Page 1 of 3
Anyone debug why this crashes ?
Posted: 03 Apr 2020 22:39
by exxos
All it does is run x& = Maddalt&(4194304,2097152) but when the program finishes , it will say program terminated (as expected) but then it bombs... but why ?
FRAM.zip
EDIT:
As far as I can tell, it crashes on RTS, so I assume its a stack issue..
Re: Anyone debug why this crashes ?
Posted: 04 Apr 2020 08:18
by mfro
Maddalt() adds/hooks memory previously not known into internal GEMDOS memory management.
If it crashes, I would assume the added memory is either not where you told GEMDOS or it was already included in the internal management before (error checking is very simplistic, as I recall).
Other than that, it's impossible to tell without knowledge what you exactly do on what exact hardware and how it was initialised before.
Re: Anyone debug why this crashes ?
Posted: 04 Apr 2020 11:22
by exxos
The alt-ram is where its supposed to be, I verified the RAM locations in STOS. So its there. So there shouldn't be any problem in allocating it. But it mostly seems to be when the installer quits, I get 4 bombs.
https://www.exxosforum.co.uk/forum/viewt ... &start=270
Re: Anyone debug why this crashes ?
Posted: 04 Apr 2020 12:40
by mfro
exxos wrote: 04 Apr 2020 11:22
The alt-ram is where its supposed to be, I verified the RAM locations in STOS. So its there. So there shouldn't be any problem in allocating it. But it mostly seems to be when the installer quits, I get 4 bombs.
I would probably start with a specialised RAM test routine to make sure it works reliably?
Re: Anyone debug why this crashes ?
Posted: 04 Apr 2020 13:13
by exxos
mfro wrote: 04 Apr 2020 12:40
I would probably start with a specialised RAM test routine to make sure it works reliably?
Works fine when I access it from STOS.
Re: Anyone debug why this crashes ?
Posted: 25 Jan 2021 15:43
by exxos
Still don't get why it doesn't work..
All I do is x& = Maddalt&(4194304,2097152)
If I run this inside GB6, it installs and lets me run tests, but then bombs when I try and exit the program.. When I just run the line on its own...
IMG_6138.JPG
I don't know how to install the FRB cookie, so maybe thats the problem ?
Re: Anyone debug why this crashes ?
Posted: 25 Jan 2021 17:59
by exxos
Tried to program this in ASM, it compiles but when I run it, I get 4 BOMBs :(
Code: Select all
move.l $200000,d1
move.l $400000,d0
move.l d1,-(sp) ; Size
move.l d0,-(sp) ; Addr
move.w #20,-(sp) ; Maddalt
trap #1 ; GEMDOS
lea 10(sp),sp
Re: Anyone debug why this crashes ?
Posted: 25 Jan 2021 18:07
by mikro
Well if that was really it, no wonder - you need to correctly terminate it - see
https://freemint.github.io/tos.hyp/en/g ... tml#Pterm0. And while you are at it, adding Mshrink() at the beginning wont hurt either ;) -
https://freemint.github.io/tos.hyp/en/g ... ml#Mshrink, for usage see:
http://tho-otto.de/hypview/hypview.cgi? ... 8&index=35
Re: Anyone debug why this crashes ?
Posted: 25 Jan 2021 18:07
by thorsten.otto
That should probably read:
Code: Select all
move.l #$200000,d1 ; <--
move.l #$400000,d0 ; <--
move.l d1,-(sp) ; Size
move.l d0,-(sp) ; Addr
move.w #20,-(sp) ; Maddalt
trap #1 ; GEMDOS
lea 10(sp),sp
clr.w -(sp)
trap #1
Re: Anyone debug why this crashes ?
Posted: 25 Jan 2021 18:12
by exxos
Thanks..
So this doesn't bomb, but doesn't seem to install alt-ram either :(
Code: Select all
move.l $200000,-(sp) ; Size
move.l $400000,-(sp) ; Addr
move.w #20,-(sp) ; Maddalt
trap #1 ; GEMDOS
lea 10(sp),sp
move.w #0,-(sp) ; Offset 0
trap #1 ; GEMDOS
addq.l #2,sp ; Correct stack