Page 8 of 10
Re: RasPiQL
Posted: Thu Sep 29, 2022 2:19 pm
by Derek_Stewart
pjw wrote:Derek_Stewart wrote:<>
Whats a Q80?
Shorthand for Q60 @ 80 MHz
No.... A Q60 has a 68060 CPU, the correct shorthand would be: Q60-clock_speed, e.g Q60-66 or in your case a Q60-80 with a 80Mhz clock speed.
you have just implied a 68080 CPU
Re: RasPiQL
Posted: Thu Sep 29, 2022 2:28 pm
by Derek_Stewart
XorA wrote:XorA wrote:Derek_Stewart wrote:
Hi,
I can only allocate maximum of 16Mb in sQLux and UQLX, how do I increase the memory size?
You need to change the code where allocation occurs.
It occurs to me when I make this change software that uses the top 8 bits of the A? registers for flags will start to fail. I am assuming that case is actually quite rare in QDOS?
I have a quick look at the C source code, but could see were to change the ram size, but judging the previous comments, the change is not straight forward and beyond my programming skills.
Re: RasPiQL
Posted: Thu Sep 29, 2022 3:27 pm
by XorA
Derek_Stewart wrote:XorA wrote:XorA wrote:
You need to change the code where allocation occurs.
It occurs to me when I make this change software that uses the top 8 bits of the A? registers for flags will start to fail. I am assuming that case is actually quite rare in QDOS?
I have a quick look at the C source code, but could see were to change the ram size, but judging the previous comments, the change is not straight forward and beyond my programming skills.
Given everyone (
https://github.com/janbredenbeek/Minerva4Q68) keeps saying 16M is the largets Minerva can address, I wonder why you want this change?
Re: RasPiQL
Posted: Thu Sep 29, 2022 5:27 pm
by Derek_Stewart
XorA wrote:Derek_Stewart wrote:XorA wrote:
It occurs to me when I make this change software that uses the top 8 bits of the A? registers for flags will start to fail. I am assuming that case is actually quite rare in QDOS?
I have a quick look at the C source code, but could see were to change the ram size, but judging the previous comments, the change is not straight forward and beyond my programming skills.
Given everyone (
https://github.com/janbredenbeek/Minerva4Q68) keeps saying 16M is the largets Minerva can address, I wonder why you want this change?
Sorry, I did not suggest the change, or even use SMSQ/E.
I was just following other suggestions.
I would leave the memory maximum size as it is, sd 16Mb is enough for most QL applications.
A better idea would to alter SMSQ/E Gold to run on a standard QL.
Or could sQLux be structured to look like a Gold/Super Card?
Re: RasPiQL
Posted: Fri Sep 30, 2022 9:02 pm
by XorA
Code: Select all
k.issp equ $4c024c02 initial ssp - no longer used as such
l vec.l k.issp ss_ramt ; 00 0 initial SSP 08 2 initial PC
So this means Minerva requires that addresses are truncated, otherwise SSP is a way out in the wilds and gets carried through the whole ROM on uqlx codebase!
Re: RasPiQL
Posted: Fri Sep 30, 2022 11:39 pm
by mk79
XorA wrote:Code: Select all
k.issp equ $4c024c02 initial ssp - no longer used as such
l vec.l k.issp ss_ramt ; 00 0 initial SSP 08 2 initial PC
So this means Minerva requires that addresses are truncated, otherwise SSP is a way out in the wilds and gets carried through the whole ROM on uqlx codebase!
Maybe I don't understand your conclusion, but this initial A7 is never referenced. It's changed early in the reset routine to a real value, so this is no show stopper.
Re: RasPiQL
Posted: Fri Sep 30, 2022 11:47 pm
by XorA
mk79 wrote:XorA wrote:Code: Select all
k.issp equ $4c024c02 initial ssp - no longer used as such
l vec.l k.issp ss_ramt ; 00 0 initial SSP 08 2 initial PC
So this means Minerva requires that addresses are truncated, otherwise SSP is a way out in the wilds and gets carried through the whole ROM on uqlx codebase!
Maybe I don't understand your conclusion, but this initial A7 is never referenced. It's changed early in the reset routine to a real value, so this is no show stopper.
it definitely is referenced. if i remove address truncation i need to zero first byte of rom to boot. I can see this address leaking into addresses later in boot.
Re: RasPiQL
Posted: Sat Oct 01, 2022 12:07 am
by XorA
XorA wrote:mk79 wrote:XorA wrote:Code: Select all
k.issp equ $4c024c02 initial ssp - no longer used as such
l vec.l k.issp ss_ramt ; 00 0 initial SSP 08 2 initial PC
So this means Minerva requires that addresses are truncated, otherwise SSP is a way out in the wilds and gets carried through the whole ROM on uqlx codebase!
Maybe I don't understand your conclusion, but this initial A7 is never referenced. It's changed early in the reset routine to a real value, so this is no show stopper.
it definitely is referenced. if i remove address truncation i need to zero first byte of rom to boot. I can see this address leaking into addresses later in boot.
i have not yet tracked down where, and may obviously be some uqlx hack gone wrong.
Re: RasPiQL
Posted: Sat Oct 01, 2022 7:16 am
by mk79
XorA wrote:i have not yet tracked down where, and may obviously be some uqlx hack gone wrong.
That's what I suspect. After reset it just executes ii_raw, which doesn't use the stack at all (or any RAM really), and immediately afterwards at the label "retry" it is set to a sensible value.
Re: RasPiQL
Posted: Tue Oct 04, 2022 9:30 am
by Peter
Generating a 16 MB version of Q68 SMSQ/E is not a problem. And more memory is quite pointless as long as UQLX only supports graphics MODE 4.