| |
tlr
Registered: Sep 2003
Posts: 1790 |
RR memory map during freeze?
I'm a bit confused about the memory map in the freeze state. Unfortunately I do not have a real RR so I can't analyse it myself. Anyone cares to enlighten me?
Jens explicitly states in Inside_Replay.txt that no RAM can be visible in the $8000-$9FFF area during freeze, but that $DE00 (or $DF00) is possible.
Quote:Freeze:
ROM is mapped to $e000-$ffff, bank 0 is active directly after Freeze. Writing to bits 0 and 1 of the $de00 register will have no effect on GAME and EXROM.
RAM can be selected and used in $df00 or $de02, respectively, but not in $8000. Banking bits work, so you have full read access to the ROM, and access to up to four RAM pages with the AllowBank bit set (minus 2 bytes if REU
compatible bit is set). Cz states this in his Inside_Replay_Essentials.txt:
Quote:Being in freeze mode allows ROM banks to be mapped by $de00 as before but of course they are mapped to $e000. RAM can only be accessed on the free I/O1 area. In spite of this it seems vice (r30505) has RAM enabled at both $8000 and $DE00 (or $DF00) during the frozen state before acking it.
From examining the 1541U2 code (3.0beta5) it looks like no RAM can be visible at $DE00 (or $DF00) during freeze which also seems to be wrong.
I was hoping that RAM at $DE00 was mapped in automatically in the freeze state so I could save regs there to avoid trashing the stack unnecessarily.
Which is it? |
|
| |
chatGPZ
Registered: Dec 2001
Posts: 11386 |
drop a testprog into the repo and i'll check it on the real hw :) (you want to check if ram can be read at $8000, and if it can be written to at either $8000 and $e000). iirc its like jens' description says.... the existing tests omit the freeze mode, so i wouldnt be surprised if that is broken in VICE. |
| |
tlr
Registered: Sep 2003
Posts: 1790 |
The really interesting bit is whether RAM is at $DE00 already from the start. Jens just states that "RAM can be selected".
I'll see if I can whip up a small test program. |
| |
tlr
Registered: Sep 2003
Posts: 1790 |
There: testprogs/C64/carts/rr-freeze/.
The test first lets you select $DE01 bits, then tries to fill $DExx and $DFxx with an increasing pattern. It will then allow you to freeze.
Upon freezing, $DExx and $DFxx will be preserved to C64 ram, then modified and preserved again. Lastly freeze mode will be acked and $9Exx + $9Fxx will be read and finally the whole lot will be dumped to screen.
I believe $DE01=$40 and $DE01=$00 are the interesting cases. |
| |
Count Zero
Registered: Jan 2003
Posts: 1932 |
Will see to check your test commit at the weekend.
*Yieks* the essentials.txt looks quite wasted :)
Cannot really remember the setup when in "freeze state" and only the "maxmachine memory setup" is active with an RR attacged. Also emulations on various hardwares will likely largely differ. That is actually one of the test cases I wanted to cover on future Action Replay & Clones Acid Ramtest v0.1 versions. Never really got on to it - as always :(
Also I am not sure about the influence the mode selection using the AllowBank bit would make but whenever that makes a difference we know whom to beat down excessively :)
Will report back whenever I find time for some more tests as well. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11386 |
i can see differences between emu and real hw.... however, it would be nice if you could explain what exactly this hexdump is trying to tell me :) what should it look like when there is ram and what should it look like when there is none? |
| |
tlr
Registered: Sep 2003
Posts: 1790 |
Quote: i can see differences between emu and real hw.... however, it would be nice if you could explain what exactly this hexdump is trying to tell me :) what should it look like when there is ram and what should it look like when there is none?
@gpz: hmm, yes maybe it's a bit to crude. :)
Basically each "sample" is presented in hex rows of two.
Initially $9Exx and $9Fxx is prefilled with an incrementing pattern. (differnt for each).
First two shows the $de00 area directly after freeze.
The next two shows the $de00 area after trying to eor that area with #$ff.
the next two shows the $de00 area read back through $9e00 in the normal memory map.
After this six more rows, but for $df00 instead.
If REU-Comp is set only the upper part is meaningful, the lower part should be unconnected space.
If REU-Comp is not set, only the lower part is meaningful.
Did that make any sense at all?
This was just a quickie to verify if RAM is present already at the freeze entry. I will try to improve the test to check for RAM/ROM/banking in a more generic way. |
| |
chatGPZ
Registered: Dec 2001
Posts: 11386 |
check the readme i added in the repo... if i interpreted the results correctly there is indeed no ram in i/o space in freeze mode. |
| |
tlr
Registered: Sep 2003
Posts: 1790 |
Interesting! Thanks for interpreting my cryptic descriptions and deciphering the test program output. :)
I noticed you tested on the Nordic Replay btw. Supposedly the control regs on that is slightly altered compared to Retro Replay just to accommodate the Atomic Power. I assume the alterations are just extra stuff. Do you know which by any chance? |
| |
Count Zero
Registered: Jan 2003
Posts: 1932 |
Huh - after querying Grp soon will try to make heads and tails of this on RR/NP/U1541v1 - looks extensible.
And wow - there IS a difference on Atomic to (new) Nordic Power (clone)?
Or is it some implementation thing? Got an original Atomic/Nordic Power here to check against the "clone Nordic Power" carts Jens produced. Will report back ASAP. |
| |
tlr
Registered: Sep 2003
Posts: 1790 |
Quote: Huh - after querying Grp soon will try to make heads and tails of this on RR/NP/U1541v1 - looks extensible.
And wow - there IS a difference on Atomic to (new) Nordic Power (clone)?
Or is it some implementation thing? Got an original Atomic/Nordic Power here to check against the "clone Nordic Power" carts Jens produced. Will report back ASAP.
What I ment is that there is a difference between RR and Nordic Replay. Jens states so here: http://wiki.icomp.de/wiki/Nordic_Replay , but he doesn't state what the difference is.
Meanwhile I extended the freeze test to try to present the detected mapping in a more sane way. Ongoing work... |
... 8 posts hidden. Click here to view all posts....
|
| Previous - 1 | 2 - Next |
