home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
AmigActive 22
/
AACD 22.iso
/
AACD
/
Resources
/
System
/
PPCUpdates
/
68060
/
68060.library.notes
< prev
next >
Wrap
Text File
|
1999-10-19
|
9KB
|
286 lines
ReleaseNotes:
Attention:
o If you use rom2fast the 68060/40.library now protects
the mapped rom as writeprotected. If somebody writes
on this memory you get an exception.
YOU MUST INSTALL CYBERGUARD TO HANDLE THAT SITUATION.
IF YOU THINK CYBERGUARD SLOWS DOWN YOUR SYSTEM DO THIS:
run >NIL: cyberguard NOZERO STACKLINES=8 SHOWDISS FILE=CNC:0/0/640/100/CyberGuard/Auto/Close
o Enforcer doesn`t work correctly with the 68060.library
because the 68060.library doesn`t like if some other tool
takes over the complete MMU control.
This is nobody`s fault. M. Sinz designed Enforcer in the
sense of the old passive 68040.library API. What API ?
it had none.
The Phase5 library does special MMU mappings for certain
features and has an API which is used by its applications.
(CyberGuard,CyberSnooper,Rom2Fast...ppc.library and so on)
If you use it be warned...a lot things won`t work anymore
or simply crash.
o You need the original Commodore math libraries and not
some custom libs found on aminet ! The reason is that
i have to patch the library base and don`t wanna check
out every mathlibrary incarnation if it supports the
math precision hook.
o Something disables the instruction cache on some machines before
it runs the startup-sequence. Can`t reproduce it here
o 46.7
o don`t init the 68060.library if it detects a lc/ec modell.
o some wrong #ifdef MMU060 instead of MMU040 had the consequence that
the utility.library isn`t patched anymore for the 64bit math functions.
The only sideeffect would have been that programs which used
it would need the emulation.
o 46.6
o Half Inst/Data Cache cleared.
There are some cpus out there(know about 2) which have the half inst
bit set by default though the user manual says clearly that these
bits are 0 after the reset.
o Release
o 46.5
o Small internal change
o Release
o 46.4
o PCMCIA fix. Because of the very early init of the 680x0rom.library
the PCMCIA card.resource wasn`t there yet and so the memory wasn`t
mapped. This update actually only has an influence for a rom/qrom
version.
o Release
o 46.3
o Release
o 46.2
o Yet another rework to support a 680x0quick.library rom module
which has no fpu code. That fpu related code is then inserted
by the libs:680x0.library and the 680x0quick.library transforms
into the RealThing(TM).
(experimenting with faster boot/quick enable of the data cache to
boot faster)
o AddMemList() now also adds the new memory to the MMU Table
o AddConfigDev() now also adds the new io area to the MMU Table,
*IF* the ConfigDev doesn`t point to ram(which may be used in
fake configdev`s rom magic)
o speeded up MMU table creation
o for the rom versions it now starts before the STRAP.
I *really* hope this doesn`t give any trouble with bad driver
software.
o cacheflush added to ColdReboot() to make sure residenttags are
in ram before the machine resets. Early Boot Screen
o removed the clr 4.w in the OS coldreboot replacement which
also caused problems with the early boot screen
o 46.1
o cachepre/post problem FINALLY fixed..puhh
o little pci board detection change..caused some problems
o 46.0
o FPUExceptionsMode function..to disable fpu exception propagation
completely. Fusion compatability mode.
o 45.3
o global fpu exception disable option for fusion
o 45.2
o now fpu exception jmp into the OS exception if the exception is
enabled. Otherwise it`s ignored->wrong results...but well..MacApps:-)
o 45.1
o Operr ignored now...it seems Netscape causes this exception but
works further even with wrong fpu data....GREAT:-(..(Fenton)
o Ignored **fake** scsi configdev entry
o PCI Board detection rewritten
o 45.0
o Converted most asm modules to C
o Internal cleanups
o reworked DMA handling
o 44.6
o silly ppc bus detection bug
o setcachemode changes
o rework of the ppc board detection..that`s probably the cause for the
ppc.library crash on non ppc boards
o 44.5
o 8MB more addressspace for the PPC/MK3 gfxcard
o fix of the MK3 256 memspace for the gfxcard
o 44.4
o sorry..had a clr.l 4.w in my new MK3 reset code..which stopped
all resident software which used coldreboot and started after
the 68060.library.
o 44.3
o CAS2 (fix?) for some mac programs which use this illegal instruction
on the amiga. I cleaned the emulation from unnecessary instructions
to emulate the read-write modify locked cycle cas,tas,cmp are using
which is illegal on an Amiga.
o 44.2
o Setcachemode now also handles non existing memory
(ATEO concept`s Z2 board)
o 44.1
o PowerUP Bus extensions
o 43.5
o CyberStormPPC/MK3/BlizzardPPC special reset code...
o Fix in a CyberGuard related internal function
o 43.4
o different 68040 fpu test
o 43.3 Release
o Yet another internal changes
o 43.2
o Internal changes for the coming rom version and to support LC version
o Small blizzard change at 0xf50000..is now mapped
o 43.1
o Map cyberstormppc/mk3/blizzardppc flash area as non cachable
o 43.0 Release
o 43.0Beta4
o Fixed some silly bug in the rom map code introduced
with some changes lately
o 43.0Beta3
o Fixed some exception handler problem
o 43.0Beta2
o BlizzardPPC support
o 43.0Beta1
o Internal changes for Cyberguard writeprotect handling
o 43.0Beta0
o Finally a reason to add a CPU check code:-)
o some private goodies addons for cyberguard
o 42.1
o CyberStormMK3 support
o 42.0
o CyberStormPPC support
o Changed Chipram to nonsynchronized...well..if problems occur
contact me.
o 41.6
o Final Release(Well..not really)
o 41.5Beta12
- Fixed Picollo64 in Zorro3 mode mapping problem..it wasn`t set
into nonserialized mode.(Niels Knoop)
Worked in Z2 mode though.
- improved CV3d mapping
o 41.5Beta11
- Hopefully fixed a rom2fast random crash problem
o 41.5Beta10
- Fast Screen Switching CyberGfx3.0 support
- Fixed gfxboard cybervision bug introduced with 64 mapping.
(Forgot an else)
o 41.5Beta9
- Call old Buserror routine on a non branch buserror instead
of ignoring the buserror.
o 41.5Beta8
- Shapeshifter CyberGfx3.0 support
- DoubTransFieee(),DoubTransTieee() replaced
o 41.5Beta7
- Added Rainbow3 support
o 41.5Beta6
- Fixed rommapping problem i added in Beta4:-)
o 41.5Beta5
- Fixed PicassoBoard detection
o 41.5Beta4
- Added optimized pow function with error correction
- Rommapping only works if nothing was mapped before
- AmiWinDemo hangs itself and system BUG fixed...
o 41.5Beta3
- Added Cybervision 3d support
o 41.5Beta2
- ixemul bug rework
o 41.5Beta1
- Added PowerUP Support
o 41.5Beta0
- Betarev counter activated
- Some checks to make it more 060Rom Exec aware if there'll be
one ever.
o 41.5
- Added a nonserialized nocache marking of known gfxboards for Z2
and Z3 space. Picasso,Spectrum,Picollo,Picollo64,Cybervision
are recognized.
- Removed some unnecessary debug code
o 41.4
- Some internal reordering to hopefully fix the Beta 68060.lib crash
problem some testers observed.
o 41.3
- mathsinbas,mathdoubbas patched now
- added the 881 precision hooks
Fixes Maxon Cinema 3.1 trapv
- sincos bug caused by wrong fpulib emulation documentation
Fixes Maxon Cinema 3.1 wrong calculations
o 41.2
- sincos library patch was broken.
Fixes Maxon Cinema 3.1 crash
o 41.1
- Branchprediction Buserror activated..sorry i really forgot to activate
it. The Cyberguard buserror handler took care of this so Term 4.6
worked there.
Something like that happens if the thing is done almost a year until
it was released with a 060 card.
Fixes Term4.6 problem.
The reason for the problem was that SAS generates a nop/skip instruction
like
cmp.w #$xyz,d0 where xyz was jumped at later..(mostly in a loop)
This caused the branch prediction buserror and the handler has to
clear the branch cache.
- Now maps the Zeropage into the Fastram. Has only an effect for people
which don`t use Cyberguard.
o 41.0
- New function SetCacheMode() for setting Cache modes for old Z2 Boards
- Now utility and mathsintrans;mathdoubtrans are patched
- automatic bumprev removed because it confused some Phase5 minds.
o 40.325
- Now uses the latest Motorola 060 emulation modules
o 40.3xx
- Activated CachePre and Post..forgot to activate it
- Cosmetic Changes to the SetRomAddress function
- Rom2Fast
o 40.300 - Release