix MFC_WrListStallAck on interpreters 2020-05-20 22:55:30 +03:00
Eladash 4405f46aec SPU MFC: Fix SN interrupts 2020-05-20 22:55:30 +03:00
Eladash 81684919f5 SPU MFC: Implement MFC_SDCRZ_CMD 2020-05-20 22:55:30 +03:00
Eladash a2653532ef SPU reservations (TSX): Remove wait flag in PUTQLLUC 2020-05-17 14:20:21 +03:00
Eladash 61f43d78df SPU: Minor cleanup of exception in stop_and_signal 2020-05-14 16:58:50 +01:00
Eladash 54dd9f4eae sys_spu: Fix sys_spu_thread_group_terminate vs sys_spu_thread_group_exit race on values 2020-05-14 16:58:50 +01:00
Eladash 9266507e4c SPU: Implement spu_channel_(4_)t::try_read 2020-05-13 19:36:44 +03:00
Eladash 7ff25588f4 sys_spu: Minor cleanup of group termination process 2020-05-13 19:36:44 +03:00
Eladash 525453794f SPU/PPU reservations: Optimizations part 1 
- Implement vm::reservation_trylock, optimized locking on reservation stores with no waiting. Always fail if reservation lock bitsa are set.
- Make SPU accurate GET transfers on non-TSX not modify reservation lock bits.
- Add some optimization regarding to unmodified data reservations writes.
 2020-05-13 11:10:13 +01:00
Eladash f95b81574f
sys_spu: Fix race in sys_spu_thread_group_destroy and other minor fixes (#8182) 
* sys_spu: Fix race in sys_spu_thread_group_destroy and other minor fixes

* SPU: Wait for all threads to have error codes if exited by sys_spu_thread_exit

On last thread in group to run.

* sys_spu: Fix sys_spu_thread_group_start

* fixup ad fix sys_spu_thread_group_terminate

idk why "- !group->running" was put in the first place but its probably no longer relevant due to other changes and was causing other issues such as not always waiting for last SPU thread to set group state to INITIALIZED.
 2020-05-11 21:24:04 +03:00
Eladash 09797c3584 sys_spu: Improve sys_spu_thread_get_exit_status 2020-05-10 03:46:11 +01:00
Eladash 1bd6cb2105 SPU/PPU debugger: use ':' instead of '=' 2020-05-05 13:46:26 +03:00
Eladash edde748519 sys_event_queue: Fix forced event queue destruction 
Add missing last existence check at sys_spu_thread_(try)receive_event and lv2_event_queue::send.
 2020-05-04 01:10:19 +03:00
Eladash 0e6937a359 SPU GETLLAR: Don't use loop detection for TSX 2020-05-02 14:57:38 +03:00
Eladash 2b75df22d9 sys_event_queue: Fix ports disconnection after queue destruction 2020-04-30 18:58:42 +03:00
Nekotekina 19acf260d8 SPU DMA: Fix PUTLLUC (TSX) 
Prevent edge case of missing store.
 2020-04-29 15:40:41 +03:00
Eladash 954e3f6e6c Fixup for cpu_flag::pause state check after #8114 2020-04-29 05:56:47 +03:00
Eladash a505d87565 Partial revert of 3be687cd18 2020-04-28 20:20:19 +03:00
Nekotekina 790fd9ce14 SPU DMA: implement cmp_rdata_avx 
Use technique similar to mov_rdata_avx with inline assembly.
 2020-04-28 17:58:26 +03:00
Eladash 7da8ba5c15 Wait for SPU event to be received in sys_spu_thread_group_exit 
Also use atomic check for group->run_state outside the mutex,
explicitly forbid group termination while we are waiting for an event to be received.
 2020-04-28 14:58:17 +03:00
Eladash a9c18964b6 Add missing cpu state check sys_spu_thread_receive_event 2020-04-28 14:58:17 +03:00
Eladash 9506676223 SPU debugger: dump Stall Stat and SRR0 2020-04-28 14:27:40 +03:00
Eladash 7d7909149f SPU: Detect reservation spinning loop 2020-04-28 14:27:40 +03:00
Eladash e423128a32 SPU: New GETLLAR technique 2020-04-28 14:27:40 +03:00
Nekotekina 3ec73b651e SPU DMA: more tuning for mov_rdata_avx 
Avoid unaligned stores.
Prefer asm path if __AVX2__ is not set.
Don't emit vzeroupper if __AVX__ is set.
 2020-04-27 18:05:52 +03:00
Nekotekina 4f71c570bd SPU DMA: disable memcpy path 
Due to update of the alternative path (SSE/AVX)
 2020-04-26 22:36:55 +03:00
Nekotekina 8ae2554505 Implement mov_rdata_avx 2020-04-26 22:36:45 +03:00
Eladash 83c7f6f149 debugger: Rephrase "Current function" to "In function" 
Takes less space which makes actual function name display a bit nicer.
Also the meaning is clearer.
 2020-04-19 10:56:42 +01:00
Eladash a3f2dfa232 sys_isolated_spu 2020-04-17 11:41:50 +01:00
Nekotekina c7fe8567b8 Experimental squashing of reservation memory area. 
Enables trivial synchronization between shared mem.
Reduces memory usage, but potentially degrades performance.
Rename an overload of vm::passive_lock to vm::range_lock.
 2020-04-16 02:25:43 +03:00
Nekotekina b1b67a13c6 Revert "Replace rotate utils with std::rotl" (partial) 
This reverts commit 4d8bfe328b.
 2020-04-14 19:45:53 +03:00
Eladash ec1e82bc9d SPU debugger: Implement blocking functions dumping 2020-04-14 17:09:58 +03:00
scribam f37adc4188 Add fallthrough attribute 2020-04-14 17:06:58 +03:00
Nekotekina 4d8bfe328b Replace rotate utils with std::rotl 
More include cleanup.
 2020-04-14 16:05:58 +03:00
Eladash cb14805d78 rsx fp/vp analyzers: Fix strict type aliasing and improve codegen 2020-04-12 16:48:43 +03:00
Nekotekina 1b68f90e42 Tweak TSC calibration 
Round to 3 digits after dot (count in MHz).
 2020-04-09 16:23:33 +03:00
Eladash cce946f10e spu: typo fix 2020-04-08 19:23:13 +03:00
Eladash c0f4fb6e35 SPU debugger: dump reservation address 2020-04-08 14:35:44 +03:00
Eladash cc8f024c6c Fixup ppu/spu_thread::dump_all() 2020-04-08 14:35:44 +03:00
Megamouse b1fdbc7fcc Move some format functions 2020-04-06 20:59:58 +02:00
Eladash 13820d6802 SPU debugger: Show channels data 2020-04-03 18:37:21 +03:00
Eladash 0beea91d5e Minor debugger fixups 2020-04-03 18:37:21 +03:00
Nick Renieris 1231274e0f CPUThread: Split dump() info to separate methods 2020-04-03 01:36:35 +01:00
Emmanuel Gil Peyrot 425e032a62 SPU: Copy with memcpy() instead of hand-rolled SSE2 
In some very unscientific benchmark:
spu_thread::do_dma_transfer() was taking 2.27% of my CPU before, now
0.07%, while __memmove_avx_unaligned_erms() was taking 1.47% and now
2.88%, which added makes about 0.8% saved.
 2020-03-22 15:51:03 +03:00
Nekotekina 6a2571d0e1 SPU: print current chunk hash in dump 2020-03-18 18:28:46 +03:00
Eladash 9e14e835e7 Fix sys_raw_spu_destroy 2020-03-16 21:06:33 +03:00
Eladash c11074a128 RawSPU: fix race condition in RunCntl stop request 2020-02-29 21:54:54 +03:00
Nekotekina 65eeee0f4c Remove cancerous lf_value<> 
Replace thread names (generic, PPU, SPU) with new shared pointers.
Devirtualize cpu_thread::get_name (used in single case).
 2020-02-28 18:54:46 +03:00
Nekotekina 972e0ab31d Remove -Wno-reorder and make it an error 2020-02-21 15:20:34 +03:00
Nekotekina 92e3eaf3ff Fix signed-unsigned comparisons and mark warning as error (part 2). 2020-02-19 22:54:58 +03:00
Nekotekina 771eff273b First part of fixing sign-compare warning (inside be_t). 2020-02-19 22:54:58 +03:00
Eladash df8d0cde4a RSX/SPU: Accurate reservation access 2020-02-19 18:11:30 +00:00
Eladash 727d783959 RawSPU: protect NPC from writes/reads in running state 2020-02-18 18:09:10 +00:00
Eladash fad8b38b28 sys_spu: protect sys_spu_image members in kernel mode 
Save relevant info in idm, set sys_spu_image segs and nsegs members to 0.
 2020-02-18 18:09:10 +00:00
Megamouse fe75311be2 move config structs to own files and clean up some headers 2020-02-17 15:08:17 +03:00
eladash f901846acb RawSPU: execute MFC proxy cmd after reading CMDStatus 
Implement MFC proxy argument sequence checking.
 2020-02-06 20:43:38 +00:00
Eladash 37513b1898 SPU reservations: Do not access violate under vm::writer_lock 
TODO: Throw exception when encountering page faults notification enabled memory
 2020-02-06 00:27:17 +00:00
Eladash f8b3c48af7
sys_spu: Implement proper SPU group flags (#7320) 
* sys_spu: Implement proper SPU group flags
 2020-02-05 20:46:05 +00:00
Eladash 049e392a97 Make preferred spu threads dynamically adjustable 2020-02-05 10:06:07 +00:00
Nekotekina c0f80cfe7a Use attributes for LIKELY/UNLIKELY 
Remove LIKELY/UNLIKELY macro.
 2020-02-05 10:42:34 +03:00
Nekotekina 6dfd97f0b6 Modernize SPU logging (spu_log variable) and remove log legacy 
Remove legacy macro (LOG_ERROR, etc)
 2020-02-01 11:52:52 +03:00
Eladash fe381b8581 SPU: Add SPU LS to debugger 2020-01-21 16:45:41 +03:00
Nekotekina ddda09607d SPU: fixup for STOP 0w0 2020-01-21 16:32:00 +03:00
Eladash 9993df9b8b RawSPU: fix race between spu start and stop 
This race could lead to spu status bits indicate RUNNING status, but cpu state being stopped.
Fix it by making sure cpu state is set before spu status.
 2020-01-21 14:08:39 +03:00
Nekotekina 98a8eeaac2 SPU: properly support STOP 0x0 instruction 2020-01-20 23:40:10 +03:00
Eladash 765bd6b6c6 SPU: Optimize gpr reset for MSVC 2020-01-11 22:56:46 +03:00
Eladash b59a825e48 Minor fixup after #6894 2019-12-30 23:46:45 +03:00
Eladash 45cff1219c Allow sys_raw_spu_create_tag to be called more than once 2019-12-30 23:46:45 +03:00
Eladash 5631382623 sys_spu: Fix SPU Thread Id 
* Removed wrong code in sys_spu_thread_group_terminate.
* SPU Thread ID is accurate, including 5th thread id "rule".
* Fixed possible use-after-free access of spu_thread::group member.
* RawSPU ID management simplified.
 2019-12-06 19:59:29 +03:00
Nekotekina 185c067d5b C-style cast cleanup V 2019-12-03 17:23:00 +03:00
Markus Stockhausen cd6b6c8a4f Lightweight putllc() for non-TSX if no data changed 
This replaces the totally messed up PR #6728

Some games make heavy use of getllar() & putllc() without even changing data.
In this case avoid unneccesary heavy locking of the PPU threads on non-TSX
hosts.
 2019-11-19 18:10:29 +03:00
Nekotekina ccac9d4777 Remove throwing and catching cpu_flag::stop 
Since there is spu_runtime::g_escape function now.
 2019-11-08 19:27:11 +03:00
Eladash f41f5054f7 sys_spu Fixup after #6864 2019-10-29 23:13:16 +03:00
Eladash b99992d570 sys_spu: Fix SNR and Inbound Mailbox state reset 
Also remove bugged ESTAT check at sys_spu_thread_write_spu_mb.
 2019-10-29 18:34:28 +03:00
Nekotekina 9ac6ef6494 SPU: cleanup former OOM handling 
Remove cpu_flag::jit_return.
It's obsolete now, and worked only in SPU ASMJIT anyway.
 2019-10-26 21:24:12 +03:00
MSuih f3ed26e9db Small warnings cleanup (#6671) 
* Ignore more warnings

These are intentional

* Signed/unsigned mismatch when comparing

* Explictly cast values

* Intentionally discard a nodiscard value

* Change ppu_tid to u32

* Do not use POSIX function name on Windows

* Qt: Use horizontalAdvance instead of width

* Change progress variables to u32
 2019-10-25 13:32:21 +03:00
Nekotekina b329bb604c SPU LLVM: implemented asynchronous compilation 
Implemented interpreter-based pre-recompiler.
Interpreter functions are build with SPU LLVM.
 2019-10-21 19:29:34 +03:00
eladash 95752607ea sys_spu: Don't reset snr config at group_start() 
Also first check for EINVAL in sys_spu_thread_set_spu_cfg
 2019-10-16 21:11:29 +03:00
Nekotekina 49e96b39dd [SPU, TSX] Fix reservation corruption in PUTLLC 
Change reservation locking logic.
 2019-10-12 15:41:24 +03:00
Eladash c2278fb879 spu: Mask SRR0 at write 2019-10-08 02:52:33 +03:00
Nekotekina feee3838eb Revert "Revert "Remove shared_cond and simplify reservation waiting"" 
This reverts commit b70c08a2e8.
 2019-09-24 05:01:00 +03:00
Nekotekina b70c08a2e8 Revert "Remove shared_cond and simplify reservation waiting" 
This reverts commit 0a96497e13.
 2019-09-14 00:02:48 +03:00
Nekotekina 2fc8844315 atomic.hpp: add atomic wait mask support 2019-09-13 15:53:34 +03:00
Nekotekina 0a96497e13 Remove shared_cond and simplify reservation waiting 
Use atomic wait for reservations
Cleanup some obsolete code
 2019-09-10 19:25:39 +03:00
Eladash 0d88f037ff Add new accuracy control for PUTLLUC accuracy setting (non-TSX) 
With the option enabled GET commands are blocked until the current PUTLLC/PUTLLUC executer on that address finishes

Additional improvements:
- Minor race fix of sys_ppu_thread_exit (wait until the writer finishes)
- Max number of ppu threads bumped to 8
 2019-08-17 00:42:46 +03:00
Eladash 49aefc0795 Prefetch MFC list elements (#5345) 
* Prefetch mfc list elemets to protect from overwriting

Also move some stuff away from command processing such as a few constant arguments setup
 2019-07-24 01:13:45 +03:00
Nekotekina cb5c26f2b5 Fix SPU Interpreter regression after #6147 2019-07-15 16:34:34 +03:00
Nekotekina 22e4ef147a SPU TSX: fix "Preferred SPU Threads" 2019-07-14 17:33:20 +03:00
Eladash d7a2d42d8f MFC: Fix Tag Status report for sync/eieio/barrier commands 2019-07-10 17:35:39 +03:00
msuih 146e43b6ec Do not use negative unsigned literals 2019-07-01 04:33:23 +03:00
Eladash 43f919c04b Fixup after #6143 (#6146) 
vm::spu max address was overflowing resulting in issues, so cast to u64 where needed. Fixes #6145.
    Use vm::get_addr instead of manually substructing vm::base(0) from pointer in texture cache code.
    Prefer std::atomic_thread_fence over _mm_?fence(), adjust usage to be more correct.
    Used sequantially consistent ordering in semaphore_release for TSX path as well.
    Improved memory ordering for sys_rsx_context_iounmap/map.
    Fixed sync bugs in HLE gcm because of not using atomic instructions.
    Use release memory barrier in lwsync for PPU LLVM, according to this xbox360 programming guide lwsync is a hw release memory barrier.
    Also use release barrier where lwsync was originally used in liblv2 sys_lwmutex and cellSync.
    Use acquire barrier for isync instruction, see https://devblogs.microsoft.com/oldnewthing/20180814-00/?p=99485
 2019-06-29 18:48:42 +03:00
Eladash 1ee7b91646 Refactoring (#6143) 
Prefer vm::ptr<>::ptr over vm::get_addr.
    Prefer vm::_ptr/base over vm::g_base_addr with offset.
    Added methods atomic_t<>::bts and atomic_t<>::btr .
    Removed obsolute rsx:🧵:Read/WriteIO32 methods.
    Removed wrong check in semaphore_release.
    Added handling for PUTRx commands for RawSPU MFC proxy.
    Prefer overloaded methods of v128 instead of _mm_... in VPKSHUS ppu interpreter precise.
    Fixed more potential overflows that may result in wrong behaviour.
    Added io/size alignment check for sys_rsx_context_iounmap.
    Added rsx::constants::local_mem_base which represents RSX local memory base address.
    Removed obsolute rsx:🧵:main_mem_addr/ioSize/ioAddress members.
 2019-06-29 01:27:49 +03:00
JohnHolmesII be521ff0ab Fix warnings related to parentheses 2019-06-25 20:36:32 -07:00
kd-11 4ff77a8555 rsx: Improve balancing of the offloader thread 
- Use two counters to avoid atomic operations
- Yield instead of sleeping because some games are very sensitive to timing
 2019-06-25 20:50:54 +03:00
Lassi Hämäläinen 499035512b Split Emu/Memory into more logical headers 
- Add vm_locking.h and vm_reservation.h and move relevant functions
  and types to these headers.
- Change include order and make vm_ptr.h, vm_var.h and vm_ref.h headers
  usable invidually and them including vm.h instead of other way around
- Because usage of vm::ptr now requires including vm_ptr.h instead of
  vm.h updated multiple #includes
- Added additional #includes to vm_reservation.h and vm_locking to
  where vm::reservation_* and locking related functions are used
 2019-06-25 17:11:10 +03:00
Nekotekina b9b591bf02 Fix SPU Loop Detection 2019-06-20 14:46:32 +03:00
Nekotekina 5d45a3e47d Implement cpu_thread::suspend_all 
Remove Accurate PUTLLC option.
Implement fallback path for SPU transactions.
 2019-06-19 20:36:12 +03:00
Nekotekina 9dc0368079 Rename cond_x16 to shared_cond 
Extend capacity from 16 to 32.
Remove redundant m_total counter.
 2019-06-04 16:37:50 +03:00
scribam 790962425c Fix some "-Wpedantic" warnings 2019-06-01 22:59:23 +03:00
Nekotekina 71b71537a0 SPU TSX: implement Accurate PUTLLC option 
Allow spurious PUTLLC failure if disabled (default).
 2019-05-25 22:23:23 +03:00
Nekotekina b839cc9d5b SPU TSX: restore busy_wait in GETLLAR 2019-05-25 21:41:11 +03:00
Nekotekina 7de3c410cf SPU/PPU: update reservation logic on TSX path transactions 
Make use of lock bits in reservation counters.
On PPU, fallback to compare_and_swap instead of desperate retry.
On SPU, lighten write set on retry by 'locking' outside of the transaction.
 2019-05-20 14:32:50 +03:00
Nekotekina 9abb303569 vm: expand reservation lock bit area to 7 bit 
This is minor change.
 2019-05-19 17:46:55 +03:00
Nekotekina f33b81545e SPU: implement recompiler gateway function in assembly 
Use GHC calling convention directly for SPU object entry points.
This may address performance degradation after #5923.
 2019-05-14 22:15:04 +03:00
Nekotekina cc8c635855 SPU: PIC support preview 
SPU ASMJIT not supported yet.
Giga mode not supported properly.
 2019-05-14 22:15:04 +03:00
eladash 3a5f4ed757 Print SPU Group ID on the debugger 2019-04-20 20:43:58 +01:00
Nekotekina 9060177dbd SPU transactions: add SSE path if AVX is not available 
This handles hypothetical situation when AVX is disabled system-wise.
Also refactored register use, to match Windows path with Linux path.
This reduces read set a little at the cost of stack use.
 2019-04-16 23:49:18 +03:00
Nekotekina 3354f068fc PPU/SPU transactions: ease cache line interference (TSX path) 
Touch memory on the same memory page, but different cache lines.
 2019-04-10 13:58:12 +03:00
Nekotekina 71b88cdc82 New SPU interpreter (SPU fast) 
Use LLVM to build SPU interpreter.
Simplify interpreter loop.
 2019-03-27 20:33:44 +03:00
Nekotekina 4b381fbbb1 Implement spu_runtime::reset 
To handle JIT: Out Of Memory error.
 2019-03-23 02:43:41 +03:00
Nekotekina f143035af1 Fix sys_spu_thread_group_join wait condition 
After waiting, thread group cannot be safely accessed
Following #5643
 2019-03-01 00:08:19 +03:00
eladash 0861226271 Make more use of the new atomic_t<>::release 2019-02-10 00:16:57 +03:00
eladash e3ee481f01 Make sys_spu_thread_group_join return once per termination 2019-02-10 00:16:57 +03:00
Nekotekina 50922faac9 Remove SPUThread::jit_dispatcher 
Use global array - save memory
Move the array to JIT memory
 2019-01-29 03:32:16 +03:00
elad fc92ae4085 SPU/PPU atomics performance and LR event fixes (#5435) 
* Fix SPU LR event setting in atomic commands according to hw test
* MFC: increment timestamp for PUT cmd in non-tsx path
* MFC: fix reservation lost test on non-tsx path in regard to the lock bit
* Reservation notification moved out of writer_lock scope to reduce its lifetime
* Use passive_lock/unlock in ppu atomic inctrustions to reduce redundancy
* Lock only once for dma transfers (non-TSX)
* Don't use RDTSC in reservation update logic
* Remove MFC cmd args passing to process_mfc_cmd
* Reorder check_state cpu_flag::memory check for faster unlocking
* Specialization for 128-byte data copy in SPU dma transfers
* Implement memory range locks and isolate PPU and SPU passive lock logic
 2019-01-15 18:31:21 +03:00
eladash f19fd23227 spu: Fix support for multiple lists when one is stalled 2019-01-15 02:33:22 +03:00
Nekotekina cfdf50dcff SPU: ensure sys_spu_thread_group_join receives correct exit status 
Following #5334
 2019-01-13 14:45:36 +03:00
eladash 653a4ef0df Set group status INIT on last thread stopped 
this fixes the group status after sys_spu_thread_exit when not joining the spu group
 2018-12-25 19:59:41 +03:00
elad 90265edfcd SPU MFC: avoid copying of the lockline onto the stack in putllc/putlluc (#5392) 2018-12-04 21:09:55 +03:00
Nekotekina f442a8a84c SPU TG: add thread group stop counter 
Fix possible race condition introduced by waiting on `running` value
 2018-11-27 23:37:26 +03:00
Nekotekina febe4d4a10 Implement class cond_x16 
Use as reservation notifier
Limited to 16 threads but allows more precise control of contention
 2018-11-26 00:23:29 +03:00
Nekotekina 0044eb44e2 Cleanup after #5310 (SPU thread groups) 
Move lambda into a cpu_stop()
Use running thread counter to synchronize with sys_spu_thread_group_join()
Use SPU_STATUS_STOPPED_BY_STOP exclusively for sys_spu_thread_exit() as before
Remove unnecessary waiting in sys_spu_thread_group_exit()
Rollback some minor unnecessary changes
Use shared_mutex in SPU TG
 2018-11-14 12:50:24 +03:00