Gillou68310
5d281a221c
Fix OSAL_SSE not defined
2021-02-24 16:49:34 +01:00
M4xw
38f97a0447
[new_dynarec] Trivial perf improvement on initial recompilation
2021-02-23 15:40:37 +01:00
Richard Goedeken
51d24dddba
Merge pull request #791 from Rosalie241/datapath
...
Allow front-ends to override user paths
2021-02-20 08:45:35 -08:00
Rosalie Wanders
3ab7a88917
Allow front-ends to override user paths
2021-02-20 10:40:50 +01:00
Richard Goedeken
23cabde67f
fix issue #812 . since CRC matching can be ambiguous, we return no match if the given CRC matches more than one entry in the database. This will cause a match to rely on MD5 instead in this case, which should be fine because every database entry has a unique MD5 hash
2021-02-19 19:15:17 -08:00
Richard Goedeken
478b94682f
properly remove executable permissions from new dynarec jit cache block in new_dynarec_cleanup()
2021-02-18 19:17:46 -08:00
Richard Goedeken
8a9e6bbd54
fix memory mapping operations for New Dynarec cache on ARM64 and x86_64 CPUs: you cannot call munmap() on a memory region which was statically allocated in the data segment of your process; thats just a bad idea
2021-02-09 22:47:06 -08:00
Bobby Smiles
7bceeb572a
Fix some warnings about wrong pointer format "%p" needs a void*
2021-02-02 19:22:30 +01:00
Bobby Smiles
53d2e92948
Fix some unsigned/signed comparison warnings
2021-02-01 09:50:50 +01:00
Bobby Smiles
16a9144c39
Change mempak formatting
...
This should better match what the underlying file system expects and is
in accordance with libdragon formatting procedure.
Previously we were writing 0x0003 (=Free page indicator) until the end
of mempak, but this should only be done in pages 5-128 slots.
The remaining content of the mempak is zeroed.
2021-02-01 09:46:52 +01:00
Bobby Smiles
f436c6ea41
refactor lba_end computation to better convey intent
2021-01-31 12:00:49 +01:00
LuigiBlood
942533a94e
Add support for 64DD saving
...
- it supports both Full disk copy and RAM Area only
(exception for MAME disk which only support full disk copy).
- User configure such option using Core/SaveDiskFormat option
2021-01-30 17:00:06 +01:00
LuigiBlood
022014d194
Fix D64 LBA RAM End Info Expand (puts invalid LBA 0x10C4 instead of LBA 0x10C3)
2021-01-30 16:53:52 +01:00
Rosalie Wanders
3e24b42c80
Add prefix to m64p_rom_save_type members
2021-01-23 12:24:51 +01:00
Richard Goedeken
36eaa7cafe
Merge pull request #798 from Rosalie241/save-slot-bindings
...
Allow customizable save slot keybindings
2021-01-22 21:00:13 -08:00
Richard Goedeken
17b73729ee
Merge pull request #788 from Rosalie241/settings
...
Add M64CMD_ROM_SET_SETTINGS
2021-01-22 19:58:35 -08:00
Bobby Smiles
bc0ea8169b
AI always works 16-bit stereo samples whatever the BITRATE is ...
...
So we can remove the "bits" parameter from audio backend set_format
method and rename it to set_frequency which is more appropriate.
BITRATE purpose is to specify the rate at which serial audio
bits are transfered to the audio DAC (BU9480F). Not to specify the
sample depth.
2021-01-17 11:43:12 +01:00
Bobby Smiles
13acbe3016
Fix missing last_read = 0 in ai_end_of_dma_event
2021-01-17 11:38:15 +01:00
LuigiBlood
497c5298cb
[Disk] Do not use magic numbers for System Area check
2020-12-26 18:59:49 +01:00
LuigiBlood
884440b754
[Disk] Make sure the wrong System Data is checked upon read of the System Area (Regression)
2020-12-26 18:17:21 +01:00
bsmiles32
6d7f88a33f
Merge pull request #803 from bsmiles32/storage_save_chunk
...
storage backend save method nows know about modified byte range
2020-12-12 15:29:09 +01:00
Tim Gates
13d3e01189
docs: fix simple typo, unconditonal -> unconditional
...
There is a small typo in src/device/r4300/new_dynarec/new_dynarec.c.
Should read `unconditional` rather than `unconditonal`.
2020-12-12 06:21:37 +11:00
carnivoroussociety
cf615aa446
Rewrote overclock implementation
...
Rename enable_overclock to count_per_op_denom_pot
Renamed cfg option
2020-12-08 05:42:03 +00:00
Bobby Smiles
e57c70fb97
storage backend save method nows know about modified byte range
...
Also, for file_storage_backend, we need to do a full save on the first
write to guaranty that file content matches current in-memory data.
Further calls to save will only update chunks being modified.
2020-12-06 23:43:46 +01:00
Bobby Smiles
6049618def
Rework flashram implementation
...
By combining the following sources :
- reverse engineering of Majoras Mask flashram functions [1]
- @sanni's cart reader [2]
- and MX29L1611 datasheet ("close enough" flash memory from Macronix)
I was able to figure a better model for flashram than what we currently
have. As this new modelling of flashram doesn't match one-to-one with
old model, meaning it will possibly "break" older savestate,
I had to bump savestate version.
[1] pending PR in https://github.com/n64decomp/majora/
[2] https://github.com/sanni/cartreader/blob/master/Cart_Reader/N64.ino
2020-12-06 10:54:08 +01:00
bsmiles32
8ae539f610
Merge pull request #762 from LuigiBlood/diskupdate
...
[Disk] Major Disk Support Changes
2020-12-06 10:53:12 +01:00
Bobby Smiles
0062351e59
Refactor load_dd_disk
...
- Extract scan_and_expand_disk_format from load_dd_disk
- Extract get_disk_format_name
2020-12-05 12:38:33 +01:00
Bobby Smiles
51886e1b1b
Call GenerateLBATOPhysTable inside dd_load_dd_disk
...
prefer dd_disk forward declaration to disk header inclusion
2020-12-05 11:06:49 +01:00
LuigiBlood
ae26f5de2b
[Disk] Change DiskSaveFormat description according to current implementation
2020-12-05 01:51:12 +01:00
Bobby Smiles
9a450b6eef
Use proper save names (path + ext) for 64DD disk saves.
2020-12-05 01:21:24 +01:00
Bobby Smiles
b098adc1ae
Refactor load_dd_disk
2020-12-05 01:13:07 +01:00
Bobby Smiles
522dd2f43c
Fix warnings about potentially uninitialized variable
...
In this case it was a false -alarm from GCC, but forcing 0 doesn't hurt.
2020-12-05 01:13:07 +01:00
LuigiBlood
60d6a9f66e
[Disk] Optimize RAM seek for D64 seek
2020-12-03 20:27:01 +01:00
LuigiBlood
269e127085
[Disk] Fix Write freeze
2020-12-03 18:13:54 +01:00
Bobby Smiles
2747266b45
Silence DD_MS_RAM read warnings
2020-12-03 12:44:17 +01:00
Bobby Smiles
e7474a44bc
Make ASIC_CUR_TK and ASIC_CUR_SECTOR read-only
2020-12-03 12:38:44 +01:00
Bobby Smiles
83949db50a
Make DD_ASIC_CUR_SECTOR more like hardware (data in upper 16 bits).
...
I had to adapt the sector usage in several places code was expecting a
sector in the [0,90[ range and not the "full" sector number in the
[0,179[ range.
This way I could get rid of bm_block intermedaite variable.
2020-12-03 12:21:03 +01:00
Bobby Smiles
87ca7bb4f5
Make DD_ASIC_CUR_TK more like hardware (data in upper 16bits)
2020-12-02 23:49:18 +01:00
Bobby Smiles
749a17ba28
Developement disk have a 0xC0 sector size for head=0,track<6
2020-12-02 23:49:02 +01:00
Bobby Smiles
1f4260f97f
Refactor disk interface to better decoralate drive access from disk geometry
2020-12-02 22:31:22 +01:00
Bobby Smiles
a986d23cb3
Move seek_track logic to disk module
2020-12-01 22:55:05 +01:00
LuigiBlood
09f25ca7a0
Merge pull request #3 from bsmiles32/diskupdate
...
Diskupdate
2020-12-01 19:25:18 +01:00
LuigiBlood
7ce6da5028
[Disk] Fix D64 RAM Save Loading + Move DD Disk Debug Info
2020-12-01 18:39:18 +01:00
LuigiBlood
8673068901
[Disk] Load Saved Disk Data
...
- Keep offset to ram for future use for saving
2020-12-01 18:30:18 +01:00
Bobby Smiles
7b73d2db68
fix GCC warning about maybe-uninitialized data
...
The warning ws a false-positive from GCC.
Still it doesn't hurt to zero these variables and avoid the warning.
2020-12-01 18:22:19 +01:00
Bobby Smiles
9b11cecb96
fix printing of already free'd filepath in savestate module
2020-12-01 18:22:19 +01:00
Bobby Smiles
8c9a072cfc
Move Disk helper functions to disk module
2020-12-01 18:04:46 +01:00
Bobby Smiles
5a906492d0
Readd empty line.
2020-12-01 17:33:53 +01:00
Bobby Smiles
d9da47de5b
Handle DD disk format differences through a specific storage backend
2020-11-29 19:28:28 +01:00
Bobby Smiles
84f30aa0fd
Remove extra data from storage_backend_interface
...
Storage backend is only meant to abstract storage accesses, not format.
Pass all the format information as regular parameters to dd controller.
2020-11-29 12:34:44 +01:00