Mise à jour MiSTer et changelog du 2022-07-27 14:42:13

PSX:

Games affected by fixes:
Gouketuji Ichizoku 2, Eggs of Steel, Dragon Ball Z, Final Fantasy VI, Final Fantasy IX, Ape Escape, Aconcagua,
Slam Dragon, Rascal, Time Bokan Series – Bokan GoGoGo, Final Fantasy Tactics, Vigilante 8, DoDonPachi, X2 – No Relief

new features:
– added pause from CD slow -> will pause the core if data delivery from e.g. SDcard is slow instead of possibly crashing the game (can be disabled in OSD)
– add function for autoincrement savestate slot when saving
– add 24bit dithering option for VGA out

changes/fixes:
– Mixed: updated sys/framework
– Mixed: remove start button from savestate save/load combo
– Mixed: Digital pad can now use left Analog Stick as virtual DPad
– Mixed: improve timing closure for spuram to savestate readback (second sdram only)
– Mixed: added timeout to still pause the core if a game requests CD or SPU but doesn’t read the requested data
– Mixed: show saving info overlay if OSD is closed
– Mixed: removed pause from heartbeat
– OSD: option for CD slow overlay: show only slow reads or slow reads and seeks
– OSD: autosave when opening OSD is now renamed and default
– OSD: show save status in OSD (not saved, saved, saving)
– CD: always accept reset cmd, even with second response irq pending
– CD: don’t ignore pending speed change time when going into seek or read
– CD: don’t ignore pending seek time when read/play is started while seek is still ongoing
– CD: remove periodic driveDelayNext where it’s not required
– CD: don’t fetch sector from HPS again if last fetch was from same sector
– CD: allow commands to execute while CD sector fetch from HPS is ongoing (fixes hangs when HPS is busy)
– MDEC: fix softreset from CPU write
– CPU: fix design flaw that didn’t allow to pause CPU for DMA when CPU is constantly doing memory accesses
– GPU: clear texture cache when drawmode texture width changes
– DMA: increased CPU to DMA working ratio when transmitting data from VRAM to RAM
– GPU: fix video out read offset when framebuffer in vram is not 4 pixel aligned
– GPU: wait for previous draw calls pixels to be fully written before next readback
– GPU: write pixels to 64bit word aggregation timeout can be skipped if pipeline is done(speedup)
– GPU: allow output of videodata while core is paused(fixes glitches when unpausing)
– GPU: restructure most GPU modules to complete VRAM request even when core is paused
– GPU: don’t reread fetched vram data after unpausing
– GPU: skip drawing line primitives that are fully outside the drawingArea

lien source