[Virtualacorn-list] VRPS falling over, big time.

Jeremy Nicoll - virtualacorn jn.ml.vac.83 at wingsandbeaks.org.uk
Wed May 27 00:57:28 BST 2009


Kev Smith <kev.smudge at ntlworld.com> wrote:

> The second one is the biggy. Adjusting the amount of RAM for RISC OS 
> causes an error in the RISC OS boot sequence and also has a knock on 
> effect by causing an error in Windows once quitting VRPC. If I change 
> bank 1 RAM from the default 16MB to 64MB I get the error "DataAbort: 
> Abort on data transfer at &02229596 (error number &80000002)" then 
> stops on the supervisor prompt. Put bank one back to 16 and all works. 
> the same problem occurs by putting back 2 to 64, leaving bank 1 at 16, 
> 32MB in each doesn't work, 32 in bank 1 with 16 in bank 2 does work, 
> but 16 and 32 doesn't, wierd, especially as the previous install on 
> the now dead machine had bank 1 at 64MB and bank 2 at 32MB.

Looking at the settings I have:

[Ram]
base = 0x10000000
size = 64Mb

[Ram]
base = 0x14000000
size = 64Mb

makes me guess that the 'base' address for the second bank probably needs to
be the right value - I'm assuming that it's contiguous with the first bank.
In your experiments did you adjust the second 'base' value each time?  

The first figure, 0x10000000 is the hexadecimal representation of 256 M (ie
256 * 1024 * 1024).  If you add 64M to that you get 320M (ie 320 * 1024 *
1024) which in hex is 0x14000000 which suggests the two values are related
this way.  I wonder if you may have told VRPC to use either overlapping
ranges of memory or two areas with a gap between.
 
-- 
Jeremy C B Nicoll - my opinions are my own.




More information about the Virtualacorn-list mailing list