A-Bus CS2

[srg320]

I found an interesting thing about the A-Bus CS2. The bios uses different addresses to read 16 and 32 bits data from the CD block: 0x25898000 and 0x25818000 respectively. The bios reads File Info and TOC data in words from 0x25898000, and sector data in long words from 0x25818000. When address bit 19 is 0 SCU does two consecutive reads from the A-Bus. Only address bits AA[5:2] and AA[14:12] are connected to YGR.

I wrote a simple test program and captured A-Bus logs:

temp16 += *(volatile uint16_t *)(0x25818000);
temp32 += *(volatile uint32_t *)(0x25818004);
*(volatile uint16_t *)(0x25818000) = 0x0123;
*(volatile uint32_t *)(0x25818004) = 0x01234567;

temp16 += *(volatile uint16_t *)(0x25898008);
temp32 += *(volatile uint32_t *)(0x2589800C);
*(volatile uint16_t *)(0x25898008) = 0x89AB;
*(volatile uint32_t *)(0x2589800C) = 0x89ABCDEF;

0x25818000 read word:

0x25818004 read long:

0x25818000 write word,0x25818004 write long:

0x25898008 read word,0x2589800C read long:

0x25898008 write word,0x2589800C write long:

I couldn't find any notes about it in SCU manuals.

 

[antime]

SCU errata 32 says that reads from the A- and B-bus areas always generate a 4-byte read, which I ran into when designing the USB cartridge. It is curious there is a region that works differently.

 

[tpu]

Some test about ABUS-CS2:
you can't set wait cycles from ASR1 register. But ADDR[18-15] set the wait cycles.

 

[Jowaco]

I’m looking for similar info and came across your post. I’ve been digging through some old SCU docs and noticed the same thing with the A-Bus CS2. It’s odd how the BIOS uses different addresses for 16-bit and 32-bit data. I didn’t find much in the manuals either, but I did see some info on skin.land where they talked about these quirks. It might be worth a look if you’re still into this topic.

 

[Knight0fDragon]

follow up to this, A17 also seems to do something to the cart port when reading from CS2.

This is derived from the netlink command needed after every call to the UART (0x2582503D)

As far as I can tell, A17 does not have a pin accessible on the Netlink modem, which means it probably alters some state like A19 does.

(A18 may also do something as well based on this info)

Now I noticed tpu said 15-18 deal with the wait cycle, but 16 is tied to the lights on the XBand modem, so I do not think this is the case, at least not without some other means first.