gbadev.org forum archive

This is a read-only mirror of the content originally found on forum.gbadev.org (now offline), salvaged from Wayback machine copies. A new forum can be found here.

Announcements And Comments > devkitARM release 27 and associated libraries.

#171626 - wintermute - Sat Dec 05, 2009 9:56 pm

devkitARM release 27 is now available.

See http://www.devkitpro.org/devkitarm/devkitarm-release-27-and-associated-libraries/ for full details.
_________________
devkitPro - professional toolchains at amateur prices
devkitPro IRC support
Personal Blog

#171634 - sverx - Mon Dec 07, 2009 3:06 pm

Just updated... but now if I rebuild and run (in no$gba) my programs, the emulator notifies that it's experiencing some 200 errors/sec. :|
It wasn't like that with dkP/libnds previous version, most of my programs had 0 errors notified...

edit: 'uint' type removed?

#171635 - vuurrobin - Mon Dec 07, 2009 4:07 pm

yea, I also notised that uint was removedt. to bad, cause I used that alot.
_________________
my blog:
http://vuurrobin.100webcustomers.com/

#171636 - wintermute - Mon Dec 07, 2009 8:46 pm

sverx wrote:
Just updated... but now if I rebuild and run (in no$gba) my programs, the emulator notifies that it's experiencing some 200 errors/sec. :|
It wasn't like that with dkP/libnds previous version, most of my programs had 0 errors notified...


Yeah, unfortunately no$gba's idea of errors is somewhat flawed - any access to the uncached RAM counts as an "error"

Quote:

edit: 'uint' type removed?


what uint type?
_________________
devkitPro - professional toolchains at amateur prices
devkitPro IRC support
Personal Blog

#171656 - sverx - Wed Dec 09, 2009 9:33 am

wintermute wrote:
Yeah, unfortunately no$gba's idea of errors is somewhat flawed - any access to the uncached RAM counts as an "error"


oh, I didn't know that. Why is libnds accessing uncached RAM now?

wintermute wrote:
what uint type?


unsigned int ... there was an uint typedef somewhere (I was using it twice in one of my programs)... not a big problem btw.

#171662 - wintermute - Wed Dec 09, 2009 5:47 pm

sverx wrote:
wintermute wrote:
Yeah, unfortunately no$gba's idea of errors is somewhat flawed - any access to the uncached RAM counts as an "error"


oh, I didn't know that. Why is libnds accessing uncached RAM now?


It always was, I've just had to rearrange the memory layout a bit for forthcoming DSi support.

Quote:

wintermute wrote:
what uint type?


unsigned int ... there was an uint typedef somewhere (I was using it twice in one of my programs)... not a big problem btw.



Had a bit of a dig around and found the type in the newlib headers so it's still there and works fine for me.

Code:

#include <nds.h>
#include <stdio.h>

//---------------------------------------------------------------------------------
int main(int argc, char **argv) {
//---------------------------------------------------------------------------------
   uint test = 100;

   consoleDemoInit();
   
   printf("%d\n",test);
   while(1) swiWaitForVBlank();

}

_________________
devkitPro - professional toolchains at amateur prices
devkitPro IRC support
Personal Blog

#171691 - sverx - Mon Dec 14, 2009 12:17 pm

wintermute wrote:
sverx wrote:
wintermute wrote:
Yeah, unfortunately no$gba's idea of errors is somewhat flawed - any access to the uncached RAM counts as an "error"

oh, I didn't know that. Why is libnds accessing uncached RAM now?

It always was, I've just had to rearrange the memory layout a bit for forthcoming DSi support.


I got lost. :|

If libnds was already accessing uncached RAM in the previous version, why does no$gba started notifying mistakes only from this latest libnds version? 8| 8| 8|

Sorry to bother...

#173228 - wintermute - Sun Mar 28, 2010 10:51 am

Sorry, just noticed this last question while searching for this thread to answer another post.

Prior to r27 the memory layout was set up for DS - the normal uncached region ran from 0x0240000 to 0x027fffff which is a mirror of the 4meg memory. The upper 4k of that region, 0x027ff000 upwards is used on both commercial games and homebrew for various system uses - card header and some global ( to both processors ) system variables. On the DSi this region is obviously in the middle of the 16meg DSi mode RAM so it was much easier to shift this to 0x02fff000 than to mess around with detection code. With no$ the 0x027 4k region isn't reported as an error but the 0x02f region is.
_________________
devkitPro - professional toolchains at amateur prices
devkitPro IRC support
Personal Blog

#173263 - sverx - Mon Mar 29, 2010 9:56 am

Wow, thanks for the answer! :) Now I see...