Problem:
Now that shutup is working I found that the address variable is incremented anyway, this happens before the board is assigned or told to shut up
Why this is a problem:
If we tell a board to shut up (or there is not sufficient space for the board) the address gets increased by 'AutoConfSize' anyway causing premature exhaustion of zorro space
Example:
Say you have a config chain that looks like this:
2MB Z2 board
8MB Z2 board (no space, shut up!)
4MB Z2 board
Kickstart will Assign space to boards 1&3 but Diagrom thinks there is no more space when it gets to board 3.
This is a bit of an edge case, not many people are going to have configurations like this because card 2 can never work but for a Z2 ram I'm working on it will attempt to offer smaller memory blocks if told to shut up, and this was preventing me from testing it properly.
I somehow feel like the fix I've done isn't ideal but couldn't think of a better way without reworking large parts of the autoconfig code as the address is incremented early on, but I welcome any feedback if there's a better/easier way to fix this :)
Hi Again @ChuckyGang :)
Problem: Now that shutup is working I found that the address variable is incremented anyway, this happens before the board is assigned or told to shut up
Why this is a problem: If we tell a board to shut up (or there is not sufficient space for the board) the address gets increased by 'AutoConfSize' anyway causing premature exhaustion of zorro space
Example: Say you have a config chain that looks like this:
Kickstart will Assign space to boards 1&3 but Diagrom thinks there is no more space when it gets to board 3.
This is a bit of an edge case, not many people are going to have configurations like this because card 2 can never work but for a Z2 ram I'm working on it will attempt to offer smaller memory blocks if told to shut up, and this was preventing me from testing it properly.
I somehow feel like the fix I've done isn't ideal but couldn't think of a better way without reworking large parts of the autoconfig code as the address is incremented early on, but I welcome any feedback if there's a better/easier way to fix this :)