Page 1 of 21
TF360 Status
Posted: 28 Jul 2019 16:47
by terriblefire
I sat down at my bench today for the first time in a month to try and see where the TF360 is at.
So far as I can tell the major functions all work but there remains some glitch when the MMU is activated.
Essentially SetPatch fails (violently crashes the machine) when running against MMULib (its been updated etc).
I believe the issue is hardware rather than software.
If MMU is disabled with a jumper SetPatch does not crash but obviously it doesnt do anything.
I believe i fixed the cache glitch issue I had before. I tried a disk image supaduper send me but it crashes in WinUAE and the TF360.
I think its time to figure out a minimalist MMU setup to try and see the issue.
Re: TF360 Status
Posted: 28 Jul 2019 20:23
by terriblefire
From a bit of digging it looks like the MMU tables arent being persisted correctly in fastram.
Re: TF360 Status
Posted: 29 Jul 2019 20:05
by Higgy
It's good that you identified something in your hunt. Hopefully that will narrow down the corrective fix.
I guess with the 060 having a big footprint, it makes a nice tea mug warmer while debugging :lol: Although this summer so far its been more cold drink weather rather than tea and a rain coat!
Re: TF360 Status
Posted: 30 Jul 2019 20:32
by terriblefire
So i dig further.
Refer to page 52 of this manual.
https://www.nxp.com/docs/en/data-sheet/MC68060UM.pdf
MMU Table Search Access I expect with TT=01 or 00 and TM = 100 or 011.. correct?
Now see below...
mmu_access_maybe.JPG
AFAICT this is a Function Code 3 access? WTF is that? and why is it at the MMU table address..
Re: TF360 Status
Posted: 30 Jul 2019 22:57
by 8 Bit Dreams
I am sure Thomas can answer this, though im afraid he's don't read this forum.. We need to ask on A1k.org get a quick answer, there is even MMU thread (Vampire) i can ask there if You like...
Re: TF360 Status
Posted: 30 Jul 2019 23:01
by terriblefire
I emailed Thomas with this.
Re: TF360 Status
Posted: 30 Jul 2019 23:07
by 8 Bit Dreams
Other method is to hook Your Logic Analyser to Chuckies A3660 and see what's going on there as A3660 is working as it should...
I assume the secret is hidden in the GAL Chips, we surely missing something...
Re: TF360 Status
Posted: 30 Jul 2019 23:33
by terriblefire
DOnt think that will help. The issue is only for fastram. Without fastram everything works fine.
Re: TF360 Status
Posted: 31 Jul 2019 05:29
by 8 Bit Dreams
Oh...that's fair enough, A3660 doesn't have any ram on it...
Re: TF360 Status
Posted: 31 Jul 2019 13:52
by terriblefire
I have a Apollo A1260 but the kit isn’t too happy being plugged in and unplugged over and over.