mmu.library and CPU libs updated

[guest11527]

Hi folks,

in a couple of days, new versions of the mmu.library and the CPU libraries will appear in the Aminet. If you received a version just a day ago, then this update will again include a couple of minor fixes, thus probably just wait a bit until downloading again.

The mmu.library 43.13 (release 43.13.1) includes a workaround for some Apollo specific tools (MuDrap) which caused a hang on software-initiated cold reboots of the system. If the mmu-driven CPU libraries are used, the MMU-Configuration file is no longer required on well-behaiving hardware - but it does not hurt to simply keep it. For P5 hardware, P5Init and the MMU-Configuration is still necessary.

The CPU libraries (new release, to appear in a couple of days) also mirror these changes, and include a couple of cosmetical fixes in their startup code. No major changes for the 68060 and 68040 lib.

The 68030.library includes one additional workaround for MMU-less (680EC30) systems for an infamous bug of the 68030 caching logic. This bug can be either fixed by the MMU or by the "transparent translation" registers. The 68030.library uses now whatever logic is available, preferably the MMU.

For details, please go into the manual.

[mfilos]

WOW! Thanks a lot Thomas!

I was curious seeing all these Aminet individual modules of MMULib package 2 days ago having the the same version numbering with the ones in the standard MMULib archive.
Will wait patiently and check it out!

Thanks again for your tremendous amount of work \o/

[lionstorm]

I saw that too on Aminet : plenty of MU tools uploaded letting me think a lot which one to get for my blizzard1260.
Thanks Thomas for continuing updateing your tools !

[Lurch]

Awesome stuff, will have to update the A1200T when I get time :-)

[klx300r]

Many thanks Thomas !

[lionstorm]

Many thanks Robert !

hum you mean Thomas right ?

[Retrofan]

Wonderful! Thanks a lot Thomas

[XDelusion]

So let me get this right... this might help Apollo 060 users with issues they might be having setting up Shapeshifter? That being a side effect and not necessarily the main goal of the project of course.

[klx300r]

hum you mean Thomas right ?

lol yes indeed I meant Thomas:)

[matthey]

@ThoR
Thanks for the update. Everything seems to be working here so far.

I noticed there is an old version of the disassembler.library in the MuTools directory of the MMULib.lha version 43.11. Is there a purpose for this or could it be removed for the next version?

I know of a problem that is not specific to the current release but maybe someone can help. I know someone who is having trouble using the X-Surf 100 board with the Mu libraries. This is his system:

4000T with 3640 68040
X-Surf 100 board in a Zorro III slot
Manufacturer=4626
Product=100
BoardAddr=$40000000
BoardSize=$00010000 (64kB)

MuScan reports:

68040 MMU detected
MMU page size is 0x1000 bytes

Memory map:
0x00000000 - 0x001FFFFF CacheInhibit
0x00200000 - 0x00BBFFFF CacheInhibit Bundled to 0x07076000
0x00BC0000 - 0x00BFFFFF CacheInhibit I/O space
0x00C00000 - 0x00D7FFFF CacheInhibit Bundled to 0x07076000
0x00D80000 - 0x00DFFFFF CacheInhibit I/O space
0x00E00000 - 0x00EFFFFF CacheInhibit Bundled to 0x07076000
0x00F00000 - 0x00FFFFFF
0x01000000 - 0x06FFFFFF CacheInhibit Bundled to 0x07076000
0x07000000 - 0x07FFFFFF CopyBack
0x08000000 - 0x3FFFFFFF CacheInhibit Bundled to 0x07076000
0x40000000 - 0x4000FFFF CacheInhibit I/O space
0x40010000 - 0xFFFFFFFF CacheInhibit Bundled to 0x07076000

The next to last entry above seems to be correct for the X-Surf 100 but a requester pops up for him saying "Datacache is enabled, please use special X-Surf device with data cache disabled!". The C= and P5 68040.library work fine. He sent an e-mail to tech support and received this reply:

Hi,

Not sure what that version of the 68040 lib is doing - it *should*
disable datacache through MMU table for the X-Surf-100.

The 68030 is a little different - we can't switch off datacache through
MMU with that CPU, so the device complains about using the special lib.
There's two possible causes:

a) CPU recognition goes wrong, and the device thinks that it's running
on a 68030 processor.

b) the 68040 lib that you're using is not disabling datacache for the
x-surf-100.

ciao,

Jens Schönfeld

I write in public if case anyone else has a similar problem or has found a workaround. Thanks for any help.