Add NES UNROM-512 mapper support, mapper defines to subtargets. #191
Conversation
asiekierka
changed the title
|
Fixes #173 |
mos-platform/nes-unrom-512/bank.c
Outdated
| } | ||
|
|
||
| __attribute__((leaf)) inline char get_chr_bank(void) { | ||
| return get_bank_state(UNROM_512_CHR_BANK_MASK) >> UNROM_512_CHR_BANK_SHIFT; |
There was a problem hiding this comment.
Nice, this is a clean pattern.
mos-platform/nes-unrom-512/bank.h
Outdated
| /** | ||
| * @brief Define this in a .c file to use 2-screen vertical mirroring. | ||
| */ | ||
| #define UNROM_512_USE_2_SCREEN_VERTICAL \ |
There was a problem hiding this comment.
Also, man this is a really nice pattern; we should have these for most settings in the SDK I think. NAR here; I made #193 for this.
| @@ -18,7 +18,7 @@ banked_call: | |||
| sta __rc17 ; __rc17 = requested PRG bank | |||
There was a problem hiding this comment.
Ah, I hadn't noticed that the RC value is live across the call, right.
Since __rc20 is caller-saved, the value on entry to banked_call needs to be saved and restored too, à la lda __rc20; pha.
There was a problem hiding this comment.
Sorry! Should be fixed.
There was a problem hiding this comment.
Aaand this clobbers the value on entry in A; the store to __rc17 needs to happen first ;)
There was a problem hiding this comment.
Sorry, I've had my head in a whole other place...
All sub-modes are supported, as is both CHR-ROM and CHR-RAM.
This also fixes "bus conflict" behaviour in the UNROM mapper implementation; well, it doesn't exactly fix it, it papers around it - as per the NESdev wiki, programmers should not rely on the "bitwise AND" behaviour of the bus during a conflict. It also marks the submapper in the NES 2.0 header as a non-bus-conflict one.
I'm not sure how to best fix it properly - given the major ROM usage increase required to do so (256 bytes, fewer if one could vary it by linker script), there's some discussion opportunity to be had here, IMO. Probably best for a new issue.