First read/write!
This commit is contained in:
@@ -123,8 +123,6 @@ cpuif_resets
|
||||
I have zero interest in implementing resynchronizers
|
||||
|
||||
! regwidth/accesswidth is sane
|
||||
! accesswidth is the same for all registers in the regblock
|
||||
- accesswidth is what implies the cpuif bus width
|
||||
! accesswidth == regwidth
|
||||
Enforce this for now. Dont feel like supporting fancy modes yet
|
||||
X regwidth < accesswidth
|
||||
@@ -132,10 +130,18 @@ cpuif_resets
|
||||
! regwidth > accesswidth
|
||||
Need to extend address decode strobes to have multiple bits
|
||||
this is where looking at endinaness matters to determine field placement
|
||||
Dont feel like supporting this yet
|
||||
! constant regwidth?
|
||||
For now, probably limit to only allow the same regwidth everywhere?
|
||||
|
||||
|
||||
! Contents of target are all internal. No external regs
|
||||
! Does not contain any mem components
|
||||
|
||||
! Do not allow unaligned addresses
|
||||
All offsets & strides shall be a multiple of the accesswidth used
|
||||
All offsets & strides shall be a multiple of the regwidth used
|
||||
|
||||
- each reg needs to be aligned to its width
|
||||
- each regfile/addrmap/stride shall be aligned to the largest regwidth it encloses
|
||||
--> Should i promote this check to the compiler? At least as a warnable condition
|
||||
Currently i think I only do the more stringent case of block alignment.
|
||||
|
||||
Reference in New Issue
Block a user