![]() |
![]() |
#1 |
Registered User
Join Date: Jan 2005
Location: Umeå
Age: 44
Posts: 939
|
VBRControl causes ~1min Enforcer startup time
I thought moving VBR to fast ram was rather standardized, but it seems this is an incorrect assumption.
I have been using VBRControl to move VBR to fast ram, but noticed that it causes Enforcer to behave rather weirdly on startup - it locks the computer for about a minute before reporting "Enforcer is on the job." (can't even move the mouse during that time). Using any other of the tools available on aminet to move VBR to fast ram does not cause this behaviour. An example of a tool not giving this effect with Enforcer is VbrMove. This is on a very standard A3000 030, 16MB RAM, Kickstart/Workbench 3.1. Anyone else have experienced this or have an idea what might cause it? |
![]() |
![]() |
#2 |
Registered User
Join Date: Jan 2005
Location: Umeå
Age: 44
Posts: 939
|
Replying to myself here as I stumbled upon the reason by accident
![]() Was working on a tool which reads the whole interrupt vector table. On one of the machines I tested it on, I got lots of enforcer hits because I was reading past the end of fastmem. It turns out this machine was running VBRControl and disabling it fixed the enforcer hits ![]() After digging some more, I found that the real reason is that VBRControl treats the interrupt vector table as it only is 256 Bytes, while it actually is 1024 Bytes plus the fact that it allocates this space from end of fastmem. Because VBRControl allocates from end of fastmem makes it quite easy to see that it only allocates 256 Bytes: Code:
8.Ram Disk:> vbrcontrol on Vector base is located at: 07FFFF00 8.Ram Disk:> work:System-Info/Scout/Scout.os3 COMMAND=memory Address Name Type Pri Lower Upper Attr $07000000 expansion memory memory 30 $07000020 $08000000 $0505 $00004000 chip memory memory -10 $00004020 $00200000 $0703 At least on the A3000, reading past of the end of fastmem is very slow, thats why Enforcer takes so long time to start in this case. Note: If the mmu.library and 68030.library from the MuTools package are loaded, the reads past fastmem end are fast and Enforcer starts without noticeable delay. |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
non-system startup code time/clock problem | modrobert | Coders. Asm / Hardware | 89 | 14 December 2018 15:37 |
Enforcer with WinUAE ? | Lord Riton | support.WinUAE | 11 | 26 September 2017 19:12 |
enforcer with FS-UAE | Lazycow | support.FS-UAE | 8 | 29 November 2015 20:22 |
Enforcer | mritter0 | Coders. General | 6 | 15 July 2014 16:09 |
MMu and Enforcer | Maur "Mef" Fontana | support.Apps | 0 | 19 May 2002 00:50 |
|
|